Merge pull request #4 from dlcs/feature/handle_pdf_err

Handle 404 when fetching PDF + bump dependencies

.github/workflows/build-image.yml

@@ -19,11 +19,11 @@ jobs:
 
       - name: Set up Docker Buildx
         id: docker-setup-buildx
-        uses: docker/setup-buildx-action@v1
+        uses: docker/setup-buildx-action@v2
 
       - name: Docker metadata
         id: docker-meta
-        uses: docker/metadata-action@v3
+        uses: docker/metadata-action@v4
         with:
           images: ghcr.io/dlcs/pdf-to-alto
           tags: |
@@ -36,15 +36,15 @@ jobs:
 
       - name: Login to GitHub Container Registry
         id: docker-login
-        uses: docker/login-action@v1
+        uses: docker/login-action@v2
         with:
           registry: ghcr.io
           username: ${{ github.actor }}
           password: ${{ secrets.GITHUB_TOKEN }}
 
       - name: Build and push
         id: docker-build-push
-        uses: docker/build-push-action@v2
+        uses: docker/build-push-action@v4
         with:
           context: .
           builder: ${{ steps.docker-setup-buildx.outputs.name }}

Dockerfile

@@ -1,29 +1,19 @@
-FROM debian:bullseye as build
-
-# avoid issue with packages requiring interaction (e.g. tzdata)
-ARG DEBIAN_FRONTEND=noninteractive
-RUN apt-get update && apt-get install -y wget cmake clang git autoconf pkg-config
-
-# Change submodule to https as we're cloning only. Avoids issues with ssh
-# 8bb209c0c21476ee904a is 0.4 with some bugfixes
-RUN mkdir /home/pdfalto && cd /home/pdfalto \
-    && git clone https://github.com/kermitt2/pdfalto.git && cd pdfalto && git checkout 8bb209c0c21476ee904a && ./install_deps.sh \
-    && git submodule set-url xpdf-4.03 https://github.com/kermitt2/xpdf-4.03.git && git submodule update --init --recursive \
-    && cmake ./ && make
-
-FROM python:3.9-slim
+FROM python:3.11-slim
 
 LABEL maintainer="Donald Gray <[email protected]>"
 LABEL org.opencontainers.image.source=https://github.com/dlcs/pdf-to-alto
 LABEL org.opencontainers.image.description="Extract ALTO from PDF"
 
-COPY --from=build  /home/pdfalto/pdfalto/pdfalto /usr/bin/pdfalto
+COPY /deps/pdfalto /usr/bin/pdfalto
 
 COPY requirements.txt /opt/app/requirements.txt
 
 WORKDIR /opt/app
 RUN pip install --no-cache-dir -r requirements.txt
-COPY . /opt/app
+
+COPY app /opt/app/app
+COPY monitor.py /opt/app/monitor.py
+COPY wait-for-localstack.sh /opt/app/wait-for-localstack.sh
 
 RUN chmod +x wait-for-localstack.sh
 

app/pdf_processor.py

@@ -82,6 +82,8 @@ def _create_work_folder(self):
     def _download_pdf(self, target_file: Path):
         try:
             download_request = requests.get(self.pdf_location, stream=True)
+            download_request.raise_for_status()
+
             with open(target_file, "wb") as file:
                 for chunk in download_request.iter_content(DOWNLOAD_CHUNK_SIZE):
                     file.write(chunk)

compose/localstack/Dockerfile

@@ -1,2 +1,2 @@
-FROM localstack/localstack
-COPY seed-resources.sh /docker-entrypoint-initaws.d/
+FROM localstack/localstack:2.2.0
+COPY seed-resources.sh /etc/localstack/init/ready.d/

deps/readme.md

@@ -0,0 +1,18 @@
+# Dependencies
+
+pdfalto is a required binary from https://github.com/kermitt2/pdfalto.git. Built using:
+
+```dockerfile
+FROM debian:bullseye as build
+
+# avoid issue with packages requiring interaction (e.g. tzdata)
+ARG DEBIAN_FRONTEND=noninteractive
+RUN apt-get update && apt-get install -y wget cmake clang git autoconf pkg-config
+
+# Change submodule to https as we're cloning only. Avoids issues with ssh
+# 8bb209c0c21476ee904a is 0.4 with some bugfixes
+RUN mkdir /home/pdfalto && cd /home/pdfalto \
+    && git clone https://github.com/kermitt2/pdfalto.git && cd pdfalto && git checkout 8bb209c0c21476ee904a && ./install_deps.sh \
+    && git submodule set-url xpdf-4.03 https://github.com/kermitt2/xpdf-4.03.git && git submodule update --init --recursive \
+    && cmake ./ && make
+```

requirements.txt

@@ -6,9 +6,9 @@ colorama==0.4.4
 idna==3.3
 jmespath==0.10.0
 logzero==1.7.0
-lxml==4.7.1
+lxml==4.9.3
 pycryptodome==3.12.0
-PyMuPDF==1.19.4
+PyMuPDF==1.22.5
 python-dateutil==2.8.2
 requests==2.27.1
 s3transfer==0.5.0