def extract_update_bundle(bundle):
past_update_dir = paths.data_path("update")
if past_update_dir.exists():
log.info("Found previous update data. Removing...")
shutil.rmtree(past_update_dir, ignore_errors=True)
log.debug("Extracting update bundle")
extraction_dir = paths.data_path("update", "extracted")
extraction_dir.mkdir(parents=True, exist_ok=True)
with zipfile.ZipFile(bundle, compression=zipfile.ZIP_LZMA) as archive:
archive.extractall(extraction_dir)
return extraction_dir
def extract_update_bundle(bundle):
past_update_dir = paths.data_path("update")
if past_update_dir.exists():
log.info("Found previous update data. Removing...")
shutil.rmtree(past_update_dir, ignore_errors=True)
log.debug("Extracting update bundle")
bundle.seek(0)
extraction_dir = paths.data_path("update", "extracted")
if extraction_dir.exists():
shutil.rmtree(extraction_dir)
extraction_dir.mkdir(parents=True, exist_ok=True)
archive_file = BytesIO(decompress(bundle.read()))
with zipfile.ZipFile(archive_file) as archive:
archive.extractall(extraction_dir)
return extraction_dir
def get_bundled_documents_folder():
bd_path = paths.data_path("bundled_documents")
if not bd_path.exists():
bd_path.mkdir(parents=True, exist_ok=True)
return bd_path
def get_tesseract_path():
return data_path(f"tesseract_ocr_{app.arch}").resolve()
def get_xpdf_config_file():
xpdf_config = data_path("xpdf_config", "default.xpdf")
if not xpdf_config.exists():
xpdf_config.parent.mkdir(parents=True, exist_ok=True)
xpdf_config.write_text(generate_xpdfrc())
return xpdf_config