So in pymupdf documentation states that PyMuPDF does not support running on multiple threads So they use multiprocessing, and they do this weird thing with segments in example code: seg_size = int(num_pages / cpu + 1) seg_from = idx * seg_size seg_to = min(seg_from + seg_size, num_pages) for i in range(seg_from, seg_to): # work through our page segment page = doc # page.get_text("rawdict") # use any page-related type of work here, eg Why not load document first->get number of pages and then pass number to handler function? instead of using segments as in example code ? Would this cause issues? def extract_text_from_page(args: Tuple[bytes, int]) -> Tuple[int, str]: pdf_stream, page_num = args # Open a new Document instance in this process doc = pymupdf.open(stream=pdf_stream) page = doc.load_page(page_num) # Load the specific page text = page.get_text(sort=True) # Extract text with sorting doc.close() # Clean up return (page_num, text) Continue reading...
