Release/3.20.1

.bumpversion.cfg

@@ -1,5 +1,5 @@
 [bumpversion]
-current_version = 3.20.0
+current_version = 3.20.1
 tag_name = {new_version}
 commit = True
 tag = True

Dockerfile

@@ -25,7 +25,7 @@ RUN apt-get -qq -y update \
     libtiff5-dev libjpeg-dev libfreetype6-dev libwebp-dev \
     libtiff-tools ghostscript librsvg2-bin jbig2dec \
     pst-utils \
-    ### tesseract 
+    ### tesseract
     tesseract-ocr-eng \
     tesseract-ocr-swa \
     tesseract-ocr-swe \
@@ -96,7 +96,7 @@ RUN apt-get -qq -y update \
     tesseract-ocr-aze \
     tesseract-ocr-bel \
     tesseract-ocr-uzb \
-    ### pdf convert: libreoffice + a bunch of fonts 
+    ### pdf convert: libreoffice + a bunch of fonts
     libreoffice fonts-opensymbol hyphen-fr hyphen-de \
     hyphen-en-us hyphen-it hyphen-ru fonts-dejavu fonts-dejavu-core fonts-dejavu-extra \
     fonts-droid-fallback fonts-dustin fonts-f500 fonts-fanwood fonts-freefont-ttf \
@@ -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 --prefer-binary -r /tmp/requirements.txt
+RUN pip3 install --no-cache-dir --no-binary "tesserocr" -r /tmp/requirements.txt
 
 # Install spaCy models
 RUN python3 -m spacy download en_core_web_sm \
@@ -136,7 +136,7 @@ RUN python3 -m spacy download en_core_web_sm \
 RUN python3 -m spacy download ru_core_news_sm \
     && python3 -m spacy download pt_core_news_sm \
     && python3 -m spacy download ro_core_news_sm \
-    && python3 -m spacy download mk_core_news_sm 
+    && python3 -m spacy download mk_core_news_sm
 RUN python3 -m spacy download el_core_news_sm \
     && python3 -m spacy download pl_core_news_sm \
     && python3 -m spacy download it_core_news_sm \

ingestors/__init__.py

@@ -1,7 +1,7 @@
 """Provides a set of ingestors based on different file types."""
 import logging
 
-__version__ = "3.20.0"
+__version__ = "3.20.1"
 
 logging.getLogger("chardet").setLevel(logging.INFO)
 logging.getLogger("PIL").setLevel(logging.INFO)

ingestors/cli.py

@@ -8,6 +8,7 @@
 from servicelayer.jobs import Job, Dataset
 from servicelayer import settings as sl_settings
 from servicelayer.archive.util import ensure_path
+from servicelayer.tags import Tags
 
 from ingestors import settings
 from ingestors.manager import Manager
@@ -112,5 +113,18 @@ def debug(path, languages=None):
         pprint(entity.to_dict())
 
 
+@cli.command()
+@click.argument(
+    "prefix",
+    default="",
+)
+def cache_clear(prefix):
+    """Delete all ingest cache entries.
+
+    Only delete entries with the given prefix (e.g: 'ocr:', 'pdf:').
+    """
+    Tags("ingest_cache").delete(prefix=prefix)
+
+
 if __name__ == "__main__":
     cli()

setup.py

@@ -5,7 +5,7 @@
 
 setup(
     name="ingest",
-    version="3.20.0",
+    version="3.20.1",
     author="Organized Crime and Corruption Reporting Project",
     packages=find_packages(exclude=["tests"]),
     package_dir={"ingestors": "ingestors"},

tests/test_image.py

@@ -10,15 +10,14 @@ def test_match(self):
     def test_ingest_on_svg(self):
         fixture_path, entity = self.fixture("image.svg")
         self.manager.ingest(fixture_path, entity)
-        # print result.to_dict()
 
         self.assertIn("TEST", entity.first("bodyText"))
-        # self.assertIn(u'1..2..3..', result.pages[0]['text'])
         self.assertEqual(entity.first("processingStatus"), self.manager.STATUS_SUCCESS)
 
-    def test_ingest_hand_written_text(self):
-        fixture_path, entity = self.fixture("some hand wirtten veird text.jpg")
+    def test_ingest_on_jpeg(self):
+        fixture_path, entity = self.fixture("jpegtest.jpg")
         self.manager.ingest(fixture_path, entity)
+        self.assertIn("Debian", entity.first("bodyText"))
+        self.assertEqual(entity.first("mimeType"), "image/jpeg")
 
-        # self.assert(u'Testing ingestors', result.pages[0]['text'])
         self.assertEqual(entity.first("processingStatus"), self.manager.STATUS_SUCCESS)