rdsm/arxiv_audio_summary
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

Test

Browse source
This commit is contained in:
Regis David Souza Mesquita 2025-03-02 15:47:53 +00:00
parent 1d90930e29
commit ef993f8693
1 changed files with 9 additions and 2 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

11
vibe/converter.py
View file

@ -12,8 +12,12 @@ from .config import ARTICLES_CACHE_DIR
logger = logging.getLogger(__name__) logger = logging.getLogger(__name__)
pdf_options = PdfFormatOption(pipeline_options=PdfPipelineOptions(generate_picture_images=True)) pipeline_options = PdfPipelineOptions()
pipeline_options.ocr_options.use_gpu = False
pipeline_options.generate_picture_images = False
pdf_options = PdfFormatOption(pipeline_options=pipeline_options)
doc_converter = DocumentConverter(format_options={InputFormat.PDF: pdf_options}) doc_converter = DocumentConverter(format_options={InputFormat.PDF: pdf_options})
doc_converter = DocumentConverter()
def fetch_and_convert_article(article): def fetch_and_convert_article(article):
""" """
@ -51,10 +55,13 @@ def fetch_and_convert_article(article):
f.write(converted_text) f.write(converted_text)
logger.info("Conversion successful for article '%s'. Cached output.", article["id"]) logger.info("Conversion successful for article '%s'. Cached output.", article["id"])
return converted_text return converted_text
except SystemExit as se:
logger.exception("Docling conversion exited with error code %s for article '%s'. Skipping conversion.", se.code, article["id"])
return ""
except Exception as e: except Exception as e:
logger.exception("Conversion failed for article '%s': %s", article["id"], e) logger.exception("Conversion failed for article '%s': %s", article["id"], e)
return "" return ""
finally: finally:
if os.path.exists(tmp_pdf_path): if os.path.exists(tmp_pdf_path):
os.unlink(tmp_pdf_path) os.unlink(tmp_pdf_path)
logger.debug("Temporary PDF file %s removed.", tmp_pdf_path) logger.debug("Temporary PDF file %s removed.", tmp_pdf_path)