{xml_escape(authors) if authors else ''}
-/ blocks so their $ delimiters are never treated as math
- placeholders = {}
- counter = [0]
-
- def mask(m):
- key = f"\x00MASK{counter[0]}\x00"
- counter[0] += 1
- placeholders[key] = m.group(0)
- return key
-
- masked = re.sub(r'|', mask, html_text, flags=re.DOTALL)
-
- def try_convert(latex):
+ text = re.sub(r'\{\\rm\s+([^}]+)\}', r'\\mathrm{\1}', text)
+ text = re.sub(r'\\rm\s+([a-zA-Z0-9]+)', r'\\mathrm{\1}', text)
+ text = re.sub(r'\\rm\b', r'\\mathrm', text)
+ return text
+
+def resolve_metadata(markdown_dir: Path, md_path: Path) -> Dict[str, Any]:
+ metadata: Dict[str, Any] = {
+ "title": md_path.stem.replace("_", " ").title(),
+ "creator": "Unknown Author",
+ "language": "en",
+ "publisher": "PDF2EPUB"
+ }
+
+ desc_file = markdown_dir / "description.json"
+ if desc_file.exists():
try:
- return latex2mathml.converter.convert(latex)
+ with open(desc_file, "r", encoding="utf-8") as f:
+ data = json.load(f).get("metadata", {})
+ if data.get("dc:title"): metadata["title"] = data["dc:title"]
+ if data.get("dc:creator"): metadata["creator"] = data["dc:creator"]
+ if data.get("dc:language"): metadata["language"] = data["dc:language"]
except Exception:
- return None
-
- # Standalone display math that Markdown wrapped in : replace the whole
- # paragraph to avoid invalid XHTML like
...
- def replace_display_paragraph(m):
- mathml = try_convert(m.group(1))
- return f'{mathml}' if mathml else m.group(0)
-
- # Remaining $$...$$ (inside phrasing content): use to stay valid
- def replace_display_inline(m):
- mathml = try_convert(m.group(1))
- return f'{mathml}' if mathml else m.group(0)
-
- # Inline $...$
- def replace_inline(m):
- mathml = try_convert(m.group(1))
- return f'{mathml}' if mathml else m.group(0)
-
- masked = re.sub(r'\s*\$\$(.*?)\$\$\s*
', replace_display_paragraph, masked, flags=re.DOTALL)
- masked = re.sub(r'\$\$(.*?)\$\$', replace_display_inline, masked, flags=re.DOTALL)
- masked = re.sub(r'(? Optional[Path]:
+ """Dynamically identifies the best candidate for cover image (returns absolute path)."""
+ if not images_dir.exists():
+ return None
+
+ # 1. Look for files with explicit cover names
+ for f in images_dir.iterdir():
+ if f.is_file() and "cover" in f.name.lower() and f.suffix.lower() in [".jpg", ".jpeg", ".png"]:
+ return f.resolve()
+
+ # 2. Search for any image generated from page 0 of the PDF (sorted numerically)
+ page_zero_candidates = sorted(
+ [f for f in images_dir.iterdir() if f.is_file() and f.name.startswith("_page_0_") and f.suffix.lower() in [".jpg", ".jpeg", ".png"]],
+ key=lambda x: x.name
+ )
+ if page_zero_candidates:
+ return page_zero_candidates[0].resolve()
- return masked
+ return None
-def get_chapter_XML(work_dir: str, md_filename: str, css_filenames: list[str], content: Optional[str] = None) -> tuple[str, list[str]]:
+def convert_to_epub(markdown_dir: Path, output_path: Path) -> None:
"""
- Convert markdown chapter to XHTML and process images.
- Returns tuple of (XHTML content, list of images referenced in chapter)
-
- Args:
- work_dir: Working directory containing markdown files
- md_filename: Name of markdown file
- css_filenames: List of CSS files to include
- content: Optional pre-loaded markdown content. If None, content is read from file
+ Compiles artifacts generated by Marker into an EPUB 3 file using Pandoc.
+ Maintains strict compatibility with main.py invocations.
"""
- work_dir_path = Path(work_dir)
-
- if content is None:
- with open(work_dir_path / md_filename, "r", encoding="utf-8") as f:
- markdown_data = f.read()
- else:
- markdown_data = content
-
- # Process markdown for images and get list of referenced images
- markdown_data, chapter_images = process_markdown_for_images(markdown_data, work_dir_path)
-
- # Convert to HTML
- html_text = markdown.markdown(
- markdown_data,
- extensions=["codehilite", "tables", "fenced_code", "footnotes"],
- extension_configs={"codehilite": {"guess_lang": False}}
- )
+ check_pandoc_installed()
- # Convert LaTeX math to MathML for EPUB readers
- html_text = convert_math_to_mathml(html_text)
+ # Normalize markdown_dir to absolute path from the start
+ markdown_dir = Path(markdown_dir).resolve()
+ if not markdown_dir.exists() or not markdown_dir.is_dir():
+ raise FileNotFoundError(f"Markdown directory not found: {markdown_dir}")
- # Generate XHTML wrapper
- xhtml = f"""
-
-
-
- {''.join(f'' for css in css_filenames)}
-
-
-{html_text}
-
-"""
+ md_files = sorted(list(markdown_dir.glob("*.md")))
+ if not md_files:
+ raise ValueError(f"No .md files found in: {markdown_dir}")
- return xhtml, chapter_images
+ primary_md = md_files[0]
+ # Normalize destination to absolute path
+ output_path = Path(output_path).resolve()
+ if output_path.is_dir() or not output_path.suffix:
+ output_path.mkdir(parents=True, exist_ok=True)
+ epub_target = (markdown_dir / f"{markdown_dir.name}.epub").resolve()
+ else:
+ output_path.parent.mkdir(parents=True, exist_ok=True)
+ epub_target = output_path.resolve()
+ # Sanitize Markdown in a temporary file
+ raw_content = primary_md.read_text(encoding="utf-8")
+ sanitized_content = sanitize_math_syntax(raw_content)
-def convert_to_epub(markdown_dir: Path, output_path: Path) -> None:
- """
- Convert markdown files and images to EPUB format.
- """
- if not markdown_dir.exists():
- raise FileNotFoundError(f"Markdown directory not found: {markdown_dir}")
+ with tempfile.NamedTemporaryFile("w", encoding="utf-8", suffix=".md", delete=False) as tmp_md, \
+ tempfile.NamedTemporaryFile("w", encoding="utf-8", suffix=".css", delete=False) as tmp_css:
- if not list(markdown_dir.glob('*.md')):
- raise ValueError(f"No markdown files found in: {markdown_dir}")
-
- # Generate EPUB file
- epub_path = markdown_dir / f"{markdown_dir.name}.epub"
- main([str(markdown_dir), str(epub_path)])
-
-def main(args):
- if len(args) < 2:
- print("\nUsage:\n python md2epub.py ")
- exit(1)
+ tmp_md_path = Path(tmp_md.name).resolve()
+ tmp_css_path = Path(tmp_css.name).resolve()
- work_dir = args[0]
- output_path = args[1]
+ tmp_md.write(sanitized_content)
+ tmp_md.flush()
- images_dir = os.path.join(work_dir, 'images/')
- css_dir = os.path.join(work_dir, 'css/')
+ tmp_css.write(DEFAULT_EPUB_CSS)
+ tmp_css.flush()
try:
- # Reading/Creating the JSON file containing the description of the eBook
- description_path = os.path.join(work_dir, "description.json")
- existing_metadata = {}
-
- if os.path.exists(description_path):
- with open(description_path, 'r', encoding='utf-8') as f:
- existing_metadata = json.load(f)
-
- # Get metadata from user
- json_data = get_metadata_from_user(existing_metadata)
-
- # Find all markdown files if not already in metadata
- if not json_data["chapters"]:
- markdown_files = [f for f in os.listdir(work_dir) if f.endswith('.md')]
- for md_file in sorted(markdown_files):
- json_data["chapters"].append({
- "markdown": md_file,
- "css": ""
- })
-
- # Save the updated description.json
- with open(description_path, 'w', encoding='utf-8') as f:
- json.dump(json_data, f, indent=2)
-
- # Review markdown files and store updated content
- chapter_contents = {}
- for chapter in json_data["chapters"]:
- md_path = Path(work_dir) / chapter["markdown"]
- should_continue, content = review_markdown(md_path)
- if not should_continue:
- print("\nConversion aborted by user.")
- return
- chapter_contents[chapter["markdown"]] = content
-
- # Get title and author
- title = json_data["metadata"].get("dc:title", "Untitled Document")
- authors = json_data["metadata"].get("dc:creator", None)
-
- # Compile list of files
- all_md_filenames = []
- all_css_filenames = json_data["default_css"][:]
- for chapter in json_data["chapters"]:
- if chapter["markdown"] not in all_md_filenames:
- all_md_filenames.append(chapter["markdown"])
- if len(chapter["css"]) and (chapter["css"] not in all_css_filenames):
- all_css_filenames.append(chapter["css"])
-
- all_image_filenames = get_all_filenames(images_dir, extensions=["gif", "jpg", "jpeg", "png"])
-
- # First process all chapters and images
- images_dir = Path(work_dir) / 'images'
- epub_images_dir = Path(work_dir) / 'epub_images'
- processed_images = {} # Store processed image data
- all_referenced_images = set()
- chapter_data = {} # Store processed chapter data
-
- # First pass: Process chapters and collect image references
- print("\nProcessing chapters and collecting image references...")
- for i, chapter in enumerate(json_data["chapters"]):
- css_files = json_data["default_css"][:]
- if chapter["css"]:
- css_files.append(chapter["css"])
-
- # Process chapter content
- chapter_xhtml, chapter_images = get_chapter_XML(
- work_dir,
- chapter["markdown"],
- css_files,
- content=chapter_contents[chapter["markdown"]]
- )
- chapter_data[chapter["markdown"]] = chapter_xhtml
- all_referenced_images.update(chapter_images)
-
- # Process and optimize images
- print("\nProcessing and optimizing images...")
- if images_dir.exists() and all_referenced_images:
- epub_images_dir.mkdir(exist_ok=True)
-
- for image in all_referenced_images:
- src_path = images_dir / image
- if src_path.exists():
- try:
- dest_path = epub_images_dir / image
- copy_and_optimize_image(src_path, dest_path)
-
- # Store processed image data
- with open(dest_path, "rb") as f:
- processed_images[image] = f.read()
- except Exception as e:
- print(f"Warning: Failed to process image {image}: {e}")
- else:
- print(f"Warning: Referenced image not found: {src_path}")
-
- # Cleanup temporary directory
- import shutil
- shutil.rmtree(epub_images_dir, ignore_errors=True)
-
- # Now create the EPUB file with all prepared content
- print("\nCreating EPUB file...")
- with zipfile.ZipFile(output_path, "w") as epub:
- # Write mimetype (must be first and uncompressed)
- epub.writestr("mimetype", "application/epub+zip")
-
- # Write container.xml
- epub.writestr("META-INF/container.xml", get_container_XML(), zipfile.ZIP_DEFLATED)
-
- # Write package.opf
- epub.writestr("OPS/package.opf",
- get_packageOPF_XML(
- md_filenames=all_md_filenames,
- image_filenames=all_image_filenames,
- css_filenames=all_css_filenames,
- description_data=json_data
- ),
- zipfile.ZIP_DEFLATED
- )
-
- # Write cover page
- coverpage_data = get_coverpage_XML(title, authors)
- epub.writestr("OPS/titlepage.xhtml", coverpage_data.encode('utf-8'), zipfile.ZIP_DEFLATED)
-
- # Write processed chapters
- print("Writing chapters...")
- for i, chapter in enumerate(json_data["chapters"]):
- print(f" Writing chapter {i+1}/{len(json_data['chapters'])}: {chapter['markdown']}")
- epub.writestr(
- f"OPS/s{i:05d}-{chapter['markdown'].split('.')[0]}.xhtml",
- chapter_data[chapter["markdown"]].encode('utf-8'),
- zipfile.ZIP_DEFLATED
- )
-
- # Write processed images
- if processed_images:
- print(f"Writing {len(processed_images)} processed images...")
- for image_name, image_data in processed_images.items():
- epub.writestr(f"OPS/images/{image_name}", image_data, zipfile.ZIP_DEFLATED)
-
- # Write TOC files
- print("Writing table of contents...")
- epub.writestr("OPS/TOC.xhtml",
- get_TOC_XML(json_data["default_css"], all_md_filenames),
- zipfile.ZIP_DEFLATED
- )
-
- epub.writestr("OPS/toc.ncx",
- get_TOCNCX_XML(
- all_md_filenames,
- uid=json_data["metadata"].get("dc:identifier", ""),
- title=json_data["metadata"].get("dc:title", "")
- ),
- zipfile.ZIP_DEFLATED
- )
-
- # Copy remaining images that weren't referenced in markdown
- remaining_images = set(all_image_filenames) - set(processed_images.keys())
- if remaining_images and os.path.exists(images_dir):
- print(f"Writing {len(remaining_images)} additional images...")
- for image in remaining_images:
- with open(os.path.join(images_dir, image), "rb") as f:
- epub.writestr(f"OPS/images/{image}", f.read(), zipfile.ZIP_DEFLATED)
-
- # Copy CSS files; write a default style.css for any that are missing
- default_css_content = b"""body { font-family: serif; line-height: 1.5; margin: 5%; }
-h1, h2, h3, h4, h5, h6 { font-family: sans-serif; }
-img { max-width: 100%; height: auto; }
-pre, code { font-family: monospace; font-size: 0.9em; }
-"""
- print(f"Writing {len(all_css_filenames)} CSS files...")
- for css in all_css_filenames:
- css_path = os.path.join(css_dir, css)
- if os.path.exists(css_path):
- with open(css_path, "rb") as f:
- epub.writestr(f"OPS/css/{css}", f.read(), zipfile.ZIP_DEFLATED)
- else:
- epub.writestr(f"OPS/css/{css}", default_css_content, zipfile.ZIP_DEFLATED)
-
- print(f"\nEPUB creation complete: {output_path}")
-
- except Exception:
- import traceback
- print(f"Error processing {work_dir}:")
- print(traceback.format_exc())
+ meta = resolve_metadata(markdown_dir, primary_md)
+ images_dir = (markdown_dir / "images").resolve()
+
+ cmd = [
+ "pandoc",
+ str(tmp_md_path),
+ "-o", str(epub_target),
+ "-f", "markdown+smart",
+ "-t", "epub3",
+ "--split-level=2",
+ "--toc",
+ "--toc-depth=3",
+ "--math-method=mathml", # Official Pandoc 3.x syntax without deprecation warning
+ f"--resource-path=.:images:{markdown_dir}:{images_dir}",
+ f"--css={tmp_css_path}",
+ "-M", f"title={meta['title']}",
+ "-M", f"author={meta['creator']}",
+ "-M", f"lang={meta['language']}",
+ "-M", f"publisher={meta['publisher']}"
+ ]
+
+ cover_img = find_cover_image(images_dir)
+ if cover_img and cover_img.exists():
+ cmd.append(f"--epub-cover-image={cover_img}")
+
+ print(f"\n[Pandoc] Starting EPUB 3 compilation: {epub_target.name}")
+ result = subprocess.run(
+ cmd,
+ cwd=str(markdown_dir),
+ capture_output=True,
+ text=True,
+ check=True
+ )
+ if result.stderr:
+ print(f"[Pandoc Warning]:\n{result.stderr}", file=sys.stderr)
+
+ print(f"[Pandoc] EPUB 3 generated successfully: {epub_target}")
+
+ except subprocess.CalledProcessError as e:
+ print(f"[Pandoc ERROR]: Compilation failed.\n{e.stderr}", file=sys.stderr)
raise
-
-if __name__ == "__main__":
- main(sys.argv[1:])
+ finally:
+ if tmp_md_path.exists(): tmp_md_path.unlink()
+ if tmp_css_path.exists(): tmp_css_path.unlink()
\ No newline at end of file
diff --git a/requirements.txt b/requirements.txt
index 4e206c3..d4546ae 100644
--- a/requirements.txt
+++ b/requirements.txt
@@ -2,7 +2,7 @@
# marker-pdf constrains Pillow<11.0.0, and Pillow only ships cp314 wheels from
# 11.3.0 onward, so on Python 3.14 pip builds Pillow from source rather than
# installing a wheel.
+torch
marker-pdf==1.10.2
-transformers==4.57.6
-markdown==3.10.2
-latex2mathml==3.81.0
\ No newline at end of file
+Pillow
+regex
\ No newline at end of file