grayhook/pdf2ai
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Use five-digit chapter indices

Browse Source 
Keep generated chapter filenames and displayed section numbers lexicographically ordered beyond 99 sections.

Co-authored-by: Cursor <cursoragent@cursor.com>
This commit is contained in:
Сергей Маринкевич
2026-07-15 18:19:54 +07:00
parent 72a7de6146
commit e3ee0aafaf
2 changed files with 9 additions and 5 deletions
Download Patch File Download Diff File Expand all files Collapse all files
README.md
+1 -1
View File
@@ -24,7 +24,7 @@ chmod +x run.sh
out/RealTek-r8169/ out/RealTek-r8169/
├── INDEX.md # оглавление со ссылками на главы ├── INDEX.md # оглавление со ссылками на главы
├── chapters/ # одна глава = один файл ├── chapters/ # одна глава = один файл
│ ├── 01-features.md │ ├── 00001-features.md
│ └── ... │ └── ...
└── assets/ # embedded PNG + page PNG для сломанных layout-страниц └── assets/ # embedded PNG + page PNG для сломанных layout-страниц
``` ```
extract.py
+8 -4
View File
@@ -24,6 +24,7 @@ MIN_CELL_LEN_FOR_DEDUP = 4
SPACED_LETTERS = re.compile(r"^([A-Z](?: [A-Z]){1,}|[A-Z])$") SPACED_LETTERS = re.compile(r"^([A-Z](?: [A-Z]){1,}|[A-Z])$")
DEFAULT_PAGE_DPI = 150 DEFAULT_PAGE_DPI = 150
DEFAULT_CHAPTER_LEVEL = 2 DEFAULT_CHAPTER_LEVEL = 2
SECTION_INDEX_WIDTH = 5
def slugify(title: str) -> str: def slugify(title: str) -> str:
@@ -320,9 +321,9 @@ def convert(
if preamble_pages: if preamble_pages:
sections.append( sections.append(
{ {
"num": "00", "num": f"{0:0{SECTION_INDEX_WIDTH}d}",
"title": "front-matter", "title": "front-matter",
"slug": "00-front-matter", "slug": f"{0:0{SECTION_INDEX_WIDTH}d}-front-matter",
"pages": preamble_pages, "pages": preamble_pages,
} }
) )
@@ -339,9 +340,12 @@ def convert(
continue continue
sections.append( sections.append(
{ {
"num": f"{index:02d}", "num": f"{index:0{SECTION_INDEX_WIDTH}d}",
"title": entry["title"], "title": entry["title"],
"slug": f"{index:02d}-{slugify(' - '.join((*entry['parents'], entry['title'])))}", "slug": (
f"{index:0{SECTION_INDEX_WIDTH}d}-"
f"{slugify(' - '.join((*entry['parents'], entry['title'])))}"
),
"pages": pages, "pages": pages,
} }
) )