julienfroidefond/stripstream-librarian
1
0
Fork 0
Code Issues 4 Pull Requests Actions Packages Projects Releases Wiki Activity

fix: improve CBR extraction with fallback and increase timeout

Browse Source 
- Try multiple entries in CBR archive until finding valid image
- Increase timeout from 12s to 30s for large files
- Better error messages for debugging
This commit is contained in:
Julien Froidefond
2026-03-07 15:02:55 +01:00
parent 217919fa77
commit 162b4712e7
1 changed files with 62 additions and 65 deletions
Download Patch File Download Diff File Expand all files Collapse all files

127
apps/api/src/pages.rs
View File

@@ -220,7 +220,7 @@ pub async fn get_page(
let start_time = std::time::Instant::now();
let bytes = tokio::time::timeout(
Duration::from_secs(12),
Duration::from_secs(30),
tokio::task::spawn_blocking(move || {
render_page(&abs_path_clone, &input_format, n, &format_clone, quality, width)
}),
@@ -364,74 +364,71 @@ fn extract_cbr_page(abs_path: &str, page_number: u32) -> Result<Vec<u8>, ApiErro
.collect();
entries.sort();
// Debug: show first few entries
if entries.len() > 0 {
debug!("First 5 entries in CBR {}: {:?}", abs_path, &entries[..entries.len().min(5)]);
}
debug!("First 10 entries in CBR {}: {:?}", abs_path, &entries[..entries.len().min(10)]);
debug!("Found {} images in CBR {}", entries.len(), abs_path);
let index = page_number as usize - 1;
let selected = entries.get(index).ok_or_else(|| {
error!("Page {} out of range in {} (total: {})", page_number, abs_path, entries.len());
ApiError::not_found("page out of range")
})?;
debug!("Extracting page {} ({}) from {}", page_number, selected, abs_path);
let page_output = std::process::Command::new("unrar")
.arg("p")
.arg("-inul")
.arg("-y")
.arg(abs_path)
.arg(selected)
.output()
.map_err(|e| {
error!("unrar extract command failed for {} page {}: {}", abs_path, selected, e);
ApiError::internal(format!("unrar extract failed: {e}"))
})?;
if !page_output.status.success() {
let stderr = String::from_utf8_lossy(&page_output.stderr);
error!("unrar could not extract page {} from {}: {}", selected, abs_path, stderr);
return Err(ApiError::internal("unrar could not extract page"));
// Try multiple entries starting from the requested page number
// Some archives have non-image files (readme.txt, etc.) at the start
let mut tried_entries: Vec<String> = Vec::new();
let start_index = page_number.saturating_sub(1) as usize;
for offset in 0..10 {
let idx = start_index + offset;
if idx >= entries.len() {
break;
}
let selected = &entries[idx];
tried_entries.push(selected.clone());
debug!("Trying to extract entry {} ({}) from CBR {}", idx, selected, abs_path);
let page_output = std::process::Command::new("unrar")
.arg("p")
.arg("-inul")
.arg("-y")
.arg(abs_path)
.arg(selected)
.output()
.map_err(|e| {
error!("unrar extract command failed for {} page {}: {}", abs_path, selected, e);
ApiError::internal(format!("unrar extract failed: {e}"))
})?;
if !page_output.status.success() {
let stderr = String::from_utf8_lossy(&page_output.stderr);
warn!("unrar could not extract page {} from {}: {}", selected, abs_path, stderr);
continue;
}
let extracted_data = &page_output.stdout;
// Validate it's actually an image
if extracted_data.len() < 1000 {
warn!("Entry {} ({}) too small ({} bytes), trying next", idx, selected, extracted_data.len());
continue;
}
let is_valid_image = extracted_data.len() > 4 && (
extracted_data.starts_with(&[0xFF, 0xD8, 0xFF]) || // JPEG
extracted_data.starts_with(&[0x89, 0x50, 0x4E, 0x47]) || // PNG
extracted_data.starts_with(b"GIF87a") || extracted_data.starts_with(b"GIF89a") || // GIF
(extracted_data.starts_with(b"RIFF") && extracted_data.len() > 12 && &extracted_data[8..12] == b"WEBP") || // WebP
extracted_data.starts_with(&[0x49, 0x49, 0x2A, 0x00]) || extracted_data.starts_with(&[0x4D, 0x4D, 0x00, 0x2A]) // TIFF
);
if !is_valid_image {
warn!("Entry {} ({}) is not a valid image, trying next", idx, selected);
continue;
}
// Found valid image!
info!("Successfully extracted valid image from CBR entry {} ({}) at offset {}", idx, selected, offset);
return Ok(extracted_data.to_vec());
}
let extracted_data = &page_output.stdout;
debug!("Extracted {} bytes from CBR page {}", extracted_data.len(), page_number);
// Verify it's actually an image by checking magic bytes
if extracted_data.len() < 1000 {
// Show first few bytes for debugging
let preview: Vec<u8> = extracted_data.iter().take(32).copied().collect();
let hex_preview: String = preview.iter().map(|b| format!("{:02x}", b)).collect();
error!("Extracted data too small ({} bytes) for page {} from {} - first bytes: {}",
extracted_data.len(), page_number, abs_path, hex_preview);
return Err(ApiError::internal("extracted data too small - not a valid image"));
}
// Check magic bytes to verify it's an image
let is_valid_image = extracted_data.len() > 4 && (
// JPEG
extracted_data.starts_with(&[0xFF, 0xD8, 0xFF]) ||
// PNG
extracted_data.starts_with(&[0x89, 0x50, 0x4E, 0x47]) ||
// GIF
extracted_data.starts_with(b"GIF87a") || extracted_data.starts_with(b"GIF89a") ||
// WebP
extracted_data.starts_with(b"RIFF") && extracted_data.len() > 12 && &extracted_data[8..12] == b"WEBP" ||
// TIFF
extracted_data.starts_with(&[0x49, 0x49, 0x2A, 0x00]) || extracted_data.starts_with(&[0x4D, 0x4D, 0x00, 0x2A])
);
if !is_valid_image {
// Show first few bytes for debugging
let preview: Vec<u8> = extracted_data.iter().take(32).copied().collect();
let hex_preview: String = preview.iter().map(|b| format!("{:02x}", b)).collect();
error!("Extracted data for page {} from {} is not a valid image format. First bytes: {} (size: {})",
page_number, abs_path, hex_preview, extracted_data.len());
return Err(ApiError::internal("extracted data is not a valid image"));
}
debug!("Successfully extracted {} bytes from CBR page {}", extracted_data.len(), page_number);
Ok(extracted_data.to_vec())
error!("Could not find valid image in CBR {} after trying: {:?}", abs_path, tried_entries);
Err(ApiError::not_found("no valid image found in archive after trying multiple entries"))
}
fn render_pdf_page(abs_path: &str, page_number: u32, width: u32) -> Result<Vec<u8>, ApiError> {