bugfix: allow loading truncated images, enable Pillow 10.4.0

Author: stchris

Dockerfile

@@ -25,7 +25,7 @@ RUN apt-get -qq -y update \
     # image processing, djvu
     imagemagick-common imagemagick mdbtools djvulibre-bin \
     libtiff5-dev libjpeg-dev libfreetype6-dev libwebp-dev \
-    libtiff-tools ghostscript librsvg2-bin jbig2dec \
+    libtiff-tools ghostscript librsvg2-bin jbig2dec libopenjp2-7-dev \
     pst-utils \
     ### tesseract
     tesseract-ocr-eng \
@@ -126,7 +126,7 @@ RUN mkdir /models/ && \
 COPY requirements.txt /tmp/
 RUN pip3 install --no-cache-dir --prefer-binary --upgrade pip
 RUN pip3 install --no-cache-dir --prefer-binary --upgrade setuptools wheel
-RUN pip3 install --no-cache-dir --no-binary "tesserocr" -r /tmp/requirements.txt
+RUN pip3 install --no-cache-dir --no-binary "tesserocr" --no-binary "Pillow" -r /tmp/requirements.txt
 
 # Install spaCy models
 RUN python3 -m spacy download en_core_web_sm \

ingestors/media/image.py

@@ -1,6 +1,6 @@
 import logging
 from io import BytesIO
-from PIL import Image, ExifTags
+from PIL import Image, ExifTags, ImageFile
 from followthemoney import model
 
 from ingestors.ingestor import Ingestor
@@ -10,6 +10,9 @@
 
 log = logging.getLogger(__name__)
 
+# from https://stackoverflow.com/a/47958486
+ImageFile.LOAD_TRUNCATED_IMAGES = True
+
 
 class ImageIngestor(Ingestor, OCRSupport, TimestampSupport):
     """Image file ingestor class. Extracts the text from images using OCR."""

ingestors/support/ocr.py

@@ -3,7 +3,7 @@
 import threading
 from hashlib import sha1
 from normality import stringify
-from PIL import Image
+from PIL import Image, ImageFile
 from io import BytesIO
 from languagecodes import list_to_alpha3 as alpha3
 
@@ -13,6 +13,7 @@
 
 log = logging.getLogger(__name__)
 TESSERACT_LOCALE = "C"
+ImageFile.LOAD_TRUNCATED_IMAGES = True
 
 
 class OCRSupport(CacheSupport):

requirements.txt

@@ -30,7 +30,7 @@ odfpy==1.4.1
 cchardet==2.1.7
 lxml==5.0.0
 olefile==0.47
-Pillow==10.3.0
+Pillow==10.4.0
 vobject==0.9.6.1
 msglite==0.30.0
 icalendar==5.0.12