This commit is contained in:
2025-11-14 14:47:19 +00:00
parent 2a85845835
commit 3758c7572a
137 changed files with 365654 additions and 11147 deletions
+33 -20
View File
@@ -39,28 +39,41 @@ class DocumentProcessor:
# Use LibreOffice for conversion
with tempfile.TemporaryDirectory() as temp_dir:
output_file = Path(temp_dir) / f"{input_file.stem}.pdf"
# Convert using LibreOffice
cmd = [
'libreoffice',
'--headless',
'--convert-to', 'pdf',
'--outdir', str(temp_dir),
str(input_file)
]
try:
subprocess.run(cmd, check=True, capture_output=True)
except subprocess.CalledProcessError as e:
raise RuntimeError(f"Conversion failed: {e.stderr.decode()}")
with tempfile.TemporaryDirectory() as profile_dir:
output_file = Path(temp_dir) / f"{input_file.stem}.pdf"
# Convert using LibreOffice with explicit profile directory
cmd = [
'/Applications/LibreOffice.app/Contents/MacOS/soffice',
'--headless',
'--invisible',
'--nodefault',
'--nolockcheck',
'--nologo',
'--norestore',
f'-env:UserInstallation=file://{profile_dir}',
'--convert-to', 'pdf',
'--outdir', str(temp_dir),
str(input_file)
]
try:
result = subprocess.run(cmd, check=True, capture_output=True, timeout=120)
except subprocess.CalledProcessError as e:
stderr_msg = e.stderr.decode() if e.stderr else "No stderr output"
stdout_msg = e.stdout.decode() if e.stdout else "No stdout output"
raise RuntimeError(f"Conversion failed: {stderr_msg}. Stdout: {stdout_msg}")
except subprocess.TimeoutExpired:
raise RuntimeError("Conversion failed: Process timed out after 120 seconds")
if not output_file.exists():
raise RuntimeError("Conversion failed: Output file not created")
if not output_file.exists():
# List all files in temp_dir for debugging
files_in_dir = list(Path(temp_dir).glob('*'))
raise RuntimeError(f"Conversion failed: Output file not created. Files in output dir: {files_in_dir}")
# Read and return the PDF content
with open(output_file, 'rb') as f:
return f.read()
# Read and return the PDF content
with open(output_file, 'rb') as f:
return f.read()
def batch_convert_directory(self, directory: str) -> List[Dict]:
"""