mirror of
https://github.com/openmlsys/openmlsys-zh.git
synced 2026-10-29 11:35:03 +08:00
9c0439f424
- Remap 129 broken image paths in v1/en_chapters to the files that exist under v1/img (old chapter numbering, renamed files, and .pdf figures that browsers cannot display inline). - Restore ch07/7.5/distribute.png and ch08/distillation.png, which were dropped in #499 because the `dist*` .gitignore rule matched them; narrow the rule to `/dist/`. - Fix the English toc blocks: chapter entries now point at real files and the root toc has part headings, so section pages are built again (previously only chapter index pages were generated). - Add tests/test_image_references.py to CI to catch missing or unrenderable figure references. Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
60 lines
2.3 KiB
Python
60 lines
2.3 KiB
Python
from __future__ import annotations
|
|
|
|
import os
|
|
import re
|
|
import unittest
|
|
from pathlib import Path
|
|
|
|
from tools.ensure_book_resources import RESOURCE_NAMES
|
|
|
|
REPO_ROOT = Path(__file__).resolve().parent.parent
|
|
CHAPTER_DIRS = (
|
|
REPO_ROOT / "v1" / "en_chapters",
|
|
REPO_ROOT / "v1" / "zh_chapters",
|
|
)
|
|
SHARED_RESOURCES = REPO_ROOT / "v1"
|
|
MARKDOWN_IMAGE_RE = re.compile(r"!\[[^\]]*\]\(\s*([^)\s]+)")
|
|
HTML_IMAGE_RE = re.compile(r"<img[^>]*?\ssrc=\"([^\"]+)\"")
|
|
# Browsers cannot display these inline, so mdBook renders a broken image.
|
|
UNRENDERABLE_SUFFIXES = {".pdf", ".eps", ".tif", ".tiff"}
|
|
|
|
|
|
def iter_markdown_files(chapter_dir: Path):
|
|
for path in sorted(chapter_dir.rglob("*.md")):
|
|
relative = path.relative_to(chapter_dir)
|
|
if relative.parts[0] in RESOURCE_NAMES:
|
|
continue
|
|
yield path
|
|
|
|
|
|
def resolve_reference(chapter_dir: Path, markdown_file: Path, reference: str) -> Path:
|
|
# Resource dirs are symlinks created at build time, so resolve them manually.
|
|
target = Path(os.path.normpath(markdown_file.parent / reference))
|
|
relative = target.relative_to(chapter_dir) if target.is_relative_to(chapter_dir) else None
|
|
if relative is not None and relative.parts and relative.parts[0] in RESOURCE_NAMES:
|
|
return SHARED_RESOURCES / relative
|
|
return target
|
|
|
|
|
|
class ImageReferenceTests(unittest.TestCase):
|
|
def test_chapter_images_exist_and_are_renderable(self) -> None:
|
|
problems: list[str] = []
|
|
for chapter_dir in CHAPTER_DIRS:
|
|
for markdown_file in iter_markdown_files(chapter_dir):
|
|
text = markdown_file.read_text(encoding="utf-8")
|
|
references = MARKDOWN_IMAGE_RE.findall(text) + HTML_IMAGE_RE.findall(text)
|
|
for reference in references:
|
|
if reference.startswith(("http://", "https://", "data:")):
|
|
continue
|
|
location = f"{markdown_file.relative_to(REPO_ROOT)}: {reference}"
|
|
if Path(reference).suffix.lower() in UNRENDERABLE_SUFFIXES:
|
|
problems.append(f"unrenderable format {location}")
|
|
elif not resolve_reference(chapter_dir, markdown_file, reference).is_file():
|
|
problems.append(f"missing file {location}")
|
|
|
|
self.assertEqual(problems, [], "\n".join(problems))
|
|
|
|
|
|
if __name__ == "__main__":
|
|
unittest.main()
|