Maintenance: Dockerfile, GHCR workflow, bug fixes, install docs

.dockerignore

@@ -0,0 +1,14 @@
+.git
+.github
+.gitignore
+.ruff_cache
+__pycache__
+**/__pycache__
+docs
+input
+*.md
+!README.md
+venv
+.venv
+Dockerfile
+.dockerignore

.github/workflows/docker.yml

@@ -0,0 +1,39 @@
+name: Docker
+
+on:
+  push:
+    tags: ["v*"]
+  workflow_dispatch:
+
+jobs:
+  publish:
+    runs-on: ubuntu-latest
+    permissions:
+      contents: read
+      packages: write
+    steps:
+      - uses: actions/checkout@v7
+
+      - name: Log in to GHCR
+        uses: docker/login-action@v3
+        with:
+          registry: ghcr.io
+          username: ${{ github.actor }}
+          password: ${{ secrets.GITHUB_TOKEN }}
+
+      - name: Extract metadata
+        id: meta
+        uses: docker/metadata-action@v5
+        with:
+          images: ghcr.io/${{ github.repository }}
+          tags: |
+            type=semver,pattern={{version}}
+            type=raw,value=latest
+
+      - name: Build and push
+        uses: docker/build-push-action@v6
+        with:
+          context: .
+          push: true
+          tags: ${{ steps.meta.outputs.tags }}
+          labels: ${{ steps.meta.outputs.labels }}

.gitignore

@@ -176,3 +176,4 @@ config.yaml
 
 # macOS
 .DS_Store
+.claude/

Dockerfile

@@ -0,0 +1,21 @@
+# CPU-only image. For GPU acceleration install a CUDA/ROCm PyTorch build
+# instead (see README) or run natively.
+FROM python:3.13-slim
+
+WORKDIR /app
+
+COPY requirements.txt .
+RUN pip install --no-cache-dir -r requirements.txt
+
+COPY main.py .
+COPY modules/ modules/
+
+# marker-pdf downloads its models on first run; mount a volume at this path
+# to cache them across container runs.
+ENV HF_HOME=/models
+VOLUME /models
+
+# PDFs are read from /data/input and results written next to them by default.
+WORKDIR /data
+
+ENTRYPOINT ["python", "/app/main.py"]

README.md

@@ -30,7 +30,15 @@ Python 3.13 is recommended.
 `marker-pdf` constrains `Pillow<11.0.0`, and Pillow only ships Python 3.14
 wheels from 11.3.0 onward. On Python 3.10–3.13 every dependency installs as a
 prebuilt wheel. On 3.14, pip has to build Pillow from source instead: this
-works, but it is slower and requires a working C toolchain.
+works, but it is slower and requires a working C toolchain plus the image
+library headers Pillow links against. On Debian/Ubuntu install them first:
+
+```bash
+sudo apt install libjpeg-dev zlib1g-dev libtiff-dev libfreetype6-dev libwebp-dev
+```
+
+Without these headers the install fails with
+`RequiredDependencyException: The headers or library files could not be found for jpeg`.
 
 ## 💻 Installation
 
@@ -74,6 +82,30 @@ print(torch.backends.mps.is_available())  # Should return True for Apple Silicon
 print(torch.version.hip)  # Should print ROCm version for AMD
 ```
 
+### 🐳 Docker
+
+A CPU-only image can be built from the included `Dockerfile`:
+
+```bash
+docker build -t pdf2epub .
+```
+
+Run it with your PDFs mounted at `/data` and a model cache volume (marker-pdf
+downloads its models on first run):
+
+```bash
+docker run -it --rm \
+  -v "$(pwd)":/data \
+  -v pdf2epub-models:/models \
+  pdf2epub input.pdf
+```
+
+`-it` is required for EPUB generation because metadata is prompted
+interactively; with `--skip-epub` it can run non-interactively.
+
+Tagged releases are also published to
+`ghcr.io/overcuriousity/pdf2epub` by the Docker workflow.
+
 ## 🚀 Usage
 
 ### Basic Usage

main.py

@@ -11,7 +11,7 @@
 def main():
     if torch.cuda.is_available():
         print("CUDA is available. Using GPU for processing.")
-    elif torch.mps.is_available():
+    elif torch.backends.mps.is_available():
         print("MPS is available. Using Apple Silicon for processing.")
     else:
         print("CUDA is not available. Using CPU for processing.")

modules/mark2epub.py

@@ -58,7 +58,12 @@ def review_markdown(markdown_path: Path) -> tuple[bool, str]:
         response = input("\nWould you like to review the markdown file before conversion? (y/n): ").lower()
         if response in ['y', 'yes']:
             try:
-                subprocess.run(['xdg-open' if os.name == 'posix' else 'start', str(markdown_path)], check=True)
+                if sys.platform == 'darwin':
+                    subprocess.run(['open', str(markdown_path)], check=True)
+                elif os.name == 'posix':
+                    subprocess.run(['xdg-open', str(markdown_path)], check=True)
+                else:
+                    os.startfile(str(markdown_path))
 
                 while True:
                     proceed = input("\nPress Enter when you're done editing (or 'q' to abort): ").lower()
@@ -333,8 +338,8 @@ def get_coverpage_XML(title, authors):
 </head>
 <body>
     <div class="cover">
-        <h1>{title}</h1>
-        <p>{authors}</p>
+        <h1>{xml_escape(title)}</h1>
+        <p>{xml_escape(authors) if authors else ''}</p>
     </div>
 </body>
 </html>"""
@@ -355,7 +360,7 @@ def get_TOC_XML(default_css_filenames, markdown_filenames):
     for i,md_filename in enumerate(markdown_filenames):
         stem = md_filename.split(".")[0]
         href = quote("s{:05d}-{}.xhtml".format(i, stem))
-        toc_xhtml += """<li><a href="{}">{}</a></li>""".format(href, stem)
+        toc_xhtml += """<li><a href="{}">{}</a></li>""".format(href, xml_escape(stem))
     toc_xhtml += """</ol>\n</nav>\n</body>\n</html>"""
 
     return toc_xhtml
@@ -377,7 +382,7 @@ def get_TOCNCX_XML(markdown_filenames, uid="", title=""):
         stem = md_filename.split(".")[0]
         src = quote("s{:05d}-{}.xhtml".format(i, stem))
         toc_ncx += """<navPoint id="navpoint-{}" playOrder="{}">\n""".format(i, i + 1)
-        toc_ncx += """<navLabel>\n<text>{}</text>\n</navLabel>""".format(stem)
+        toc_ncx += """<navLabel>\n<text>{}</text>\n</navLabel>""".format(xml_escape(stem))
         toc_ncx += """<content src="{}"/>""".format(src)
         toc_ncx += """ </navPoint>"""
     toc_ncx += """</navMap>\n</ncx>"""