ciberfauna/guara
1
0 Fork 0
Code Issues Pull Requests Actions Packages Projects Releases 19 Wiki Activity

feat: explorador de arquivos, miniaturas, metadados e recursos periciais (0.3.0)
Release / Cria a release (push) Canceled after 0s
Details
Release / CLI (linux-aarch64) (push) Canceled after 0s
Details
Release / CLI (linux-x86_64) (push) Canceled after 0s
Details
Release / CLI (macos-aarch64) (push) Canceled after 0s
Details
Release / CLI (macos-x86_64) (push) Canceled after 0s
Details
Release / CLI (windows-x86_64) (push) Canceled after 0s
Details
Release / App desktop (linux-x86_64) (push) Canceled after 0s
Details
Release / App desktop (macos-aarch64) (push) Canceled after 0s
Details
Release / App desktop (macos-x86_64) (push) Canceled after 0s
Details
Release / App desktop (windows-x86_64) (push) Canceled after 0s
Details
Release / Publica a release (push) Canceled after 0s
Details

Browse Source 
- Explorador com grade/lista virtualizada, navegação por tipo, data EXIF, pasta e
  câmera, seleção estilo gerenciador de arquivos e visualização rápida.
- Miniaturas no núcleo, metadados (EXIF, PDF, Office, MP4), datas MACB do NTFS/FAT/exFAT.
- Hashes MD5/SHA-1/SHA-256, busca de texto no disco, exportação CSV, relatório HTML,
  detecção de duplicadas, tema, correções de interface e textos sem travessões.

Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
This commit is contained in:
Lucas RayanandClaude Sonnet 5.5 committed 2026-10-18 16:16:20 -03:00
1 parent 1457caa453
commit 45d535bdb7
34 files changed
+4433 -365
Download Patch File Download Diff File Expand all files Collapse all files

No files matched your search

CHANGELOG.md
+13
View File
@@ -2,6 +2,19 @@
Formato baseado em [Keep a Changelog](https://keepachangelog.com/pt-BR/1.1.0/); versões seguem SemVer.
## [0.3.0] - 2026-10-18
### Adicionado
- **Explorador de arquivos:** grade de miniaturas e lista com rolagem virtual (milhares de itens), agrupamento por dia/mês, navegação lateral por tipo, data da foto, pasta, câmera e confiança, seleção com Ctrl/Shift/teclado, painel de detalhes e visualização rápida em tela cheia.
- **Miniaturas rápidas** geradas no núcleo (JPEG com redução na decodificação, PNG, GIF, BMP, WebP), com orientação EXIF e uso da miniatura embutida quando a imagem está danificada.
- **Metadados:** EXIF (câmera, lente, datas, GPS), PDF, Office/OpenDocument e MP4/MOV; datas de criação, modificação e acesso do NTFS, FAT e exFAT.
- **Perícia:** hashes MD5, SHA-1 e SHA-256, `guara hash`, `guara search` (texto no disco inteiro em UTF-8 e UTF-16), `guara export` (CSV), `guara report` (relatório HTML com cadeia de custódia) e detecção de duplicadas. Tudo também na interface gráfica.
- Nova identidade visual: logo de pixels que se juntam num lobo-guará de silhueta lisa.
### Alterado
- A barra inferior e o painel de progresso não são mais recriados a cada atualização.
- Textos sem travessões.
## [0.2.0] - 2026-10-18
### Adicionado
Cargo.lock
Generated
+218 -2
View File
@@ -2,6 +2,12 @@
# It is not intended for manual editing.
version = 3
[[package]]
name = "adler2"
version = "2.0.1"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "320119579fcad9c21884f5c4861d16174d0e06250625266f50fe6898340abefa"
[[package]]
name = "anstream"
version = "1.0.0"
@@ -58,6 +64,12 @@ version = "1.1.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "d92bec98840b8f03a5ff5413de5293bfcd8bf96467cf5452609f939ec6f5de16"
[[package]]
name = "autocfg"
version = "1.5.1"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "f2032f911046de80f0a198e0901378627c33f59ea0ac00e363d481118bd70a53"
[[package]]
name = "bitflags"
version = "2.13.2"
@@ -82,6 +94,18 @@ dependencies = [
"objc2",
]
[[package]]
name = "bytemuck"
version = "1.25.2"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "95832e849adfb21180ccb6826a99da14e5d266ae5c2e668e1602cf234f153797"
[[package]]
name = "byteorder-lite"
version = "0.1.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "8f1fe948ff07f4bd06c30984e69f5b4899c516a3ef74f34df92a2df2ab535495"
[[package]]
name = "cfg-if"
version = "1.0.5"
@@ -140,6 +164,12 @@ version = "1.1.1"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "1c133bc6a41be0d194c306b5506d15e6feeea7b1d6604bd3f8310dfb2ca96486"
[[package]]
name = "color_quant"
version = "1.1.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "3d7b894f5411737b7867f4827955924d7c254fc9f4d91a6aad6b097804b1018b"
[[package]]
name = "colorchoice"
version = "1.0.5"
@@ -155,6 +185,15 @@ dependencies = [
"libc",
]
[[package]]
name = "crc32fast"
version = "1.5.2"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "01a7799fd6b852db0e61728dde9a204c423b44d689dbd432522543614b490e78"
dependencies = [
"cfg-if",
]
[[package]]
name = "crypto-common"
version = "0.1.7"
@@ -214,6 +253,26 @@ version = "2.5.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "da7c62ceae207dd37ea5b845da6a0696c799f85e97da1ab5b7910be3c1c80223"
[[package]]
name = "fdeflate"
version = "0.3.7"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "1e6853b52649d4ac5c0bd02320cddc5ba956bdb407c4b75a2c6b75bf51500f8c"
dependencies = [
"simd-adler32",
]
[[package]]
name = "flate2"
version = "1.1.10"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "6e634e2e0ebac1ee034020da1ca582e17ffe4e0f5e985823721e168928136dcb"
dependencies = [
"crc32fast",
"miniz_oxide 0.9.1",
"zlib-rs",
]
[[package]]
name = "generic-array"
version = "0.14.7"
@@ -246,9 +305,19 @@ dependencies = [
"r-efi",
]
[[package]]
name = "gif"
version = "0.14.2"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "ee8cfcc411d9adbbaba82fb72661cc1bcca13e8bba98b364e62b2dba8f960159"
dependencies = [
"color_quant",
"weezl",
]
[[package]]
name = "guara-cli"
version = "0.2.0"
version = "0.3.0"
dependencies = [
"clap",
"ctrlc",
@@ -260,11 +329,17 @@ dependencies = [
[[package]]
name = "guara-core"
version = "0.2.0"
version = "0.3.0"
dependencies = [
"flate2",
"image",
"jpeg-decoder",
"jpeg-encoder",
"libc",
"md-5",
"serde",
"serde_json",
"sha1",
"sha2",
"tempfile",
"thiserror",
@@ -283,6 +358,32 @@ version = "1.0.3"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "df3b46402a9d5adb4c86a0cf463f42e19994e3ee891101b1841f30a545cb49a9"
[[package]]
name = "image"
version = "0.25.10"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "85ab80394333c02fe689eaf900ab500fbd0c2213da414687ebf995a65d5a6104"
dependencies = [
"bytemuck",
"byteorder-lite",
"color_quant",
"gif",
"image-webp",
"moxcms",
"num-traits",
"png",
]
[[package]]
name = "image-webp"
version = "0.2.4"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "525e9ff3e1a4be2fbea1fdf0e98686a6d98b4d8f937e1bf7402245af1909e8c3"
dependencies = [
"byteorder-lite",
"quick-error",
]
[[package]]
name = "is_terminal_polyfill"
version = "1.70.2"
@@ -295,6 +396,18 @@ version = "1.0.18"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "8f42a60cbdf9a97f5d2305f08a87dc4e09308d1276d28c869c684d7777685682"
[[package]]
name = "jpeg-decoder"
version = "0.3.2"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "00810f1d8b74be64b13dbf3db89ac67740615d6c891f0e7b6179326533011a07"
[[package]]
name = "jpeg-encoder"
version = "0.6.1"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "b454d911ac55068f53495488d8ccd0646eaa540c033a28ee15b07838afafb01f"
[[package]]
name = "libc"
version = "0.2.191"
@@ -313,12 +426,52 @@ version = "0.4.34"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "f9f8bd3e56ce4dfc153cf470fffbfa98c7620958b312ca5c3a4b8d5181fd13c6"
[[package]]
name = "md-5"
version = "0.10.6"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "d89e7ee0cfbedfc4da3340218492196241d89eefb6dab27de5df917a6d2e78cf"
dependencies = [
"cfg-if",
"digest",
]
[[package]]
name = "memchr"
version = "2.8.3"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "cf8baf1c55e62ffcace7a9f06f4bd9cd3f0c4beb022d3b367256b91b87513d98"
[[package]]
name = "miniz_oxide"
version = "0.8.9"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "1fa76a2c86f704bdb222d66965fb3d63269ce38518b83cb0575fca855ebb6316"
dependencies = [
"adler2",
"simd-adler32",
]
[[package]]
name = "miniz_oxide"
version = "0.9.1"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "b63fbc4a50860e98e7b2aa7804ded1db5cbc3aff9193adaff57a6931bf7c4b4c"
dependencies = [
"adler2",
"simd-adler32",
]
[[package]]
name = "moxcms"
version = "0.8.1"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "bb85c154ba489f01b25c0d36ae69a87e4a1c73a72631fc6c0eb6dde34a73e44b"
dependencies = [
"num-traits",
"pxfm",
]
[[package]]
name = "nix"
version = "0.31.3"
@@ -331,6 +484,15 @@ dependencies = [
"libc",
]
[[package]]
name = "num-traits"
version = "0.2.19"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "071dfc062690e90b734c0b2273ce72ad0ffa95f0c74596bc250dcfd960262841"
dependencies = [
"autocfg",
]
[[package]]
name = "objc2"
version = "0.6.5"
@@ -358,6 +520,19 @@ version = "1.70.2"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "384b8ab6d37215f3c5301a95a4accb5d64aa607f1fcb26a11b5303878451b4fe"
[[package]]
name = "png"
version = "0.18.1"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "60769b8b31b2a9f263dae2776c37b1b28ae246943cf719eb6946a1db05128a61"
dependencies = [
"bitflags",
"crc32fast",
"fdeflate",
"flate2",
"miniz_oxide 0.8.9",
]
[[package]]
name = "proc-macro2"
version = "1.0.107"
@@ -367,6 +542,18 @@ dependencies = [
"unicode-ident",
]
[[package]]
name = "pxfm"
version = "0.1.30"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "d55d956fa96f5ec02be2e13af0e20391a5aa83d6a074e3ad368959d0fab299ea"
[[package]]
name = "quick-error"
version = "2.0.1"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "a993555f31e5a609f617c12db6250dedcac1b0a85076912c436e6fc9b2c8e6a3"
[[package]]
name = "quote"
version = "1.0.47"
@@ -438,6 +625,17 @@ dependencies = [
"zmij",
]
[[package]]
name = "sha1"
version = "0.10.7"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "a978451301f4db1d02937a4ab3ccce137717b81826e79b7d49ffe3244a13c3b8"
dependencies = [
"cfg-if",
"cpufeatures",
"digest",
]
[[package]]
name = "sha2"
version = "0.10.9"
@@ -449,6 +647,12 @@ dependencies = [
"digest",
]
[[package]]
name = "simd-adler32"
version = "0.3.10"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "3a219298ac11a56ea9a6d2120044824d6f01aeb034955e7af7bc16858527deea"
[[package]]
name = "strsim"
version = "0.11.1"
@@ -541,6 +745,12 @@ version = "0.11.1+wasi-snapshot-preview1"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "ccf3ec651a847eb01de73ccad15eb7d99f80485de043efb2f370cd654f4ea44b"
[[package]]
name = "weezl"
version = "0.1.12"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "a28ac98ddc8b9274cb41bb4d9d4d5c425b6020c50c46f25559911905610b4a88"
[[package]]
name = "windows-link"
version = "0.2.1"
@@ -629,6 +839,12 @@ version = "0.52.6"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "589f6da84c646204747d1270a2a5661ea66ed1cced2631d546fdfb155959f9ec"
[[package]]
name = "zlib-rs"
version = "0.6.8"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "b268e58e7c693d7c271f93ffc4ba3b380412554231c85bf61ca7af91042a4112"
[[package]]
name = "zmij"
version = "1.0.23"
Cargo.toml
+1 -1
View File
@@ -3,7 +3,7 @@ resolver = "2"
members = ["crates/guara-core", "crates/guara-cli"]
[workspace.package]
version = "0.2.0"
version = "0.3.0"
edition = "2021"
license = "MIT OR Apache-2.0"
repository = "https://github.com/lucasrguerra/guara"
README.md
+27 -4
View File
@@ -17,7 +17,7 @@ O Guará junta, numa única varredura, o que ferramentas como Recuva, PhotoRec,
- A origem é aberta **somente para leitura**. Nada no núcleo escreve no dispositivo recuperado.
- A recuperação **recusa destinos no mesmo disco** da origem (Linux/macOS).
- **Imagem forense** (`guara image`) tolerante a setores ruins, com mapa de erros e SHA-256: trabalhe sempre na cópia de um disco com defeito.
- Cada recuperação grava `guara-manifesto.json` com o **SHA-256** de todos os arquivos.
- Cada recuperação grava `guara-manifesto.json` com **MD5, SHA-1 e SHA-256** de todos os arquivos.
- **Volumes cifrados** (BitLocker, LUKS) são **detectados e avisados**; a v1 não desbloqueia nem varre dentro deles.
## Instalação
@@ -48,6 +48,10 @@ guara scan disco.img --recover /mnt/outro-disco/Recuperados
guara recover s.json /mnt/outro-disco/Recuperados --category image
guara resume s.json # continua após Ctrl+C
sudo guara image /dev/sdb /mnt/outro-disco/sdb.img # imagem com mapa de erros
sudo guara hash /dev/sdb # MD5, SHA-1 e SHA-256 do dispositivo
sudo guara search /dev/sdb "contrato" "cpf" # texto em todo o disco (UTF-8 e UTF-16)
guara export s.json achados.csv --category image # planilha dos achados
guara report s.json relatorio.html --case 2026-001 --examiner "Nome" --hash-source
guara ui # interface no navegador (127.0.0.1 + token)
```
@@ -55,9 +59,28 @@ guara ui # interface no navegador (127.0.0.1 + t
## Interface gráfica
App desktop (Tauri) com barra lateral de dispositivos, mapa do disco, resultados ao vivo, filtros por tipo, busca, confiança por arquivo, pausar/continuar e salvar sessão. Tema **Auto / Claro / Escuro** à sua escolha (lembrado também ao reiniciar como administrador) e layout responsivo. O mesmo front-end roda em `guara ui`.
App desktop (Tauri) no estilo de um gerenciador de arquivos, pensado para quando o Guará acha milhares de itens:
**Pré-visualização antes de recuperar** (lida direto do disco, sem gravar nada): imagens (com zoom), **vídeo (MP4/MOV/WebM)** e áudio com player, **PDF** (página a página), texto, documentos Word/Excel/PowerPoint/OpenDocument (texto extraído), conteúdo de ZIP e visão hexadecimal para o resto. Use ↑/↓ para percorrer a lista, Espaço/Enter ou duplo clique para abrir em tela cheia. Quando um arquivo não abre, o aviso ajuda a julgar se ele está danificado. A reprodução de vídeo depende dos codecs do seu sistema (no Linux, instale os plugins `gstreamer1.0-libav` e `gstreamer1.0-plugins-good`).
- **Grade de miniaturas** (gerada pelo núcleo, rápida mesmo com milhares de fotos) ou **lista**, com controle de tamanho, agrupamento por dia ou mês e rolagem virtual que aguenta dezenas de milhares de arquivos.
- **Navegação lateral** por tipo, **data real da foto (EXIF)**, pasta, **câmera**, confiança, com localização GPS e ocultando duplicadas.
- **Seleção como no Finder/Explorer:** clique, Ctrl/Cmd, Shift, setas, Ctrl+A. Espaço, Enter ou duplo clique abrem a **visualização rápida** em tela cheia com navegação por ← →.
- **Painel de detalhes** com pré-visualização, datas (tirada, modificada, criada, acessada) e metadados.
- Pausar/continuar, salvar sessão, tema **Auto / Claro / Escuro** (lembrado também ao reiniciar como administrador) e layout responsivo. O mesmo front-end roda em `guara ui`.
**Pré-visualização antes de recuperar** (lida direto do disco, sem gravar nada): imagens, **vídeo (MP4/MOV/WebM)** e áudio com player, **PDF** página a página, texto, documentos Word/Excel/PowerPoint/OpenDocument (texto extraído), conteúdo de ZIP e visão hexadecimal para o resto. Quando um arquivo não abre, o aviso ajuda a julgar se ele está danificado. A reprodução de vídeo depende dos codecs do seu sistema (no Linux, instale os plugins `gstreamer1.0-libav` e `gstreamer1.0-plugins-good`).
## Recursos periciais
Pensados para quem precisa documentar o que fez:
- **Hashes MD5, SHA-1 e SHA-256** de cada arquivo recuperado, da imagem forense e do dispositivo inteiro (`guara hash`).
- **Metadados:** EXIF (câmera, lente, número de série, data, **GPS**), informações de PDF (autor, programa, datas), propriedades de documentos Office e OpenDocument (autor, última modificação por, empresa) e dados de vídeo MP4/MOV (data de gravação, duração, resolução).
- **Linha do tempo:** datas de criação, modificação e acesso lidas do NTFS, FAT e exFAT.
- **Pesquisa de texto no disco inteiro** (`guara search`), inclusive espaço livre e fora de partições, em UTF-8 e UTF-16, mostrando posição, partição e contexto.
- **Exportação CSV** (abre no Excel) e **relatório HTML** pronto para imprimir (`guara report`), com identificação do caso, origem, partições, resumo, hashes e lista dos arquivos.
- **Detecção de duplicadas** por assinatura de conteúdo.
Escopo: o Guará trabalha com **mídias de armazenamento** (HD, SSD, pendrive, cartões e imagens de disco). Ele **não** desbloqueia celulares nem contorna senhas ou travas de segurança, e não substitui ferramentas de extração de dispositivos móveis.
## Limites conhecidos (v0.1)
@@ -70,7 +93,7 @@ Seja realista sobre o que dá para esperar:
- **exFAT:** coberto por teste com imagem sintética; falta validar em mais mídias reais.
- **Cifragem:** somente detecção.
- **Windows/macOS:** o código compila e é testado em CI, mas os testes com imagens reais rodam só no Linux. Relatos de mídias reais são muito bem-vindos.
- **GUI:** no Linux/macOS o app reinicia inteiro como administrador (via `pkexec`/`osascript`) em vez de usar um processo auxiliar separado — melhoria planejada.
- **GUI:** no Linux/macOS o app reinicia inteiro como administrador (via `pkexec`/`osascript`) em vez de usar um processo auxiliar separado (melhoria planejada).
## Desenvolvimento
SECURITY.md
+1 -1
View File
@@ -1,3 +1,3 @@
# Segurança
Guará abre dispositivos em modo somente leitura e processa dados potencialmente corrompidos ou maliciosos (sistemas de arquivos e arquivos forjados). Vulnerabilidades — corrupção de memória em parsers, escrita indevida na origem, escape de caminho na recuperação — devem ser reportadas em privado pela aba **Security → Report a vulnerability** do repositório no GitHub. Não abra issue pública antes da correção.
Guará abre dispositivos em modo somente leitura e processa dados potencialmente corrompidos ou maliciosos (sistemas de arquivos e arquivos forjados). Vulnerabilidades (corrupção de memória em parsers, escrita indevida na origem, escape de caminho na recuperação) devem ser reportadas em privado pela aba **Security → Report a vulnerability** do repositório no GitHub. Não abra issue pública antes da correção.
apps/gui/src-tauri/Cargo.toml
+1 -1
View File
@@ -3,7 +3,7 @@
[package]
name = "guara-gui"
description = "Guará: interface gráfica de recuperação de arquivos"
version = "0.2.0"
version = "0.3.0"
edition = "2021"
license = "MIT OR Apache-2.0"
repository = "https://github.com/lucasrguerra/guara"
apps/gui/src-tauri/src/main.rs
+22
View File
@@ -101,6 +101,28 @@ fn elevated_exe() -> Result<std::path::PathBuf, String> {
/// com suporte a Range para vídeo/áudio/PDF, lendo direto da origem.
fn serve_file(eng: &Engine, req: &tauri::http::Request<Vec<u8>>) -> tauri::http::Response<Vec<u8>> {
use tauri::http::Response;
if let Some(rest) = req.uri().path().strip_prefix("/thumb/") {
let mut it = rest.split('/');
let (job, id) = (it.next().unwrap_or(""), it.next().and_then(|s| s.parse::<u64>().ok()));
let size = req
.uri()
.query()
.and_then(|q| q.split('&').find_map(|kv| kv.strip_prefix("s=")))
.and_then(|v| v.parse::<u32>().ok())
.unwrap_or(256);
let err = |code: u16| Response::builder().status(code).header("Access-Control-Allow-Origin", "*").body(Vec::new()).unwrap();
let Some(id) = id else { return err(400) };
return match eng.thumb(job, id, size) {
Ok(b) => Response::builder()
.status(200)
.header("Content-Type", "image/jpeg")
.header("Cache-Control", "private, max-age=3600")
.header("Access-Control-Allow-Origin", "*")
.body(b.as_ref().clone())
.unwrap_or_else(|_| err(500)),
Err(code) => err(code),
};
}
const CAP: u64 = 8 * 1024 * 1024;
let mut it = req.uri().path().trim_start_matches("/file/").split('/');
let (job, id) = (it.next().unwrap_or(""), it.next().and_then(|s| s.parse::<u64>().ok()));
apps/gui/src-tauri/tauri.conf.json
+1 -1
View File
@@ -1,7 +1,7 @@
{
"$schema": "https://schema.tauri.app/config/2",
"productName": "Guara",
"version": "0.2.0",
"version": "0.3.0",
"identifier": "dev.guara.app",
"build": {
"frontendDist": "../ui"
apps/gui/ui/app.css
+122
View File
@@ -332,3 +332,125 @@ dialog.lightbox[open] { display: grid; }
.hero-logo { width: min(440px, 80%); height: auto; display: block; margin: 0 0 22px -4px; }
.pill.accent { background: var(--accent-soft); color: var(--accent); }
/* ------------------------------------------------------------ explorador de arquivos */
.main.fixed { overflow: hidden; }
.page.wide { max-width: none; padding: 14px 18px 0; }
.page.fill { height: 100%; display: flex; flex-direction: column; }
.xp-top { display: flex; gap: 16px; align-items: flex-end; flex-wrap: wrap; margin-bottom: 8px; }
.xp-top h1 { font-size: 24px; }
.xp-top .sub { font-size: 12.5px; }
.xp-host { flex: 1; min-height: 0; padding-bottom: 12px; }
.progress-card.slim { padding: 14px 18px; border-radius: 16px; gap: 10px; margin-bottom: 10px; }
.stats.compact { gap: 22px; }
.stats.compact .stat b { font-size: 20px; }
.done-bar { display: flex; gap: 14px; align-items: center; flex-wrap: wrap; padding: 8px 14px; border-radius: 12px; background: var(--accent-soft); margin-bottom: 10px; font-size: 13.5px; }
.done-bar .grow { flex: 1; }
.xp { position: relative; display: grid; grid-template-columns: 236px minmax(0, 1fr) 340px; grid-template-rows: auto minmax(0, 1fr); gap: 10px 14px; height: 100%; }
.xp-bar { grid-column: 1 / -1; display: flex; gap: 10px; align-items: center; min-width: 0; }
.xp-crumb { display: flex; align-items: baseline; gap: 8px; min-width: 0; flex: 1; white-space: nowrap; overflow: hidden; }
.xp-crumb b { font-size: 15px; overflow: hidden; text-overflow: ellipsis; }
.xp-crumb small { color: var(--text-3); font-variant-numeric: tabular-nums; }
.xp-q { flex: 0 1 300px !important; min-width: 120px; }
.xp-bar .theme-switch button { display: inline-grid; place-items: center; padding: 5px 9px; }
.xp-bar input[type="range"] { width: 110px; accent-color: var(--accent); }
.xp-navbtn { display: none; }
.xp-nav { overflow-y: auto; border-radius: var(--radius); background: var(--surface); box-shadow: var(--shadow); padding: 8px 6px 14px; min-height: 0; }
.xp-sec { padding: 4px 0 6px; }
.xp-sec h4 { margin: 8px 10px 4px; font-size: 11.5px; letter-spacing: .02em; color: var(--text-3); font-weight: 600; }
.xp-node { display: flex; align-items: center; gap: 4px; width: 100%; border: 0; background: transparent; cursor: pointer; text-align: left; padding: 6px 8px 6px calc(8px + var(--d, 0) * 14px); border-radius: 9px; font-size: 13.5px; color: var(--text); }
.xp-node:hover { background: color-mix(in srgb, var(--text) 6%, transparent); }
.xp-node.on { background: var(--accent-soft); color: var(--accent); font-weight: 600; }
.xp-node .l { flex: 1; min-width: 0; overflow: hidden; text-overflow: ellipsis; white-space: nowrap; }
.xp-node small { color: var(--text-3); font-variant-numeric: tabular-nums; font-weight: 400; }
.xp-node .tw { display: inline-grid; place-items: center; width: 18px; height: 18px; border-radius: 5px; flex: none; transition: transform .15s; }
.xp-node .tw:hover { background: color-mix(in srgb, var(--text) 10%, transparent); }
.xp-main { position: relative; display: flex; flex-direction: column; min-width: 0; min-height: 0; border-radius: var(--radius); background: var(--surface); box-shadow: var(--shadow); overflow: hidden; }
.xp-chips { display: flex; gap: 6px; flex-wrap: wrap; align-items: center; padding: 0 12px; }
.xp-chips:not(:empty) { padding: 10px 12px 2px; }
.xp-chips .chip { display: inline-flex; gap: 6px; align-items: center; background: var(--accent-soft); color: var(--accent); font-weight: 600; }
.xp-scroll { position: relative; flex: 1; min-height: 0; overflow-y: auto; outline: none; }
.xp-sizer { position: relative; width: 100%; }
.xp-row { position: absolute; left: 0; right: 0; display: flex; gap: 10px; padding: 0 12px; box-sizing: border-box; }
.xp-row.head { align-items: flex-end; }
.xp-head { display: flex; align-items: baseline; gap: 10px; width: 100%; padding: 0 2px 6px; border-bottom: 1px solid var(--line); }
.xp-head b::first-letter { text-transform: uppercase; }
.xp-head b { font-size: 14px; } .xp-head small { color: var(--text-3); } .xp-hsel { margin-left: auto; border: 0; background: transparent; color: var(--accent); cursor: pointer; font-size: 12.5px; opacity: 0; }
.xp-head:hover .xp-hsel, .xp-hsel:focus-visible { opacity: 1; }
.xp-tile { position: relative; flex: none; border-radius: 12px; cursor: pointer; user-select: none; -webkit-user-select: none; }
.xp-thumb { position: relative; border-radius: 12px; overflow: hidden; background: var(--surface-2); box-shadow: inset 0 0 0 1px var(--line); }
.xp-thumb img { width: 100%; height: 100%; object-fit: cover; display: block; background: var(--surface-2); }
.xp-tile[aria-selected="true"] .xp-thumb { box-shadow: 0 0 0 3px var(--accent); }
.xp-tile[data-cur="1"]:focus-visible .xp-thumb, .xp-tile[data-cur="1"] .xp-thumb { outline: 2px solid color-mix(in srgb, var(--accent) 50%, transparent); outline-offset: 2px; }
.xp-ico { width: 100%; height: 100%; display: grid; place-content: center; text-align: center; gap: 2px; color: #fff; }
.xp-ico span { font-size: 34px; line-height: 1; } .xp-ico b { font-size: 12px; letter-spacing: .04em; opacity: .9; }
.c-document { background: linear-gradient(145deg, #3b82f6, #1d4ed8); } .c-video { background: linear-gradient(145deg, #8b5cf6, #5b21b6); }
.c-audio { background: linear-gradient(145deg, #ec4899, #9d174d); } .c-archive { background: linear-gradient(145deg, #f59e0b, #b45309); }
.c-database { background: linear-gradient(145deg, #14b8a6, #0f766e); } .c-other { background: linear-gradient(145deg, #64748b, #334155); } .c-image { background: linear-gradient(145deg, #0ea5e9, #0369a1); }
.c-bad { background: linear-gradient(145deg, #6b7280, #374151); }
.xp-chk { position: absolute; top: 7px; left: 7px; width: 22px; height: 22px; border-radius: 50%; background: rgba(0, 0, 0, .35); box-shadow: inset 0 0 0 2px #fff; opacity: 0; transition: opacity .12s; backdrop-filter: blur(4px); }
.xp-tile:hover .xp-chk, .xp-tile[aria-selected="true"] .xp-chk, .xp-lrow:hover .xp-chk, .xp-lrow[aria-selected="true"] .xp-chk { opacity: 1; }
.xp-tile[aria-selected="true"] .xp-chk, .xp-lrow[aria-selected="true"] .xp-chk { background: var(--accent) url("data:image/svg+xml,%3Csvg xmlns='http://www.w3.org/2000/svg' viewBox='0 0 24 24' fill='none' stroke='white' stroke-width='3.2' stroke-linecap='round' stroke-linejoin='round'%3E%3Cpath d='M6 12.5l4 4 8-9'/%3E%3C/svg%3E") center / 14px no-repeat; }
.xp-dot { position: absolute; right: 7px; bottom: 7px; width: 10px; height: 10px; border-radius: 50%; box-shadow: 0 0 0 2px rgba(255, 255, 255, .85); }
.xp-dot.ok { background: var(--ok); } .xp-dot.warn { background: var(--warn); } .xp-dot.bad { background: var(--bad); }
.xp-gps { position: absolute; left: 7px; bottom: 6px; color: #fff; filter: drop-shadow(0 1px 2px rgba(0, 0, 0, .6)); }
.xp-cap { padding: 5px 2px 0; min-width: 0; }
.xp-cap b { display: block; font-size: 12.5px; font-weight: 500; white-space: nowrap; overflow: hidden; text-overflow: ellipsis; }
.xp-cap small { display: block; font-size: 11.5px; color: var(--text-3); white-space: nowrap; overflow: hidden; text-overflow: ellipsis; }
.xp-lrow { position: relative; display: grid; grid-template-columns: 26px 40px minmax(0, 1fr) 132px 76px 58px; gap: 12px; align-items: center; width: 100%; height: 50px; padding: 0 10px; border-radius: 10px; cursor: pointer; user-select: none; }
.xp-lrow:hover { background: var(--surface-2); }
.xp-lrow[aria-selected="true"] { background: var(--accent-soft); }
.xp-lrow .xp-chk { position: static; opacity: .0; }
.xp-lth { width: 40px; height: 40px; border-radius: 8px; overflow: hidden; background: var(--surface-2); display: grid; place-items: center; }
.xp-lth img { width: 100%; height: 100%; object-fit: cover; }
.xp-lth .xp-ico span { font-size: 18px; } .xp-lth .xp-ico b { display: none; }
.xp-lico { font-size: 18px; width: 100%; height: 100%; display: grid; place-items: center; color: #fff; }
.xp-lnm { min-width: 0; } .xp-lnm b { display: block; font-weight: 500; white-space: nowrap; overflow: hidden; text-overflow: ellipsis; } .xp-lnm small { display: block; color: var(--text-3); white-space: nowrap; overflow: hidden; text-overflow: ellipsis; }
.xp-lcol { color: var(--text-2); font-size: 13px; white-space: nowrap; } .xp-lcol.num { text-align: right; font-variant-numeric: tabular-nums; }
.xp-actions { display: flex; gap: 10px; align-items: center; justify-content: center; flex-wrap: wrap; padding: 10px 14px; border-top: 1px solid var(--line); background: var(--surface); }
.xp-actions .info { color: var(--text-2); font-size: 13.5px; margin-right: auto; }
.xp-insp { overflow-y: auto; border-radius: var(--radius); background: var(--surface); box-shadow: var(--shadow); padding: 14px; display: flex; flex-direction: column; gap: 10px; min-height: 0; }
.xp-insp .viewer { max-height: 38vh; min-height: 150px; }
.xp-ititle { display: flex; gap: 8px; align-items: flex-start; justify-content: space-between; } .xp-ititle b { overflow-wrap: anywhere; font-size: 15px; }
.xp-meta h5 { margin: 6px 0 6px; font-size: 12px; color: var(--text-3); font-weight: 600; }
.xp-multi { display: grid; gap: 12px; padding: 20px 4px; text-align: center; } .xp-multi b { font-size: 18px; }
.xp-insp-x { display: none; align-self: flex-end; }
.xp-veil { display: none; }
.kv dd { overflow-wrap: anywhere; }
@media (max-width: 1280px) { .xp { grid-template-columns: 208px minmax(0, 1fr) 300px; } }
@media (max-width: 1100px) {
.xp { grid-template-columns: minmax(0, 1fr); }
.xp-bar { flex-wrap: wrap; }
.xp-navbtn { display: inline-grid; }
.xp-nav { position: fixed; inset: 0 auto 0 0; width: min(86vw, 310px); z-index: 40; border-radius: 0 var(--radius) var(--radius) 0; transform: translateX(-104%); transition: transform .26s cubic-bezier(.2, .8, .2, 1); }
.xp.show-nav .xp-nav { transform: none; }
.xp-insp { position: fixed; left: 0; right: 0; bottom: 0; max-height: 80vh; z-index: 40; border-radius: 20px 20px 0 0; transform: translateY(104%); transition: transform .26s cubic-bezier(.2, .8, .2, 1); padding-bottom: calc(14px + env(safe-area-inset-bottom)); }
.xp.show-insp .xp-insp { transform: none; }
.xp-insp-x { display: inline-grid; }
.xp-veil { position: fixed; inset: 0; z-index: 30; background: rgba(0, 0, 0, .32); opacity: 0; pointer-events: none; transition: opacity .2s; display: block; }
.xp.show-nav .xp-veil, .xp.show-insp .xp-veil { opacity: 1; pointer-events: auto; }
.xp-q { flex: 1 1 140px !important; }
.xp-crumb { flex: 1 1 100%; order: -1; }
}
@media (max-width: 860px) {
.page.wide { padding: 10px 10px 0; }
.xp-lrow { grid-template-columns: 26px 40px minmax(0, 1fr) 64px; } .xp-lrow .xp-lcol { display: none; }
.xp-bar input[type="range"] { display: none !important; }
}
.fld { display: grid; gap: 4px; margin: 10px 0; font-size: 13px; color: var(--text-2); }
.fld textarea { resize: vertical; font: inherit; }
.hits { background: var(--surface); border-radius: var(--radius); box-shadow: var(--shadow); overflow: hidden; }
.hit { display: grid; grid-template-columns: 130px 80px 90px minmax(0, 1fr); gap: 12px; padding: 9px 14px; border-bottom: 1px solid var(--line); font-size: 13px; align-items: start; }
.hit.head { background: var(--surface-2); color: var(--text-3); font-size: 12px; }
.hit .ctx { font-family: var(--mono); font-size: 12px; overflow-wrap: anywhere; color: var(--text-2); }
.hit mark { background: var(--accent-soft); color: var(--accent); padding: 0 2px; border-radius: 3px; font-weight: 600; }
@media (max-width: 860px) { .hit { grid-template-columns: 100px 1fr; } .hit span:nth-child(2), .hit span:nth-child(3) { display: none; } }
apps/gui/ui/app.js
+133 -229
View File
@@ -1,7 +1,7 @@
// Guará — interface. Fala com o núcleo por `rpc(método, parâmetros)`:
// Guará: interface. Fala com o núcleo por `rpc(método, parâmetros)`:
// • app Tauri: invoke('rpc') • modo `guara ui`: POST /rpc/<método> com token.
import { renderPreview, kindOf, KIND_LABEL } from './preview.js';
import { createExplorer } from './explorer.js';
const tauri = window.__TAURI__;
const params = new URLSearchParams(location.search);
@@ -25,6 +25,7 @@ async function rpc(method, p = {}) {
const isWin = /Windows/i.test(navigator.userAgent);
const fileBase = tauri ? (isWin ? 'http://guara.localhost' : 'guara://localhost') : '';
const thumbUrl = (job, id, s = 256) => `${fileBase}/thumb/${job}/${id}?s=${s}` + (tauri ? '' : `&t=${encodeURIComponent(token || '')}`);
const fileUrl = (job, id) => `${fileBase}/file/${job}/${id}` + (tauri ? '' : `?t=${encodeURIComponent(token || '')}`);
// Atualiza o HTML só quando mudou (evita piscar e perder o foco/hover).
@@ -63,7 +64,7 @@ const dur = (s) => {
s = Math.round(s);
return s < 60 ? `${s}s` : s < 3600 ? `${Math.floor(s / 60)}min ${s % 60}s` : `${Math.floor(s / 3600)}h ${Math.floor((s % 3600) / 60)}min`;
};
const date = (t) => (t ? new Date(t * 1000).toLocaleString('pt-BR', { dateStyle: 'medium', timeStyle: 'short' }) : '—');
const date = (t) => (t ? new Date(t * 1000).toLocaleString('pt-BR', { dateStyle: 'medium', timeStyle: 'short' }) : 'sem data');
const CATS = [
['image', 'Fotos'], ['document', 'Documentos'], ['video', 'Vídeos'], ['audio', 'Áudio'],
['archive', 'Compactados'], ['database', 'Bancos de dados'], ['other', 'Outros'],
@@ -103,6 +104,23 @@ function ask({ title, text, value, placeholder, ok = 'OK', pick }) {
});
}
function askForm({ title, text, fields, ok = 'OK' }) {
return new Promise((resolve) => {
const d = $('#dlg');
d.innerHTML = `<form method="dialog"><h2>${esc(title)}</h2><p>${text || ''}</p>
${fields.map((f) => `<label class="fld"><span>${esc(f.label)}</span>${f.multiline ? `<textarea class="input" name="${f.name}" rows="3" placeholder="${esc(f.placeholder || '')}">${esc(f.value || '')}</textarea>` : `<input class="input" name="${f.name}" value="${esc(f.value || '')}" placeholder="${esc(f.placeholder || '')}" autocomplete="off">`}</label>`).join('')}
<menu><button class="btn" value="cancel">Cancelar</button><button class="btn primary" value="ok">${esc(ok)}</button></menu></form>`;
d.onclose = () => {
if (d.returnValue !== 'ok') return resolve(null);
const out = {};
d.querySelectorAll('[name]').forEach((el) => (out[el.name] = el.value.trim()));
resolve(out);
};
d.showModal();
d.querySelector('[name]')?.focus();
});
}
function info({ title, html }) {
const d = $('#dlg');
d.innerHTML = `<form method="dialog"><h2>${esc(title)}</h2>${html}<menu><button class="btn primary" value="ok">Entendi</button></menu></form>`;
@@ -153,7 +171,7 @@ function renderSide() {
dl.innerHTML = S.devices.map((d, i) => `
<li><button class="dev" data-i="${i}" aria-current="${S.device?.path === d.path}">
<span class="ico">${icon(d.kind)}</span>
<span><span class="t">${esc(d.name)}</span><span class="s">${esc(d.kind)} · ${d.size ? human(d.size) : '—'}${d.readable ? '' : ' · requer admin'}</span></span>
<span><span class="t">${esc(d.name)}</span><span class="s">${esc(d.kind)} · ${d.size ? human(d.size) : '?'}${d.readable ? '' : ' · requer admin'}</span></span>
</button></li>`).join('');
dl.querySelectorAll('.dev').forEach((b) => (b.onclick = () => { openDevice(S.devices[+b.dataset.i]); setMenu(false); }));
const sl = $('#sessionList');
@@ -197,12 +215,13 @@ $('#loadSessionBtn').onclick = async () => {
// ------------------------------------------------------------------ início
function renderHome() {
main.classList.remove('fixed');
stopPoll();
S.device = null; S.job = null;
renderSide();
main.innerHTML = `<div class="page">
<section class="hero"><img class="hero-logo" src="guara-logo.svg" alt="Guará"><h1>Encontre o que você achou que tinha perdido.</h1>
<p>Escolha um disco, SSD ou pendrive ao lado. O Guará lê tudo — arquivos apagados, partições perdidas e espaço não particionado — sem escrever nada no disco de origem.</p></section>
<p>Escolha um disco, SSD ou pendrive ao lado. O Guará lê tudo: arquivos apagados, partições perdidas e espaço não particionado, sem escrever nada no disco de origem.</p></section>
<div class="cards">
<div class="card"><h3>Uma varredura, tudo junto</h3><p>Metadados do sistema de arquivos e busca por assinatura na mesma passada, sem duplicatas.</p></div>
<div class="card"><h3>Confiança por arquivo</h3><p>Cada achado mostra a chance de abrir corretamente: alta, média ou baixa.</p></div>
@@ -214,6 +233,7 @@ function renderHome() {
// ------------------------------------------------------------------ dispositivo
async function openDevice(d) {
main.classList.remove('fixed');
stopPoll();
S.device = d; S.job = null; S.layout = null;
renderSide();
@@ -252,7 +272,7 @@ function renderDevice(err) {
const notes = [];
if (!d.readable) notes.push(['bad', 'Este disco exige permissão de administrador. ' + (tauri ? '<button class="btn sm" id="elevate" style="margin-left:6px">Reiniciar como administrador</button>' : 'Feche e rode de novo com <code>sudo guara ui</code> (Linux/macOS) ou como administrador (Windows).')]);
if (err) notes.push(['bad', esc(err)]);
if (d.mounted_at?.length) notes.push(['warn', `Este disco está em uso (montado em <code>${esc(d.mounted_at.join(', '))}</code>). O Guará só lê, mas evite gravar nele enquanto recupera — qualquer escrita pode sobrescrever o que você quer salvar.`]);
if (d.mounted_at?.length) notes.push(['warn', `Este disco está em uso (montado em <code>${esc(d.mounted_at.join(', '))}</code>). O Guará só lê, mas evite gravar nele enquanto recupera, porque qualquer escrita pode sobrescrever o que você quer salvar.`]);
if (d.trim) notes.push(['warn', 'Este SSD usa TRIM: o sistema avisa o disco quais blocos foram apagados e o próprio SSD pode zerá-los. Arquivos apagados há bastante tempo tendem a não ser recuperáveis.']);
(l?.warnings || []).forEach((w) => notes.push(['warn', esc(w)]));
@@ -272,6 +292,14 @@ function renderDevice(err) {
<div class="mode"><h3>Criar imagem do disco</h3><p>Copia o disco inteiro para um arquivo, tolerando setores ruins. Ideal para discos com defeito: você trabalha na cópia.</p>
<button class="btn lg" id="goImage" ${d.readable ? '' : 'disabled'}>Criar imagem…</button></div>
</div>
<h2 style="margin-top:26px">Ferramentas de perícia</h2>
<div class="modes">
<div class="mode"><h3>Pesquisar texto no disco</h3><p>Procura palavras em todo o disco, inclusive em espaço livre e fora de partições (UTF-8 e UTF-16). Útil quando não há estrutura de arquivo.</p>
<input class="input" id="srchTerms" placeholder="termos separados por vírgula" aria-label="Termos a pesquisar" style="margin-bottom:8px">
<button class="btn lg" id="goSearch" ${d.readable ? '' : 'disabled'}>Pesquisar</button></div>
<div class="mode"><h3>Hashes da origem</h3><p>Calcula MD5, SHA-1 e SHA-256 do disco inteiro para registrar a integridade e a cadeia de custódia no relatório.</p>
<button class="btn lg" id="goHash" ${d.readable ? '' : 'disabled'}>Calcular hashes</button></div>
</div>
<details class="adv"><summary>Opções avançadas</summary>
<label class="check"><input type="checkbox" id="optAlloc"><span><b>Incluir áreas ainda em uso</b><br><span class="sub">Também procura cópias antigas dentro de arquivos existentes. Bem mais lento.</span></span></label>
<div class="row" id="optCats">${CATS.map(([k, n]) => `<label class="check"><input type="checkbox" value="${k}"> ${n}</label>`).join('')}</div>
@@ -287,6 +315,9 @@ function renderDevice(err) {
$('#goDeep')?.addEventListener('click', () => startScan(opts('deep')));
$('#goQuick')?.addEventListener('click', () => startScan(opts('quick')));
$('#goImage')?.addEventListener('click', startImage);
$('#goSearch')?.addEventListener('click', startSearch);
$('#srchTerms')?.addEventListener('keydown', (e) => { if (e.key === 'Enter') startSearch(); });
$('#goHash')?.addEventListener('click', startHash);
}
async function startScan(options) {
@@ -312,80 +343,96 @@ async function startImage() {
} catch (e) { toast(e.message); }
}
async function startHash() {
try {
const { job } = await rpc('hash_start', { path: S.device.path });
watchAux(job, 'Calculando hashes da origem', (st) => {
const r = st.report || {};
S.sourceHashes = { ...r, path: S.device.path };
return `<p>Hashes do conteúdo inteiro de <code>${esc(S.device.path)}</code>:</p><dl class="kv"><dt>MD5</dt><dd class="mono">${esc(r.md5)}</dd><dt>SHA-1</dt><dd class="mono">${esc(r.sha1)}</dd><dt>SHA-256</dt><dd class="mono">${esc(r.sha256)}</dd></dl><p class="sub">Eles serão incluídos nos relatórios gerados nesta sessão.</p>`;
});
} catch (e) { toast(e.message); }
}
async function startSearch() {
const terms = ($('#srchTerms')?.value || '').split(',').map((t) => t.trim()).filter(Boolean);
if (!terms.length) { toast('Digite pelo menos um termo.'); $('#srchTerms')?.focus(); return; }
try {
const { job } = await rpc('search_start', { path: S.device.path, terms, ignore_case: true, utf16: true });
openSearch(job, terms);
} catch (e) { toast(e.message); }
}
function openSearch(job, terms) {
stopPoll();
EX?.destroy(); EX = null;
S.job = null;
main.classList.remove('fixed');
main.innerHTML = `<div class="page"><button class="back" id="backBtn">‹ ${esc(S.device?.name || 'Dispositivos')}</button>
<h1 id="shTitle">Procurando…</h1><div class="sub">Termos: ${terms.map((t) => `<code>${esc(t)}</code>`).join(', ')}</div>
<div id="shProg" style="margin:14px 0"></div><div class="hits" id="hits"></div></div>`;
$('#backBtn').onclick = () => (S.device ? openDevice(S.device) : renderHome());
const termRe = new RegExp(terms.map((t) => esc(t).replace(/[.*+?^${}()|[\]\\]/g, '\\$&')).join('|'), 'gi');
const tick2 = async () => {
let st, res;
try { st = await rpc('status', { job }); res = await rpc('search_results', { job, offset: 0, limit: 1000 }); } catch { return false; }
const run = st.status === 'running';
const p = st.progress || {};
const pct = p.total ? Math.floor((p.done / p.total) * 100) : 0;
$('#shTitle').textContent = run ? 'Procurando…' : `${res.total.toLocaleString('pt-BR')} ocorrência${res.total === 1 ? '' : 's'}`;
setHTML($('#shProg'), run ? `<div class="bar"><i style="width:${pct}%"></i></div><div class="sub" style="margin-top:6px">${pct}% · ${res.total.toLocaleString('pt-BR')} até agora</div><button class="btn sm danger" data-stop="1" style="margin-top:8px">Parar</button>` : st.status === 'error' ? `<div class="notice bad">${warnSvg}<div>${esc(st.error)}</div></div>` : '');
setHTML($('#hits'), res.items.length ? `<div class="hit head"><span>Posição</span><span>Partição</span><span>Codificação</span><span>Contexto</span></div>` + res.items.map((h) => `<div class="hit"><span class="mono" title="${h.offset} bytes (0x${h.offset.toString(16)})">${h.offset.toLocaleString('pt-BR')}</span><span>${h.partition ? '#' + h.partition : 'fora'}</span><span>${esc(h.encoding)}</span><span class="ctx">${esc(h.context).replace(termRe, (m) => `<mark>${m}</mark>`)}</span></div>`).join('') : (run ? '' : '<div class="empty"><h3>Nada encontrado</h3><p>Esses termos não aparecem em nenhum ponto do disco (nem em UTF-8 nem em UTF-16).</p></div>'));
return run;
};
$('#shProg').onclick = (e) => { if (e.target.dataset.stop) rpc('job_stop', { job }); };
(async () => { while (await tick2()) await new Promise((r) => setTimeout(r, 700)); })();
}
// ------------------------------------------------------------------ varredura
let EX = null; // explorador de arquivos (grade, navegação, detalhes)
function openJob(s) {
stopPoll();
EX?.destroy();
EX = null;
S.job = s.job;
S.filters = { category: null, minConfidence: null, query: '', sort: 'confidence' };
S.items = []; S.total = 0; S.selected = new Set(); S.allFiltered = false; S.current = null; S.status = null;
S.status = null;
renderSide();
renderScan();
tick();
S.poll = setInterval(tick, 500);
main.scrollTop = 0;
}
function stopPoll() { clearInterval(S.poll); S.poll = null; }
let lastFound = -1, lastFilterKey = '';
async function tick() {
if (!S.job) return;
let st;
try { st = await rpc('status', { job: S.job }); } catch (e) { stopPoll(); toast(e.message); return; }
S.status = st;
const key = JSON.stringify(S.filters);
if (st.found !== lastFound || key !== lastFilterKey) {
lastFound = st.found; lastFilterKey = key;
await loadResults(true);
}
paintStatus();
if (st.status !== 'running') {
EX?.onStatus(st);
// continua consultando enquanto a varredura roda ou as datas/EXIF ainda estão sendo lidas
if (st.status !== 'running' && !st.enriching) {
stopPoll();
if (st.status === 'error') toast(st.error);
await loadResults(true);
paintStatus();
} else if (!S.poll) {
S.poll = setInterval(tick, 700);
}
}
async function loadResults(reset) {
const f = S.filters;
const query = { sort: f.sort, offset: reset ? 0 : S.items.length, limit: 200 };
if (f.category) query.category = f.category;
if (f.minConfidence) query.min_confidence = f.minConfidence;
if (f.query) query.query = f.query;
const r = await rpc('results', { job: S.job, query });
S.items = reset ? r.items : S.items.concat(r.items);
S.total = r.total;
paintList();
}
function renderScan() {
main.innerHTML = `<div class="page">
<button class="back" id="backBtn">‹ Dispositivos</button>
<div class="dev-head"><div><h1 id="scanTitle">Procurando…</h1><div class="sub" id="scanSub"></div></div></div>
main.classList.add('fixed');
main.innerHTML = `<div class="page wide fill">
<div class="xp-top"><button class="back" id="backBtn">‹ Dispositivos</button><div><h1 id="scanTitle">Procurando…</h1><div class="sub" id="scanSub"></div></div></div>
<div id="progressArea"></div>
<div id="warnArea"></div>
<div class="toolbar">
<div class="chips" id="chips" role="group" aria-label="Filtrar por tipo"></div>
</div>
<div class="toolbar" style="margin-top:0">
<input class="input" id="q" type="search" placeholder="Buscar por nome ou pasta" value="${esc(S.filters.query)}" aria-label="Buscar">
<select class="select" id="conf" aria-label="Confiança mínima"><option value="">Qualquer confiança</option><option value="medium">Média ou alta</option><option value="high">Só alta</option></select>
<select class="select" id="sort" aria-label="Ordenar"><option value="confidence">Mais confiáveis</option><option value="size">Maiores</option><option value="name">Nome</option></select>
</div>
<div class="split"><div class="table" id="table"></div><aside class="preview" id="preview" aria-live="polite"></aside></div>
</div><div id="barHost"></div>`;
<div class="xp-host" id="xpHost"></div>
</div>`;
$('#backBtn').onclick = () => (S.device ? openDevice(S.device) : renderHome());
$('#progressArea').onclick = onProgressClick;
$('#chips').onclick = (e) => { const c = e.target.closest('.chip'); if (c) { S.filters.category = c.dataset.c || null; resetSel(); tick(); } };
let t;
$('#q').oninput = (e) => { clearTimeout(t); t = setTimeout(() => { S.filters.query = e.target.value; resetSel(); tick(); }, 250); };
$('#conf').onchange = (e) => { S.filters.minConfidence = e.target.value || null; resetSel(); tick(); };
$('#sort').onchange = (e) => { S.filters.sort = e.target.value; tick(); };
wireTable();
paintPreview();
EX = createExplorer({ host: $('#xpHost'), job: S.job, rpc, esc, human, fileUrl, thumbUrl, onRecover: recover, toast });
}
const resetSel = () => { S.selected = new Set(); S.allFiltered = false; };
function paintStatus() {
const st = S.status;
@@ -397,26 +444,27 @@ function paintStatus() {
$('#scanSub').innerHTML = `<span class="mono">${esc(st.target)}</span>`;
const rate = p.total && st.elapsed > 1 && p.done ? p.done / st.elapsed : 0;
const eta = rate && running ? dur((p.total - p.done) / rate) : null;
setHTML($('#progressArea'), `<div class="progress-card">
<div class="row spread"><div><b>${esc(p.phase || (running ? 'Iniciando' : ''))}</b>${pct != null ? ` <span class="sub">· ${pct}%</span>` : ''}</div>
<div class="row">${running ? '<button class="btn sm" data-act="pause">Pausar</button><button class="btn sm danger" data-act="stop">Parar</button>' : st.status === 'paused' ? '<button class="btn sm primary" data-act="resume">Continuar</button>' : ''}
${!running && st.found ? '<button class="btn sm" data-act="save">Salvar sessão…</button>' : ''}</div></div>
<div class="bar ${running && pct == null ? 'indet' : ''}"><i style="width:${pct ?? (running ? 35 : 100)}%"></i></div>
<div class="stats">
<div class="stat"><b>${st.found.toLocaleString('pt-BR')}</b><span>arquivos encontrados</span></div>
<div class="stat"><b>${dur(st.elapsed)}</b><span>decorrido${eta ? ' · faltam ~' + eta : ''}</span></div>
<div class="stat"><b style="color:var(--ok)">${st.confidence.high}</b><span>confiança alta</span></div>
<div class="stat"><b style="color:var(--warn)">${st.confidence.medium}</b><span>média</span></div>
<div class="stat"><b style="color:var(--bad)">${st.confidence.low}</b><span>baixa</span></div>
</div></div>`);
const enr = st.enriching ? `<span class="sub">Lendo datas e metadados das fotos… ${st.enriching.done.toLocaleString('pt-BR')}/${st.enriching.total.toLocaleString('pt-BR')}</span>` : '';
const buttons = running ? '<button class="btn sm" data-act="pause">Pausar</button><button class="btn sm danger" data-act="stop">Parar</button>'
: st.status === 'paused' ? '<button class="btn sm primary" data-act="resume">Continuar</button>' : '';
const tools = !running && st.found ? '<button class="btn sm" data-act="csv">Exportar CSV</button><button class="btn sm" data-act="report">Relatório</button><button class="btn sm" data-act="save">Salvar sessão…</button>' : '';
if (running || st.status === 'paused') {
setHTML($('#progressArea'), `<div class="progress-card slim">
<div class="row spread"><div><b>${esc(p.phase || 'Iniciando')}</b>${pct != null ? ` <span class="sub">· ${pct}%</span>` : ''}</div><div class="row">${buttons}</div></div>
<div class="bar ${running && pct == null ? 'indet' : ''}"><i style="width:${pct ?? (running ? 35 : 100)}%"></i></div>
<div class="stats compact">
<div class="stat"><b>${st.found.toLocaleString('pt-BR')}</b><span>arquivos</span></div>
<div class="stat"><b>${dur(st.elapsed)}</b><span>decorrido${eta ? ' · faltam ~' + eta : ''}</span></div>
<div class="stat"><b style="color:var(--ok)">${st.confidence.high}</b><span>alta</span></div>
<div class="stat"><b style="color:var(--warn)">${st.confidence.medium}</b><span>média</span></div>
<div class="stat"><b style="color:var(--bad)">${st.confidence.low}</b><span>baixa</span></div>
</div></div>`);
} else {
setHTML($('#progressArea'), `<div class="done-bar"><span><b>${st.found.toLocaleString('pt-BR')}</b> arquivos em ${dur(st.elapsed)}</span>
<span class="sub">confiança alta ${st.confidence.high} · média ${st.confidence.medium} · baixa ${st.confidence.low}</span>${enr}<span class="grow"></span>${tools}</div>`);
}
setHTML($('#warnArea'), (st.warnings || []).map((w) => `<div class="notice">${warnSvg}<div>${esc(w)}</div></div>`).join('')
+ (st.error ? `<div class="notice bad">${warnSvg}<div>${esc(st.error)}</div></div>` : ''));
// chips
const total = st.found;
setHTML($('#chips'), [`<button class="chip" data-c="" aria-pressed="${!S.filters.category}">Tudo<small>${total}</small></button>`]
.concat(CATS.filter(([k]) => st.categories[k]).map(([k, n]) => `<button class="chip" data-c="${k}" aria-pressed="${S.filters.category === k}">${n}<small>${st.categories[k]}</small></button>`)).join(''));
const sel = $('#conf'); if (sel && sel.value !== (S.filters.minConfidence || '')) sel.value = S.filters.minConfidence || '';
paintBar();
}
async function onProgressClick(e) {
@@ -425,171 +473,27 @@ async function onProgressClick(e) {
if (act === 'pause') { await rpc('scan_pause', { job: S.job }); toast('Pausando…'); }
else if (act === 'stop') await rpc('scan_stop', { job: S.job });
else if (act === 'resume') { await rpc('scan_resume', { job: S.job }); stopPoll(); S.poll = setInterval(tick, 500); tick(); }
else if (act === 'save') {
else if (act === 'csv') {
const path = await ask({ title: 'Exportar para CSV', text: 'Grava os arquivos da vista atual (com os filtros aplicados). O arquivo abre direto no Excel.', value: (S.defaults.home || '') + (S.defaults.sep || '/') + 'guara-achados.csv', ok: 'Exportar' });
if (path) { try { const r = await rpc('export_csv', { job: S.job, path, query: EX?.query() }); toast(`${r.rows.toLocaleString('pt-BR')} linhas gravadas`); } catch (err) { toast(err.message); } }
} else if (act === 'report') {
const f = await askForm({ title: 'Relatório de recuperação', text: 'Gera um relatório HTML pronto para imprimir, com a origem examinada, os achados e os hashes dos arquivos recuperados.', ok: 'Gerar relatório', fields: [
{ name: 'path', label: 'Salvar em', value: (S.defaults.home || '') + (S.defaults.sep || '/') + 'guara-relatorio.html' },
{ name: 'case_id', label: 'Número do caso' }, { name: 'examiner', label: 'Responsável pelo exame' }, { name: 'organization', label: 'Organização' },
{ name: 'notes', label: 'Observações', multiline: true } ] });
if (f) {
try {
const hashes = S.sourceHashes && S.sourceHashes.path === (S.status?.target) ? { md5: S.sourceHashes.md5, sha1: S.sourceHashes.sha1, sha256: S.sourceHashes.sha256 } : null;
await rpc('export_report', { job: S.job, path: f.path, case: { case_id: f.case_id, examiner: f.examiner, organization: f.organization, notes: f.notes }, hashes });
toast('Relatório gravado');
} catch (err) { toast(err.message); }
}
} else if (act === 'save') {
const path = await ask({ title: 'Salvar sessão', text: 'Guarda os resultados para recuperar depois, sem varrer de novo.', value: (S.defaults.home || '') + (S.defaults.sep || '/') + 'guara-sessao.json', ok: 'Salvar' });
if (path) { try { await rpc('session_save', { job: S.job, path }); toast('Sessão salva'); } catch (err) { toast(err.message); } }
}
}
function rowHtml(i) {
const [cl, cc] = CONF[i.confidence];
const on = S.allFiltered || S.selected.has(i.id);
return `<div class="tr" role="row" data-id="${i.id}" aria-selected="${on || S.current?.id === i.id}" tabindex="0">
<input type="checkbox" ${on ? 'checked' : ''} aria-label="Selecionar ${esc(i.name)}">
<div class="nm"><b>${esc(i.name)}</b><small>${esc(i.dir || (i.origin === 'carved' ? 'Encontrado por assinatura' : '—'))}</small></div>
<div class="num">${human(i.size)}</div><div class="num" style="text-align:left">${i.origin === 'carved' ? 'assinatura' : esc(i.origin)}</div>
<div class="num"><span class="pill ${cc}">${cl}</span></div></div>`;
}
function paintList() {
const el = $('#table');
if (!el) return;
const st = S.status;
if (!S.items.length) {
setHTML(el, `<div class="empty"><h3>${st && st.status === 'running' ? 'Procurando…' : 'Nada por aqui'}</h3><p>${st && st.status === 'running' ? 'Os arquivos aparecem aqui assim que são encontrados.' : (S.filters.category || S.filters.query || S.filters.minConfidence) ? 'Nenhum arquivo combina com os filtros.' : 'Não encontramos arquivos recuperáveis neste disco.'}</p></div>`);
paintBar();
return;
}
const allSel = S.items.every((i) => S.selected.has(i.id)) || S.allFiltered;
setHTML(el, `<div class="thead"><input type="checkbox" id="selAll" aria-label="Selecionar todos" ${allSel ? 'checked' : ''}><span>Nome</span><span style="text-align:right">Tamanho</span><span>Origem</span><span style="text-align:right">Confiança</span></div>`
+ S.items.map(rowHtml).join('')
+ (S.items.length < S.total ? `<div class="more"><button class="btn" id="moreBtn">Mostrar mais (${(S.total - S.items.length).toLocaleString('pt-BR')})</button></div>` : ''));
paintBar();
}
// Eventos da tabela por delegação (o HTML é recriado, o listener não).
function wireTable() {
const el = $('#table');
el.addEventListener('click', (e) => {
if (e.target.id === 'moreBtn') return loadResults(false);
const row = e.target.closest('.tr');
if (e.target.id === 'selAll') {
if (e.target.checked) S.items.forEach((i) => S.selected.add(i.id)); else resetSel();
el._h = null; paintList();
return;
}
if (!row) return;
const id = +row.dataset.id;
if (e.target.matches('input[type=checkbox]')) {
if (S.allFiltered) { S.allFiltered = false; S.items.forEach((i) => S.selected.add(i.id)); }
e.target.checked ? S.selected.add(id) : S.selected.delete(id);
row.setAttribute('aria-selected', e.target.checked);
el._h = null;
paintBar();
return;
}
focusItem(S.items.find((i) => i.id === id));
});
el.addEventListener('dblclick', (e) => { if (e.target.closest('.tr')) openLightbox(); });
}
let stopPv = null;
function focusItem(item, { scroll = false } = {}) {
if (!item) return;
S.current = item;
document.querySelectorAll('.tr').forEach((r) => {
const on = +r.dataset.id === item.id || S.selected.has(+r.dataset.id) || S.allFiltered;
r.setAttribute('aria-selected', on);
if (scroll && +r.dataset.id === item.id) r.scrollIntoView({ block: 'nearest' });
});
paintPreview();
}
function detailsHtml(i) {
const [cl, cc] = CONF[i.confidence];
return `<dl class="kv"><dt>Nome</dt><dd>${esc(i.name)}</dd><dt>Pasta</dt><dd>${esc(i.dir || '—')}</dd><dt>Tamanho</dt><dd>${human(i.size)}</dd>
<dt>Modificado</dt><dd>${date(i.modified)}</dd><dt>Origem</dt><dd>${i.origin === 'carved' ? 'Assinatura de formato' : esc(i.origin.toUpperCase())}</dd>
<dt>Confiança</dt><dd><span class="pill ${cc}">${cl}</span></dd>${i.note ? `<dt>Aviso</dt><dd>${esc(i.note)}</dd>` : ''}</dl>`;
}
function paintPreview() {
const el = $('#preview');
if (!el) return;
stopPv?.(); stopPv = null;
const i = S.current;
if (!i) { el.innerHTML = '<div class="viewer"><div class="pv-note">Selecione um arquivo para pré-visualizar</div></div>'; return; }
el.innerHTML = `<div class="viewer" id="pvHost"></div><div class="sub" style="font-size:12.5px">${esc(KIND_LABEL[kindOf(i)])}</div>${detailsHtml(i)}
<div class="row"><button class="btn sm" id="pvFull">Tela cheia</button><button class="btn primary sm" id="recOne">Recuperar este arquivo</button></div>`;
stopPv = renderPreview($('#pvHost'), i, { url: fileUrl(S.job, i.id) });
$('#pvFull').onclick = openLightbox;
$('#recOne').onclick = () => recover({ ids: [i.id] });
}
// ------------------------------------------------------------------ tela cheia
let stopLb = null;
function openLightbox() {
const i = S.current;
if (!i) return;
const d = $('#lightbox');
const idx = S.items.findIndex((x) => x.id === i.id);
d.innerHTML = `<div class="lb-head"><button class="icon-btn" id="lbPrev" aria-label="Anterior" ${idx <= 0 ? 'disabled' : ''}>‹</button><button class="icon-btn" id="lbNext" aria-label="Próximo" ${idx >= S.items.length - 1 ? 'disabled' : ''}>›</button>
<div class="ttl"><b>${esc(i.name)}</b><small>${esc(KIND_LABEL[kindOf(i)])} · ${human(i.size)} · confiança ${CONF[i.confidence][0]}</small></div>
<button class="icon-btn" id="lbClose" aria-label="Fechar"><svg viewBox="0 0 24 24" width="20" height="20"><path d="M6 6l12 12M18 6L6 18" stroke="currentColor" stroke-width="2" stroke-linecap="round"/></svg></button></div>
<div class="lb-body"><div class="viewer large" id="lbHost"></div></div>
<div class="lb-foot"><span class="sub">${i.note ? esc(i.note) : '← → para navegar · Esc para fechar'}</span><button class="btn primary sm" id="lbRec">Recuperar este arquivo</button></div>`;
stopLb?.();
stopLb = renderPreview($('#lbHost', d), i, { url: fileUrl(S.job, i.id), large: true });
const go = (n) => { const t = S.items[idx + n]; if (t) { focusItem(t, { scroll: true }); openLightbox(); } };
$('#lbPrev', d).onclick = () => go(-1);
$('#lbNext', d).onclick = () => go(1);
$('#lbClose', d).onclick = () => d.close();
$('#lbRec', d).onclick = () => { d.close(); recover({ ids: [i.id] }); };
d.onclose = () => { stopLb?.(); stopLb = null; };
d.onkeydown = (e) => { if (e.key === 'ArrowLeft') go(-1); else if (e.key === 'ArrowRight') go(1); };
if (!d.open) d.showModal();
}
// Teclado: ↑/↓ percorre a lista; Espaço/Enter abre a tela cheia.
document.addEventListener('keydown', (e) => {
if (!S.job || document.querySelector('dialog[open]') || /^(INPUT|SELECT|TEXTAREA)$/.test(document.activeElement?.tagName) && document.activeElement.type !== 'checkbox') return;
const idx = S.items.findIndex((x) => x.id === S.current?.id);
if (e.key === 'ArrowDown' || e.key === 'ArrowUp') {
e.preventDefault();
const t = S.items[Math.min(S.items.length - 1, Math.max(0, idx + (e.key === 'ArrowDown' ? 1 : -1)))];
focusItem(t, { scroll: true });
} else if ((e.key === ' ' || e.key === 'Enter') && S.current && !e.target.closest?.('button')) {
e.preventDefault();
openLightbox();
}
});
function selCount() { return S.allFiltered ? S.total : S.selected.size; }
function paintBar() {
const host = $('#barHost');
if (!host) return;
const st = S.status;
const n = selCount();
const canAll = st && st.found > 0 && st.status !== 'error';
if (!n && !canAll) { host.innerHTML = ''; host._h = null; return; }
const moreOffer = !S.allFiltered && S.selected.size > 0 && S.selected.size < S.total && S.items.length === S.selected.size
? '<button class="btn sm" data-bar="all">Selecionar todos os ' + S.total.toLocaleString('pt-BR') + '</button>' : '';
const inner = `<span class="info">${n ? `${n.toLocaleString('pt-BR')} selecionado(s)` : `${S.total.toLocaleString('pt-BR')} arquivo(s) na lista`}</span>${moreOffer}
${n ? '<button class="btn primary" data-bar="sel">Recuperar selecionados</button>' : '<button class="btn primary" data-bar="rec">Recuperar todos da lista</button>'}`;
let bar = host.firstElementChild;
if (!bar) {
// A barra é criada uma vez (com animação de entrada) e depois só tem o texto atualizado.
host.innerHTML = '<div class="actionbar enter"></div>';
bar = host.firstElementChild;
bar.onclick = (e) => {
const act = e.target.closest('[data-bar]')?.dataset.bar;
if (act === 'all') { S.allFiltered = true; $('#table')._h = null; paintList(); }
else if (act === 'sel') recover(S.allFiltered ? { filter: currentFilter() } : { ids: [...S.selected] });
else if (act === 'rec') recover({ filter: currentFilter() });
};
}
setHTML(bar, inner);
}
function currentFilter() {
const f = S.filters;
const q = { sort: f.sort, limit: 1 };
if (f.category) q.category = f.category;
if (f.minConfidence) q.min_confidence = f.minConfidence;
if (f.query) q.query = f.query;
return q;
}
async function recover(sel) {
const dest = await ask({
title: 'Recuperar para…',
apps/gui/ui/explorer.js
+630
View File
@@ -0,0 +1,630 @@
// Explorador de arquivos recuperados: grade de miniaturas (virtualizada, aguenta
// dezenas de milhares de itens), navegação por tipo/data/pasta/câmera, seleção
// estilo gerenciador de arquivos (Ctrl/Shift/teclado), painel de detalhes e
// visualização rápida em tela cheia.
import { renderPreview, kindOf, KIND_LABEL } from './preview.js';
const CAT = {
image: ['Fotos', '🖼'], document: ['Documentos', '📄'], video: ['Vídeos', '🎞'], audio: ['Áudio', '🎵'],
archive: ['Compactados', '🗜'], database: ['Bancos de dados', '🗄'], other: ['Outros', '📦'],
};
const CONF = { high: ['alta', 'ok'], medium: ['média', 'warn'], low: ['baixa', 'bad'] };
const MONTHS = ['janeiro', 'fevereiro', 'março', 'abril', 'maio', 'junho', 'julho', 'agosto', 'setembro', 'outubro', 'novembro', 'dezembro'];
const SORTS = [
['date', 'Data (mais recentes)'], ['date_asc', 'Data (mais antigas)'], ['name', 'Nome'], ['size', 'Tamanho'], ['confidence', 'Confiança'],
];
const GAP = 10;
const PAD = 12;
const svg = (d, w = 18) => `<svg viewBox="0 0 24 24" width="${w}" height="${w}" fill="none" stroke="currentColor" stroke-width="1.8" stroke-linecap="round" stroke-linejoin="round">${d}</svg>`;
const I = {
grid: svg('<rect x="4" y="4" width="7" height="7" rx="1.5"/><rect x="13" y="4" width="7" height="7" rx="1.5"/><rect x="4" y="13" width="7" height="7" rx="1.5"/><rect x="13" y="13" width="7" height="7" rx="1.5"/>'),
list: svg('<path d="M8 6h12M8 12h12M8 18h12M4 6h.01M4 12h.01M4 18h.01"/>'),
filter: svg('<path d="M4 5h16l-6 8v6l-4-2v-4z"/>'),
chev: svg('<path d="M9 6l6 6-6 6"/>', 14),
x: svg('<path d="M6 6l12 12M18 6L6 18"/>', 14),
pin: svg('<path d="M12 21s-6-5.3-6-10a6 6 0 1 1 12 0c0 4.7-6 10-6 10z"/><circle cx="12" cy="11" r="2"/>', 14),
warn: svg('<path d="M12 3l10 18H2z"/><path d="M12 10v5M12 18v.01"/>', 22),
close: svg('<path d="M6 6l12 12M18 6L6 18"/>', 20),
};
export function createExplorer(deps) {
const { host, job, rpc, esc, human, fileUrl, thumbUrl, onRecover, toast } = deps;
const E = {
filters: { category: null, year: null, month: null, undated: false, folder: null, camera: null, origin: null, minConf: null, gps: false, dedupe: false, query: '' },
sort: 'date',
view: 'grid',
tile: 150,
items: [],
total: 0,
sel: new Set(),
allFiltered: false,
anchor: -1,
curId: null,
facets: null,
expanded: new Set(),
rows: [],
height: 0,
rendered: new Map(),
reloadToken: 0,
lastReload: 0,
lastFound: -1,
wasEnriching: false,
status: null,
};
try {
E.view = localStorage.getItem('guara-view') || 'grid';
E.tile = +localStorage.getItem('guara-tile') || 150;
} catch { /* sem armazenamento */ }
host.innerHTML = `<div class="xp" data-view="${E.view}">
<div class="xp-bar">
<button class="icon-btn xp-navbtn" id="xpNavBtn" aria-label="Filtros">${I.filter}</button>
<div class="xp-crumb" id="xpCrumb"></div>
<input class="input xp-q" id="xpQ" type="search" placeholder="Buscar por nome, pasta ou câmera" aria-label="Buscar">
<select class="select" id="xpSort" aria-label="Ordenar">${SORTS.map(([k, n]) => `<option value="${k}">${n}</option>`).join('')}</select>
<div class="theme-switch" role="group" aria-label="Modo de exibição"><button data-view="grid" aria-label="Grade" title="Grade de miniaturas">${I.grid}</button><button data-view="list" aria-label="Lista" title="Lista">${I.list}</button></div>
<input type="range" id="xpSize" min="100" max="300" step="4" aria-label="Tamanho das miniaturas">
</div>
<aside class="xp-nav" id="xpNav" aria-label="Navegação"></aside>
<section class="xp-main">
<div class="xp-chips" id="xpChips"></div>
<div class="xp-scroll" id="xpScroll" tabindex="0" role="listbox" aria-multiselectable="true" aria-label="Arquivos encontrados"><div class="xp-sizer" id="xpSizer"></div></div>
<div class="xp-actions" id="xpActions"></div>
</section>
<aside class="xp-insp" id="xpInsp" aria-label="Detalhes"></aside>
<div class="xp-veil" id="xpVeil"></div>
</div>`;
const $ = (s) => host.querySelector(s);
const root = $('.xp'), scroll = $('#xpScroll'), sizer = $('#xpSizer'), nav = $('#xpNav'), insp = $('#xpInsp'), actions = $('#xpActions');
$('#xpSort').value = E.sort;
$('#xpSize').value = E.tile;
// ------------------------------------------------------------ dados
const toQuery = () => {
const f = E.filters;
const q = {};
if (f.category) q.category = f.category;
if (f.year) q.year = f.year;
if (f.month) q.month = f.month;
if (f.undated) q.undated = true;
if (f.folder) q.folder = f.folder;
if (f.camera) q.camera = f.camera;
if (f.origin) q.origin = f.origin;
if (f.minConf) q.min_confidence = f.minConf;
if (f.gps) q.has_gps = true;
if (f.dedupe) q.dedupe = true;
if (f.query) q.query = f.query;
return q;
};
const anyFilter = () => Object.entries(E.filters).some(([k, v]) => (k === 'query' ? !!v : !!v));
async function reload() {
const token = ++E.reloadToken;
E.lastReload = Date.now();
const base = { ...toQuery(), sort: E.sort };
let items = [], total = 0;
for (;;) {
let r;
try { r = await rpc('results', { job, query: { ...base, offset: items.length, limit: 2000 } }); } catch (e) { toast(e.message); return; }
if (token !== E.reloadToken) return;
items = items.concat(r.items);
total = r.total;
if (!r.items.length || items.length >= total || items.length >= 30000) break;
if (items.length === 2000) { E.items = items; E.total = total; relayout(); paintAll(); }
}
E.items = items;
E.total = total;
relayout();
paintAll();
}
async function loadFacets() {
try { E.facets = await rpc('facets', { job }); } catch { return; }
paintNav();
}
const reloadAll = () => { reload(); loadFacets(); };
// ------------------------------------------------------------ layout virtual
// Agrupa por dia; se houver poucos itens por dia, agrupa por mês.
let groupBy = 'day';
const dayKey = (t) => (t == null ? 'Sem data' : new Date(t * 1000).toISOString().slice(0, groupBy === 'day' ? 10 : 7));
const dayLabel = (k) => {
if (k === 'Sem data') return k;
return groupBy === 'day'
? new Date(k + 'T00:00:00Z').toLocaleDateString('pt-BR', { timeZone: 'UTC', weekday: 'long', day: 'numeric', month: 'long', year: 'numeric' })
: new Date(k + '-01T00:00:00Z').toLocaleDateString('pt-BR', { timeZone: 'UTC', month: 'long', year: 'numeric' });
};
const fmtDate = (t) => (t == null ? 'sem data' : new Date(t * 1000).toLocaleString('pt-BR', { timeZone: 'UTC', dateStyle: 'medium', timeStyle: 'short' }));
const fmtDay = (t) => (t == null ? '' : new Date(t * 1000).toLocaleDateString('pt-BR', { timeZone: 'UTC', day: '2-digit', month: 'short', year: 'numeric' }));
function geometry() {
const W = Math.max(200, scroll.clientWidth - PAD * 2);
if (E.view === 'list') return { W, cols: 1, tw: W, cap: 0, rowH: 54 };
const T = E.tile;
const cols = Math.max(1, Math.floor((W + GAP) / (T + GAP)));
const tw = Math.floor((W - GAP * (cols - 1)) / cols);
const cap = tw >= 128 ? 42 : 0;
return { W, cols, tw, cap, rowH: tw + cap + GAP };
}
let G = geometry();
function relayout() {
G = geometry();
const grouped = E.view === 'grid' && (E.sort === 'date' || E.sort === 'date_asc');
if (grouped) {
groupBy = 'day';
const days = new Set(E.items.map((it) => dayKey(it.when)));
if (E.items.length / Math.max(1, days.size) < 2.5) groupBy = 'month';
}
const rows = [];
let y = PAD;
let cur = null, rowItems = [];
const flush = () => {
if (rowItems.length) { rows.push({ t: 'row', y, h: G.rowH, idx: rowItems }); y += G.rowH; rowItems = []; }
};
E.items.forEach((it, i) => {
if (grouped) {
const k = dayKey(it.when);
if (k !== cur) {
flush();
cur = k;
rows.push({ t: 'head', y, h: 40, key: k, start: i });
y += 40;
}
}
rowItems.push(i);
if (rowItems.length === G.cols) flush();
});
flush();
if (grouped) {
// contagem por grupo
for (let r = 0; r < rows.length; r++) {
if (rows[r].t !== 'head') continue;
let n = 0;
for (let i = rows[r].start; i < E.items.length && dayKey(E.items[i].when) === rows[r].key; i++) n++;
rows[r].count = n;
}
}
E.rows = rows;
E.height = y + PAD;
sizer.style.height = `${E.height}px`;
for (const el of E.rendered.values()) el.remove();
E.rendered.clear();
renderWindow();
}
function rowAt(top) {
let lo = 0, hi = E.rows.length - 1, ans = 0;
while (lo <= hi) {
const mid = (lo + hi) >> 1;
if (E.rows[mid].y + E.rows[mid].h < top) lo = mid + 1; else { ans = mid; hi = mid - 1; }
}
return ans;
}
function renderWindow() {
if (!E.rows.length) { for (const el of E.rendered.values()) el.remove(); E.rendered.clear(); return; }
const top = Math.max(0, scroll.scrollTop - 400), bottom = scroll.scrollTop + scroll.clientHeight + 400;
const a = rowAt(top);
let b = a;
while (b < E.rows.length && E.rows[b].y < bottom) b++;
for (const [ri, el] of E.rendered) {
if (ri < a || ri >= b) { el.remove(); E.rendered.delete(ri); }
}
const frag = document.createDocumentFragment();
for (let ri = a; ri < b; ri++) {
if (E.rendered.has(ri)) continue;
const row = E.rows[ri];
const el = document.createElement('div');
el.className = `xp-row ${row.t}`;
el.style.top = `${row.y}px`;
el.style.height = `${row.h}px`;
el.dataset.ri = ri;
el.innerHTML = row.t === 'head' ? headHtml(row) : row.idx.map((i) => (E.view === 'list' ? listHtml(i) : tileHtml(i))).join('');
E.rendered.set(ri, el);
frag.appendChild(el);
}
sizer.appendChild(frag);
}
// ------------------------------------------------------------ itens
const isSel = (id) => E.allFiltered || E.sel.has(id);
function headHtml(row) {
return `<div class="xp-head"><b>${esc(dayLabel(row.key))}</b><small>${row.count} ${row.count === 1 ? 'item' : 'itens'}</small><button class="xp-hsel" data-hsel="${row.start}" data-n="${row.count}" aria-label="Selecionar este grupo">Selecionar grupo</button></div>`;
}
function thumbSrc(it, px) {
const s = px <= 200 ? 256 : px <= 330 ? 384 : 640;
return thumbUrl(job, it.id, s);
}
function tileHtml(i) {
const it = E.items[i];
const [cl, cc] = CONF[it.confidence];
const on = isSel(it.id);
const [catName, catIco] = CAT[it.category] || CAT.other;
const media = it.category === 'image'
? `<img loading="lazy" decoding="async" alt="" src="${thumbSrc(it, G.tw * (window.devicePixelRatio > 1 ? 1.5 : 1))}" data-th="1">`
: `<div class="xp-ico c-${it.category}"><span>${catIco}</span><b>${esc((it.ext || '?').toUpperCase().slice(0, 5))}</b></div>`;
return `<div class="xp-tile" role="option" data-i="${i}" data-id="${it.id}" aria-selected="${on}" ${E.curId === it.id ? 'data-cur="1"' : ''} style="width:${G.tw}px;height:${G.tw + G.cap}px" title="${esc(it.name)}">
<div class="xp-thumb" style="height:${G.tw}px">${media}
<span class="xp-chk" data-chk="1" aria-hidden="true"></span>
<span class="xp-dot ${cc}" title="Confiança ${cl}"></span>
${it.gps ? `<span class="xp-gps" title="Tem localização GPS">${I.pin}</span>` : ''}
</div>
${G.cap ? `<div class="xp-cap"><b>${esc(it.name)}</b><small>${human(it.size)}${it.when != null ? ' · ' + esc(fmtDay(it.when)) : ''}</small></div>` : ''}
</div>`;
}
function listHtml(i) {
const it = E.items[i];
const [cl, cc] = CONF[it.confidence];
const on = isSel(it.id);
const [, catIco] = CAT[it.category] || CAT.other;
const th = it.category === 'image' ? `<img loading="lazy" decoding="async" alt="" src="${thumbUrl(job, it.id, 128)}" data-th="1">` : `<span class="xp-lico c-${it.category}">${catIco}</span>`;
return `<div class="xp-lrow" role="option" data-i="${i}" data-id="${it.id}" aria-selected="${on}" ${E.curId === it.id ? 'data-cur="1"' : ''}>
<span class="xp-chk" data-chk="1" aria-hidden="true"></span><div class="xp-lth">${th}</div>
<div class="xp-lnm"><b>${esc(it.name)}</b><small>${esc(it.dir || (it.origin === 'carved' ? 'Encontrado por assinatura' : '/'))}</small></div>
<span class="xp-lcol">${esc(fmtDate(it.when))}</span><span class="xp-lcol num">${human(it.size)}</span><span class="pill ${cc}">${cl}</span>
</div>`;
}
// imagens que falham viram um aviso (arquivo possivelmente danificado)
scroll.addEventListener('error', (e) => {
const img = e.target;
if (img.tagName !== 'IMG' || !img.dataset.th) return;
const wrap = img.closest('.xp-thumb, .xp-lth');
img.remove();
wrap?.insertAdjacentHTML('afterbegin', `<div class="xp-ico c-bad" title="Não foi possível gerar a miniatura (arquivo possivelmente danificado)"><span>${I.warn}</span><b>?</b></div>`);
}, true);
// ------------------------------------------------------------ seleção
const selCount = () => (E.allFiltered ? E.total : E.sel.size);
const idxOf = (id) => E.items.findIndex((x) => x.id === id);
function paintSel() {
const set = (el) => {
const id = +el.dataset.id;
el.setAttribute('aria-selected', String(isSel(id)));
if (E.curId === id) el.dataset.cur = '1'; else delete el.dataset.cur;
};
scroll.querySelectorAll('.xp-tile, .xp-lrow').forEach(set);
paintActions();
}
function selectRange(a, b) {
const [lo, hi] = a < b ? [a, b] : [b, a];
for (let i = lo; i <= hi; i++) E.sel.add(E.items[i].id);
}
function setCursor(i, { extend = false, toggle = false, reveal = false } = {}) {
if (i < 0 || i >= E.items.length) return;
const it = E.items[i];
if (E.allFiltered) { E.allFiltered = false; E.sel = new Set(E.items.map((x) => x.id)); }
if (extend && E.anchor >= 0) {
E.sel = new Set();
selectRange(E.anchor, i);
} else if (toggle) {
if (E.sel.has(it.id)) E.sel.delete(it.id); else E.sel.add(it.id);
E.anchor = i;
} else {
E.sel = new Set([it.id]);
E.anchor = i;
}
E.curId = it.id;
if (reveal) revealIndex(i);
paintSel();
schedInspector();
if (matchMedia('(max-width: 1100px)').matches && !extend && !toggle) root.classList.add('show-insp');
}
function revealIndex(i) {
const ri = E.rows.findIndex((r) => r.t === 'row' && r.idx.includes(i));
if (ri < 0) return;
const r = E.rows[ri];
if (r.y < scroll.scrollTop + 4) scroll.scrollTop = r.y - 8;
else if (r.y + r.h > scroll.scrollTop + scroll.clientHeight) scroll.scrollTop = r.y + r.h - scroll.clientHeight + 8;
renderWindow();
}
scroll.addEventListener('scroll', () => renderWindow(), { passive: true });
new ResizeObserver(() => { if (Math.abs(geometry().W - G.W) > 1 || geometry().cols !== G.cols) relayout(); else renderWindow(); }).observe(scroll);
scroll.addEventListener('click', (e) => {
const hs = e.target.closest('[data-hsel]');
if (hs) {
const start = +hs.dataset.hsel, n = +hs.dataset.n;
if (E.allFiltered) { E.allFiltered = false; E.sel = new Set(); }
for (let i = start; i < start + n; i++) E.sel.add(E.items[i].id);
paintSel();
return;
}
const el = e.target.closest('.xp-tile, .xp-lrow');
if (!el) return;
const i = +el.dataset.i;
if (e.target.closest('[data-chk]')) setCursor(i, { toggle: true });
else setCursor(i, { extend: e.shiftKey, toggle: e.ctrlKey || e.metaKey });
});
scroll.addEventListener('dblclick', (e) => {
const el = e.target.closest('.xp-tile, .xp-lrow');
if (el) { setCursor(+el.dataset.i); openLightbox(+el.dataset.i); }
});
scroll.addEventListener('keydown', (e) => {
if (!E.items.length) return;
let i = idxOf(E.curId);
if (i < 0) i = 0;
const cols = G.cols;
const move = { ArrowRight: 1, ArrowLeft: -1, ArrowDown: cols, ArrowUp: -cols, PageDown: cols * 4, PageUp: -cols * 4 }[e.key];
if (move != null) {
e.preventDefault();
setCursor(Math.max(0, Math.min(E.items.length - 1, idxOf(E.curId) < 0 ? 0 : i + move)), { extend: e.shiftKey, reveal: true });
} else if (e.key === 'Home') { e.preventDefault(); setCursor(0, { extend: e.shiftKey, reveal: true }); }
else if (e.key === 'End') { e.preventDefault(); setCursor(E.items.length - 1, { extend: e.shiftKey, reveal: true }); }
else if (e.key === ' ' || e.key === 'Enter') { e.preventDefault(); if (E.curId == null) setCursor(0); openLightbox(idxOf(E.curId)); }
else if ((e.ctrlKey || e.metaKey) && e.key.toLowerCase() === 'a') { e.preventDefault(); E.allFiltered = true; paintSel(); }
else if (e.key === 'Escape') { E.sel = new Set(); E.allFiltered = false; paintSel(); }
});
// ------------------------------------------------------------ barra de ações
function paintActions() {
const n = selCount();
const html = n
? `<span class="info"><b>${n.toLocaleString('pt-BR')}</b> selecionado${n === 1 ? '' : 's'}${E.allFiltered ? ' (todos desta vista)' : ''}</span>
<button class="btn sm" data-a="clear">Limpar seleção</button>
${!E.allFiltered && E.total > n ? `<button class="btn sm" data-a="all">Selecionar todos os ${E.total.toLocaleString('pt-BR')}</button>` : ''}
<button class="btn primary" data-a="rec">Recuperar ${n === 1 ? 'selecionado' : 'selecionados'}</button>`
: `<span class="info">${E.total.toLocaleString('pt-BR')} ${E.total === 1 ? 'item' : 'itens'} nesta vista${E.total > E.items.length ? ` (mostrando ${E.items.length.toLocaleString('pt-BR')})` : ''}</span>
${E.total ? '<button class="btn sm" data-a="all">Selecionar tudo</button><button class="btn primary" data-a="recall">Recuperar tudo desta vista</button>' : ''}`;
if (actions._h !== html) { actions._h = html; actions.innerHTML = html; }
}
actions.addEventListener('click', (e) => {
const a = e.target.closest('[data-a]')?.dataset.a;
if (a === 'clear') { E.sel = new Set(); E.allFiltered = false; paintSel(); }
else if (a === 'all') { E.allFiltered = true; paintSel(); }
else if (a === 'rec') onRecover(E.allFiltered ? { filter: toQuery() } : { ids: [...E.sel] });
else if (a === 'recall') onRecover({ filter: toQuery() });
});
// ------------------------------------------------------------ navegação lateral
function paintNav() {
const f = E.facets;
if (!f) return;
const F = E.filters;
const node = (label, count, attrs, active, depth = 0, extra = '') =>
`<button class="xp-node${active ? ' on' : ''}" ${attrs} style="--d:${depth}" aria-pressed="${!!active}">${extra}<span class="l">${label}</span><small>${count.toLocaleString('pt-BR')}</small></button>`;
const sec = (title, body) => (body ? `<div class="xp-sec"><h4>${title}</h4>${body}</div>` : '');
const cats = Object.keys(CAT).filter((k) => f.categories[k]).map((k) => node(`${CAT[k][1]} ${CAT[k][0]}`, f.categories[k], `data-f="category" data-v="${k}"`, F.category === k)).join('');
const years = f.years.map((y) => {
const open = E.expanded.has('y' + y.year) || F.year === y.year;
return node(String(y.year), y.count, `data-f="year" data-v="${y.year}"`, F.year === y.year && !F.month, 0,
`<span class="tw" data-tw="y${y.year}" aria-label="Expandir" style="${open ? 'transform:rotate(90deg)' : ''}">${I.chev}</span>`)
+ (open ? y.months.map((m) => node(MONTHS[m.month - 1], m.count, `data-f="month" data-y="${y.year}" data-v="${m.month}"`, F.year === y.year && F.month === m.month, 1)).join('') : '');
}).join('') + (f.undated ? node('Sem data', f.undated, 'data-f="undated" data-v="1"', F.undated) : '');
const folders = f.folders.filter((d) => {
// só mostra se todos os ancestrais estiverem expandidos
const parts = d.path.split('/');
for (let k = 1; k < parts.length; k++) if (!E.expanded.has('d' + parts.slice(0, k).join('/')) && !(F.folder || '').startsWith(parts.slice(0, k).join('/'))) return false;
return true;
}).map((d) => {
const hasKids = f.folders.some((o) => o.path.startsWith(d.path + '/'));
const open = E.expanded.has('d' + d.path) || (F.folder || '').startsWith(d.path + '/');
return node(esc(d.name), d.count, `data-f="folder" data-v="${esc(d.path)}"`, F.folder === d.path, d.depth,
hasKids ? `<span class="tw" data-tw="d${esc(d.path)}" style="${open ? 'transform:rotate(90deg)' : ''}">${I.chev}</span>` : '<span class="tw"></span>');
}).join('')
+ (f.root ? node('/ (raiz)', f.root, 'data-f="folder" data-v="@raiz"', F.folder === '@raiz') : '')
+ (f.carved ? node('Por assinatura', f.carved, 'data-f="folder" data-v="@carved"', F.folder === '@carved') : '');
const cams = f.cameras.map((c) => node(esc(c.name), c.count, `data-f="camera" data-v="${esc(c.name)}"`, F.camera === c.name)).join('');
const conf = ['high', 'medium', 'low'].map((k) => node(`Confiança ${CONF[k][0]}`, f.confidence[k], `data-f="minConf" data-v="${k}"`, F.minConf === k)).join('');
const extra = (f.gps ? node('Com localização (GPS)', f.gps, 'data-f="gps" data-v="1"', F.gps) : '')
+ (f.duplicates ? node('Ocultar duplicadas', f.duplicates, 'data-f="dedupe" data-v="1"', F.dedupe) : '');
const html = `<div class="xp-sec">${node('Tudo', f.total, 'data-f="all"', !anyFilter())}</div>`
+ sec('Tipos', cats) + sec('Datas', years) + sec('Pastas', folders) + sec('Câmeras', cams) + sec('Confiança', conf) + sec('Outros', extra);
if (nav._h !== html) { nav._h = html; const st = nav.scrollTop; nav.innerHTML = html; nav.scrollTop = st; }
}
nav.addEventListener('click', (e) => {
const tw = e.target.closest('[data-tw]');
if (tw) {
const k = tw.dataset.tw;
if (E.expanded.has(k)) E.expanded.delete(k); else E.expanded.add(k);
nav._h = null; paintNav();
e.stopPropagation();
return;
}
const b = e.target.closest('.xp-node');
if (!b) return;
const F = E.filters;
const f = b.dataset.f, v = b.dataset.v;
if (f === 'all') clearFilters(false);
else if (f === 'year') { const same = F.year === +v && !F.month; F.year = same ? null : +v; F.month = null; F.undated = false; E.expanded.add('y' + v); }
else if (f === 'month') { const same = F.year === +b.dataset.y && F.month === +v; F.year = same ? null : +b.dataset.y; F.month = same ? null : +v; F.undated = false; }
else if (f === 'undated') { F.undated = !F.undated; F.year = F.month = null; }
else if (f === 'gps') F.gps = !F.gps;
else if (f === 'dedupe') F.dedupe = !F.dedupe;
else F[f] = F[f] === v ? null : v;
if (f === 'folder' && F.folder && !F.folder.startsWith('@')) E.expanded.add('d' + F.folder);
nav._h = null;
root.classList.remove('show-nav');
changed();
});
function clearFilters(refresh = true) {
E.filters = { category: null, year: null, month: null, undated: false, folder: null, camera: null, origin: null, minConf: null, gps: false, dedupe: false, query: '' };
$('#xpQ').value = '';
if (refresh) changed(); else changed();
}
function changed() {
E.sel = new Set(); E.allFiltered = false; E.curId = null; E.anchor = -1;
scroll.scrollTop = 0;
paintCrumb(); paintChips(); paintNav(); paintInspector();
reload();
}
function paintCrumb() {
const F = E.filters, parts = [];
if (F.category) parts.push(CAT[F.category][0]);
if (F.year) parts.push(String(F.year));
if (F.month) parts.push(MONTHS[F.month - 1]);
if (F.undated) parts.push('Sem data');
if (F.folder) parts.push(F.folder === '@carved' ? 'Por assinatura' : F.folder === '@raiz' ? 'Raiz' : F.folder.split('/').join(' › '));
if (F.camera) parts.push(F.camera);
$('#xpCrumb').innerHTML = `<b>${parts.length ? esc(parts.join(' › ')) : 'Todos os arquivos'}</b><small>${E.total.toLocaleString('pt-BR')}</small>`;
}
function paintChips() {
const F = E.filters, chips = [];
const add = (label, key) => chips.push(`<button class="chip" data-x="${key}">${label} ${I.x}</button>`);
if (F.category) add(CAT[F.category][0], 'category');
if (F.year) add(String(F.year), 'year');
if (F.month) add(MONTHS[F.month - 1], 'month');
if (F.undated) add('Sem data', 'undated');
if (F.folder) add(F.folder === '@carved' ? 'Por assinatura' : F.folder === '@raiz' ? 'Raiz' : esc(F.folder), 'folder');
if (F.camera) add(esc(F.camera), 'camera');
if (F.minConf) add(`Confiança ${CONF[F.minConf][0]}`, 'minConf');
if (F.gps) add('Com GPS', 'gps');
if (F.dedupe) add('Sem duplicadas', 'dedupe');
if (F.query) add(`“${esc(F.query)}”`, 'query');
$('#xpChips').innerHTML = chips.length ? chips.join('') + '<button class="btn sm" data-x="all">Limpar filtros</button>' : '';
}
$('#xpChips').addEventListener('click', (e) => {
const k = e.target.closest('[data-x]')?.dataset.x;
if (!k) return;
if (k === 'all') return clearFilters();
E.filters[k] = k === 'query' ? '' : k === 'undated' || k === 'gps' || k === 'dedupe' ? false : null;
if (k === 'year') E.filters.month = null;
if (k === 'query') $('#xpQ').value = '';
nav._h = null;
changed();
});
// ------------------------------------------------------------ barra superior
let qt;
$('#xpQ').addEventListener('input', (e) => { clearTimeout(qt); qt = setTimeout(() => { E.filters.query = e.target.value.trim(); changed(); }, 250); });
$('#xpSort').addEventListener('change', (e) => { E.sort = e.target.value; reload(); });
const paintView = () => {
root.dataset.view = E.view;
host.querySelectorAll('[data-view]').forEach((b) => { if (b.tagName === 'BUTTON') b.setAttribute('aria-pressed', String(b.dataset.view === E.view)); });
$('#xpSize').style.display = E.view === 'grid' ? '' : 'none';
};
host.querySelectorAll('button[data-view]').forEach((b) => (b.onclick = () => {
E.view = b.dataset.view;
try { localStorage.setItem('guara-view', E.view); } catch { /* */ }
paintView(); relayout();
}));
$('#xpSize').addEventListener('input', (e) => {
E.tile = +e.target.value;
try { localStorage.setItem('guara-tile', String(E.tile)); } catch { /* */ }
relayout();
});
$('#xpNavBtn').onclick = () => root.classList.toggle('show-nav');
$('#xpVeil').onclick = () => root.classList.remove('show-nav', 'show-insp');
// ------------------------------------------------------------ painel de detalhes
let inspT, stopPv = null, metaToken = 0;
const schedInspector = () => { clearTimeout(inspT); inspT = setTimeout(paintInspector, 90); };
function timesHtml(it) {
const rows = [['Tirada em', it.taken], ['Modificado', it.modified], ['Criado', it.created], ['Acessado', it.accessed]].filter(([, v]) => v != null);
return rows.map(([k, v]) => `<dt>${k}</dt><dd>${esc(fmtDate(v))}</dd>`).join('');
}
function paintInspector() {
stopPv?.(); stopPv = null;
const n = selCount();
const it = E.items[idxOf(E.curId)];
const close = `<button class="icon-btn xp-insp-x" data-i="close" aria-label="Fechar">${I.close}</button>`;
if (n > 1) {
const sizeSum = E.allFiltered ? null : [...E.sel].reduce((a, id) => a + (E.items[idxOf(id)]?.size || 0), 0);
insp.innerHTML = `${close}<div class="xp-multi"><b>${n.toLocaleString('pt-BR')} itens selecionados</b>${sizeSum != null ? `<span class="sub">${human(sizeSum)} no total</span>` : ''}
<div class="row"><button class="btn primary" data-i="rec">Recuperar selecionados</button><button class="btn" data-i="clear">Limpar seleção</button></div></div>`;
return;
}
if (!it) {
insp.innerHTML = `${close}<div class="xp-empty-insp"><div class="viewer"><div class="pv-note">Selecione um arquivo para pré-visualizar.<br><span class="sub">Setas navegam, Espaço abre em tela cheia.</span></div></div></div>`;
return;
}
const [cl, cc] = CONF[it.confidence];
const url = kindOf(it) === 'image' ? thumbUrl(job, it.id, 640) : fileUrl(job, it.id);
insp.innerHTML = `${close}<div class="viewer" id="xpPv"></div>
<div class="xp-ititle"><b>${esc(it.name)}</b><span class="pill ${cc}">${cl}</span></div>
<div class="sub" style="font-size:12.5px">${esc(KIND_LABEL[kindOf(it)])} · ${human(it.size)}</div>
${it.note ? `<div class="pv-note bad" style="text-align:left">${esc(it.note)}</div>` : ''}
<dl class="kv"><dt>Pasta</dt><dd>${esc(it.dir || (it.origin === 'carved' ? 'Encontrado por assinatura' : '/'))}</dd>${timesHtml(it)}${it.camera ? `<dt>Câmera</dt><dd>${esc(it.camera)}</dd>` : ''}${it.px ? `<dt>Resolução</dt><dd>${it.px[0]} × ${it.px[1]} px</dd>` : ''}</dl>
<div class="xp-meta" id="xpMeta"></div>
<div class="row"><button class="btn sm" data-i="full">Tela cheia</button><button class="btn primary sm" data-i="rec1">Recuperar este arquivo</button></div>`;
stopPv = renderPreview($('#xpPv'), it, { url, thumb: kindOf(it) === 'image' });
const my = ++metaToken;
rpc('metadata', { job, id: it.id }).then((m) => {
if (my !== metaToken || !$('#xpMeta')) return;
const rows = (m.fields || []).filter(([k]) => !['Dimensões'].includes(k));
if (!rows.length) return;
$('#xpMeta').innerHTML = `<h5>Metadados${m.kind ? ` · ${esc(m.kind)}` : ''}</h5><dl class="kv">${rows.map(([k, v]) => `<dt>${esc(k)}</dt><dd>${esc(v)}</dd>`).join('')}</dl>`
+ (m.gps ? `<div class="sub" style="font-size:12.5px">Coordenadas: ${m.gps[0].toFixed(5)}, ${m.gps[1].toFixed(5)}</div>` : '');
}).catch(() => {});
}
insp.addEventListener('click', (e) => {
const a = e.target.closest('[data-i]')?.dataset.i;
if (a === 'full') openLightbox(idxOf(E.curId));
else if (a === 'rec1') onRecover({ ids: [E.curId] });
else if (a === 'rec') onRecover(E.allFiltered ? { filter: toQuery() } : { ids: [...E.sel] });
else if (a === 'clear') { E.sel = new Set(); E.allFiltered = false; paintSel(); paintInspector(); }
else if (a === 'close') root.classList.remove('show-insp');
});
// ------------------------------------------------------------ tela cheia
let stopLb = null;
function openLightbox(i) {
const it = E.items[i];
if (!it) return;
const d = document.querySelector('#lightbox');
const [cl] = CONF[it.confidence];
d.innerHTML = `<div class="lb-head"><button class="icon-btn" id="lbPrev" aria-label="Anterior" ${i <= 0 ? 'disabled' : ''}>‹</button><button class="icon-btn" id="lbNext" aria-label="Próximo" ${i >= E.items.length - 1 ? 'disabled' : ''}>›</button>
<div class="ttl"><b>${esc(it.name)}</b><small>${esc(KIND_LABEL[kindOf(it)])} · ${human(it.size)}${it.when != null ? ' · ' + esc(fmtDate(it.when)) : ''} · confiança ${cl} · ${i + 1} de ${E.items.length}</small></div>
<button class="icon-btn" id="lbClose" aria-label="Fechar">${I.close}</button></div>
<div class="lb-body"><div class="viewer large" id="lbHost"></div></div>
<div class="lb-foot"><span class="sub">${it.note ? esc(it.note) : '← → para navegar · Esc para fechar'}</span><button class="btn primary sm" id="lbRec">Recuperar este arquivo</button></div>`;
stopLb?.();
stopLb = renderPreview(d.querySelector('#lbHost'), it, { url: fileUrl(job, it.id), large: true });
const go = (n) => { const t = i + n; if (t >= 0 && t < E.items.length) { setCursor(t, { reveal: true }); openLightbox(t); } };
d.querySelector('#lbPrev').onclick = () => go(-1);
d.querySelector('#lbNext').onclick = () => go(1);
d.querySelector('#lbClose').onclick = () => d.close();
d.querySelector('#lbRec').onclick = () => { d.close(); onRecover({ ids: [it.id] }); };
d.onclose = () => { stopLb?.(); stopLb = null; scroll.focus({ preventScroll: true }); };
d.onkeydown = (e) => { if (e.key === 'ArrowLeft') go(-1); else if (e.key === 'ArrowRight') go(1); };
if (!d.open) d.showModal();
}
// ------------------------------------------------------------ paint geral
function paintAll() { paintCrumb(); paintChips(); paintActions(); renderWindow(); }
paintView(); paintCrumb(); paintActions(); paintInspector();
reloadAll();
return {
/** Chamado a cada atualização de status da varredura. */
onStatus(st) {
E.status = st;
const enriching = !!st.enriching;
const now = Date.now();
const foundChanged = st.found !== E.lastFound;
const justDone = st.status !== 'running' && E.running;
E.running = st.status === 'running';
if ((foundChanged && now - E.lastReload > 1500) || justDone || (E.wasEnriching && !enriching) || (enriching && now - E.lastReload > 2500)) {
E.lastFound = st.found;
reloadAll();
}
E.wasEnriching = enriching;
},
query: () => toQuery(),
destroy() { E.reloadToken++; stopPv?.(); stopLb?.(); host.innerHTML = ''; },
focus() { scroll.focus({ preventScroll: true }); },
};
}
apps/gui/ui/preview.js
+5 -5
View File
@@ -90,7 +90,7 @@ function note(host, html, cls = '') {
function failure(host, item, err, withHex) {
host.innerHTML = '';
note(host, `<b>Não foi possível pré-visualizar.</b><br>${esc(err?.message || err || '')} O arquivo pode estar danificado, incompleto ou parcialmente sobrescrito — isso também indica que a chance de recuperação é menor.`, 'bad');
note(host, `<b>Não foi possível pré-visualizar.</b><br>${esc(err?.message || err || '')} O arquivo pode estar danificado, incompleto ou parcialmente sobrescrito, o que também indica que a chance de recuperação é menor.`, 'bad');
return withHex ? hex(host, item, host.dataset.url, false) : null;
}
@@ -121,7 +121,7 @@ async function text(host, item, url, large) {
if (item.size > b.length) note(host, `Mostrando o início (${human(b.length)} de ${human(item.size)}).`);
}
async function image(host, item, url, state) {
async function image(host, item, url, state, thumb) {
await new Promise((resolve) => {
const img = new Image();
img.alt = `Pré-visualização de ${item.name}`;
@@ -133,7 +133,7 @@ async function image(host, item, url, state) {
host.appendChild(img);
host.classList.add('zoomable');
img.onclick = () => host.classList.toggle('actual');
note(host, `${img.naturalWidth} × ${img.naturalHeight} px — clique para alternar entre ajustar e tamanho real.`);
if (!thumb) note(host, `${img.naturalWidth} × ${img.naturalHeight} px. Clique para alternar entre ajustar e tamanho real.`);
resolve();
};
img.onerror = () => { if (!state.dead) failure(host, item, 'A imagem não pôde ser decodificada.', true); resolve(); };
@@ -272,7 +272,7 @@ async function zip(host, item, url, state) {
// ------------------------------------------------------------------ API
/** Desenha a pré-visualização de `item` dentro de `host`. Devolve uma função de limpeza. */
export function renderPreview(host, item, { url, large = false }) {
export function renderPreview(host, item, { url, large = false, thumb = false }) {
const state = { dead: false, cleanup: [] };
host.dataset.url = url;
host.classList.remove('zoomable', 'actual');
@@ -281,7 +281,7 @@ export function renderPreview(host, item, { url, large = false }) {
(async () => {
try {
if (!item.size) { host.innerHTML = ''; note(host, 'Arquivo vazio.'); return; }
if (kind === 'image') await image(host, item, url, state);
if (kind === 'image') await image(host, item, url, state, thumb);
else if (kind === 'video') media(host, item, url, 'video', state);
else if (kind === 'audio') media(host, item, url, 'audio', state);
else if (kind === 'pdf') await pdf(host, item, url, state, large);
crates/guara-cli/src/main.rs
+160 -4
View File
@@ -12,7 +12,7 @@ use std::sync::Arc;
use std::time::Duration;
#[derive(Parser)]
#[command(name = "guara", version, about = "Guará — recuperação de arquivos (somente leitura na origem)", long_about = None)]
#[command(name = "guara", version, about = "Guará: recuperação de arquivos (somente leitura na origem)", long_about = None)]
struct Cli {
#[command(subcommand)]
cmd: Cmd,
@@ -50,6 +50,50 @@ enum Cmd {
},
/// Cria uma imagem do disco (tolerante a setores ruins) para trabalhar na cópia
Image { path: String, dest: String },
/// Exporta os achados de uma sessão para CSV (abre no Excel)
Export {
session: String,
/// Arquivo CSV de saída
csv: String,
#[command(flatten)]
filter: FilterArgs,
},
/// Gera o relatório HTML de recuperação (identificação do caso, origem, achados, hashes)
Report {
session: String,
/// Arquivo HTML de saída
out: String,
#[arg(long, default_value = "")]
case: String,
#[arg(long, default_value = "")]
examiner: String,
#[arg(long, default_value = "")]
org: String,
#[arg(long, default_value = "")]
notes: String,
/// Calcula MD5/SHA-1/SHA-256 da origem e inclui no relatório (pode demorar)
#[arg(long)]
hash_source: bool,
},
/// Procura texto em todo o disco, inclusive espaço livre (UTF-8 e UTF-16)
Search {
path: String,
/// Termos a procurar
#[arg(required = true)]
terms: Vec<String>,
/// Diferencia maiúsculas de minúsculas
#[arg(long)]
case_sensitive: bool,
/// Não procura em UTF-16
#[arg(long)]
no_utf16: bool,
#[arg(long, default_value_t = 200)]
max: usize,
#[arg(long)]
json: bool,
},
/// Calcula MD5, SHA-1 e SHA-256 do dispositivo ou imagem (verificação de integridade)
Hash { path: String },
/// Abre a interface gráfica no navegador (modo local)
Ui {
#[arg(long, default_value_t = 0)]
@@ -202,7 +246,9 @@ fn run(cli: Cli) -> Result<(), String> {
"a varredura dessa sessão não terminou; use `guara resume` primeiro".into(),
);
}
recover_job(&eng, &job, &dest, &filter)
recover_job(&eng, &job, &dest, &filter)?;
eng.save_session(&job, &session).map_err(s)?;
Ok(())
}
Cmd::Image { path, dest } => {
let job = eng.start_image(&path, &dest).map_err(s)?;
@@ -220,6 +266,107 @@ fn run(cli: Cli) -> Result<(), String> {
}
Ok(())
}
Cmd::Export {
session,
csv,
filter,
} => {
let job = eng.load_session(&session).map_err(s)?;
let q = query_from(&filter)?;
let r = eng
.rpc("export_csv", json!({ "job": job, "path": csv, "query": q }))
.map_err(s)?;
println!("{} linha(s) gravadas em {csv}", r["rows"]);
Ok(())
}
Cmd::Report {
session,
out,
case,
examiner,
org,
notes,
hash_source,
} => {
let job = eng.load_session(&session).map_err(s)?;
let mut hashes = Value::Null;
if hash_source {
let st = eng.status(&job).map_err(s)?;
let target = st["target"].as_str().unwrap_or("").to_string();
let hj = eng.start_hash(&target).map_err(s)?;
let hst = wait(&eng, &hj, None)?;
eprintln!();
if hst["status"] == "error" {
return Err(hst["error"].as_str().unwrap_or("falha").to_string());
}
hashes = hst["report"].clone();
}
let case = json!({ "case_id": case, "examiner": examiner, "organization": org, "notes": notes });
eng.rpc(
"export_report",
json!({ "job": job, "path": out, "case": case, "hashes": hashes }),
)
.map_err(s)?;
println!("Relatório gravado em {out}");
Ok(())
}
Cmd::Search {
path,
terms,
case_sensitive,
no_utf16,
max,
json,
} => {
let opts = json!({ "path": path, "terms": terms, "ignore_case": !case_sensitive, "utf16": !no_utf16, "max_hits": max });
let job = eng.rpc("search_start", opts).map_err(s)?["job"]
.as_str()
.unwrap_or("")
.to_string();
let st = wait(&eng, &job, None)?;
eprintln!();
if st["status"] == "error" {
return Err(st["error"].as_str().unwrap_or("falha").to_string());
}
let res = eng
.rpc("search_results", json!({ "job": job, "limit": 2000 }))
.map_err(s)?;
if json {
println!("{}", serde_json::to_string_pretty(&res).unwrap());
return Ok(());
}
for h in res["items"].as_array().into_iter().flatten() {
println!(
"{:>14} {} {:<6} {:<8} {}",
h["offset"],
h["partition"]
.as_u64()
.map(|p| format!("part.{p}"))
.unwrap_or_else(|| "fora".into()),
h["encoding"].as_str().unwrap_or(""),
h["term"].as_str().unwrap_or(""),
h["context"].as_str().unwrap_or("")
);
}
println!("{} ocorrência(s).", res["total"]);
Ok(())
}
Cmd::Hash { path } => {
let job = eng.start_hash(&path).map_err(s)?;
let st = wait(&eng, &job, None)?;
eprintln!();
if st["status"] == "error" {
return Err(st["error"].as_str().unwrap_or("falha").to_string());
}
let r = &st["report"];
println!(
"MD5 {}\nSHA-1 {}\nSHA-256 {}",
r["md5"].as_str().unwrap_or(""),
r["sha1"].as_str().unwrap_or(""),
r["sha256"].as_str().unwrap_or("")
);
Ok(())
}
Cmd::Ui { port, no_open } => serve::run(port, !no_open),
}
}
@@ -246,6 +393,7 @@ fn query_from(f: &FilterArgs) -> Result<Query, String> {
sort: Some("confidence".into()),
offset: 0,
limit: 2000,
..Default::default()
})
}
@@ -290,6 +438,9 @@ fn drive_scan(
}
if let Some(dest) = recover {
recover_job(eng, job, dest, filter)?;
if let Some(p) = &out.session {
eng.save_session(job, p).map_err(s)?;
}
} else if st["found"].as_u64().unwrap_or(0) > 0 && out.session.is_some() {
eprintln!(
"Para recuperar: guara recover {} <pasta-de-destino>",
@@ -391,8 +542,13 @@ fn wait(eng: &Engine, job: &str, session: Option<&str>) -> Result<Value, String>
.and_then(|v| v.checked_div(t))
.map_or(" … ".to_string(), |v| format!("{v:>3}%"));
let found = st["found"].as_u64().unwrap_or(0);
let found = if st["hits"].as_u64().unwrap_or(0) > 0 {
st["hits"].as_u64().unwrap_or(0)
} else {
found
};
let line = format!(
"{pct} {} · {found} arquivo(s) encontrado(s) · {:.0}s",
"{pct} {} · {found} resultado(s) · {:.0}s",
p["phase"].as_str().unwrap_or(""),
st["elapsed"].as_f64().unwrap_or(0.0)
);
@@ -443,7 +599,7 @@ fn print_layout(v: &Value) {
fn print_summary(st: &Value) {
println!(
"Concluído em {:.1}s — {} arquivo(s) encontrado(s).",
"Concluído em {:.1}s: {} arquivo(s) encontrado(s).",
st["elapsed"].as_f64().unwrap_or(0.0),
st["found"]
);
crates/guara-cli/src/serve.rs
+42
View File
@@ -11,6 +11,7 @@ const CSS: &str = include_str!("../../../apps/gui/ui/app.css");
const JS: &str = include_str!("../../../apps/gui/ui/app.js");
const PREVIEW_JS: &str = include_str!("../../../apps/gui/ui/preview.js");
const LOGO: &str = include_str!("../../../apps/gui/ui/guara-logo.svg");
const EXPLORER_JS: &str = include_str!("../../../apps/gui/ui/explorer.js");
const ICON: &str = include_str!("../../../apps/gui/ui/guara.svg");
const PDF_JS: &str = include_str!("../../../apps/gui/ui/vendor/pdf.min.mjs");
const PDF_WORKER: &str = include_str!("../../../apps/gui/ui/vendor/pdf.worker.min.mjs");
@@ -85,6 +86,12 @@ fn handle(mut req: Request, eng: &Engine, token: &str, hosts: &[String]) {
(Method::Get, "/guara-logo.svg") => {
let _ = req.respond(Response::from_string(LOGO).with_header(ct("image/svg+xml")));
}
(Method::Get, "/explorer.js") => {
let _ = req.respond(
Response::from_string(EXPLORER_JS)
.with_header(ct("text/javascript; charset=utf-8")),
);
}
(Method::Get, "/guara.svg") => {
let _ = req.respond(Response::from_string(ICON).with_header(ct("image/svg+xml")));
}
@@ -98,6 +105,41 @@ fn handle(mut req: Request, eng: &Engine, token: &str, hosts: &[String]) {
Response::from_string(PDF_WORKER).with_header(ct("text/javascript; charset=utf-8")),
);
}
(Method::Get, p) if p.starts_with("/thumb/") => {
let tok_ok = query
.split('&')
.any(|kv| kv.strip_prefix("t=") == Some(token));
if !tok_ok {
let _ = req.respond(Response::from_string("token inválido").with_status_code(401));
return;
}
let mut it = p.trim_start_matches("/thumb/").split('/');
let (job, id) = (
it.next().unwrap_or(""),
it.next().and_then(|s| s.parse::<u64>().ok()),
);
let size = query
.split('&')
.find_map(|kv| kv.strip_prefix("s="))
.and_then(|v| v.parse::<u32>().ok())
.unwrap_or(256);
let Some(id) = id else {
let _ = req.respond(Response::from_string("id inválido").with_status_code(400));
return;
};
match eng.thumb(job, id, size) {
Ok(bytes) => {
let resp = Response::from_data(bytes.as_ref().clone())
.with_header(ct("image/jpeg"))
.with_header(header("Cache-Control", "private, max-age=3600"));
let _ = req.respond(resp);
}
Err(code) => {
let _ =
req.respond(Response::from_string("sem miniatura").with_status_code(code));
}
}
}
(Method::Get, p) if p.starts_with("/file/") => {
let tok_ok = query
.split('&')
crates/guara-core/Cargo.toml
+6
View File
@@ -11,6 +11,12 @@ rust-version.workspace = true
serde = { version = "1", features = ["derive"] }
serde_json = "1"
sha2 = "0.10"
sha1 = "0.10"
md-5 = "0.10"
jpeg-decoder = { version = "0.3", default-features = false }
jpeg-encoder = "0.6"
image = { version = "0.25", default-features = false, features = ["png", "gif", "bmp", "webp"] }
flate2 = { version = "1", default-features = false, features = ["rust_backend"] }
thiserror = "2"
[target.'cfg(unix)'.dependencies]
crates/guara-core/src/explorer.rs
+368
View File
@@ -0,0 +1,368 @@
//! Navegação estilo gerenciador de arquivos: filtros, ordenação, facetas
//! (tipos, datas, pastas, câmeras) e cache de miniaturas.
use crate::types::*;
use serde::{Deserialize, Serialize};
use serde_json::{json, Value};
use std::collections::{BTreeMap, HashMap, VecDeque};
use std::sync::{Arc, Condvar, Mutex};
#[derive(Debug, Clone, Serialize, Deserialize)]
pub struct Query {
#[serde(default)]
pub category: Option<Category>,
#[serde(default)]
pub min_confidence: Option<Confidence>,
#[serde(default)]
pub query: Option<String>,
#[serde(default)]
pub sort: Option<String>,
#[serde(default)]
pub offset: usize,
#[serde(default = "default_limit")]
pub limit: usize,
/// Pasta (prefixo), "@carved" (achados por assinatura) ou "@raiz".
#[serde(default)]
pub folder: Option<String>,
#[serde(default)]
pub year: Option<i64>,
#[serde(default)]
pub month: Option<u32>,
/// Só arquivos sem data conhecida.
#[serde(default)]
pub undated: Option<bool>,
#[serde(default)]
pub camera: Option<String>,
#[serde(default)]
pub origin: Option<Origin>,
#[serde(default)]
pub min_size: Option<u64>,
#[serde(default)]
pub max_size: Option<u64>,
/// Megapixels mínimos.
#[serde(default)]
pub min_mp: Option<f64>,
#[serde(default)]
pub has_gps: Option<bool>,
/// Mostra só a primeira ocorrência de cada conteúdo repetido.
#[serde(default)]
pub dedupe: Option<bool>,
}
fn default_limit() -> usize {
200
}
impl Default for Query {
fn default() -> Self {
serde_json::from_str("{}").unwrap()
}
}
impl Query {
pub fn needle(&self) -> Option<String> {
self.query
.as_deref()
.map(|x| x.to_lowercase())
.filter(|x| !x.is_empty())
}
/// Itens que atendem ao filtro (na ordem de achado), já sem duplicadas se pedido.
pub fn select<'a>(&self, found: &'a [Found]) -> Vec<&'a Found> {
let needle = self.needle();
let mut v: Vec<&Found> = found.iter().filter(|f| self.matches(f, &needle)).collect();
if self.dedupe == Some(true) {
let mut seen = std::collections::HashSet::new();
v.retain(|f| f.sig.map(|g| seen.insert((g, f.size))).unwrap_or(true));
}
v
}
pub fn matches(&self, f: &Found, needle: &Option<String>) -> bool {
if self.category.is_some_and(|c| f.category != c) {
return false;
}
if self.min_confidence.is_some_and(|c| f.confidence < c) {
return false;
}
if self.origin.is_some_and(|o| f.origin != o) {
return false;
}
if self.min_size.is_some_and(|m| f.size < m) || self.max_size.is_some_and(|m| f.size > m) {
return false;
}
if let Some(n) = needle {
let hay =
format!("{} {}", f.full_path(), f.camera.as_deref().unwrap_or("")).to_lowercase();
if !hay.contains(n.as_str()) {
return false;
}
}
if let Some(folder) = &self.folder {
let ok = match folder.as_str() {
"@carved" => f.origin == Origin::Carved,
"@raiz" => {
f.origin != Origin::Carved
&& f.dir.as_deref().map(str::is_empty).unwrap_or(true)
}
p => f
.dir
.as_deref()
.is_some_and(|d| d == p || d.starts_with(&format!("{p}/"))),
};
if !ok {
return false;
}
}
if self.year.is_some() || self.month.is_some() {
let Some(t) = f.when() else { return false };
let (y, m, _) = crate::meta::ymd(t);
if self.year.is_some_and(|yy| yy != y) || self.month.is_some_and(|mm| mm != m) {
return false;
}
}
if self.undated == Some(true) && f.when().is_some() {
return false;
}
if let Some(c) = &self.camera {
if f.camera.as_deref() != Some(c.as_str()) {
return false;
}
}
if let Some(mp) = self.min_mp {
match f.px {
Some((w, h)) if (w as f64 * h as f64) / 1e6 >= mp => {}
_ => return false,
}
}
if self.has_gps == Some(true) && !f.gps {
return false;
}
true
}
}
pub fn sort_items(items: &mut [&Found], sort: Option<&str>) {
match sort {
Some("name") => items.sort_by_key(|f| f.name.to_lowercase()),
Some("size") => items.sort_by_key(|f| std::cmp::Reverse(f.size)),
Some("date") => items.sort_by(|a, b| b.when().cmp(&a.when()).then(a.id.cmp(&b.id))),
Some("date_asc") => items.sort_by(|a, b| match (a.when(), b.when()) {
(Some(x), Some(y)) => x.cmp(&y).then(a.id.cmp(&b.id)),
(Some(_), None) => std::cmp::Ordering::Less,
(None, Some(_)) => std::cmp::Ordering::Greater,
_ => a.id.cmp(&b.id),
}),
Some("confidence") => {
items.sort_by(|a, b| b.confidence.cmp(&a.confidence).then(b.size.cmp(&a.size)))
}
_ => {}
}
}
pub fn item_json(f: &Found) -> Value {
json!({
"id": f.id, "name": f.name, "dir": f.dir, "ext": f.ext, "category": f.category,
"size": f.size, "origin": f.origin, "confidence": f.confidence,
"modified": f.modified, "created": f.created, "accessed": f.accessed,
"taken": f.taken, "when": f.when(), "camera": f.camera, "px": f.px, "gps": f.gps,
"note": f.note,
})
}
/// Contagens para a navegação lateral.
pub fn facets(found: &[Found]) -> Value {
let mut cats: HashMap<&'static str, usize> = HashMap::new();
let mut years: BTreeMap<i64, (usize, BTreeMap<u32, usize>)> = BTreeMap::new();
let mut cams: BTreeMap<String, usize> = BTreeMap::new();
let mut folders: BTreeMap<String, usize> = BTreeMap::new();
let (mut undated, mut carved, mut root, mut gps) = (0usize, 0usize, 0usize, 0usize);
let mut conf = [0usize; 3];
let mut seen_sig = std::collections::HashSet::new();
let mut dups = 0usize;
for f in found {
if let Some(g) = f.sig {
if !seen_sig.insert((g, f.size)) {
dups += 1;
}
}
*cats.entry(cat_key(f.category)).or_default() += 1;
conf[f.confidence as usize] += 1;
match f.when() {
Some(t) => {
let (y, m, _) = crate::meta::ymd(t);
let e = years.entry(y).or_default();
e.0 += 1;
*e.1.entry(m).or_default() += 1;
}
None => undated += 1,
}
if let Some(c) = &f.camera {
*cams.entry(c.clone()).or_default() += 1;
}
if f.gps {
gps += 1;
}
if f.origin == Origin::Carved {
carved += 1;
} else if let Some(d) = f.dir.as_deref().filter(|d| !d.is_empty()) {
let mut acc = String::new();
for part in d.split('/') {
if !acc.is_empty() {
acc.push('/');
}
acc.push_str(part);
*folders.entry(acc.clone()).or_default() += 1;
}
} else {
root += 1;
}
}
let cams_v: Vec<Value> = {
let mut v: Vec<_> = cams.into_iter().collect();
v.sort_by_key(|x| std::cmp::Reverse(x.1));
v.into_iter()
.map(|(n, c)| json!({ "name": n, "count": c }))
.collect()
};
json!({
"total": found.len(),
"categories": cats,
"confidence": { "low": conf[0], "medium": conf[1], "high": conf[2] },
"years": years.iter().rev().map(|(y, (n, ms))| json!({
"year": y, "count": n,
"months": ms.iter().rev().map(|(m, c)| json!({"month": m, "count": c})).collect::<Vec<_>>()
})).collect::<Vec<_>>(),
"undated": undated,
"cameras": cams_v,
"folders": folders.iter().map(|(p, c)| json!({
"path": p, "name": p.rsplit('/').next().unwrap_or(p), "depth": p.matches('/').count(), "count": c
})).collect::<Vec<_>>(),
"carved": carved,
"root": root,
"gps": gps,
"duplicates": dups,
})
}
pub fn cat_key(c: Category) -> &'static str {
match c {
Category::Image => "image",
Category::Document => "document",
Category::Video => "video",
Category::Audio => "audio",
Category::Archive => "archive",
Category::Database => "database",
Category::Other => "other",
}
}
// ------------------------------------------------------------------ cache de miniaturas
type Key = (String, u64, u32);
pub struct ThumbCache {
map: HashMap<Key, Arc<Vec<u8>>>,
order: VecDeque<Key>,
bytes: usize,
}
impl ThumbCache {
const CAP: usize = 128 * 1024 * 1024;
pub fn new() -> Self {
Self {
map: HashMap::new(),
order: VecDeque::new(),
bytes: 0,
}
}
pub fn get(&self, k: &Key) -> Option<Arc<Vec<u8>>> {
self.map.get(k).cloned()
}
pub fn put(&mut self, k: Key, v: Arc<Vec<u8>>) {
self.bytes += v.len();
if self.map.insert(k.clone(), v).is_none() {
self.order.push_back(k);
}
while self.bytes > Self::CAP {
let Some(old) = self.order.pop_front() else {
break;
};
if let Some(v) = self.map.remove(&old) {
self.bytes -= v.len();
}
}
}
}
impl Default for ThumbCache {
fn default() -> Self {
Self::new()
}
}
/// Limita quantas miniaturas são decodificadas ao mesmo tempo (memória/CPU).
pub struct Gate {
n: Mutex<usize>,
cv: Condvar,
max: usize,
}
pub struct Permit<'a>(&'a Gate);
impl Gate {
pub fn new(max: usize) -> Self {
Self {
n: Mutex::new(0),
cv: Condvar::new(),
max: max.max(1),
}
}
pub fn acquire(&self) -> Permit<'_> {
let mut n = self.n.lock().unwrap();
while *n >= self.max {
n = self.cv.wait(n).unwrap();
}
*n += 1;
Permit(self)
}
}
impl Drop for Permit<'_> {
fn drop(&mut self) {
*self.0.n.lock().unwrap() -= 1;
self.0.cv.notify_one();
}
}
/// Assinatura rápida: tamanho + FNV-1a do início (32 KiB) e do fim (32 KiB).
pub fn quick_sig(src: &dyn crate::source::Source, f: &Found) -> Option<u64> {
use crate::recover::read_found_at;
if f.size == 0 {
return None;
}
let mut h: u64 = 0xcbf29ce484222325 ^ f.size;
let mut feed = |off: u64| {
let mut buf = vec![0u8; 32 * 1024];
let mut got = 0;
while got < buf.len() {
match read_found_at(src, f, off + got as u64, &mut buf[got..]) {
Ok(0) | Err(_) => break,
Ok(n) => got += n,
}
}
for &b in &buf[..got] {
h ^= b as u64;
h = h.wrapping_mul(0x100000001b3);
}
got
};
let a = feed(0);
if a == 0 {
return None;
}
if f.size > 64 * 1024 {
feed(f.size - 32 * 1024);
}
Some(h)
}
crates/guara-core/src/fs/exfat.rs
+10
View File
@@ -174,6 +174,14 @@ fn walk(
let mdate = u16::from_le_bytes([e[14], e[15]]);
let mtime = u16::from_le_bytes([e[12], e[13]]);
let mod_unix = super::dos_to_unix(mdate, mtime);
let cre_unix = super::dos_to_unix(
u16::from_le_bytes([e[10], e[11]]),
u16::from_le_bytes([e[8], e[9]]),
);
let acc_unix = super::dos_to_unix(
u16::from_le_bytes([e[18], e[19]]),
u16::from_le_bytes([e[16], e[17]]),
);
let mut stream: Option<&[u8]> = None;
let mut name_u16: Vec<u8> = vec![];
for s in 1..=secondary.min(19) {
@@ -259,6 +267,8 @@ fn walk(
origin: Origin::Exfat,
data: DataLoc::Extents { extents: ex },
modified: mod_unix,
created: cre_unix,
accessed: acc_unix,
base: conf,
note,
},
crates/guara-core/src/fs/fat.rs
+17 -3
View File
@@ -237,6 +237,11 @@ impl Walk<'_, '_, '_, '_> {
u16::from_le_bytes([e[24], e[25]]),
u16::from_le_bytes([e[22], e[23]]),
);
let ctime = super::dos_to_unix(
u16::from_le_bytes([e[16], e[17]]),
u16::from_le_bytes([e[14], e[15]]),
);
let atime = super::dos_to_unix(u16::from_le_bytes([e[18], e[19]]), 0);
self.count += 1;
if self.count % 512 == 0 {
(self.ctx.progress)(self.count, 0);
@@ -277,7 +282,7 @@ impl Walk<'_, '_, '_, '_> {
if size == 0 || !self.fat.valid(start) {
continue;
}
self.deleted_file(&name, path, start, size, mtime);
self.deleted_file(&name, path, start, size, (ctime, mtime, atime));
}
}
@@ -291,7 +296,14 @@ impl Walk<'_, '_, '_, '_> {
v
}
fn deleted_file(&mut self, name: &str, dir: &str, start: u32, size: u64, mtime: Option<i64>) {
fn deleted_file(
&mut self,
name: &str,
dir: &str,
start: u32,
size: u64,
times: (Option<i64>, Option<i64>, Option<i64>),
) {
let fat = self.fat;
let cb = fat.cluster_bytes();
let need = size.div_ceil(cb);
@@ -336,7 +348,9 @@ impl Walk<'_, '_, '_, '_> {
data: DataLoc::Extents {
extents: trim_extents(extents, size),
},
modified: mtime,
modified: times.1,
created: times.0,
accessed: times.2,
base: conf,
note,
};
crates/guara-core/src/fs/mod.rs
+9
View File
@@ -47,6 +47,8 @@ pub struct NewFound<'a> {
pub origin: Origin,
pub data: DataLoc,
pub modified: Option<i64>,
pub created: Option<i64>,
pub accessed: Option<i64>,
pub base: Confidence,
pub note: Option<&'a str>,
}
@@ -98,6 +100,13 @@ pub fn make_found(ctx: &Ctx, n: NewFound) -> Found {
confidence: conf,
data: n.data,
modified: n.modified,
created: n.created,
accessed: n.accessed,
taken: None,
camera: None,
px: None,
gps: false,
sig: None,
note,
}
}
crates/guara-core/src/fs/ntfs.rs
+10
View File
@@ -22,6 +22,8 @@ struct Rec {
parent: u64,
name: Option<(String, u8)>,
mtime: Option<i64>,
ctime: Option<i64>,
atime: Option<i64>,
data: Option<Data>,
has_attr_list: bool,
}
@@ -199,6 +201,8 @@ pub fn scan(ctx: &mut Ctx, base: u64) -> VolumeScan {
origin: Origin::Ntfs,
data,
modified: rec.mtime,
created: rec.ctime,
accessed: rec.atime,
base: conf,
note: if rec.has_attr_list {
Some("registro com lista de atributos: pode estar incompleto")
@@ -320,6 +324,8 @@ fn parse_record(r: &mut [u8], bps: u64, _base: u64, _cl: u64) -> Option<Rec> {
parent: 0,
name: None,
mtime: None,
ctime: None,
atime: None,
data: None,
has_attr_list: false,
};
@@ -341,7 +347,11 @@ fn parse_record(r: &mut [u8], bps: u64, _base: u64, _cl: u64) -> Option<Rec> {
0x10 if !nonres => {
let vo = le16(r, p + 20) as usize;
if p + vo + 16 <= r.len() {
rec.ctime = filetime_to_unix(le64(r, p + vo));
rec.mtime = filetime_to_unix(le64(r, p + vo + 8));
if p + vo + 32 <= r.len() {
rec.atime = filetime_to_unix(le64(r, p + vo + 24));
}
}
}
0x20 => rec.has_attr_list = true,
crates/guara-core/src/hashing.rs
+45
View File
@@ -0,0 +1,45 @@
//! Hashes de integridade usados em perícia: MD5, SHA-1 e SHA-256 numa única passada.
use md5::Md5;
use serde::{Deserialize, Serialize};
use sha1::Sha1;
use sha2::{Digest, Sha256};
#[derive(Debug, Clone, Default, PartialEq, Eq, Serialize, Deserialize)]
pub struct Hashes {
pub md5: String,
pub sha1: String,
pub sha256: String,
}
#[derive(Default)]
pub struct MultiHasher {
md5: Md5,
sha1: Sha1,
sha256: Sha256,
}
impl MultiHasher {
pub fn new() -> Self {
Self::default()
}
pub fn update(&mut self, data: &[u8]) {
self.md5.update(data);
self.sha1.update(data);
self.sha256.update(data);
}
pub fn finalize(self) -> Hashes {
let hex = |b: &[u8]| b.iter().map(|x| format!("{x:02x}")).collect::<String>();
Hashes {
md5: hex(&self.md5.finalize()),
sha1: hex(&self.sha1.finalize()),
sha256: hex(&self.sha256.finalize()),
}
}
}
pub fn hash_bytes(data: &[u8]) -> Hashes {
let mut h = MultiHasher::new();
h.update(data);
h.finalize()
}
crates/guara-core/src/image.rs
+10 -7
View File
@@ -2,9 +2,9 @@
//! erros e hash. Serve para trabalhar sempre sobre a cópia.
use crate::error::{Error, Result};
use crate::hashing::MultiHasher;
use crate::source::Source;
use serde::{Deserialize, Serialize};
use sha2::{Digest, Sha256};
use std::io::Write;
use std::path::Path;
@@ -12,6 +12,10 @@ use std::path::Path;
pub struct ImageReport {
pub bytes: u64,
pub sha256: String,
#[serde(default)]
pub md5: String,
#[serde(default)]
pub sha1: String,
/// (offset, tamanho) dos trechos ilegíveis, preenchidos com zeros.
pub bad_ranges: Vec<(u64, u64)>,
pub cancelled: bool,
@@ -38,7 +42,7 @@ pub fn create_image(
})?;
let size = src.size();
let mut rep = ImageReport::default();
let mut hasher = Sha256::new();
let mut hasher = MultiHasher::new();
let mut buf = vec![0u8; BLOCK];
let mut pos = 0u64;
while pos < size {
@@ -72,11 +76,10 @@ pub fn create_image(
}
out.flush()?;
rep.bytes = pos;
rep.sha256 = hasher
.finalize()
.iter()
.map(|b| format!("{b:02x}"))
.collect();
let h = hasher.finalize();
rep.sha256 = h.sha256;
rep.md5 = h.md5;
rep.sha1 = h.sha1;
let map = dest.with_extension("map.json");
std::fs::write(map, serde_json::to_vec_pretty(&rep)?)?;
Ok(rep)
crates/guara-core/src/lib.rs
+7 -1
View File
@@ -1,4 +1,4 @@
//! Guará — núcleo de recuperação de arquivos.
//! Guará: núcleo de recuperação de arquivos.
//!
//! Princípio central: a origem é aberta **somente para leitura**; nada neste
//! crate escreve no dispositivo que está sendo recuperado.
@@ -6,15 +6,21 @@
pub mod carve;
pub mod device;
pub mod error;
pub mod explorer;
pub mod fs;
pub mod hashing;
pub mod image;
pub mod meta;
pub mod partition;
pub mod recover;
pub mod report;
pub mod scan;
pub mod search;
pub mod service;
pub mod settings;
pub mod signatures;
pub mod source;
pub mod thumb;
pub mod types;
pub use error::{Error, Result};
crates/guara-core/src/meta.rs
+952
View File
@@ -0,0 +1,952 @@
//! Extração de metadados de arquivos recuperados: EXIF (câmera, datas, GPS),
//! informações de PDF, propriedades de documentos Office/OpenDocument e
//! dados de vídeo MP4/MOV. Tudo é leitura posicional, sem gravar em disco.
use crate::recover::read_found_at;
use crate::source::Source;
use crate::types::Found;
use serde::{Deserialize, Serialize};
use std::io::Read;
#[derive(Debug, Clone, Default, Serialize, Deserialize)]
pub struct Meta {
pub kind: String,
pub fields: Vec<(String, String)>,
/// (latitude, longitude) em graus decimais.
pub gps: Option<(f64, f64)>,
/// Data em que a foto foi tirada (EXIF), s desde 1970 (hora da câmera lida como UTC).
#[serde(default)]
pub taken: Option<i64>,
#[serde(default)]
pub camera: Option<String>,
#[serde(default)]
pub px: Option<(u32, u32)>,
/// Orientação EXIF (1..8).
#[serde(default)]
pub orientation: u16,
}
impl Meta {
fn add(&mut self, k: &str, v: impl Into<String>) {
let v = v.into();
let v = v.trim().to_string();
if !v.is_empty() {
self.fields.push((k.to_string(), v));
}
}
}
struct Fr<'a> {
src: &'a dyn Source,
f: &'a Found,
}
impl Fr<'_> {
fn read(&self, off: u64, len: usize) -> Vec<u8> {
let mut out = vec![0u8; len.min(self.f.size.saturating_sub(off) as usize)];
let mut got = 0;
while got < out.len() {
match read_found_at(self.src, self.f, off + got as u64, &mut out[got..]) {
Ok(0) | Err(_) => break,
Ok(n) => got += n,
}
}
out.truncate(got);
out
}
}
pub fn extract(src: &dyn Source, f: &Found) -> Meta {
let fr = Fr { src, f };
let mut m = Meta::default();
match f.ext.to_ascii_lowercase().as_str() {
"jpg" | "jpeg" => {
m.kind = "Imagem JPEG".into();
jpeg(&fr, &mut m);
}
"png" => {
m.kind = "Imagem PNG".into();
let h = fr.read(0, 32);
if h.len() >= 24 && &h[12..16] == b"IHDR" {
m.px = Some((be32(&h[16..]), be32(&h[20..])));
m.add(
"Dimensões",
format!("{} × {} px", be32(&h[16..]), be32(&h[20..])),
);
}
}
"gif" => {
m.kind = "Imagem GIF".into();
let h = fr.read(0, 16);
if h.len() >= 10 {
m.px = Some((le16(&h[6..]) as u32, le16(&h[8..]) as u32));
m.add(
"Dimensões",
format!("{} × {} px", le16(&h[6..]), le16(&h[8..])),
);
}
}
"bmp" => {
m.kind = "Imagem BMP".into();
let h = fr.read(0, 32);
if h.len() >= 26 {
m.px = Some((le32(&h[18..]), (le32(&h[22..]) as i32).unsigned_abs()));
m.add(
"Dimensões",
format!(
"{} × {} px",
le32(&h[18..]),
(le32(&h[22..]) as i32).unsigned_abs()
),
);
}
}
"pdf" => {
m.kind = "Documento PDF".into();
pdf(&fr, &mut m);
}
"docx" | "xlsx" | "pptx" => {
m.kind = "Documento Office".into();
ooxml(&fr, &mut m);
}
"odt" | "ods" | "odp" => {
m.kind = "Documento OpenDocument".into();
odf(&fr, &mut m);
}
"mp4" | "mov" | "m4v" | "m4a" | "3gp" => {
m.kind = "Mídia MP4/MOV".into();
mp4(&fr, &mut m);
}
_ => {}
}
m
}
fn be16(b: &[u8]) -> u16 {
u16::from_be_bytes([b[0], b[1]])
}
fn be32(b: &[u8]) -> u32 {
u32::from_be_bytes([b[0], b[1], b[2], b[3]])
}
fn le16(b: &[u8]) -> u16 {
u16::from_le_bytes([b[0], b[1]])
}
fn le32(b: &[u8]) -> u32 {
u32::from_le_bytes([b[0], b[1], b[2], b[3]])
}
// ------------------------------------------------------------------ EXIF
fn jpeg(fr: &Fr, m: &mut Meta) {
let d = fr.read(0, 256 * 1024);
jpeg_bytes(&d, m);
}
/// Lê os segmentos de cabeçalho de um JPEG (já em memória): dimensões e EXIF.
pub fn jpeg_bytes(d: &[u8], m: &mut Meta) {
let mut p = 2usize;
while p + 4 <= d.len() && d[p] == 0xFF {
let mk = d[p + 1];
if mk == 0xDA || mk == 0xD9 {
break;
}
let len = be16(&d[p + 2..]) as usize;
if len < 2 || p + 2 + len > d.len() {
break;
}
if mk == 0xE1 && d[p + 4..].starts_with(b"Exif\0\0") {
exif(&d[p + 10..p + 2 + len], m);
}
if (mk == 0xC0 || mk == 0xC2) && p + 9 <= d.len() {
m.px = Some((be16(&d[p + 7..]) as u32, be16(&d[p + 5..]) as u32));
m.add(
"Dimensões",
format!("{} × {} px", be16(&d[p + 7..]), be16(&d[p + 5..])),
);
}
p += 2 + len;
}
}
/// Miniatura JPEG embutida no EXIF (IFD1), útil quando a imagem principal está danificada.
pub fn exif_thumbnail(d: &[u8]) -> Option<Vec<u8>> {
let mut p = 2usize;
while p + 4 <= d.len() && d[p] == 0xFF {
let mk = d[p + 1];
if mk == 0xDA || mk == 0xD9 {
return None;
}
let len = be16(&d[p + 2..]) as usize;
if len < 2 || p + 2 + len > d.len() {
return None;
}
if mk == 0xE1 && d[p + 4..].starts_with(b"Exif\0\0") {
let tiff = &d[p + 10..p + 2 + len];
if tiff.len() < 8 {
return None;
}
let le = match &tiff[0..2] {
b"II" => true,
b"MM" => false,
_ => return None,
};
let t = Tiff { d: tiff, le };
let ifd0 = t.u32(4)? as usize;
let n = t.u16(ifd0)? as usize;
let ifd1 = t.u32(ifd0 + 2 + n * 12)? as usize;
if ifd1 == 0 {
return None;
}
let (mut off, mut length) = (None, None);
for e in t.ifd(ifd1) {
match e.tag {
0x0201 => off = t.u32(e.pos),
0x0202 => length = t.u32(e.pos),
_ => {}
}
}
let (o, l) = (off? as usize, length? as usize);
return tiff.get(o..o + l).map(|b| b.to_vec());
}
p += 2 + len;
}
None
}
struct Tiff<'a> {
d: &'a [u8],
le: bool,
}
struct Entry {
tag: u16,
typ: u16,
count: u32,
pos: usize, // onde estão os dados do valor dentro do TIFF
}
impl Tiff<'_> {
fn u16(&self, o: usize) -> Option<u16> {
let b = self.d.get(o..o + 2)?;
Some(if self.le {
u16::from_le_bytes([b[0], b[1]])
} else {
u16::from_be_bytes([b[0], b[1]])
})
}
fn u32(&self, o: usize) -> Option<u32> {
let b = self.d.get(o..o + 4)?;
Some(if self.le {
u32::from_le_bytes([b[0], b[1], b[2], b[3]])
} else {
u32::from_be_bytes([b[0], b[1], b[2], b[3]])
})
}
fn ifd(&self, off: usize) -> Vec<Entry> {
let mut v = vec![];
let Some(n) = self.u16(off) else { return v };
for i in 0..(n as usize).min(200) {
let e = off + 2 + i * 12;
let (Some(tag), Some(typ), Some(count)) =
(self.u16(e), self.u16(e + 2), self.u32(e + 4))
else {
break;
};
let size = match typ {
1 | 2 | 6 | 7 => 1,
3 | 8 => 2,
4 | 9 | 11 => 4,
5 | 10 | 12 => 8,
_ => 1,
} * count as usize;
let pos = if size <= 4 {
e + 8
} else {
self.u32(e + 8).unwrap_or(0) as usize
};
v.push(Entry {
tag,
typ,
count,
pos,
});
}
v
}
fn ascii(&self, e: &Entry) -> String {
let end = (e.pos + e.count as usize).min(self.d.len());
let b = self.d.get(e.pos..end).unwrap_or(&[]);
String::from_utf8_lossy(b)
.trim_end_matches('\0')
.trim()
.to_string()
}
fn rational(&self, e: &Entry, i: usize) -> Option<f64> {
let o = e.pos + i * 8;
let (n, d) = (self.u32(o)?, self.u32(o + 4)?);
(d != 0).then(|| n as f64 / d as f64)
}
fn num(&self, e: &Entry) -> Option<u32> {
match e.typ {
3 => self.u16(e.pos).map(u32::from),
4 => self.u32(e.pos),
_ => None,
}
}
}
fn exif(d: &[u8], m: &mut Meta) {
if d.len() < 8 {
return;
}
let le = match &d[0..2] {
b"II" => true,
b"MM" => false,
_ => return,
};
let t = Tiff { d, le };
let Some(ifd0) = t.u32(4) else { return };
let mut exif_ptr = None;
let mut gps_ptr = None;
let (mut make, mut model) = (String::new(), String::new());
for e in t.ifd(ifd0 as usize) {
match e.tag {
0x010F => {
make = t.ascii(&e);
m.add("Fabricante da câmera", make.clone());
}
0x0110 => {
model = t.ascii(&e);
m.add("Modelo da câmera", model.clone());
}
0x0112 => m.orientation = t.num(&e).unwrap_or(1) as u16,
0x0131 => m.add("Software", t.ascii(&e)),
0x0132 => m.add("Data (EXIF)", exif_date(&t.ascii(&e))),
0x013B => m.add("Autor", t.ascii(&e)),
0x8298 => m.add("Copyright", t.ascii(&e)),
0x8769 => exif_ptr = t.u32(e.pos),
0x8825 => gps_ptr = t.u32(e.pos),
_ => {}
}
}
let first = make.split_whitespace().next().unwrap_or("").to_lowercase();
let cam = if model.is_empty() {
make.clone()
} else if make.is_empty() || (!first.is_empty() && model.to_lowercase().starts_with(&first)) {
model.clone()
} else {
format!("{make} {model}")
};
if !cam.trim().is_empty() {
m.camera = Some(cam.trim().to_string());
}
if let Some(o) = exif_ptr {
let (mut w, mut h) = (None, None);
for e in t.ifd(o as usize) {
match e.tag {
0x9003 => {
m.taken = exif_epoch(&t.ascii(&e));
m.add("Foto tirada em", exif_date(&t.ascii(&e)));
}
0x9004 => m.add("Digitalizada em", exif_date(&t.ascii(&e))),
0xA434 => m.add("Lente", t.ascii(&e)),
0xA431 => m.add("Nº de série do corpo", t.ascii(&e)),
0xA430 => m.add("Proprietário da câmera", t.ascii(&e)),
0xA002 => w = t.num(&e),
0xA003 => h = t.num(&e),
0x829A => {
if let Some(v) = t.rational(&e, 0) {
m.add(
"Exposição",
if v > 0.0 && v < 1.0 {
format!("1/{:.0} s", 1.0 / v)
} else {
format!("{v:.1} s")
},
);
}
}
0x829D => {
if let Some(v) = t.rational(&e, 0) {
m.add("Abertura", format!("f/{v:.1}"));
}
}
0x8827 => {
if let Some(v) = t.num(&e) {
m.add("ISO", v.to_string());
}
}
0x920A => {
if let Some(v) = t.rational(&e, 0) {
m.add("Distância focal", format!("{v:.0} mm"));
}
}
_ => {}
}
}
if let (Some(w), Some(h)) = (w, h) {
if m.px.is_none() {
m.px = Some((w, h));
}
if !m.fields.iter().any(|(k, _)| k == "Dimensões") {
m.add("Dimensões", format!("{w} × {h} px"));
}
}
}
if let Some(o) = gps_ptr {
let (mut lat, mut lon, mut lat_ref, mut lon_ref, mut alt) = (None, None, 'N', 'E', None);
for e in t.ifd(o as usize) {
let dms = |e: &Entry| -> Option<f64> {
Some(t.rational(e, 0)? + t.rational(e, 1)? / 60.0 + t.rational(e, 2)? / 3600.0)
};
match e.tag {
0x0001 => lat_ref = t.ascii(&e).chars().next().unwrap_or('N'),
0x0002 => lat = dms(&e),
0x0003 => lon_ref = t.ascii(&e).chars().next().unwrap_or('E'),
0x0004 => lon = dms(&e),
0x0006 => alt = t.rational(&e, 0),
0x001D => m.add("Data do GPS", t.ascii(&e).replace(':', "-")),
_ => {}
}
}
if let (Some(a), Some(b)) = (lat, lon) {
let a = if lat_ref == 'S' { -a } else { a };
let b = if lon_ref == 'W' { -b } else { b };
if a.abs() <= 90.0 && b.abs() <= 180.0 && !(a == 0.0 && b == 0.0) {
m.gps = Some((a, b));
m.add("Localização (GPS)", format!("{a:.6}, {b:.6}"));
if let Some(h) = alt {
m.add("Altitude", format!("{h:.0} m"));
}
}
}
}
}
/// "2023:07:14 18:22:05" -> segundos desde 1970 (hora da câmera lida como UTC)
fn exif_epoch(s: &str) -> Option<i64> {
let b = s.as_bytes();
if b.len() < 19 || b[4] != b':' || b[7] != b':' {
return None;
}
let n = |a: usize, c: usize| s.get(a..c)?.parse::<i64>().ok();
let (y, mo, d, h, mi, se) = (
n(0, 4)?,
n(5, 7)?,
n(8, 10)?,
n(11, 13)?,
n(14, 16)?,
n(17, 19)?,
);
if !(1..=12).contains(&mo) || !(1..=31).contains(&d) || y < 1970 {
return None;
}
Some(crate::fs::days_from_civil(y, mo, d) * 86400 + h * 3600 + mi * 60 + se)
}
/// "2023:07:14 18:22:05" -> "14/07/2023 18:22:05"
fn exif_date(s: &str) -> String {
let b = s.as_bytes();
if b.len() >= 19 && b[4] == b':' && b[7] == b':' {
format!("{}/{}/{} {}", &s[8..10], &s[5..7], &s[0..4], &s[11..19])
} else {
s.to_string()
}
}
// ------------------------------------------------------------------ PDF
fn pdf(fr: &Fr, m: &mut Meta) {
let size = fr.f.size;
let head = fr.read(0, 96 * 1024);
let tail = if size > 192 * 1024 {
fr.read(size - 96 * 1024, 96 * 1024)
} else {
Vec::new()
};
let mut all = head.clone();
all.extend_from_slice(&tail);
if head.starts_with(b"%PDF-") {
m.add(
"Versão do PDF",
String::from_utf8_lossy(&head[5..8.min(head.len())]).to_string(),
);
}
for (key, label) in [
("/Title", "Título"),
("/Author", "Autor"),
("/Subject", "Assunto"),
("/Creator", "Criado com"),
("/Producer", "Gerado por"),
] {
if let Some(v) = pdf_string(&all, key) {
m.add(label, v);
}
}
for (key, label) in [
("/CreationDate", "Criado em"),
("/ModDate", "Modificado em"),
] {
if let Some(v) = pdf_string(&all, key) {
m.add(label, pdf_date(&v));
}
}
let pages = find_all(&all, b"/Count ").into_iter().filter_map(|i| {
let s: String = all[i + 7..]
.iter()
.take(8)
.map(|&c| c as char)
.take_while(|c| c.is_ascii_digit())
.collect();
s.parse::<u32>().ok()
});
if let Some(n) = pages.max() {
m.add("Páginas (aprox.)", n.to_string());
}
if !find_all(&all, b"/Encrypt").is_empty() {
m.add("Criptografado", "sim");
}
}
fn find_all(h: &[u8], n: &[u8]) -> Vec<usize> {
let mut v = vec![];
let mut i = 0;
while i + n.len() <= h.len() {
if &h[i..i + n.len()] == n {
v.push(i);
i += n.len();
} else {
i += 1;
}
}
v
}
fn pdf_string(d: &[u8], key: &str) -> Option<String> {
let k = key.as_bytes();
for at in find_all(d, k) {
let mut p = at + k.len();
// evita /Creator casar com /CreatorXyz
if d.get(p).map(|c| c.is_ascii_alphanumeric()).unwrap_or(false) {
continue;
}
while d.get(p).map(|c| c.is_ascii_whitespace()).unwrap_or(false) {
p += 1;
}
match d.get(p)? {
b'(' => {
let mut depth = 0;
let mut out = vec![];
let mut i = p;
while i < d.len() && out.len() < 400 {
let c = d[i];
match c {
b'\\' if i + 1 < d.len() => {
i += 1;
if d[i].is_ascii_digit() && d[i] < b'8' {
// escape octal \ddd (1 a 3 dígitos)
let mut v = 0u32;
let mut n = 0;
while n < 3 && i < d.len() && d[i] >= b'0' && d[i] < b'8' {
v = v * 8 + (d[i] - b'0') as u32;
i += 1;
n += 1;
}
out.push((v & 0xFF) as u8);
continue;
}
out.push(match d[i] {
b'n' => b'\n',
b'r' => b'\r',
b't' => b'\t',
o => o,
});
}
b'(' => {
depth += 1;
if depth > 1 {
out.push(c);
}
}
b')' => {
depth -= 1;
if depth == 0 {
break;
}
out.push(c);
}
_ => out.push(c),
}
i += 1;
}
return Some(pdf_text(&out));
}
b'<' if d.get(p + 1) != Some(&b'<') => {
let hex: Vec<u8> = d[p + 1..]
.iter()
.take_while(|&&c| c != b'>')
.copied()
.filter(|c| c.is_ascii_hexdigit())
.collect();
let bytes: Vec<u8> = hex
.chunks(2)
.filter(|c| c.len() == 2)
.map(|c| u8::from_str_radix(std::str::from_utf8(c).unwrap(), 16).unwrap_or(0))
.collect();
return Some(pdf_text(&bytes));
}
_ => {}
}
}
None
}
fn pdf_text(b: &[u8]) -> String {
if b.len() >= 2 && b[0] == 0xFE && b[1] == 0xFF {
let u: Vec<u16> = b[2..]
.chunks_exact(2)
.map(|c| u16::from_be_bytes([c[0], c[1]]))
.collect();
return String::from_utf16_lossy(&u);
}
// PDFDocEncoding ~ Latin-1 na prática
b.iter().map(|&c| c as char).collect()
}
/// "D:20230714182205-03'00'" -> "14/07/2023 18:22:05"
fn pdf_date(s: &str) -> String {
let t = s.trim_start_matches("D:");
let digits: String = t.chars().take_while(|c| c.is_ascii_digit()).collect();
if digits.len() >= 8 {
let g = |a: usize, b: usize| digits.get(a..b).unwrap_or("00");
return format!(
"{}/{}/{} {}:{}:{}",
g(6, 8),
g(4, 6),
g(0, 4),
g(8, 10),
g(10, 12),
g(12, 14)
);
}
s.to_string()
}
// ------------------------------------------------------------------ ZIP (Office/ODF)
struct ZipEnt {
name: String,
method: u16,
csize: u64,
usize: u64,
off: u64,
}
fn zip_dir(fr: &Fr) -> Vec<ZipEnt> {
let size = fr.f.size;
let tail_len = size.min(66_000);
let tail = fr.read(size - tail_len, tail_len as usize);
let Some(e) = (0..tail.len().saturating_sub(21))
.rev()
.find(|&i| tail[i..].starts_with(b"PK\x05\x06"))
else {
return vec![];
};
let n = le16(&tail[e + 10..]) as usize;
let cd_size = le32(&tail[e + 12..]) as usize;
let cd_off = le32(&tail[e + 16..]) as u64;
if cd_size > 8 << 20 || cd_off + cd_size as u64 > size {
return vec![];
}
let cd = fr.read(cd_off, cd_size);
let mut out = vec![];
let mut p = 0;
while out.len() < n.min(5000) && p + 46 <= cd.len() && &cd[p..p + 4] == b"PK\x01\x02" {
let nl = le16(&cd[p + 28..]) as usize;
let xl = le16(&cd[p + 30..]) as usize;
let cl = le16(&cd[p + 32..]) as usize;
if p + 46 + nl > cd.len() {
break;
}
out.push(ZipEnt {
name: String::from_utf8_lossy(&cd[p + 46..p + 46 + nl]).to_string(),
method: le16(&cd[p + 10..]),
csize: le32(&cd[p + 20..]) as u64,
usize: le32(&cd[p + 24..]) as u64,
off: le32(&cd[p + 42..]) as u64,
});
p += 46 + nl + xl + cl;
}
out
}
fn zip_read(fr: &Fr, e: &ZipEnt) -> Option<Vec<u8>> {
if e.usize > 4 << 20 {
return None;
}
let lh = fr.read(e.off, 30);
if lh.len() < 30 || &lh[0..4] != b"PK\x03\x04" {
return None;
}
let start = e.off + 30 + le16(&lh[26..]) as u64 + le16(&lh[28..]) as u64;
let data = fr.read(start, e.csize as usize);
match e.method {
0 => Some(data),
8 => {
let mut out = Vec::new();
flate2::read::DeflateDecoder::new(&data[..])
.take(4 << 20)
.read_to_end(&mut out)
.ok()?;
Some(out)
}
_ => None,
}
}
/// Texto do primeiro elemento `<tag ...>texto</tag>`.
fn xml_tag(xml: &str, tag: &str) -> Option<String> {
let open = format!("<{tag}");
let mut from = 0;
while let Some(i) = xml[from..].find(&open) {
let s = from + i + open.len();
match xml[s..].chars().next() {
Some('>') | Some(' ') | Some('/') => {
let gt = xml[s..].find('>')? + s;
if xml[..gt].ends_with('/') {
return None;
}
let end = xml[gt + 1..].find(&format!("</{tag}>"))? + gt + 1;
return Some(unescape(&xml[gt + 1..end]));
}
_ => from = s,
}
}
None
}
fn unescape(s: &str) -> String {
s.replace("&lt;", "<")
.replace("&gt;", ">")
.replace("&quot;", "\"")
.replace("&apos;", "'")
.replace("&amp;", "&")
}
/// "2023-07-14T18:22:05Z" -> "14/07/2023 18:22:05 (UTC)"
fn iso_date(s: &str) -> String {
let b = s.as_bytes();
if b.len() >= 19 && b[4] == b'-' && b[10] == b'T' {
let utc = if s.ends_with('Z') { " (UTC)" } else { "" };
return format!(
"{}/{}/{} {}{}",
&s[8..10],
&s[5..7],
&s[0..4],
&s[11..19],
utc
);
}
s.to_string()
}
fn ooxml(fr: &Fr, m: &mut Meta) {
let ents = zip_dir(fr);
let get = |n: &str| {
ents.iter()
.find(|e| e.name == n)
.and_then(|e| zip_read(fr, e))
.map(|b| String::from_utf8_lossy(&b).to_string())
};
if let Some(x) = get("docProps/core.xml") {
for (tag, label) in [
("dc:title", "Título"),
("dc:subject", "Assunto"),
("dc:creator", "Autor"),
("cp:lastModifiedBy", "Última modificação por"),
("cp:keywords", "Palavras-chave"),
] {
if let Some(v) = xml_tag(&x, tag) {
m.add(label, v);
}
}
for (tag, label) in [
("dcterms:created", "Criado em"),
("dcterms:modified", "Modificado em"),
] {
if let Some(v) = xml_tag(&x, tag) {
m.add(label, iso_date(&v));
}
}
}
if let Some(x) = get("docProps/app.xml") {
for (tag, label) in [
("Application", "Aplicativo"),
("Company", "Empresa"),
("Pages", "Páginas"),
("Words", "Palavras"),
] {
if let Some(v) = xml_tag(&x, tag) {
m.add(label, v);
}
}
}
}
fn odf(fr: &Fr, m: &mut Meta) {
let ents = zip_dir(fr);
let Some(x) = ents
.iter()
.find(|e| e.name == "meta.xml")
.and_then(|e| zip_read(fr, e))
.map(|b| String::from_utf8_lossy(&b).to_string())
else {
return;
};
for (tag, label) in [
("dc:title", "Título"),
("meta:initial-creator", "Criado por"),
("dc:creator", "Última modificação por"),
("meta:generator", "Aplicativo"),
] {
if let Some(v) = xml_tag(&x, tag) {
m.add(label, v);
}
}
for (tag, label) in [
("meta:creation-date", "Criado em"),
("dc:date", "Modificado em"),
] {
if let Some(v) = xml_tag(&x, tag) {
m.add(label, iso_date(&v));
}
}
}
// ------------------------------------------------------------------ MP4 / MOV
fn mp4(fr: &Fr, m: &mut Meta) {
let size = fr.f.size;
let mut p = 0u64;
for _ in 0..64 {
let h = fr.read(p, 16);
if h.len() < 8 {
return;
}
let mut sz = be32(&h) as u64;
if sz == 1 && h.len() >= 16 {
sz = u64::from_be_bytes(h[8..16].try_into().unwrap());
} else if sz == 0 {
sz = size - p;
}
if sz < 8 {
return;
}
if &h[4..8] == b"moov" {
let body = fr.read(p + 8, (sz - 8).min(4 << 20) as usize);
moov(&body, m);
return;
}
p += sz;
if p >= size {
return;
}
}
}
fn boxes(d: &[u8]) -> Vec<(&[u8], &[u8])> {
let mut v = vec![];
let mut p = 0;
while p + 8 <= d.len() {
let sz = be32(&d[p..]) as usize;
if sz < 8 || p + sz > d.len() {
break;
}
v.push((&d[p + 4..p + 8], &d[p + 8..p + sz]));
p += sz;
}
v
}
fn moov(d: &[u8], m: &mut Meta) {
const EPOCH_1904: u64 = 2_082_844_800;
for (ty, body) in boxes(d) {
match ty {
b"mvhd" if body.len() >= 24 => {
let v = body[0];
let (created, ts, dur) = if v == 1 && body.len() >= 32 {
(
u64::from_be_bytes(body[4..12].try_into().unwrap()),
be32(&body[20..]) as u64,
u64::from_be_bytes(body[24..32].try_into().unwrap()),
)
} else {
(
be32(&body[4..]) as u64,
be32(&body[12..]) as u64,
be32(&body[16..]) as u64,
)
};
if created > EPOCH_1904 {
m.add(
"Gravado em (metadado)",
fmt_unix((created - EPOCH_1904) as i64),
);
}
if ts > 0 && dur > 0 {
let s = dur / ts;
m.add(
"Duração",
format!("{}:{:02}:{:02}", s / 3600, s % 3600 / 60, s % 60),
);
}
}
b"trak" => {
for (t2, b2) in boxes(body) {
if t2 == b"tkhd" && b2.len() >= 84 {
let off = if b2[0] == 1 { 88 } else { 76 };
if b2.len() >= off + 8 {
let (w, h) = (be32(&b2[off..]) >> 16, be32(&b2[off + 4..]) >> 16);
if w > 0 && h > 0 && !m.fields.iter().any(|(k, _)| k == "Dimensões") {
m.add("Dimensões", format!("{w} × {h} px"));
}
}
}
}
}
_ => {}
}
}
}
pub fn fmt_unix(t: i64) -> String {
let days = t.div_euclid(86400);
let rem = t.rem_euclid(86400);
let (y, mo, d) = civil_from_days(days);
format!(
"{d:02}/{mo:02}/{y} {:02}:{:02}:{:02} (UTC)",
rem / 3600,
rem % 3600 / 60,
rem % 60
)
}
/// (ano, mês, dia) de um instante em segundos desde 1970.
pub fn ymd(t: i64) -> (i64, u32, u32) {
let (y, m, d) = civil_from_days(t.div_euclid(86400));
(y, m as u32, d as u32)
}
/// Inverso de `days_from_civil` (Howard Hinnant).
fn civil_from_days(z: i64) -> (i64, i64, i64) {
let z = z + 719468;
let era = if z >= 0 { z } else { z - 146096 } / 146097;
let doe = z - era * 146097;
let yoe = (doe - doe / 1460 + doe / 36524 - doe / 146096) / 365;
let y = yoe + era * 400;
let doy = doe - (365 * yoe + yoe / 4 - yoe / 100);
let mp = (5 * doy + 2) / 153;
let d = doy - (153 * mp + 2) / 5 + 1;
let m = if mp < 10 { mp + 3 } else { mp - 9 };
(if m <= 2 { y + 1 } else { y }, m, d)
}
crates/guara-core/src/recover.rs
+13 -11
View File
@@ -3,10 +3,10 @@
use crate::error::{Error, Result};
use crate::fs::sanitize_name;
use crate::hashing::{Hashes, MultiHasher};
use crate::source::{read_vec, Source};
use crate::types::*;
use serde::{Deserialize, Serialize};
use sha2::{Digest, Sha256};
use std::collections::HashSet;
use std::io::Write;
use std::path::{Path, PathBuf};
@@ -17,6 +17,10 @@ pub struct RecoveredFile {
pub path: String,
pub size: u64,
pub sha256: String,
#[serde(default)]
pub md5: String,
#[serde(default)]
pub sha1: String,
pub confidence: Confidence,
pub origin: Origin,
pub partial: bool,
@@ -111,7 +115,7 @@ pub fn recover(
}
used.insert(path.clone());
match write_file(src, f, &path) {
Ok(hash) => {
Ok(h) => {
report.total_bytes += f.size;
report.recovered.push(RecoveredFile {
id: f.id,
@@ -121,7 +125,9 @@ pub fn recover(
.to_string_lossy()
.replace('\\', "/"),
size: f.size,
sha256: hash,
sha256: h.sha256,
md5: h.md5,
sha1: h.sha1,
confidence: f.confidence,
origin: f.origin,
partial: f.confidence == Confidence::Low,
@@ -135,7 +141,7 @@ pub fn recover(
"ferramenta": "Guará",
"versao": env!("CARGO_PKG_VERSION"),
"gerado_em": std::time::SystemTime::now().duration_since(std::time::UNIX_EPOCH).map(|d| d.as_secs()).unwrap_or(0),
"algoritmo_hash": "SHA-256",
"algoritmos_hash": ["MD5", "SHA-1", "SHA-256"],
"relatorio": &report,
});
std::fs::write(
@@ -165,12 +171,12 @@ fn target_rel(f: &Found) -> PathBuf {
}
/// Escreve o conteúdo e devolve o SHA-256.
pub fn write_file(src: &dyn Source, f: &Found, path: &Path) -> Result<String> {
pub fn write_file(src: &dyn Source, f: &Found, path: &Path) -> Result<Hashes> {
let mut out = std::fs::OpenOptions::new()
.write(true)
.create_new(true)
.open(path)?;
let mut hasher = Sha256::new();
let mut hasher = MultiHasher::new();
match &f.data {
DataLoc::Resident { bytes } => {
out.write_all(bytes)?;
@@ -204,11 +210,7 @@ pub fn write_file(src: &dyn Source, f: &Found, path: &Path) -> Result<String> {
let _ = out.set_modified(std::time::UNIX_EPOCH + std::time::Duration::from_secs(m as u64));
}
out.flush()?;
Ok(hasher
.finalize()
.iter()
.map(|b| format!("{b:02x}"))
.collect())
Ok(hasher.finalize())
}
/// Lê `buf.len()` bytes do conteúdo lógico de um achado a partir de `offset`
crates/guara-core/src/report.rs
+306
View File
@@ -0,0 +1,306 @@
//! Exportação para análise e relatório de recuperação (cadeia de custódia):
//! CSV dos achados e um relatório HTML autocontido, pronto para imprimir.
use crate::hashing::Hashes;
use crate::meta::fmt_unix;
use crate::recover::RecoverReport;
use crate::types::*;
use serde::{Deserialize, Serialize};
#[derive(Debug, Clone, Default, Serialize, Deserialize)]
pub struct CaseInfo {
#[serde(default)]
pub case_id: String,
#[serde(default)]
pub examiner: String,
#[serde(default)]
pub organization: String,
#[serde(default)]
pub notes: String,
}
pub struct ReportInput<'a> {
pub case: &'a CaseInfo,
pub target: &'a str,
pub options: &'a ScanOptions,
pub layout: Option<&'a DiskLayout>,
pub found: &'a [Found],
pub warnings: &'a [String],
pub started_unix: Option<i64>,
pub elapsed: f64,
pub recoveries: &'a [RecoverReport],
pub source_hashes: Option<&'a Hashes>,
}
fn cell(s: &str) -> String {
// aspas e escapa; neutraliza "injeção de fórmula" em planilhas
let mut t = s.replace('"', "\"\"");
if t.starts_with(['=', '+', '-', '@', '\t', '\r']) {
t.insert(0, '\'');
}
format!("\"{t}\"")
}
fn iso(t: Option<i64>) -> String {
match t {
Some(t) => {
let (y, m, d) = crate::meta::ymd(t);
let r = t.rem_euclid(86400);
format!(
"{y:04}-{m:02}-{d:02} {:02}:{:02}:{:02}",
r / 3600,
r % 3600 / 60,
r % 60
)
}
None => String::new(),
}
}
/// CSV com `;` (abre direto no Excel em português) e BOM UTF-8.
pub fn csv(found: &[&Found]) -> String {
let mut out = String::from("\u{feff}id;nome;pasta;extensao;categoria;tamanho_bytes;origem;confianca;modificado_utc;criado_utc;acessado_utc;foto_tirada;camera;largura;altura;gps;deslocamento_no_disco;observacao\r\n");
for f in found {
let (w, h) =
f.px.map(|(w, h)| (w.to_string(), h.to_string()))
.unwrap_or_default();
let off = f.first_offset().map(|o| o.to_string()).unwrap_or_default();
let cat = crate::explorer::cat_key(f.category);
let row = [
f.id.to_string(),
cell(&f.name),
cell(f.dir.as_deref().unwrap_or("")),
cell(&f.ext),
cat.to_string(),
f.size.to_string(),
format!("{:?}", f.origin).to_lowercase(),
format!("{:?}", f.confidence).to_lowercase(),
iso(f.modified),
iso(f.created),
iso(f.accessed),
iso(f.taken),
cell(f.camera.as_deref().unwrap_or("")),
w,
h,
if f.gps { "sim".into() } else { String::new() },
off,
cell(f.note.as_deref().unwrap_or("")),
];
out.push_str(&row.join(";"));
out.push_str("\r\n");
}
out
}
fn esc(s: &str) -> String {
s.replace('&', "&amp;")
.replace('<', "&lt;")
.replace('>', "&gt;")
.replace('"', "&quot;")
}
fn human(n: u64) -> String {
const U: [&str; 5] = ["B", "KB", "MB", "GB", "TB"];
let mut v = n as f64;
let mut i = 0;
while v >= 1024.0 && i < 4 {
v /= 1024.0;
i += 1;
}
if i == 0 {
format!("{n} B")
} else {
format!("{v:.1} {}", U[i])
}
}
const MAX_ROWS: usize = 20_000;
pub fn html(r: &ReportInput) -> String {
let now = std::time::SystemTime::now()
.duration_since(std::time::UNIX_EPOCH)
.map(|d| d.as_secs() as i64)
.unwrap_or(0);
let mut o = String::new();
o.push_str(
r#"<!doctype html><html lang="pt-BR"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1"><title>Guará: relatório de recuperação</title><style>
:root{color-scheme:light}*{box-sizing:border-box}body{margin:0;font:14px/1.5 -apple-system,"Segoe UI",Inter,Roboto,sans-serif;color:#10201f;background:#f3f7f7}
main{max-width:1100px;margin:0 auto;padding:36px 28px 60px}header{border-bottom:3px solid #0f766e;padding-bottom:14px;margin-bottom:22px}
h1{margin:0;font-size:28px;letter-spacing:-.02em}h2{margin:30px 0 10px;font-size:17px;color:#0f766e}small,.mut{color:#5b6f6d}
table{width:100%;border-collapse:collapse;background:#fff;border-radius:10px;overflow:hidden;box-shadow:0 1px 3px rgba(0,0,0,.08)}
th,td{padding:7px 10px;text-align:left;border-bottom:1px solid #e4eded;vertical-align:top;font-size:13px}th{background:#e7f3f2;color:#0b4f4a;font-weight:600}
td.n{text-align:right;font-variant-numeric:tabular-nums}.mono{font-family:ui-monospace,Menlo,Consolas,monospace;font-size:12px;word-break:break-all}
.kv td:first-child{width:230px;color:#5b6f6d}.pill{display:inline-block;padding:1px 8px;border-radius:99px;font-size:12px;font-weight:600}
.alta{background:#d7f5e0;color:#146c2e}.media{background:#fff0cf;color:#8a5a00}.baixa{background:#fde0e0;color:#a32121}
.note{background:#fff7e0;border-left:4px solid #e0a100;padding:10px 14px;border-radius:6px;margin:8px 0}
footer{margin-top:34px;color:#5b6f6d;font-size:12px}@media print{body{background:#fff}main{padding:0}table{box-shadow:none}h2{break-after:avoid}tr{break-inside:avoid}}
</style></head><body><main>"#,
);
o.push_str("<header><h1>Relatório de recuperação de arquivos</h1><div class=\"mut\">Gerado pelo Guará ");
o.push_str(env!("CARGO_PKG_VERSION"));
o.push_str(&format!(" em {}</div></header>", fmt_unix(now)));
o.push_str("<h2>Identificação do caso</h2><table class=\"kv\">");
for (k, v) in [
("Número do caso", &r.case.case_id),
("Responsável pelo exame", &r.case.examiner),
("Organização", &r.case.organization),
] {
o.push_str(&format!(
"<tr><td>{k}</td><td>{}</td></tr>",
esc(if v.is_empty() { "não informado" } else { v })
));
}
if !r.case.notes.is_empty() {
o.push_str(&format!(
"<tr><td>Observações</td><td>{}</td></tr>",
esc(&r.case.notes).replace('\n', "<br>")
));
}
o.push_str("</table>");
o.push_str("<h2>Origem examinada</h2><table class=\"kv\">");
o.push_str(&format!(
"<tr><td>Dispositivo / imagem</td><td class=\"mono\">{}</td></tr>",
esc(r.target)
));
if let Some(l) = r.layout {
o.push_str(&format!("<tr><td>Tamanho</td><td>{} ({} bytes)</td></tr><tr><td>Tabela de partições</td><td>{}</td></tr>", human(l.size), l.size, esc(&l.table.to_uppercase())));
}
if let Some(h) = r.source_hashes {
o.push_str(&format!(
"<tr><td>Hash da origem</td><td class=\"mono\">MD5 {}<br>SHA-1 {}<br>SHA-256 {}</td></tr>",
esc(&h.md5),
esc(&h.sha1),
esc(&h.sha256)
));
} else {
o.push_str("<tr><td>Hash da origem</td><td class=\"mut\">não calculado nesta sessão (use <span class=\"mono\">guara hash</span> ou <span class=\"mono\">guara image</span>)</td></tr>");
}
if let Some(t) = r.started_unix {
o.push_str(&format!(
"<tr><td>Início da varredura</td><td>{}</td></tr>",
fmt_unix(t)
));
}
o.push_str(&format!(
"<tr><td>Duração da varredura</td><td>{:.1} s</td></tr><tr><td>Modo</td><td>{}{}</td></tr><tr><td>Garantia de leitura</td><td>A origem foi aberta somente para leitura; nenhuma escrita foi feita nela.</td></tr></table>",
r.elapsed,
if r.options.mode == ScanMode::Deep { "profunda (metadados + assinatura)" } else { "rápida (metadados)" },
if r.options.include_allocated { ", incluindo áreas em uso" } else { "" }
));
if let Some(l) = r.layout {
o.push_str("<h2>Partições</h2><table><tr><th>#</th><th>Sistema</th><th>Início (bytes)</th><th>Tamanho</th><th>Rótulo</th><th>Observação</th></tr>");
for p in &l.partitions {
let obs = [
p.lost.then_some("encontrada por varredura".to_string()),
p.encrypted
.as_ref()
.map(|e| format!("cifrada ({e}), não examinada")),
]
.into_iter()
.flatten()
.collect::<Vec<_>>()
.join("; ");
o.push_str(&format!(
"<tr><td>{}</td><td>{}</td><td class=\"n\">{}</td><td class=\"n\">{}</td><td>{}</td><td>{}</td></tr>",
p.index + 1,
esc(p.fs.name()),
p.start,
human(p.len),
esc(p.label.as_deref().unwrap_or("")),
esc(&obs)
));
}
for g in &l.gaps {
o.push_str(&format!("<tr><td>-</td><td>espaço não particionado</td><td class=\"n\">{}</td><td class=\"n\">{}</td><td></td><td></td></tr>", g.start, human(g.len)));
}
o.push_str("</table>");
}
// Resumo
let mut by_cat = std::collections::BTreeMap::new();
let mut conf = [0usize; 3];
let mut bytes = 0u64;
for f in r.found {
*by_cat.entry(f.category.label_pt()).or_insert(0usize) += 1;
conf[f.confidence as usize] += 1;
bytes += f.size;
}
o.push_str(&format!("<h2>Resumo dos achados</h2><table class=\"kv\"><tr><td>Arquivos encontrados</td><td><b>{}</b> ({})</td></tr>", r.found.len(), human(bytes)));
o.push_str(&format!(
"<tr><td>Confiança</td><td><span class=\"pill alta\">alta {}</span> <span class=\"pill media\">média {}</span> <span class=\"pill baixa\">baixa {}</span></td></tr>",
conf[2], conf[1], conf[0]
));
o.push_str(&format!(
"<tr><td>Por tipo</td><td>{}</td></tr></table>",
by_cat
.iter()
.map(|(k, v)| format!("{k}: {v}"))
.collect::<Vec<_>>()
.join(" · ")
));
for w in r.warnings {
o.push_str(&format!("<div class=\"note\">{}</div>", esc(w)));
}
if !r.recoveries.is_empty() {
o.push_str("<h2>Arquivos recuperados (integridade)</h2>");
for rec in r.recoveries {
o.push_str(&format!(
"<p>Destino: <span class=\"mono\">{}</span>, {} arquivo(s), {}{}</p>",
esc(&rec.destination),
rec.recovered.len(),
human(rec.total_bytes),
if rec.failed.is_empty() {
String::new()
} else {
format!(", {} falha(s)", rec.failed.len())
}
));
o.push_str("<table><tr><th>Arquivo</th><th>Tamanho</th><th>Hashes</th></tr>");
for f in rec.recovered.iter().take(MAX_ROWS) {
o.push_str(&format!(
"<tr><td class=\"mono\">{}</td><td class=\"n\">{}</td><td class=\"mono\">MD5 {}<br>SHA-1 {}<br>SHA-256 {}</td></tr>",
esc(&f.path),
human(f.size),
f.md5,
f.sha1,
f.sha256
));
}
o.push_str("</table>");
}
}
o.push_str("<h2>Relação de arquivos encontrados</h2>");
if r.found.len() > MAX_ROWS {
o.push_str(&format!("<div class=\"note\">Mostrando os primeiros {MAX_ROWS} de {}. Use a exportação CSV para a lista completa.</div>", r.found.len()));
}
o.push_str("<table><tr><th>Nome</th><th>Pasta</th><th>Tamanho</th><th>Modificado (UTC)</th><th>Criado (UTC)</th><th>Origem</th><th>Confiança</th><th>Posição no disco</th></tr>");
for f in r.found.iter().take(MAX_ROWS) {
let (cl, cc) = match f.confidence {
Confidence::High => ("alta", "alta"),
Confidence::Medium => ("média", "media"),
Confidence::Low => ("baixa", "baixa"),
};
o.push_str(&format!(
"<tr><td>{}</td><td class=\"mono\">{}</td><td class=\"n\">{}</td><td>{}</td><td>{}</td><td>{}</td><td><span class=\"pill {cc}\">{cl}</span></td><td class=\"n mono\">{}</td></tr>",
esc(&f.name),
esc(f.dir.as_deref().unwrap_or("")),
human(f.size),
iso(f.modified),
iso(f.created),
match f.origin {
Origin::Carved => "assinatura",
Origin::Ntfs => "NTFS",
Origin::Fat => "FAT",
Origin::Exfat => "exFAT",
},
f.first_offset().map(|o| o.to_string()).unwrap_or_default()
));
}
o.push_str("</table><footer>Este relatório descreve o que o Guará encontrou e recuperou. Arquivos de confiança baixa podem estar incompletos ou sobrescritos. Datas dos metadados do sistema de arquivos estão em UTC. Guará é software livre, sem garantias.</footer></main></body></html>");
o
}
crates/guara-core/src/scan.rs
+7
View File
@@ -245,6 +245,13 @@ impl CarveEvents for Handler<'_> {
extents: vec![Extent::new(off, hit.len)],
},
modified: None,
created: None,
accessed: None,
taken: None,
camera: None,
px: None,
gps: false,
sig: None,
note: if hit.confidence == Confidence::Low {
Some("sem rodapé/estrutura completa: pode estar truncado".into())
} else {
crates/guara-core/src/search.rs
+194
View File
@@ -0,0 +1,194 @@
//! Busca de texto bruta em todo o disco (inclusive espaço livre e não particionado),
//! em UTF-8 e UTF-16LE, útil quando não há estrutura de arquivo para recuperar.
use crate::source::Source;
use serde::{Deserialize, Serialize};
#[derive(Debug, Clone, Serialize, Deserialize)]
pub struct SearchOpts {
pub terms: Vec<String>,
/// Ignora maiúsculas/minúsculas (letras ASCII).
#[serde(default = "yes")]
pub ignore_case: bool,
/// Procura também a codificação UTF-16 (Windows, Office, bancos de dados).
#[serde(default = "yes")]
pub utf16: bool,
#[serde(default = "default_max")]
pub max_hits: usize,
}
fn yes() -> bool {
true
}
fn default_max() -> usize {
20_000
}
#[derive(Debug, Clone, Serialize, Deserialize)]
pub struct Hit {
pub offset: u64,
pub term: String,
pub encoding: String,
pub context: String,
}
const CHUNK: usize = 8 * 1024 * 1024;
const CTX: usize = 48;
struct Pat {
term: String,
enc: &'static str,
bytes: Vec<u8>,
skip: [usize; 256],
}
fn make_pats(o: &SearchOpts) -> Vec<Pat> {
let mut v = vec![];
for t in o.terms.iter().filter(|t| !t.is_empty()) {
let folded = if o.ignore_case {
t.to_lowercase()
} else {
t.clone()
};
let mut add = |enc: &'static str, bytes: Vec<u8>| {
if bytes.is_empty() {
return;
}
let mut skip = [bytes.len(); 256];
for (i, &b) in bytes[..bytes.len() - 1].iter().enumerate() {
skip[b as usize] = bytes.len() - 1 - i;
}
v.push(Pat {
term: t.clone(),
enc,
bytes,
skip,
});
};
add("UTF-8", folded.as_bytes().to_vec());
if o.utf16 {
add(
"UTF-16",
folded
.encode_utf16()
.flat_map(|c| c.to_le_bytes())
.collect(),
);
}
}
v
}
fn find_all(hay: &[u8], p: &Pat, mut f: impl FnMut(usize) -> bool) {
let n = p.bytes.len();
if hay.len() < n {
return;
}
let mut i = 0;
while i + n <= hay.len() {
let last = hay[i + n - 1];
if last == p.bytes[n - 1] && hay[i..i + n] == p.bytes[..] {
if !f(i) {
return;
}
i += 1;
} else {
i += p.skip[last as usize];
}
}
}
fn snippet(buf: &[u8], at: usize, len: usize, utf16: bool) -> String {
let a = at.saturating_sub(CTX);
let b = (at + len + CTX).min(buf.len());
let s: String = if utf16 {
let a = a & !1;
let u: Vec<u16> = buf[a..b]
.chunks_exact(2)
.map(|c| u16::from_le_bytes([c[0], c[1]]))
.collect();
String::from_utf16_lossy(&u)
} else {
String::from_utf8_lossy(&buf[a..b]).into_owned()
};
s.chars()
.map(|c| if c.is_control() { '·' } else { c })
.collect()
}
/// Varre `src` inteira. Retorna o total de ocorrências. `on_hit` pode devolver false para parar.
pub fn search(
src: &dyn Source,
o: &SearchOpts,
cancelled: &(dyn Fn() -> bool + Sync),
progress: &mut dyn FnMut(u64, u64),
on_hit: &mut dyn FnMut(Hit) -> bool,
) -> usize {
let pats = make_pats(o);
if pats.is_empty() {
return 0;
}
let overlap = pats.iter().map(|p| p.bytes.len()).max().unwrap_or(1) + 2 * CTX;
let size = src.size();
let mut pos = 0u64;
let mut hits = 0usize;
let mut buf = vec![0u8; CHUNK + overlap];
'outer: while pos < size {
if cancelled() {
break;
}
let want = ((size - pos) as usize).min(CHUNK + overlap);
let mut got = 0;
while got < want {
match src.read_at(pos + got as u64, &mut buf[got..want]) {
Ok(0) => break,
Ok(n) => got += n,
Err(_) => {
// setor ruim: pula este trecho
for b in &mut buf[got..want] {
*b = 0;
}
got = want;
}
}
}
if got == 0 {
break;
}
let fresh = got.min(CHUNK); // início de ocorrências aceitas neste bloco
let folded: Vec<u8>;
let hay: &[u8] = if o.ignore_case {
folded = buf[..got].to_ascii_lowercase();
&folded
} else {
&buf[..got]
};
let mut chunk_hits: Vec<(usize, usize)> = vec![]; // (offset no bloco, índice do padrão)
for (pi, p) in pats.iter().enumerate() {
find_all(hay, p, |i| {
if i < fresh {
chunk_hits.push((i, pi));
}
true
});
}
chunk_hits.sort();
for (i, pi) in chunk_hits {
let p = &pats[pi];
hits += 1;
let h = Hit {
offset: pos + i as u64,
term: p.term.clone(),
encoding: p.enc.to_string(),
context: snippet(&buf[..got], i, p.bytes.len(), p.enc == "UTF-16"),
};
if !on_hit(h) || hits >= o.max_hits {
break 'outer;
}
}
pos += CHUNK as u64;
progress(pos.min(size), size);
}
progress(size, size);
hits
}
crates/guara-core/src/service.rs
+324 -95
View File
@@ -33,6 +33,8 @@ pub enum JobKind {
Scan,
Recover,
Image,
Search,
Hash,
}
struct State {
@@ -47,6 +49,11 @@ struct State {
report: Option<Value>,
started: Instant,
finished_in: Option<f64>,
/// Leitura de datas EXIF/dimensões em andamento: (feitos, total).
enrich: Option<(u64, u64)>,
recoveries: Vec<recover::RecoverReport>,
started_unix: Option<i64>,
hits: Vec<Value>,
}
pub struct Job {
@@ -88,6 +95,13 @@ impl Job {
report: None,
started: Instant::now(),
finished_in: None,
enrich: None,
recoveries: vec![],
started_unix: std::time::SystemTime::now()
.duration_since(std::time::UNIX_EPOCH)
.ok()
.map(|d| d.as_secs() as i64),
hits: vec![],
}),
})
}
@@ -128,6 +142,8 @@ impl Sink for JobSink {
pub struct Engine {
jobs: Mutex<HashMap<String, Arc<Job>>>,
counter: AtomicU64,
thumbs: Mutex<crate::explorer::ThumbCache>,
gate: crate::explorer::Gate,
}
impl Default for Engine {
@@ -141,6 +157,13 @@ impl Engine {
Self {
jobs: Mutex::new(HashMap::new()),
counter: AtomicU64::new(1),
thumbs: Mutex::new(crate::explorer::ThumbCache::new()),
gate: crate::explorer::Gate::new(
std::thread::available_parallelism()
.map(|n| n.get())
.unwrap_or(2)
.min(6),
),
}
}
@@ -231,53 +254,63 @@ impl Engine {
"error": s.error,
"report": s.report,
"elapsed": elapsed,
"hits": s.hits.len(),
"enriching": s.enrich.map(|(d, t)| json!({ "done": d, "total": t })),
}))
}
pub fn results(&self, id: &str, q: &Query) -> Result<Value> {
let j = self.job(id)?;
let s = j.st.lock().unwrap();
let needle = q
.query
.as_deref()
.map(|x| x.to_lowercase())
.filter(|x| !x.is_empty());
let mut items: Vec<&Found> = s
.found
.iter()
.filter(|f| q.category.map(|c| f.category == c).unwrap_or(true))
.filter(|f| q.min_confidence.map(|c| f.confidence >= c).unwrap_or(true))
.filter(|f| {
needle
.as_ref()
.map(|n| f.full_path().to_lowercase().contains(n.as_str()))
.unwrap_or(true)
})
.collect();
match q.sort.as_deref() {
Some("name") => items.sort_by_key(|f| f.name.to_lowercase()),
Some("size") => items.sort_by_key(|f| std::cmp::Reverse(f.size)),
Some("confidence") => {
items.sort_by(|a, b| b.confidence.cmp(&a.confidence).then(b.size.cmp(&a.size)))
}
_ => {}
}
let mut items: Vec<&Found> = q.select(&s.found);
crate::explorer::sort_items(&mut items, q.sort.as_deref());
let total = items.len();
let page: Vec<Value> = items
.into_iter()
.skip(q.offset)
.take(q.limit.clamp(1, 2000))
.map(|f| {
json!({
"id": f.id, "name": f.name, "dir": f.dir, "ext": f.ext, "category": f.category,
"size": f.size, "origin": f.origin, "confidence": f.confidence,
"modified": f.modified, "note": f.note,
})
})
.take(q.limit.clamp(1, 5000))
.map(crate::explorer::item_json)
.collect();
Ok(json!({ "total": total, "items": page }))
}
/// Contagens para a navegação lateral (tipos, datas, pastas, câmeras).
pub fn facets(&self, id: &str) -> Result<Value> {
let j = self.job(id)?;
let s = j.st.lock().unwrap();
Ok(crate::explorer::facets(&s.found))
}
/// Metadados detalhados (EXIF, PDF, Office, vídeo…) de um arquivo.
pub fn metadata(&self, id: &str, file: u64) -> Result<crate::meta::Meta> {
let (f, src) = self.open_file(id, file)?;
Ok(crate::meta::extract(src.as_ref(), &f))
}
/// Miniatura JPEG (cache em memória; decodificações limitadas em paralelo).
pub fn thumb(&self, job: &str, id: u64, size: u32) -> std::result::Result<Arc<Vec<u8>>, u16> {
let size = size.clamp(48, 640);
let key = (job.to_string(), id, size);
if let Some(v) = self.thumbs.lock().unwrap().get(&key) {
return if v.is_empty() { Err(422) } else { Ok(v) };
}
let (f, src) = self.open_file(job, id).map_err(|_| 404u16)?;
if f.category != Category::Image {
return Err(415);
}
let _permit = self.gate.acquire();
if let Some(v) = self.thumbs.lock().unwrap().get(&key) {
return if v.is_empty() { Err(422) } else { Ok(v) };
}
let out = Arc::new(crate::thumb::thumbnail(src.as_ref(), &f, size).unwrap_or_default());
self.thumbs.lock().unwrap().put(key, out.clone());
if out.is_empty() {
Err(422)
} else {
Ok(out)
}
}
/// Achado + origem para leitura posicional (streaming de pré-visualização).
pub fn open_file(&self, id: &str, file: u64) -> Result<(Found, Arc<dyn Source>)> {
let j = self.job(id)?;
@@ -310,25 +343,7 @@ impl Engine {
.iter()
.filter_map(|i| s.index.get(i).map(|&k| s.found[k].clone()))
.collect(),
(None, Some(q)) => {
let needle = q
.query
.as_deref()
.map(|x| x.to_lowercase())
.filter(|x| !x.is_empty());
s.found
.iter()
.filter(|f| q.category.map(|c| f.category == c).unwrap_or(true))
.filter(|f| q.min_confidence.map(|c| f.confidence >= c).unwrap_or(true))
.filter(|f| {
needle
.as_ref()
.map(|n| f.full_path().to_lowercase().contains(n.as_str()))
.unwrap_or(true)
})
.cloned()
.collect()
}
(None, Some(q)) => q.select(&s.found).into_iter().cloned().collect(),
(None, None) => s.found.clone(),
}
};
@@ -344,6 +359,7 @@ impl Engine {
None,
);
let j2 = job.clone();
let parent = sj.clone();
std::thread::spawn(move || {
let cancelled = || j2.cancelled();
let mut prog = |done: usize, total: usize, bytes: u64| {
@@ -360,6 +376,7 @@ impl Engine {
match res {
Ok(rep) => {
s.report = serde_json::to_value(&rep).ok();
parent.st.lock().unwrap().recoveries.push(rep.clone());
s.status = if rep.cancelled {
Status::Stopped
} else {
@@ -422,6 +439,156 @@ impl Engine {
Ok(job.id.clone())
}
// ---------------------------------------------------------------- exportação e relatório
/// CSV dos achados que atendem ao filtro (ou todos). Devolve quantas linhas gravou.
pub fn export_csv(&self, id: &str, path: &str, q: Option<&Query>) -> Result<usize> {
let j = self.job(id)?;
let s = j.st.lock().unwrap();
let items: Vec<&Found> = match q {
Some(q) => q.select(&s.found),
None => s.found.iter().collect(),
};
std::fs::write(path, crate::report::csv(&items))?;
Ok(items.len())
}
/// Relatório HTML de recuperação (identificação do caso, origem, achados, hashes).
pub fn export_report(
&self,
id: &str,
path: &str,
case: crate::report::CaseInfo,
hashes: Option<crate::hashing::Hashes>,
) -> Result<()> {
let j = self.job(id)?;
let s = j.st.lock().unwrap();
let elapsed = s
.finished_in
.unwrap_or_else(|| s.started.elapsed().as_secs_f64());
let html = crate::report::html(&crate::report::ReportInput {
case: &case,
target: &j.target,
options: &j.opts,
layout: s.layout.as_ref(),
found: &s.found,
warnings: &s.warnings,
started_unix: s.started_unix,
elapsed,
recoveries: &s.recoveries,
source_hashes: hashes.as_ref(),
});
std::fs::write(path, html)?;
Ok(())
}
// ---------------------------------------------------------------- busca de texto e hash da origem
pub fn start_search(&self, path: &str, opts: crate::search::SearchOpts) -> Result<String> {
if opts.terms.iter().all(|t| t.trim().is_empty()) {
return Err(Error::msg("informe pelo menos um termo"));
}
let src = open_path(Path::new(path)).map_err(|e| open_error(path, e))?;
let job = self.register(
JobKind::Search,
path.to_string(),
ScanOptions::default(),
Some(src.clone()),
);
let j2 = job.clone();
std::thread::spawn(move || {
let layout = crate::partition::analyze(src.as_ref(), false);
let cancelled = || j2.cancelled();
let mut prog = |d: u64, t: u64| {
j2.st.lock().unwrap().progress = Progress {
phase: "Procurando texto no disco".into(),
done: d,
total: t,
};
};
let mut on_hit = |h: crate::search::Hit| {
let part = layout
.partitions
.iter()
.find(|p| h.offset >= p.start && h.offset < p.start + p.len)
.map(|p| p.index + 1);
let mut s = j2.st.lock().unwrap();
s.hits.push(json!({
"offset": h.offset, "term": h.term, "encoding": h.encoding,
"context": h.context, "partition": part,
}));
true
};
crate::search::search(src.as_ref(), &opts, &cancelled, &mut prog, &mut on_hit);
let mut s = j2.st.lock().unwrap();
s.status = if j2.stop_requested.load(Ordering::Relaxed) {
Status::Stopped
} else {
Status::Done
};
s.finished_in = Some(s.started.elapsed().as_secs_f64());
});
Ok(job.id.clone())
}
pub fn search_results(&self, id: &str, offset: usize, limit: usize) -> Result<Value> {
let j = self.job(id)?;
let s = j.st.lock().unwrap();
Ok(json!({
"total": s.hits.len(),
"items": s.hits.iter().skip(offset).take(limit.clamp(1, 2000)).cloned().collect::<Vec<_>>(),
}))
}
/// Calcula MD5/SHA-1/SHA-256 da origem inteira (verificação de integridade).
pub fn start_hash(&self, path: &str) -> Result<String> {
let src = open_path(Path::new(path)).map_err(|e| open_error(path, e))?;
let job = self.register(
JobKind::Hash,
path.to_string(),
ScanOptions::default(),
None,
);
let j2 = job.clone();
std::thread::spawn(move || {
let size = src.size();
let mut h = crate::hashing::MultiHasher::new();
let mut buf = vec![0u8; 4 << 20];
let mut pos = 0u64;
let mut err = None;
while pos < size && !j2.cancelled() {
let want = ((size - pos) as usize).min(buf.len());
match crate::source::read_exact_at(src.as_ref(), pos, &mut buf[..want]) {
Ok(()) => {
h.update(&buf[..want]);
pos += want as u64;
}
Err(e) => {
err = Some(format!("falha de leitura em {pos}: {e}. Para discos com defeito use a imagem forense."));
break;
}
}
j2.st.lock().unwrap().progress = Progress {
phase: "Calculando hashes da origem".into(),
done: pos,
total: size,
};
}
let mut s = j2.st.lock().unwrap();
if let Some(e) = err {
s.error = Some(e);
s.status = Status::Error;
} else if pos < size {
s.status = Status::Stopped;
} else {
s.report = serde_json::to_value(h.finalize()).ok();
s.status = Status::Done;
}
s.finished_in = Some(s.started.elapsed().as_secs_f64());
});
Ok(job.id.clone())
}
// ---------------------------------------------------------------- sessões
pub fn save_session(&self, id: &str, path: &str) -> Result<()> {
@@ -436,6 +603,8 @@ impl Engine {
layout: s.layout.clone(),
warnings: s.warnings.clone(),
done: s.status == Status::Done,
recoveries: s.recoveries.clone(),
started_unix: s.started_unix,
};
std::fs::write(path, serde_json::to_vec(&data)?)?;
Ok(())
@@ -458,6 +627,8 @@ impl Engine {
s.found = data.found;
s.layout = data.layout;
s.warnings = data.warnings;
s.recoveries = data.recoveries;
s.started_unix = data.started_unix.or(s.started_unix);
s.checkpoint = data.checkpoint;
s.status = if data.done {
Status::Done
@@ -465,6 +636,9 @@ impl Engine {
Status::Paused
};
}
if data.done {
spawn_enrich(job.clone());
}
Ok(job.id.clone())
}
@@ -506,6 +680,49 @@ impl Engine {
json!({})
}
"status" => self.status(&s("job")?)?,
"facets" => self.facets(&s("job")?)?,
"export_csv" => {
let q: Option<Query> = p
.get("query")
.cloned()
.filter(|v| !v.is_null())
.map(serde_json::from_value)
.transpose()?;
json!({ "rows": self.export_csv(&s("job")?, &s("path")?, q.as_ref())? })
}
"export_report" => {
let case: crate::report::CaseInfo = p
.get("case")
.cloned()
.filter(|v| !v.is_null())
.map(serde_json::from_value)
.transpose()?
.unwrap_or_default();
let hashes: Option<crate::hashing::Hashes> = p
.get("hashes")
.cloned()
.filter(|v| !v.is_null())
.map(serde_json::from_value)
.transpose()?;
self.export_report(&s("job")?, &s("path")?, case, hashes)?;
json!({})
}
"search_start" => {
let opts: crate::search::SearchOpts = serde_json::from_value(p.clone())?;
json!({ "job": self.start_search(&s("path")?, opts)? })
}
"search_results" => self.search_results(
&s("job")?,
p.get("offset").and_then(|v| v.as_u64()).unwrap_or(0) as usize,
p.get("limit").and_then(|v| v.as_u64()).unwrap_or(200) as usize,
)?,
"hash_start" => json!({ "job": self.start_hash(&s("path")?)? }),
"metadata" => json!(self.metadata(
&s("job")?,
p.get("id")
.and_then(|v| v.as_u64())
.ok_or_else(|| Error::msg("id ausente"))?
)?),
"results" => {
let q: Query = p
.get("query")
@@ -556,38 +773,7 @@ impl Engine {
}
}
#[derive(Debug, Clone, Serialize, Deserialize)]
pub struct Query {
#[serde(default)]
pub category: Option<Category>,
#[serde(default)]
pub min_confidence: Option<Confidence>,
#[serde(default)]
pub query: Option<String>,
#[serde(default)]
pub sort: Option<String>,
#[serde(default)]
pub offset: usize,
#[serde(default = "default_limit")]
pub limit: usize,
}
fn default_limit() -> usize {
200
}
impl Default for Query {
fn default() -> Self {
Self {
category: None,
min_confidence: None,
query: None,
sort: None,
offset: 0,
limit: 200,
}
}
}
pub use crate::explorer::Query;
#[derive(Serialize, Deserialize)]
struct SessionFile {
@@ -599,19 +785,13 @@ struct SessionFile {
layout: Option<DiskLayout>,
warnings: Vec<String>,
done: bool,
#[serde(default)]
recoveries: Vec<recover::RecoverReport>,
#[serde(default)]
started_unix: Option<i64>,
}
fn cat_key(c: Category) -> &'static str {
match c {
Category::Image => "image",
Category::Document => "document",
Category::Video => "video",
Category::Audio => "audio",
Category::Archive => "archive",
Category::Database => "database",
Category::Other => "other",
}
}
use crate::explorer::cat_key;
fn open_error(path: &str, e: std::io::Error) -> Error {
if e.kind() == std::io::ErrorKind::PermissionDenied {
@@ -653,6 +833,55 @@ fn spawn_scan(job: Arc<Job>, cp: Option<Checkpoint>) {
}
}
s.finished_in = Some(s.started.elapsed().as_secs_f64());
let done = s.status == Status::Done;
drop(s);
if done {
spawn_enrich(job);
}
});
}
/// Lê EXIF/dimensões das imagens achadas (em segundo plano) para permitir
/// navegar por data, câmera e resolução.
fn spawn_enrich(job: Arc<Job>) {
std::thread::spawn(move || {
let Some(src) = job.src.clone() else { return };
let todo: Vec<(u64, Found)> = {
let s = job.st.lock().unwrap();
s.found
.iter()
.filter(|f| {
f.sig.is_none()
|| (f.category == Category::Image && f.px.is_none() && f.taken.is_none())
})
.map(|f| (f.id, f.clone()))
.collect()
};
if todo.is_empty() {
return;
}
let total = todo.len() as u64;
for (i, (id, f)) in todo.into_iter().enumerate() {
if job.cancelled() && job.stop_requested.load(Ordering::Relaxed) {
break;
}
let sig = crate::explorer::quick_sig(src.as_ref(), &f);
let m = (f.category == Category::Image && f.px.is_none() && f.taken.is_none())
.then(|| crate::meta::extract(src.as_ref(), &f));
let mut s = job.st.lock().unwrap();
if let Some(&k) = s.index.get(&id) {
let t = &mut s.found[k];
t.sig = sig;
if let Some(m) = m {
t.taken = m.taken;
t.camera = m.camera.clone();
t.px = m.px.or(t.px);
t.gps = m.gps.is_some();
}
}
s.enrich = Some((i as u64 + 1, total));
}
job.st.lock().unwrap().enrich = None;
});
}
crates/guara-core/src/thumb.rs
+126
View File
@@ -0,0 +1,126 @@
//! Miniaturas para a grade de arquivos. Decodifica com redução na própria
//! decodificação (JPEG) para ser rápido mesmo com milhares de fotos e, se a
//! imagem estiver danificada, tenta a miniatura embutida no EXIF.
use crate::recover::read_found_at;
use crate::source::Source;
use crate::types::Found;
use image::{imageops, ImageBuffer, Rgb, RgbImage};
const MAX_READ: u64 = 64 * 1024 * 1024;
fn read_all(src: &dyn Source, f: &Found) -> Option<Vec<u8>> {
if f.size == 0 || f.size > MAX_READ {
return None;
}
let mut v = vec![0u8; f.size as usize];
let mut got = 0;
while got < v.len() {
match read_found_at(src, f, got as u64, &mut v[got..]) {
Ok(0) | Err(_) => break,
Ok(n) => got += n,
}
}
v.truncate(got);
(!v.is_empty()).then_some(v)
}
/// Miniatura JPEG com o maior lado = `max` (None se não for imagem ou não decodificar).
pub fn thumbnail(src: &dyn Source, f: &Found, max: u32) -> Option<Vec<u8>> {
let data = read_all(src, f)?;
let img = match f.ext.to_ascii_lowercase().as_str() {
"jpg" | "jpeg" => decode_jpeg(&data, max).or_else(|| {
let t = crate::meta::exif_thumbnail(&data)?;
decode_jpeg(&t, max)
})?,
"png" | "gif" | "bmp" | "webp" => decode_other(&data, max)?,
_ => return None,
};
encode(&img)
}
fn decode_jpeg(data: &[u8], max: u32) -> Option<RgbImage> {
let mut m = crate::meta::Meta::default();
crate::meta::jpeg_bytes(data, &mut m);
let orientation = m.orientation;
let mut dec = jpeg_decoder::Decoder::new(std::io::Cursor::new(data));
dec.read_info().ok()?;
let _ = dec.scale(max.min(65535) as u16, max.min(65535) as u16);
let px = dec.decode().ok()?;
let info = dec.info()?;
let (w, h) = (info.width as u32, info.height as u32);
let rgb: Vec<u8> = match info.pixel_format {
jpeg_decoder::PixelFormat::RGB24 => px,
jpeg_decoder::PixelFormat::L8 => px.iter().flat_map(|&g| [g, g, g]).collect(),
jpeg_decoder::PixelFormat::L16 => px
.chunks_exact(2)
.flat_map(|c| [c[0], c[0], c[0]])
.collect(),
jpeg_decoder::PixelFormat::CMYK32 => px
.chunks_exact(4)
.flat_map(|c| {
// Adobe CMYK vem invertido
let k = c[3] as u32;
[
(c[0] as u32 * k / 255) as u8,
(c[1] as u32 * k / 255) as u8,
(c[2] as u32 * k / 255) as u8,
]
})
.collect(),
};
let img: RgbImage = ImageBuffer::from_raw(w, h, rgb)?;
Some(orient(fit(img, max), orientation))
}
fn decode_other(data: &[u8], max: u32) -> Option<RgbImage> {
let dynimg = image::load_from_memory(data).ok()?;
let rgba = dynimg.to_rgba8();
// Achata a transparência sobre cinza claro.
let mut rgb = RgbImage::new(rgba.width(), rgba.height());
for (x, y, p) in rgba.enumerate_pixels() {
let a = p[3] as u32;
let mix = |c: u8| ((c as u32 * a + 240 * (255 - a)) / 255) as u8;
rgb.put_pixel(x, y, Rgb([mix(p[0]), mix(p[1]), mix(p[2])]));
}
Some(fit(rgb, max))
}
fn fit(img: RgbImage, max: u32) -> RgbImage {
let (w, h) = (img.width(), img.height());
if w.max(h) <= max {
return img;
}
let (nw, nh) = if w >= h {
(max, (h as u64 * max as u64 / w as u64).max(1) as u32)
} else {
((w as u64 * max as u64 / h as u64).max(1) as u32, max)
};
imageops::resize(&img, nw, nh, imageops::FilterType::Triangle)
}
fn orient(img: RgbImage, o: u16) -> RgbImage {
match o {
2 => imageops::flip_horizontal(&img),
3 => imageops::rotate180(&img),
4 => imageops::flip_vertical(&img),
5 => imageops::flip_horizontal(&imageops::rotate90(&img)),
6 => imageops::rotate90(&img),
7 => imageops::flip_horizontal(&imageops::rotate270(&img)),
8 => imageops::rotate270(&img),
_ => img,
}
}
fn encode(img: &RgbImage) -> Option<Vec<u8>> {
let mut out = Vec::new();
let enc = jpeg_encoder::Encoder::new(&mut out, 80);
enc.encode(
img.as_raw(),
img.width() as u16,
img.height() as u16,
jpeg_encoder::ColorType::Rgb,
)
.ok()?;
Some(out)
}
crates/guara-core/src/types.rs
+25
View File
@@ -132,9 +132,34 @@ pub struct Found {
pub data: DataLoc,
/// Segundos desde 1970-01-01 UTC.
pub modified: Option<i64>,
#[serde(default)]
pub created: Option<i64>,
#[serde(default)]
pub accessed: Option<i64>,
/// Data em que a foto foi tirada (EXIF; hora da câmera lida como UTC), s desde 1970.
#[serde(default)]
pub taken: Option<i64>,
#[serde(default)]
pub camera: Option<String>,
/// Largura × altura em pixels, quando conhecidas.
#[serde(default)]
pub px: Option<(u32, u32)>,
/// Há coordenadas GPS nos metadados.
#[serde(default)]
pub gps: bool,
/// Assinatura rápida do conteúdo (tamanho + início + fim) para achar duplicadas.
#[serde(default)]
pub sig: Option<u64>,
pub note: Option<String>,
}
impl Found {
/// Data mais representativa: quando foi tirada > modificada > criada.
pub fn when(&self) -> Option<i64> {
self.taken.or(self.modified).or(self.created)
}
}
impl Found {
pub fn first_offset(&self) -> Option<u64> {
match &self.data {
crates/guara-core/tests/forensics.rs
+507
View File
@@ -0,0 +1,507 @@
//! Recursos periciais: hashes, metadados (EXIF/PDF/Office/MP4), miniaturas,
//! busca de texto, exportações e navegação (facetas/duplicadas).
use guara_core::explorer::{self, Query};
use guara_core::hashing::hash_bytes;
use guara_core::source::MemSource;
use guara_core::*;
use std::io::Write;
fn resident(name: &str, bytes: Vec<u8>) -> Found {
let ext = name.rsplit('.').next().unwrap_or("").to_string();
Found {
id: 1,
name: name.into(),
dir: None,
category: signatures::category_for_ext(&ext),
ext,
size: bytes.len() as u64,
origin: Origin::Carved,
confidence: Confidence::High,
data: DataLoc::Resident { bytes },
modified: None,
created: None,
accessed: None,
taken: None,
camera: None,
px: None,
gps: false,
sig: None,
note: None,
}
}
// ------------------------------------------------------------------ hashes
#[test]
fn hashes_conferem_com_vetores_conhecidos() {
let h = hash_bytes(b"abc");
assert_eq!(h.md5, "900150983cd24fb0d6963f7d28e17f72");
assert_eq!(h.sha1, "a9993e364706816aba3e25717850c26c9cd0d89d");
assert_eq!(
h.sha256,
"ba7816bf8f01cfea414140de5dae2223b00361a396177a9cb410ff61f20015ad"
);
}
// ------------------------------------------------------------------ EXIF
fn ifd(entries: &[(u16, u16, u32, Vec<u8>)], start: usize) -> Vec<u8> {
let n = entries.len();
let mut table = (n as u16).to_le_bytes().to_vec();
let mut data: Vec<u8> = vec![];
let data_off = start + 2 + 12 * n + 4;
for (tag, typ, cnt, val) in entries {
table.extend(tag.to_le_bytes());
table.extend(typ.to_le_bytes());
table.extend(cnt.to_le_bytes());
if val.len() <= 4 {
let mut v = val.clone();
v.resize(4, 0);
table.extend(v);
} else {
if data.len() % 2 == 1 {
data.push(0);
}
table.extend(((data_off + data.len()) as u32).to_le_bytes());
data.extend(val);
}
}
table.extend([0, 0, 0, 0]);
table.extend(data);
table
}
fn rat(p: &[(u32, u32)]) -> Vec<u8> {
p.iter()
.flat_map(|(a, b)| a.to_le_bytes().into_iter().chain(b.to_le_bytes()))
.collect()
}
fn jpeg_with_exif(orientation: u16, w: usize, h: usize) -> Vec<u8> {
// pixels: gradiente
let mut px = Vec::with_capacity(w * h * 3);
for y in 0..h {
for x in 0..w {
px.extend([(x * 255 / w) as u8, (y * 255 / h) as u8, 128]);
}
}
let mut jpg = vec![];
jpeg_encoder::Encoder::new(&mut jpg, 85)
.encode(&px, w as u16, h as u16, jpeg_encoder::ColorType::Rgb)
.unwrap();
let asc = |s: &str| {
let mut v = s.as_bytes().to_vec();
v.push(0);
v
};
let make = asc("Canon");
let model = asc("Canon EOS 80D");
let dt = asc("2023:07:14 18:22:05");
let e0 = |exif: u32, gps: u32| {
ifd(
&[
(0x010F, 2, make.len() as u32, make.clone()),
(0x0110, 2, model.len() as u32, model.clone()),
(0x0112, 3, 1, orientation.to_le_bytes().to_vec()),
(0x8769, 4, 1, exif.to_le_bytes().to_vec()),
(0x8825, 4, 1, gps.to_le_bytes().to_vec()),
],
8,
)
};
let first = e0(0, 0);
let exif_off = 8 + first.len();
let exif_ifd = ifd(&[(0x9003, 2, dt.len() as u32, dt.clone())], exif_off);
let gps_off = exif_off + exif_ifd.len();
let gps_ifd = ifd(
&[
(1, 2, 2, b"S\0".to_vec()),
(2, 5, 3, rat(&[(23, 1), (33, 1), (1800, 100)])),
(3, 2, 2, b"W\0".to_vec()),
(4, 5, 3, rat(&[(46, 1), (38, 1), (0, 1)])),
],
gps_off,
);
let mut tiff = b"II*\0".to_vec();
tiff.extend(8u32.to_le_bytes());
tiff.extend(e0(exif_off as u32, gps_off as u32));
tiff.extend(exif_ifd);
tiff.extend(gps_ifd);
let mut app1 = b"Exif\0\0".to_vec();
app1.extend(tiff);
let mut out = vec![0xFF, 0xD8, 0xFF, 0xE1];
out.extend(((app1.len() + 2) as u16).to_be_bytes());
out.extend(app1);
out.extend(&jpg[2..]);
out
}
#[test]
fn exif_camera_data_e_gps() {
let f = resident("foto.jpg", jpeg_with_exif(1, 64, 48));
let m = meta::extract(&MemSource(vec![]), &f);
assert_eq!(m.camera.as_deref(), Some("Canon EOS 80D"));
assert_eq!(m.px, Some((64, 48)));
// 2023-07-14 18:22:05 lido como UTC
assert_eq!(m.taken, Some(1_689_358_925));
let (lat, lon) = m.gps.unwrap();
assert!((lat + 23.5550).abs() < 0.001, "{lat}");
assert!((lon + 46.6333).abs() < 0.001, "{lon}");
assert!(m
.fields
.iter()
.any(|(k, v)| k == "Foto tirada em" && v == "14/07/2023 18:22:05"));
}
#[test]
fn miniatura_respeita_orientacao_e_tamanho() {
let src = MemSource(vec![]);
// 64x48 com orientação 6 (girar 90 graus) => retrato 48x64, reduzido para caber em 32
let f = resident("foto.jpg", jpeg_with_exif(6, 64, 48));
let t = thumb::thumbnail(&src, &f, 32).expect("miniatura");
let mut dec = jpeg_decoder::Decoder::new(&t[..]);
dec.decode().unwrap();
let i = dec.info().unwrap();
assert_eq!((i.width, i.height), (24, 32));
// PNG também
let mut img = ::image::RgbImage::new(100, 50);
for p in img.pixels_mut() {
*p = ::image::Rgb([200, 30, 30]);
}
let mut png = std::io::Cursor::new(vec![]);
img.write_to(&mut png, ::image::ImageFormat::Png).unwrap();
let f = resident("x.png", png.into_inner());
let t = thumb::thumbnail(&src, &f, 40).expect("miniatura png");
let mut dec = jpeg_decoder::Decoder::new(&t[..]);
dec.decode().unwrap();
assert_eq!(
(dec.info().unwrap().width, dec.info().unwrap().height),
(40, 20)
);
// lixo não vira miniatura
assert!(thumb::thumbnail(&src, &resident("x.jpg", vec![1, 2, 3, 4, 5]), 32).is_none());
}
// ------------------------------------------------------------------ PDF, Office, MP4
#[test]
fn metadados_de_pdf() {
let pdf = b"%PDF-1.7\n1 0 obj\n<< /Type /Catalog /Pages 2 0 R >>\nendobj\n2 0 obj\n<< /Type /Pages /Count 12 /Kids [] >>\nendobj\n3 0 obj\n<< /Title (Relat\\363rio Final) /Author (Maria \\(Perita\\) Silva) /Creator (LibreOffice Writer) /Producer <FEFF00500044004600310032> /CreationDate (D:20230714182205-03'00') >>\nendobj\ntrailer\n%%EOF\n".to_vec();
let m = meta::extract(&MemSource(vec![]), &resident("a.pdf", pdf));
let get = |k: &str| {
m.fields
.iter()
.find(|(a, _)| a == k)
.map(|(_, v)| v.clone())
};
assert_eq!(get("Título").as_deref(), Some("Relatório Final"));
assert_eq!(get("Autor").as_deref(), Some("Maria (Perita) Silva"));
assert_eq!(get("Criado com").as_deref(), Some("LibreOffice Writer"));
assert_eq!(get("Gerado por").as_deref(), Some("PDF12"));
assert_eq!(get("Criado em").as_deref(), Some("14/07/2023 18:22:05"));
assert_eq!(get("Páginas (aprox.)").as_deref(), Some("12"));
}
fn zip_store(entries: &[(&str, &str)], deflate: bool) -> Vec<u8> {
let mut out = vec![];
let mut cd = vec![];
for (name, body) in entries {
let (method, data): (u16, Vec<u8>) = if deflate {
let mut e = flate2::write::DeflateEncoder::new(vec![], flate2::Compression::default());
e.write_all(body.as_bytes()).unwrap();
(8, e.finish().unwrap())
} else {
(0, body.as_bytes().to_vec())
};
let off = out.len() as u32;
let crc = signatures::crc32(body.as_bytes());
let mut lh = b"PK\x03\x04".to_vec();
lh.extend([20, 0, 0, 0]);
lh.extend(method.to_le_bytes());
lh.extend([0, 0, 0, 0]);
lh.extend(crc.to_le_bytes());
lh.extend((data.len() as u32).to_le_bytes());
lh.extend((body.len() as u32).to_le_bytes());
lh.extend((name.len() as u16).to_le_bytes());
lh.extend([0, 0]);
lh.extend(name.as_bytes());
out.extend(&lh);
out.extend(&data);
let mut c = b"PK\x01\x02".to_vec();
c.extend([20, 0, 20, 0, 0, 0]);
c.extend(method.to_le_bytes());
c.extend([0, 0, 0, 0]);
c.extend(crc.to_le_bytes());
c.extend((data.len() as u32).to_le_bytes());
c.extend((body.len() as u32).to_le_bytes());
c.extend((name.len() as u16).to_le_bytes());
c.extend([0; 12]);
c.extend(off.to_le_bytes());
c.extend(name.as_bytes());
cd.extend(c);
}
let cd_off = out.len() as u32;
out.extend(&cd);
out.extend(b"PK\x05\x06\0\0\0\0");
out.extend((entries.len() as u16).to_le_bytes());
out.extend((entries.len() as u16).to_le_bytes());
out.extend((cd.len() as u32).to_le_bytes());
out.extend(cd_off.to_le_bytes());
out.extend([0, 0]);
out
}
#[test]
fn metadados_de_documentos_office() {
let core = r#"<?xml version="1.0"?><cp:coreProperties xmlns:cp="x" xmlns:dc="y" xmlns:dcterms="z"><dc:title>Tese &amp; Defesa</dc:title><dc:creator>Ana Souza</dc:creator><cp:lastModifiedBy>Bruno Lima</cp:lastModifiedBy><dcterms:created xsi:type="dcterms:W3CDTF">2023-07-14T18:22:05Z</dcterms:created></cp:coreProperties>"#;
let app = "<Properties><Application>Microsoft Office Word</Application><Company>ACME</Company></Properties>";
for deflate in [false, true] {
let z = zip_store(
&[
("[Content_Types].xml", "<Types/>"),
("docProps/core.xml", core),
("docProps/app.xml", app),
],
deflate,
);
let m = meta::extract(&MemSource(vec![]), &resident("t.docx", z));
let get = |k: &str| {
m.fields
.iter()
.find(|(a, _)| a == k)
.map(|(_, v)| v.clone())
};
assert_eq!(
get("Título").as_deref(),
Some("Tese & Defesa"),
"deflate={deflate}"
);
assert_eq!(get("Autor").as_deref(), Some("Ana Souza"));
assert_eq!(get("Última modificação por").as_deref(), Some("Bruno Lima"));
assert_eq!(
get("Criado em").as_deref(),
Some("14/07/2023 18:22:05 (UTC)")
);
assert_eq!(get("Aplicativo").as_deref(), Some("Microsoft Office Word"));
assert_eq!(get("Empresa").as_deref(), Some("ACME"));
}
}
#[test]
fn metadados_de_mp4() {
fn bx(ty: &[u8; 4], body: &[u8]) -> Vec<u8> {
let mut v = ((body.len() + 8) as u32).to_be_bytes().to_vec();
v.extend(ty);
v.extend(body);
v
}
let created_1904: u32 = 1_689_358_925 + 2_082_844_800;
let mut mvhd = vec![0, 0, 0, 0];
mvhd.extend(created_1904.to_be_bytes());
mvhd.extend(created_1904.to_be_bytes());
mvhd.extend(1000u32.to_be_bytes()); // timescale
mvhd.extend(125_000u32.to_be_bytes()); // 125 s
mvhd.extend(vec![0u8; 80]);
let mut tkhd = vec![0u8; 84];
tkhd[76..80].copy_from_slice(&(1920u32 << 16).to_be_bytes());
tkhd[80..84].copy_from_slice(&(1080u32 << 16).to_be_bytes());
let moov = bx(
b"moov",
&[bx(b"mvhd", &mvhd), bx(b"trak", &bx(b"tkhd", &tkhd))].concat(),
);
let mut file = bx(b"ftyp", b"isom\0\0\0\0isom");
file.extend(moov);
file.extend(bx(b"mdat", &[0u8; 32]));
let m = meta::extract(&MemSource(vec![]), &resident("v.mp4", file));
let get = |k: &str| {
m.fields
.iter()
.find(|(a, _)| a == k)
.map(|(_, v)| v.clone())
};
assert_eq!(get("Duração").as_deref(), Some("0:02:05"));
assert_eq!(
get("Gravado em (metadado)").as_deref(),
Some("14/07/2023 18:22:05 (UTC)")
);
assert_eq!(get("Dimensões").as_deref(), Some("1920 × 1080 px"));
}
// ------------------------------------------------------------------ busca de texto
#[test]
fn busca_acha_utf8_utf16_e_na_fronteira_dos_blocos() {
let mb = 1024 * 1024;
let mut disk = vec![0u8; 20 * mb];
let put = |d: &mut Vec<u8>, at: usize, b: &[u8]| d[at..at + b.len()].copy_from_slice(b);
put(&mut disk, 1000, b"... Segredo-123 ...");
let u16s: Vec<u8> = "segredo-123"
.encode_utf16()
.flat_map(|c| c.to_le_bytes())
.collect();
put(&mut disk, 3 * mb + 6, &u16s);
// exatamente cruzando o limite do primeiro bloco de 8 MiB
put(&mut disk, 8 * mb - 4, b"SEGREDO-123");
let opts = search::SearchOpts {
terms: vec!["segredo-123".into()],
ignore_case: true,
utf16: true,
max_hits: 100,
};
let mut hits = vec![];
let n = search::search(
&MemSource(disk.clone()),
&opts,
&|| false,
&mut |_, _| {},
&mut |h| {
hits.push(h);
true
},
);
assert_eq!(n, 3, "{hits:?}");
assert_eq!(hits[0].offset, 1004);
assert_eq!(hits[0].encoding, "UTF-8");
assert!(hits[0].context.contains("Segredo-123"));
assert_eq!(hits[1].offset, 3 * mb as u64 + 6);
assert_eq!(hits[1].encoding, "UTF-16");
assert_eq!(hits[2].offset, 8 * mb as u64 - 4);
// sem ignorar maiúsculas só acha a grafia exata
let opts = search::SearchOpts {
terms: vec!["Segredo-123".into()],
ignore_case: false,
utf16: false,
max_hits: 100,
};
let n = search::search(
&MemSource(disk),
&opts,
&|| false,
&mut |_, _| {},
&mut |_| true,
);
assert_eq!(n, 1);
}
// ------------------------------------------------------------------ exportações
#[test]
fn csv_escapa_e_neutraliza_formulas() {
let mut f = resident("=SOMA(A1).xlsx", vec![1]);
f.note = Some("diz \"oi\"; ok".into());
f.camera = Some("Cam;X".into());
let csv = report::csv(&[&f]);
assert!(csv.starts_with('\u{feff}'));
assert!(csv.contains("\"'=SOMA(A1).xlsx\""), "{csv}");
assert!(csv.contains("\"diz \"\"oi\"\"; ok\""));
assert_eq!(csv.lines().count(), 2);
}
#[test]
fn relatorio_html_escapa_conteudo_e_traz_o_caso() {
let f = resident("<script>alert(1)</script>.txt", vec![1, 2, 3]);
let case = report::CaseInfo {
case_id: "2026-001".into(),
examiner: "Perita <Ana>".into(),
organization: "Lab".into(),
notes: "obs".into(),
};
let opts = ScanOptions::default();
let hashes = hash_bytes(b"disco");
let html = report::html(&report::ReportInput {
case: &case,
target: "/dev/sdb",
options: &opts,
layout: None,
found: &[f],
warnings: &["aviso <b>".to_string()],
started_unix: Some(1_689_358_925),
elapsed: 1.5,
recoveries: &[],
source_hashes: Some(&hashes),
});
assert!(!html.contains("<script>alert"), "HTML não escapado");
assert!(html.contains("&lt;script&gt;alert(1)&lt;/script&gt;.txt"));
assert!(html.contains("2026-001") && html.contains("Perita &lt;Ana&gt;"));
assert!(html.contains(&hashes.sha256));
assert!(html.contains("somente para leitura"));
}
// ------------------------------------------------------------------ navegação
#[test]
fn facetas_filtros_e_duplicadas() {
let mk = |id: u64,
name: &str,
dir: Option<&str>,
taken: Option<i64>,
cam: Option<&str>,
sig: Option<u64>| {
let mut f = resident(name, vec![0; 10]);
f.id = id;
f.dir = dir.map(String::from);
f.origin = if dir.is_some() {
Origin::Fat
} else {
Origin::Carved
};
f.taken = taken;
f.camera = cam.map(String::from);
f.sig = sig;
f
};
let jul23 = 1_689_358_925; // 2023-07-14
let jan24 = 1_705_000_000; // 2024-01-11
let found = vec![
mk(
1,
"a.jpg",
Some("DCIM/100"),
Some(jul23),
Some("Canon"),
Some(7),
),
mk(
2,
"b.jpg",
Some("DCIM/100"),
Some(jan24),
Some("Canon"),
Some(8),
),
mk(3, "c.jpg", Some("Viagens"), Some(jul23), None, Some(7)), // duplicada de a.jpg
mk(4, "d.jpg", None, None, None, None),
];
let f = explorer::facets(&found);
assert_eq!(f["total"], 4);
assert_eq!(f["years"][0]["year"], 2024);
assert_eq!(f["years"][1]["year"], 2023);
assert_eq!(f["years"][1]["count"], 2);
assert_eq!(f["undated"], 1);
assert_eq!(f["carved"], 1);
assert_eq!(f["duplicates"], 1);
assert_eq!(f["cameras"][0]["name"], "Canon");
let dcim = f["folders"]
.as_array()
.unwrap()
.iter()
.find(|d| d["path"] == "DCIM")
.unwrap();
assert_eq!(dcim["count"], 2);
let q = |json: &str| -> Query { serde_json::from_str(json).unwrap() };
assert_eq!(q(r#"{"year":2023}"#).select(&found).len(), 2);
assert_eq!(q(r#"{"year":2023,"month":7}"#).select(&found).len(), 2);
assert_eq!(q(r#"{"year":2023,"month":8}"#).select(&found).len(), 0);
assert_eq!(q(r#"{"folder":"DCIM"}"#).select(&found).len(), 2);
assert_eq!(q(r#"{"folder":"@carved"}"#).select(&found).len(), 1);
assert_eq!(q(r#"{"camera":"Canon"}"#).select(&found).len(), 2);
assert_eq!(q(r#"{"undated":true}"#).select(&found).len(), 1);
assert_eq!(q(r#"{"dedupe":true}"#).select(&found).len(), 3);
assert_eq!(q(r#"{"query":"viagens"}"#).select(&found).len(), 1);
let mut items = q("{}").select(&found);
explorer::sort_items(&mut items, Some("date"));
assert_eq!(items[0].id, 2); // mais recente primeiro
assert_eq!(items.last().unwrap().id, 4); // sem data por último
}
crates/guara-core/tests/recovery.rs
+9
View File
@@ -32,6 +32,15 @@ fn check(fs: &str, only_carvable: bool) {
assert!(seen.insert(o), "[{fs}] achado duplicado no offset {o}");
}
}
if fs != "ext4" {
for f in res.found.iter().filter(|f| f.origin != Origin::Carved) {
assert!(
f.modified.is_some() && f.created.is_some(),
"[{fs}] {} sem datas de modificação/criação",
f.name
);
}
}
assert_eq!(before, sha256_file(&img), "a origem foi modificada!");
}
scripts/make-photo-fixture.py
Executable
+112
View File
@@ -0,0 +1,112 @@
#!/usr/bin/env python3
"""Gera uma imagem FAT32 com muitas fotos JPEG (EXIF de câmeras/datas/GPS) já apagadas.
Uso: make-photo-fixture.py <saida.img> [quantidade=300] [copias=1]
(`copias` repete cada foto N vezes com outro nome: é um jeito rápido de gerar milhares de arquivos.)
Serve para testar a navegação por data/câmera/pasta e as miniaturas.
"""
import os, random, shutil, struct, subprocess, sys, tempfile
out = sys.argv[1]
count = int(sys.argv[2]) if len(sys.argv) > 2 else 300
copies = int(sys.argv[3]) if len(sys.argv) > 3 else 1
rnd = random.Random(42)
def ifd_bytes(entries, start):
n = len(entries)
table = struct.pack('<H', n)
data = b''
data_off = start + 2 + 12 * n + 4
for tag, typ, cnt, val in entries:
if isinstance(val, int):
vb = struct.pack('<I', val) if typ == 4 else struct.pack('<H', val) + b'\0\0'
else:
vb = val
if len(vb) <= 4:
table += struct.pack('<HHI', tag, typ, cnt) + vb.ljust(4, b'\0')
else:
if len(data) % 2:
data += b'\0'
table += struct.pack('<HHII', tag, typ, cnt, data_off + len(data))
data += vb
return table + struct.pack('<I', 0) + data
def rat(*pairs):
return b''.join(struct.pack('<II', a, b) for a, b in pairs)
def make_exif(make, model, dt, gps, orient):
def build(exif_off, gps_off):
e0 = [(0x010F, 2, len(make) + 1, make.encode() + b'\0'), (0x0110, 2, len(model) + 1, model.encode() + b'\0'),
(0x0112, 3, 1, orient), (0x8769, 4, 1, exif_off)]
if gps:
e0.append((0x8825, 4, 1, gps_off))
return ifd_bytes(e0, 8)
ex = [(0x9003, 2, 20, dt.encode() + b'\0')]
first = build(0, 0)
exif_off = 8 + len(first)
exif_ifd = ifd_bytes(ex, exif_off)
gps_off = exif_off + len(exif_ifd)
gps_ifd = b''
if gps:
lat, lon = gps
def dms(v):
v = abs(v); d = int(v); m = int((v - d) * 60); s = round(((v - d) * 60 - m) * 60 * 100)
return rat((d, 1), (m, 1), (s, 100))
g = [(1, 2, 2, b'S\0' if lat < 0 else b'N\0'), (2, 5, 3, dms(lat)), (3, 2, 2, b'W\0' if lon < 0 else b'E\0'), (4, 5, 3, dms(lon))]
gps_ifd = ifd_bytes(g, gps_off)
body = build(exif_off, gps_off) + exif_ifd + gps_ifd
tiff = b'II*\0' + struct.pack('<I', 8) + body
return b'Exif\0\0' + tiff
def add_exif(jpg_bytes, exif):
assert jpg_bytes[:2] == b'\xff\xd8'
seg = b'\xff\xe1' + struct.pack('>H', len(exif) + 2) + exif
return jpg_bytes[:2] + seg + jpg_bytes[2:]
cams = [('Canon', 'Canon EOS 80D'), ('Apple', 'iPhone 13'), ('samsung', 'SM-G991B'), ('NIKON CORPORATION', 'NIKON D3500')]
places = [(-23.5505, -46.6333), (-8.0476, -34.8770), (-22.9068, -43.1729), (38.7223, -9.1393)]
folders = ['DCIM/100CANON', 'DCIM/Camera', 'WhatsApp/Images', 'Viagens/Recife', 'Viagens/Lisboa', 'Familia']
tmp = tempfile.mkdtemp()
files = []
for i in range(count):
cam = cams[i % 4]
y = rnd.choice([2021, 2022, 2023, 2024]); mo = rnd.randint(1, 12); d = rnd.randint(1, 28)
dt = f'{y}:{mo:02d}:{d:02d} {rnd.randint(7, 22):02d}:{rnd.randint(0,59):02d}:{rnd.randint(0,59):02d}'
folder = folders[i % len(folders)]
name = f'IMG_{i:04d}.jpg'
p = os.path.join(tmp, name)
w, h = rnd.choice([(480, 320), (320, 480), (640, 360), (400, 400)])
subprocess.run(['convert', '-seed', str(i), '-size', f'{w}x{h}', 'plasma:fractal', '-gravity', 'center', '-pointsize', '34', '-fill', 'white',
'-stroke', 'black', '-strokewidth', '1', '-annotate', '0', f'{y}-{mo:02d}-{d:02d}\n#{i}', '-quality', '82', p], check=True, stderr=subprocess.DEVNULL)
raw = open(p, 'rb').read()
if i % 9 != 0: # alguns sem EXIF
gps = places[i % 4] if i % 3 == 0 else None
raw = add_exif(raw, make_exif(cam[0], cam[1], dt, gps, rnd.choice([1, 1, 1, 6, 8])))
open(p, 'wb').write(raw)
files.append((folder, name, p))
if copies > 1:
base = list(files)
for c in range(1, copies):
for folder, name, p in base:
files.append((folder, name.replace('IMG_', f'IMG{c:03d}_'), p))
mb = max(64, len(files) * 40 // 1024 + 48)
if os.path.exists(out):
os.remove(out)
subprocess.run(['truncate', '-s', f'{mb}M', out], check=True)
subprocess.run(['mkfs.vfat', '-F', '32', out], check=True, stdout=subprocess.DEVNULL, stderr=subprocess.DEVNULL)
made = set()
for folder, name, p in files:
cur = ''
for part in folder.split('/'):
cur += '/' + part
if cur not in made:
subprocess.run(['mmd', '-i', out, '::' + cur], check=True, stderr=subprocess.DEVNULL)
made.add(cur)
for folder, name, p in files:
subprocess.run(['mcopy', '-i', out, p, f'::/{folder}/{name}'], check=True)
for folder, name, p in files:
subprocess.run(['mdel', '-i', out, f'::/{folder}/{name}'], check=True)
shutil.rmtree(tmp)
print(f'ok: {out} ({len(files)} fotos apagadas)')