p2md 0.2.0__py3-none-any.whl

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
p2md/__init__.py ADDED
@@ -0,0 +1,22 @@
1
+ # -*- coding: utf-8 -*-
2
+ """p2md —— 扫描书 / 长截图转结构化 Markdown(本地、无 LLM)。
3
+
4
+ 三段流水线(每步产物落盘、JSON 协议解耦):
5
+ ① ocr OCR → output/ocr/(macOS 用 Swift Vision,其他平台用 RapidOCR)
6
+ ② detect 图/标题/正文三分检测 → output/blocks + output/figs
7
+ ③ build 组装章节 Markdown → output/content/article_N.md + images/
8
+
9
+ 功能边界:只产出 Markdown 与裁图,不生成 PDF,不依赖 document-factory。
10
+ """
11
+
12
+ from .api import ConvertError, ConvertResult, convert, convert_images
13
+
14
+ __version__ = "0.2.0"
15
+
16
+ __all__ = [
17
+ "ConvertError",
18
+ "ConvertResult",
19
+ "convert",
20
+ "convert_images",
21
+ "__version__",
22
+ ]
p2md/__main__.py ADDED
@@ -0,0 +1,6 @@
1
+ # -*- coding: utf-8 -*-
2
+ """python -m p2md 入口"""
3
+ from .cli import main
4
+
5
+ if __name__ == '__main__':
6
+ main()
p2md/api.py ADDED
@@ -0,0 +1,138 @@
1
+ # -*- coding: utf-8 -*-
2
+ """库接口:把一个图片集从页图转成章节 Markdown,一次调用走完 ocr → detect → build。
3
+
4
+ - CLI(``main.py``)面向"一个项目一本书",路径写死在项目根下;
5
+ - 库调用时源页图可来自任意目录,中间产物落在调用方给定的工作目录,
6
+ 最终 Markdown 可复制到指定输出目录。
7
+
8
+ 约定:
9
+ - 中间产物(ocr/blocks/figs/md)落 ``work_root``,不污染项目根;
10
+ - 最终章节 Markdown 从 build 的 article 目录复制到 ``output_dir``(给了才复制),
11
+ 图片一并复制到 ``<output_dir>/images/``,md 里的 ``images/...`` 引用即可直接解析。
12
+
13
+ 错误处理:底层阶段用 ``SystemExit`` 表示"前置条件不满足"(缺图、缺 blocks 等),
14
+ 这里统一转成 ``ConvertError``,避免库调用把整个进程带走。
15
+ """
16
+
17
+ from __future__ import annotations
18
+
19
+ import shutil
20
+ from dataclasses import dataclass, field
21
+ from pathlib import Path
22
+ from typing import Callable, List, Optional
23
+
24
+
25
+ class ConvertError(RuntimeError):
26
+ """转换失败(输入缺失、工具不可用、阶段报错)。调用方据此判 failed/invalid。"""
27
+
28
+
29
+ @dataclass
30
+ class ConvertResult:
31
+ """一次图片集转换的结果。``markdowns`` 是最终可用的章节文件(已含图片相对引用)。"""
32
+
33
+ markdowns: List[Path] = field(default_factory=list)
34
+ images: List[Path] = field(default_factory=list)
35
+ pages: int = 0
36
+ work_root: Optional[Path] = None
37
+ warnings: List[str] = field(default_factory=list)
38
+
39
+
40
+ def _call(stage: str, fn, *args, **kwargs):
41
+ """把阶段里的 SystemExit / 异常收敛成 ConvertError,并带上是哪一步挂的。"""
42
+ try:
43
+ return fn(*args, **kwargs)
44
+ except ConvertError:
45
+ raise
46
+ except SystemExit as exc: # 底层用 SystemExit 表达前置条件不满足
47
+ raise ConvertError(f"{stage}: {exc}") from exc
48
+ except Exception as exc: # noqa: BLE001 - 统一收敛给调用方判定
49
+ raise ConvertError(f"{stage}: {type(exc).__name__}: {exc}") from exc
50
+
51
+
52
+ def convert_images(project_dir, *, assets_dir=None, output_dir=None, work_root=None,
53
+ start=None, end=None, force=False, chapters=None,
54
+ log: Optional[Callable[[str], None]] = None) -> ConvertResult:
55
+ """把一个图片集转成章节 Markdown。
56
+
57
+ Args:
58
+ project_dir: 项目根(读 ``config.yaml`` / ``.env``)。
59
+ assets_dir: 源页图目录,默认 ``<project_dir>/assets``。
60
+ output_dir: 最终 Markdown 落地目录;不给就只留在 work_root 的 article 目录。
61
+ work_root: 中间产物目录,默认 ``<project_dir>/.derived/ocr/<assets 目录名>``。
62
+ start/end: 页码范围(None=自动)。
63
+ force: 忽略 OCR 缓存重跑。
64
+ chapters: 人工章节结构(透传 build)。
65
+ log: 日志回调,默认不打印(库调用不该刷屏)。
66
+
67
+ Returns:
68
+ ConvertResult。
69
+ """
70
+ from . import build_md, detect as detect_mod, middleware, ocr
71
+ from .project import Project
72
+
73
+ say = log or (lambda _msg: None)
74
+ root = Path(project_dir).expanduser().resolve()
75
+ if not root.is_dir():
76
+ raise ConvertError(f"项目目录不存在: {root}")
77
+
78
+ src = Path(assets_dir).expanduser().resolve() if assets_dir else root / "assets"
79
+ if not src.is_dir():
80
+ raise ConvertError(f"图片集源目录不存在: {src}")
81
+
82
+ work = (Path(work_root).expanduser().resolve() if work_root
83
+ else root / ".derived" / "ocr" / src.name)
84
+ work.mkdir(parents=True, exist_ok=True)
85
+
86
+ project = Project(str(root), output_dir=str(work), assets_dir=str(src))
87
+
88
+ say(f"[ocr] 源 {src} → 中间产物 {work}")
89
+ img_dir, _start, _end = _call("ocr", ocr.run, project, str(src),
90
+ start, end, force_ocr=force)
91
+
92
+ say("[middleware] 逐页修正")
93
+ _call("middleware", middleware.run_all, project)
94
+
95
+ say("[detect] 版面检测")
96
+ _call("detect", detect_mod.detect, project, img_dir, None)
97
+
98
+ say("[build] 组装章节")
99
+ md_dir, article_dir, written = _call("build", build_md.build, project, chapters)
100
+
101
+ bodies = sorted(Path(article_dir).glob("*.md"))
102
+ images = sorted((Path(article_dir) / "images").glob("*")) \
103
+ if (Path(article_dir) / "images").is_dir() else []
104
+
105
+ pages_json = project.get_ocr_pages_json_dir()
106
+ pages = len(list(pages_json.glob("*.json"))) if pages_json.is_dir() else 0
107
+
108
+ result = ConvertResult(markdowns=bodies, images=images, pages=pages, work_root=work)
109
+
110
+ if output_dir:
111
+ dest = Path(output_dir).expanduser().resolve()
112
+ dest.mkdir(parents=True, exist_ok=True)
113
+ for path in bodies:
114
+ shutil.copyfile(path, dest / path.name)
115
+ if images:
116
+ img_dest = dest / "images"
117
+ img_dest.mkdir(parents=True, exist_ok=True)
118
+ for path in images:
119
+ shutil.copyfile(path, img_dest / path.name)
120
+ result.markdowns = [dest / p.name for p in bodies]
121
+ result.images = [dest / "images" / p.name for p in images]
122
+ say(f"[done] {len(bodies)} 章 → {dest}")
123
+
124
+ if not result.markdowns:
125
+ result.warnings.append("没有产出任何 Markdown(检查页图与章节结构)")
126
+ return result
127
+
128
+
129
+ def convert(project_dir, *, assets_dir=None, output_dir=None, work_root=None,
130
+ start=None, end=None, force=False, chapters=None,
131
+ log: Optional[Callable[[str], None]] = None) -> ConvertResult:
132
+ """:func:`convert_images` 的短别名:``p2md.convert(project, ...)`` 一句话转完。"""
133
+ return convert_images(project_dir, assets_dir=assets_dir, output_dir=output_dir,
134
+ work_root=work_root, start=start, end=end, force=force,
135
+ chapters=chapters, log=log)
136
+
137
+
138
+ __all__ = ["ConvertError", "ConvertResult", "convert", "convert_images"]