display progress for extracting pdf images as well as for analyzing them

dlibrary/dlibrary.py

@@ -382,16 +382,20 @@ class Collator:
 
     def link_pdf(self, src):
         with fitz.open(src) as pdf:
-            images = pdf_images(pdf, self.args.force_convert_pdf)
+            images = pdf_images(pdf, self.args.pdf_strategy)
             if images is None:
                 print(f'Failed to enumerate page images in PDF, skipping {src}')
                 return None
 
             self.dest.mkdir(parents=True, exist_ok=True)
+
+            print(f'0 pages collated...', end='')
             for (idx, image) in enumerate(images, start=self.index):
                 file_path = self.dest / f'{idx:04d}.{image["ext"]}'
                 with open(file_path, 'wb') as f:
                     f.write(image["image"])
+                print(f'\x1b[2K\r{idx+1-self.index} pages collated...', end='')
+            print()
 
             self.index += pdf.page_count
             return True
@@ -569,13 +573,13 @@ def display_sixel_page(page):
     finally:
         sixel_output_unref(output)
 
-def pdf_images(pdf, force=False):
+def pdf_images(pdf, strategy):
     images_by_page = [(page.get_images(), is_single_image(page)) for page in pdf]
     if all(len(images) == 1 and single for (images, single) in images_by_page):
         return (extract_image(pdf, images[0][0]) for (images, _) in images_by_page)
 
     print("Checking PDF images the quick way failed, trying the slow way")
-    print(f'0/{pdf.page_count} pages processed...', end='')
+    print(f'0/{pdf.page_count} pages analyzed...', end='')
     image_extractors = []
     for (idx, page) in enumerate(pdf):
         page_images = page.get_image_info(xrefs=True)
@@ -587,25 +591,30 @@ def pdf_images(pdf, force=False):
             image_extractors.append(lambda p=pdf, x=xref: extract_image(p, x))
         else:
             print(f'\nPage {idx+1}: {len(page_images)} images, {len([img for img in page_images if img["xref"] == 0])} non-xref images, {len(relevant_blocks(page))} total relevant objects')
-            if force:
+            choice = strategy
-                print(f'Converting page {idx+1}')
+            while True:
-                choice = 'c'
+                if choice.lower().startswith('n'):
-            else:
+                    return None
-                shown = False
+                if choice.lower().startswith('c'):
-                while True:
+                    if choice == strategy:
-                    choice = input(f'[N]ope out of this PDF / [c]onvert page to image{"" if xref is None else " / e[x]tract embedded image without text"}{"" if shown else " / [s]how page before deciding"}? [N/c{"" if xref is None else "/x"}{"" if shown else "/s"}] ')
+                        print(f'Converting page {idx+1}')
-                    if not shown and choice != '' and choice[0].lower() == 's':
+                    image_extractors.append(lambda p=page: { 'ext': 'png', 'image': p.get_pixmap(dpi=PDF_CONVERSION_DPI).tobytes('png') })
-                        display_sixel_page(page)
+                    break
-                        shown = True
+                if xref is not None and (choice.lower().startswith('x') or choice.lower() == 'extract'):
-                    else:
+                    if choice == strategy:
-                        break
+                        print(f'Extracting image from page {idx+1} without text')
-            if xref is not None and choice != '' and choice[0].lower() == 'x':
+                    image_extractors.append(lambda p=pdf, x=xref: extract_image(p, x))
-                image_extractors.append(lambda p=pdf, x=xref: extract_image(p, x))
+                    break
-            elif choice != '' and choice[0].lower() == 'c':
+                if choice.lower().startswith('d'):
-                image_extractors.append(lambda p=page: { 'ext': 'png', 'image': p.get_pixmap(dpi=PDF_CONVERSION_DPI).tobytes('png') })
+                    if choice == strategy:
-            else:
+                        print(f'Dropping page {idx+1}')
-                return None
+                    break
-        print(f'\x1b[2K\r{idx+1}/{pdf.page_count} pages processed...', end=('' if idx+1 < pdf.page_count else '\n'))
+
+                if choice.lower().startswith('s'):
+                    display_sixel_page(page)
+
+                choice = input(f'[N]ope out / [c]onvert page{"" if xref is None else " / e[x]tract image"} / [d]rop page / [s]how page? [n/c{"" if xref is None else "/x"}/d/s] ')
+        print(f'\x1b[2K\r{idx+1}/{pdf.page_count} pages analyzed...', end=('' if idx+1 < pdf.page_count else '\n'))
 
     return (extractor() for extractor in image_extractors)
 
@@ -1182,9 +1191,17 @@ parser_collate.add_argument(
     help="only collate works that are explicitly specified",
 )
 parser_collate.add_argument(
-    '--force-convert-pdf',
+    '-p', '--pdf-strategy',
-    action='store_true',
+    choices=[
-    help="convert a PDF page to a 300dpi image if there isn't a single image we can extract directly",
+        'ask', '?',
+        'show-ask', 's',
+        'convert', 'c',
+        'extract', 'x',
+        'drop', 'd',
+        'nope', 'n'
+    ],
+    default='show-ask',
+    help="how to handle PDF pages that aren't a single image with no text",
 )
 parser_collate.add_argument(
     'expression',