Allow multipage CCITT group 4 TIFF images

src/img2pdf.py

@@ -861,6 +861,30 @@ def read_images(rawdata, colorspace, first_frame_only=False):
         return [(color, ndpi, imgformat, pngidat, imgwidthpx, imgheightpx,
                  palette, False)]
 
+    # If our input is not JPEG or PNG, then we might have a format that
+    # supports multiple frames (like TIFF or GIF), so we need a loop to
+    # iterate through all frames of the image.
+    #
+    # Each frame gets compressed using PNG compression *except* if:
+    #
+    #  * The image is monochrome => encode using CCITT group 4
+    #
+    #  * The image is CMYK => zip plain RGB data
+    #
+    #  * We are handling a CCITT encoded TIFF frame => embed data
+
+    result = []
+    img_page_count = 0
+    # loop through all frames of the image (example: multipage TIFF)
+    while True:
+        try:
+            imgdata.seek(img_page_count)
+        except EOFError:
+            break
+
+        if first_frame_only and img_page_count > 0:
+            break
+
         # We can directly copy the data out of a CCITT Group 4 encoded TIFF, if it
         # only contains a single strip
         if imgformat == ImageFormat.TIFF \
@@ -878,7 +902,6 @@ def read_images(rawdata, colorspace, first_frame_only=False):
             offset, length = ccitt_payload_location_from_pil(imgdata)
             im.seek(offset)
             rawdata = im.read(length)
-        im.close()
             fillorder = imgdata.tag_v2.get(TiffImagePlugin.FILLORDER)
             if fillorder is None:
                 # no FillOrder: nothing to do
@@ -896,22 +919,10 @@ def read_images(rawdata, colorspace, first_frame_only=False):
             else:
                 raise ValueError("unsupported FillOrder: %d" % fillorder)
             logging.debug("read_images() embeds a TIFF")
-        return [(color, ndpi, ImageFormat.CCITTGroup4, rawdata, imgwidthpx,
+            result.append((color, ndpi, ImageFormat.CCITTGroup4, rawdata,
-                 imgheightpx, [], inverted)]
+                           imgwidthpx, imgheightpx, [], inverted))
-
+            img_page_count += 1
-    # Everything else has to be encoded
+            continue
-
-    result = []
-    img_page_count = 0
-    # loop through all frames of the image (example: multipage TIFF)
-    while True:
-        try:
-            imgdata.seek(img_page_count)
-        except EOFError:
-            break
-
-        if first_frame_only and img_page_count > 0:
-            break
 
         logging.debug("Converting frame: %d" % img_page_count)