Ebook extractor

[CorentinB]

Need to handle many/all ebook formats.

[vbanos]

epub ebooks are standard zip files containing HTML. This is a python function that works in production. The same could be done as a Zeno epub extractor.

def extract_epub_outlinks(epub_data: bytes) -> list[str]:                       
      """Extract outlinks from epub which are zip files containing HTML.          
      try using URL pattern matching from text.                                   
      """                                                                         
      outlinks: set[str] = set()                                                  
      try:                                                                        
          zip_stream = io.BytesIO(epub_data)                                      
          with zipfile.ZipFile(zip_stream) as zf:                                 
              for filename in zf.namelist():                                      
                  if filename.endswith((".htm", ".html", ".xhtml")):              
                      with zf.open(filename) as file:                             
                          content = file.read().decode('utf-8')                   
                          outlinks.update(extract_text_outlinks(content))         
       except Exception:                                                           
          logging.error('Could not parse epub.', exc_info=True)                   
      return list(outlinks)