container: Grab trained OCR models from GitHub

Grab Tesseract's trained models from GitHub, instead of from the Alpine
Linux repos. Over the past few months, the models in the Alpine Linux
repos did not remain stable, leading to CI issues.

Since the models are already pre-trained and available through
Tesseract's repo on GitHub, we can use the release tarball that they
offer to install them in the container image, which is basically what
the upstream packages are doing as well.

In order to make sure that we have no regressions, at the time of this
commit we ensured that the hashes of the models offered through the
Alpine Linux repos and the models offered from the GitHub release are
the same. Also, in order to detect future regressions or foul play, we
check the downloaded models against a known checksum. Given that these
models change every few years, updating the checksum should not be an
issue.

Fix #357

container/Dockerfile

@@ -1,5 +1,7 @@
 FROM alpine:latest
 
+ARG TESSDATA_CHECKSUM=990fffb9b7a9b52dc9a2d053a9ef6852ca2b72bd8dfb22988b0b990a700fd3c7
+
 # Install dependencies
 RUN apk -U upgrade && \
     apk add \
@@ -11,71 +13,25 @@ RUN apk -U upgrade && \
     poppler-data \
     python3 \
     py3-magic \
-    tesseract-ocr \
-    tesseract-ocr-data-afr \
-    tesseract-ocr-data-ara \
-    tesseract-ocr-data-aze \
-    tesseract-ocr-data-bel \
-    tesseract-ocr-data-ben \
-    tesseract-ocr-data-bul \
-    tesseract-ocr-data-cat \
-    tesseract-ocr-data-ces \
-    tesseract-ocr-data-chi_sim \
-    tesseract-ocr-data-chi_tra \
-    tesseract-ocr-data-chr \
-    tesseract-ocr-data-dan \
-    tesseract-ocr-data-deu \
-    tesseract-ocr-data-grc \
-    tesseract-ocr-data-enm \
-    tesseract-ocr-data-epo \
-    tesseract-ocr-data-equ \
-    tesseract-ocr-data-est \
-    tesseract-ocr-data-eus \
-    tesseract-ocr-data-fin \
-    tesseract-ocr-data-fra \
-    tesseract-ocr-data-frk \
-    tesseract-ocr-data-frm \
-    tesseract-ocr-data-glg \
-    tesseract-ocr-data-grc \
-    tesseract-ocr-data-heb \
-    tesseract-ocr-data-hin \
-    tesseract-ocr-data-hrv \
-    tesseract-ocr-data-hun \
-    tesseract-ocr-data-ind \
-    tesseract-ocr-data-isl \
-    tesseract-ocr-data-ita \
-    tesseract-ocr-data-ita_old \
-    tesseract-ocr-data-jpn \
-    tesseract-ocr-data-kan \
-    tesseract-ocr-data-kat \
-    tesseract-ocr-data-kor \
-    tesseract-ocr-data-lav \
-    tesseract-ocr-data-lit \
-    tesseract-ocr-data-mal \
-    tesseract-ocr-data-mkd \
-    tesseract-ocr-data-mlt \
-    tesseract-ocr-data-msa \
-    tesseract-ocr-data-nld \
-    tesseract-ocr-data-nor \
-    tesseract-ocr-data-pol \
-    tesseract-ocr-data-por \
-    tesseract-ocr-data-ron \
-    tesseract-ocr-data-rus \
-    tesseract-ocr-data-slk \
-    tesseract-ocr-data-slv \
-    tesseract-ocr-data-spa \
-    tesseract-ocr-data-spa_old \
-    tesseract-ocr-data-sqi \
-    tesseract-ocr-data-srp \
-    tesseract-ocr-data-swa \
-    tesseract-ocr-data-swe \
-    tesseract-ocr-data-tam \
-    tesseract-ocr-data-tel \
-    tesseract-ocr-data-tgl \
-    tesseract-ocr-data-tha \
-    tesseract-ocr-data-tur \
-    tesseract-ocr-data-ukr \
-    tesseract-ocr-data-vie
+    tesseract-ocr
+
+# Download the trained models from the latest GitHub release of Tesseract, and
+# store them under /usr/share/tessdata. This is basically what distro packages
+# do under the hood.
+#
+# Because the GitHub release contains more files than just the trained models,
+# we use `find` to fetch only the '*.traineddata' files in the top directory.
+#
+# Before we untar the models, we also check if the checksum is the expected one.
+RUN mkdir tessdata && cd tessdata \
+    && TESSDATA_VERSION=$(wget -O- -nv https://api.github.com/repos/tesseract-ocr/tessdata/releases/latest \
+        | sed -n 's/^.*"tag_name": "\([0-9.]\+\)".*$/\1/p') \
+    && apk --purge del jq \
+    && wget https://github.com/tesseract-ocr/tessdata/archive/$TESSDATA_VERSION/tessdata-$TESSDATA_VERSION.tar.gz \
+    && echo "$TESSDATA_CHECKSUM  tessdata-$TESSDATA_VERSION.tar.gz" | sha256sum -c \
+    && tar -xzvf tessdata-$TESSDATA_VERSION.tar.gz -C . \
+    && find . -name '*.traineddata' -maxdepth 2 -exec cp {} /usr/share/tessdata \; \
+    && cd .. && rm -r tessdata
 
 COPY dangerzone.py /usr/local/bin/
 RUN chmod +x /usr/local/bin/dangerzone.py