Use -slim python image

OCR/dev-dockerfile

@@ -1,16 +1,15 @@
-FROM python:3.10-bullseye
+FROM python:3.10-slim
 
-RUN apt-get update && apt-get install ffmpeg libsm6 libxext6 tesseract-ocr-eng tesseract-ocr -y
+RUN apt-get update && apt-get install libgl1 tesseract-ocr-eng tesseract-ocr -y
 
 # Use larger "best" training data, rather than "fast"
 # Python one-liner because we don't have curl or wget
-RUN python3 -c 'from urllib.request import urlopen; print(urlopen("https://github.com/tesseract-ocr/tessdata_best/raw/refs/heads/main/eng.traineddata").read())' > /usr/share/tesseract-ocr/4.00/tessdata/eng.traineddata
+RUN python3 -c 'from urllib.request import urlopen; print(urlopen("https://github.com/tesseract-ocr/tessdata_best/raw/refs/heads/main/eng.traineddata").read())' > /usr/share/tesseract-ocr/5/tessdata/eng.traineddata
 
 RUN pip install poetry
 
 COPY ./pyproject.toml /ocr/pyproject.toml
 COPY ./poetry.lock /ocr/poetry.lock
 
 WORKDIR /ocr
-RUN poetry install
-
+RUN poetry install && poetry cache list | xargs -n1 poetry cache clear --all

OCR/ocr/services/tesseract_ocr.py

@@ -36,6 +36,7 @@ def _guess_tessdata_path(wanted_lang="eng") -> bytes:
             "/usr/local/share/tesseract/tessdata",
             "/usr/share/tesseract/tessdata",
             "/usr/share/tesseract-ocr/4.00/tessdata",
+            "/usr/share/tesseract-ocr/5/tessdata",
             "/opt/homebrew/share/tessdata",
             "/opt/local/share/tessdata",
         ]