Bump Python image to 3.9

Author: catileptic

Dockerfile

@@ -1,30 +1,29 @@
-FROM ubuntu:20.04
+FROM python:3.9-bookworm
 ENV DEBIAN_FRONTEND noninteractive
 
 LABEL org.opencontainers.image.title "FollowTheMoney File Ingestors"
 LABEL org.opencontainers.image.licenses MIT
 LABEL org.opencontainers.image.source https://github.com/alephdata/ingest-file
 
 # Enable non-free archive for `unrar`.
-# RUN echo "deb http://http.us.debian.org/debian stretch non-free" >/etc/apt/sources.list.d/nonfree.list
-RUN apt-get -qq -y update \
-    && apt-get -qq -y install build-essential locales ca-certificates \
+RUN echo "deb http://http.us.debian.org/debian stable non-free" >/etc/apt/sources.list.d/nonfree.list \
+    && apt-get -qq -y update \
+    && apt-get -qq -y install build-essential locales \
     # python deps (mostly to install their dependencies)
-    python3-pip python3-dev python3-pil \
+    python3-dev \
     # tesseract
-    tesseract-ocr libtesseract-dev libleptonica-dev pkg-config\
+    tesseract-ocr libtesseract-dev libleptonica-dev \
     # libraries
-    libxslt1-dev libpq-dev libldap2-dev libsasl2-dev \
-    zlib1g-dev libicu-dev libxml2-dev \
+    libldap2-dev libsasl2-dev \
     # package tools
     unrar p7zip-full \
     # audio & video metadata
     libmediainfo-dev \
     # image processing, djvu
-    imagemagick-common imagemagick mdbtools djvulibre-bin \
-    libtiff5-dev libjpeg-dev libfreetype6-dev libwebp-dev \
+    mdbtools djvulibre-bin \
+    libtiff5-dev \
     libtiff-tools ghostscript librsvg2-bin jbig2dec \
-    pst-utils \
+    pst-utils libgif-dev \
     ### tesseract
     tesseract-ocr-eng \
     tesseract-ocr-swa \
@@ -98,7 +97,7 @@ RUN apt-get -qq -y update \
     tesseract-ocr-uzb \
     ### 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 \
+    hyphen-en-us hyphen-it hyphen-ru fonts-dejavu fonts-dejavu-extra \
     fonts-droid-fallback fonts-dustin fonts-f500 fonts-fanwood fonts-freefont-ttf \
     fonts-liberation fonts-lmodern fonts-lyx fonts-sil-gentium fonts-texgyre \
     fonts-tlwg-purisa \
@@ -121,11 +120,7 @@ RUN groupadd -g 1000 -r app \
 RUN mkdir /models/ && \
     curl -o "/models/model_type_prediction.ftz" "https://public.data.occrp.org/develop/models/types/type-08012020-7a69d1b.ftz"
 
-# Having updated pip/setuptools seems to break the test run for some reason (12/01/2022)
-# RUN pip3 install --no-cache-dir -U pip setuptools
 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
 
 # Install spaCy models
@@ -148,14 +143,15 @@ RUN python3 -m spacy download el_core_news_sm \
 
 COPY . /ingestors
 WORKDIR /ingestors
-RUN pip3 install --no-cache-dir --config-settings editable_mode=compat --use-pep517 -e /ingestors
+RUN pip install --no-cache-dir --config-settings editable_mode=compat --use-pep517 -e /ingestors
 RUN chown -R app:app /ingestors
 
 ENV ARCHIVE_TYPE=file \
     ARCHIVE_PATH=/data \
     FTM_STORE_URI=postgresql://aleph:aleph@postgres/aleph \
     REDIS_URL=redis://redis:6379/0 \
-    TESSDATA_PREFIX=/usr/share/tesseract-ocr/4.00/tessdata
+    TESSDATA_PREFIX=/usr/share/tesseract-ocr/5/tessdata \
+    LD_PRELOAD=/usr/lib/aarch64-linux-gnu/libgomp.so.1
 
 # USER app
 CMD ingestors process

Makefile

@@ -1,5 +1,5 @@
 INGEST=ghcr.io/alephdata/ingest-file
-COMPOSE=docker-compose
+COMPOSE=docker compose
 DOCKER=$(COMPOSE) run --rm ingest-file
 
 .PHONY: build
@@ -36,7 +36,7 @@ format-check:
 	black --check .
 
 test: services
-	$(DOCKER) pytest --cov=ingestors --cov-report html --cov-report term
+	PYTHONDEVMODE=1 PYTHONTRACEMALLOC=1 $(DOCKER) pytest --cov=ingestors --cov-report html --cov-report term
 
 restart: build
 	$(COMPOSE) up --force-recreate --no-deps --detach ingest-file

docker-compose.yml

@@ -1,5 +1,3 @@
-version: "3.2"
-
 services:
   postgres:
     image: postgres:10.0
@@ -12,6 +10,9 @@ services:
     image: redis:alpine
     command: ["redis-server", "--save", "3600", "10"]
 
+  rabbitmq:
+    image: rabbitmq:3.9-management-alpine
+
   ingest-file:
     build:
       context: .
@@ -22,7 +23,7 @@ services:
       - /data:mode=777
     environment:
       FTM_STORE_URI: postgresql://ingest:ingest@postgres/ingest
-      LOG_FORMAT: TEXT  # TEXT or JSON
+      LOG_FORMAT: TEXT # TEXT or JSON
     volumes:
       - "./ingestors:/ingestors/ingestors"
       - "./tests:/ingestors/tests"
@@ -33,3 +34,4 @@ services:
     depends_on:
       - postgres
       - redis
+      - rabbitmq

requirements.txt

@@ -3,15 +3,17 @@ normality==2.5.0
 pantomime==0.6.1
 followthemoney==3.5.9
 followthemoney-store[postgresql]==3.1.0
-servicelayer[google,amazon]==1.22.2
+servicelayer[google,amazon]==1.23.0
 languagecodes==1.1.1
 countrytagger==0.1.2
 pyicu==2.12
 google-cloud-vision==3.7.2
-tesserocr==2.6.2
-spacy==3.6.1
+tesserocr==2.7.1
+spacy==3.6.1 # pinned because spacy 3.8 requires numpy >2 which breaks fasttext (see https://groups.google.com/g/fasttext-library/c/4EOM0-S6xHU)
+numpy<2.0.0 # pinned because otherwise spacy requires an incompatible numpy
 fingerprints==1.1.1
 fasttext==0.9.2
+pika==1.3.2
 
 # Development
 pytest==8.2.0
@@ -38,4 +40,5 @@ cryptography==41.0.7
 requests[security]==2.31.0
 pymupdf==1.21.1
 
+prometheus-client==0.17.1
 sentry_sdk==2.0.1

tests/test_audio.py

@@ -10,10 +10,12 @@ def test_audio(self):
         self.assertEqual(entity.first("processingStatus"), self.manager.STATUS_SUCCESS)
         self.assertEqual(entity.first("title"), "Core Media Audio")
         self.assertEqual(entity.first("generator"), "com.apple.VoiceMemos (iOS 11.4)")
-        self.assertEqual(
-            entity.first("authoredAt"),
+        # with the change to Python 3.9 and the debian base image we are seeing a change
+        # in the amount of data we are parsing here
+        assert entity.get("authoredAt") == [
+            datetime.datetime(2018, 6, 20, 12, 9, 28).isoformat(),
             datetime.datetime(2018, 6, 20, 12, 9, 42).isoformat(),
-        )
+        ]
         self.assertEqual(entity.first("duration"), "2808")
         self.assertEqual(entity.first("samplingRate"), "44100")
         self.assertEqual(entity.schema.name, "Audio")