Use PyMuPDF wheels for non-ARM architectures.

This removes the need to build the PyMuPDF project by ourselves, but
only when on non-ARM architectures since the wheels for these are not
provided yet.

Changes the `Dockerfile` and `build-image.py` script, introducing a new
`ARCH` flag to conditionally build the wheels.

Dockerfile

@@ -2,12 +2,23 @@
 # Build PyMuPDF
 
 FROM alpine:latest as pymupdf-build
-
+ARG ARCH
 ARG REQUIREMENTS_TXT
 
+RUN mkdir -p /usr/lib/python3.12/site-packages/PyMuPDFb.libs
 # Install PyMuPDF via hash-checked requirements file
 COPY ${REQUIREMENTS_TXT} /tmp/requirements.txt
-RUN apk --no-cache add linux-headers g++ linux-headers gcc make python3-dev py3-pip clang-dev
+
+# PyMuPDF provides non-arm musl wheels only.
+# Only install build-dependencies if we are actually building the wheel
+RUN case "$ARCH" in \
+ "arm64") \
+ # This is required for copying later, but is created only in the pre-built wheels
+ mkdir -p /usr/lib/python3.12/site-packages/PyMuPDFb.libs/ \
+ && apk --no-cache add linux-headers g++ linux-headers gcc make python3-dev py3-pip clang-dev ;; \
+ *) \
+ apk --no-cache add py3-pip ;; \
+ esac
 RUN pip install -vv --break-system-packages --require-hashes -r /tmp/requirements.txt
 
 
@@ -63,6 +74,7 @@ RUN apk --no-cache -U upgrade && \
 
 COPY --from=pymupdf-build /usr/lib/python3.12/site-packages/fitz/ /usr/lib/python3.12/site-packages/fitz
 COPY --from=pymupdf-build /usr/lib/python3.12/site-packages/pymupdf/ /usr/lib/python3.12/site-packages/pymupdf
+COPY --from=pymupdf-build /usr/lib/python3.12/site-packages/PyMuPDFb.libs/ /usr/lib/python3.12/site-packages/PyMuPDFb.libs
 COPY --from=tessdata-dl /usr/share/tessdata/ /usr/share/tessdata
 COPY --from=h2orestart-dl /libreoffice_ext/ /libreoffice_ext
 

install/common/build-image.py

@@ -14,6 +14,8 @@
 elif platform.system() == "Linux":
  CONTAINER_RUNTIME = "podman"
 
+ARCH = platform.machine()
+
 
 def main():
  parser = argparse.ArgumentParser()
@@ -37,6 +39,8 @@ def main():
  )
  args = parser.parse_args()
 
+ print(f"Building for architecture '{ARCH}'")
+
  print("Exporting container pip dependencies")
  with ContainerPipDependencies():
  print("Pulling base image")
@@ -57,6 +61,8 @@ def main():
  BUILD_CONTEXT,
  "--build-arg",
  f"REQUIREMENTS_TXT={REQUIREMENTS_TXT}",
+ "--build-arg",
+ f"ARCH={ARCH}",
  "-f",
  "Dockerfile",
  "--tag",
@@ -121,7 +127,12 @@ def __enter__(self):
  # XXX Export container dependencies and exclude pymupdfb since it is not needed in container
  req_txt_pymupdfb_stripped = container_requirements_txt.split("pymupdfb")[0]
  with open(Path(BUILD_CONTEXT) / REQUIREMENTS_TXT, "w") as f:
- f.write(req_txt_pymupdfb_stripped)
+ if ARCH == "arm64":
+ # PyMuPDF needs to be built on ARM64 machines
+ # But is already provided as a prebuilt-wheel on other architectures
+ f.write(req_txt_pymupdfb_stripped)
+ else:
+ f.write(container_requirements_txt)
 
  def __exit__(self, exc_type, exc_value, exc_tb):
  print("Leaving the context...")