dsh-convert-core 0.1.0-alpha.1
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.
- package/LICENSE +21 -0
- package/README.md +149 -0
- package/bin/media-to-md.mjs +168 -0
- package/bin/tita-convert.mjs +85 -0
- package/docs/media-to-md.md +143 -0
- package/lib/adapters/anydoc.d.ts +73 -0
- package/lib/adapters/anydoc.d.ts.map +1 -0
- package/lib/adapters/anydoc.js +295 -0
- package/lib/adapters/anydoc.js.map +1 -0
- package/lib/adapters/iddoc.d.ts +28 -0
- package/lib/adapters/iddoc.d.ts.map +1 -0
- package/lib/adapters/iddoc.js +153 -0
- package/lib/adapters/iddoc.js.map +1 -0
- package/lib/adapters/index.d.ts +41 -0
- package/lib/adapters/index.d.ts.map +1 -0
- package/lib/adapters/index.js +115 -0
- package/lib/adapters/index.js.map +1 -0
- package/lib/adapters/scanned.d.ts +28 -0
- package/lib/adapters/scanned.d.ts.map +1 -0
- package/lib/adapters/scanned.js +121 -0
- package/lib/adapters/scanned.js.map +1 -0
- package/lib/adapters/text.d.ts +29 -0
- package/lib/adapters/text.d.ts.map +1 -0
- package/lib/adapters/text.js +135 -0
- package/lib/adapters/text.js.map +1 -0
- package/lib/adapters/types.d.ts +65 -0
- package/lib/adapters/types.d.ts.map +1 -0
- package/lib/adapters/types.js +11 -0
- package/lib/adapters/types.js.map +1 -0
- package/lib/contract.d.ts +185 -0
- package/lib/contract.d.ts.map +1 -0
- package/lib/contract.js +59 -0
- package/lib/contract.js.map +1 -0
- package/lib/delivery.d.ts +36 -0
- package/lib/delivery.d.ts.map +1 -0
- package/lib/delivery.js +97 -0
- package/lib/delivery.js.map +1 -0
- package/lib/http.d.ts +34 -0
- package/lib/http.d.ts.map +1 -0
- package/lib/http.js +372 -0
- package/lib/http.js.map +1 -0
- package/lib/markdown.d.ts +32 -0
- package/lib/markdown.d.ts.map +1 -0
- package/lib/markdown.js +145 -0
- package/lib/markdown.js.map +1 -0
- package/lib/media/binaries.d.ts +26 -0
- package/lib/media/binaries.d.ts.map +1 -0
- package/lib/media/binaries.js +60 -0
- package/lib/media/binaries.js.map +1 -0
- package/lib/media/convert.d.ts +19 -0
- package/lib/media/convert.d.ts.map +1 -0
- package/lib/media/convert.js +210 -0
- package/lib/media/convert.js.map +1 -0
- package/lib/media/index.d.ts +27 -0
- package/lib/media/index.d.ts.map +1 -0
- package/lib/media/index.js +27 -0
- package/lib/media/index.js.map +1 -0
- package/lib/media/markdown.d.ts +40 -0
- package/lib/media/markdown.d.ts.map +1 -0
- package/lib/media/markdown.js +89 -0
- package/lib/media/markdown.js.map +1 -0
- package/lib/media/media.d.ts +45 -0
- package/lib/media/media.d.ts.map +1 -0
- package/lib/media/media.js +167 -0
- package/lib/media/media.js.map +1 -0
- package/lib/media/models.d.ts +31 -0
- package/lib/media/models.d.ts.map +1 -0
- package/lib/media/models.js +87 -0
- package/lib/media/models.js.map +1 -0
- package/lib/media/scanned.d.ts +18 -0
- package/lib/media/scanned.d.ts.map +1 -0
- package/lib/media/scanned.js +179 -0
- package/lib/media/scanned.js.map +1 -0
- package/lib/media/types.d.ts +154 -0
- package/lib/media/types.d.ts.map +1 -0
- package/lib/media/types.js +23 -0
- package/lib/media/types.js.map +1 -0
- package/lib/server.d.ts +35 -0
- package/lib/server.d.ts.map +1 -0
- package/lib/server.js +64 -0
- package/lib/server.js.map +1 -0
- package/lib/service.d.ts +170 -0
- package/lib/service.d.ts.map +1 -0
- package/lib/service.js +562 -0
- package/lib/service.js.map +1 -0
- package/package.json +59 -0
- package/scripts/asr.py +55 -0
- package/scripts/ocr.py +40 -0
- package/scripts/pdf_pages.py +49 -0
- package/scripts/vendor.mjs +76 -0
- package/src/adapters/anydoc.ts +356 -0
- package/src/adapters/iddoc.ts +171 -0
- package/src/adapters/index.ts +147 -0
- package/src/adapters/scanned.ts +139 -0
- package/src/adapters/text.ts +146 -0
- package/src/adapters/types.ts +76 -0
- package/src/contract.ts +230 -0
- package/src/delivery.ts +124 -0
- package/src/http.ts +394 -0
- package/src/markdown.ts +141 -0
- package/src/media/binaries.ts +67 -0
- package/src/media/convert.ts +232 -0
- package/src/media/index.ts +43 -0
- package/src/media/markdown.ts +105 -0
- package/src/media/media.ts +204 -0
- package/src/media/models.ts +124 -0
- package/src/media/scanned.ts +200 -0
- package/src/media/types.ts +171 -0
- package/src/server.ts +85 -0
- package/src/service.ts +639 -0
- package/web/app.js +382 -0
- package/web/index.html +217 -0
- package/web/styles.css +263 -0
- package/web/vendor/icons.js +25 -0
- package/web/vendor/vue.global.prod.js +14 -0
|
@@ -0,0 +1,167 @@
|
|
|
1
|
+
/**
|
|
2
|
+
* 媒体探测、关键帧与字幕 —— 全部走 ffmpeg/ffprobe,不含任何模型。
|
|
3
|
+
*/
|
|
4
|
+
import { execFile } from 'node:child_process';
|
|
5
|
+
import { mkdir, readdir, readFile } from 'node:fs/promises';
|
|
6
|
+
import { basename, dirname, extname, join } from 'node:path';
|
|
7
|
+
import { promisify } from 'node:util';
|
|
8
|
+
import { MediaToMarkdownError } from './types.js';
|
|
9
|
+
import { runFfmpeg } from './binaries.js';
|
|
10
|
+
const execFileAsync = promisify(execFile);
|
|
11
|
+
export async function probeMedia(file, signal) {
|
|
12
|
+
let stdout;
|
|
13
|
+
try {
|
|
14
|
+
({ stdout } = await execFileAsync('ffprobe', [
|
|
15
|
+
'-v', 'error', '-print_format', 'json', '-show_format', '-show_streams', file,
|
|
16
|
+
], { maxBuffer: 32 * 1024 * 1024, ...(signal ? { signal } : {}) }));
|
|
17
|
+
}
|
|
18
|
+
catch (cause) {
|
|
19
|
+
throw new MediaToMarkdownError('MEDIA_UNREADABLE', `ffprobe 无法读取该文件:${cause instanceof Error ? cause.message : String(cause)}`);
|
|
20
|
+
}
|
|
21
|
+
const parsed = JSON.parse(stdout);
|
|
22
|
+
const streams = parsed.streams ?? [];
|
|
23
|
+
const video = streams.find(stream => stream.codec_type === 'video');
|
|
24
|
+
const audio = streams.find(stream => stream.codec_type === 'audio');
|
|
25
|
+
const subtitle = streams.find(stream => stream.codec_type === 'subtitle');
|
|
26
|
+
const rate = video?.r_frame_rate?.split('/');
|
|
27
|
+
const fps = rate && Number(rate[1]) > 0 ? Number(rate[0]) / Number(rate[1]) : undefined;
|
|
28
|
+
return {
|
|
29
|
+
durationSeconds: Number(parsed.format?.duration ?? 0),
|
|
30
|
+
sizeBytes: Number(parsed.format?.size ?? 0),
|
|
31
|
+
hasVideo: video !== undefined,
|
|
32
|
+
hasAudio: audio !== undefined,
|
|
33
|
+
hasSubtitleTrack: subtitle !== undefined,
|
|
34
|
+
...(video?.width ? { width: video.width } : {}),
|
|
35
|
+
...(video?.height ? { height: video.height } : {}),
|
|
36
|
+
...(fps !== undefined && Number.isFinite(fps) ? { fps: Number(fps.toFixed(2)) } : {}),
|
|
37
|
+
...(video?.codec_name ? { videoCodec: video.codec_name } : {}),
|
|
38
|
+
...(audio?.codec_name ? { audioCodec: audio.codec_name } : {}),
|
|
39
|
+
};
|
|
40
|
+
}
|
|
41
|
+
async function readdirSafe(dir) {
|
|
42
|
+
try {
|
|
43
|
+
return await readdir(dir);
|
|
44
|
+
}
|
|
45
|
+
catch {
|
|
46
|
+
return [];
|
|
47
|
+
}
|
|
48
|
+
}
|
|
49
|
+
/**
|
|
50
|
+
* 均匀采样 + 场景变化两路合并,再按内容 sha1 去重。
|
|
51
|
+
*
|
|
52
|
+
* 只靠 `scene` 检测会漏内容:口播/图文类视频几乎没有硬切,阈值再低也可能只出 1 帧
|
|
53
|
+
* (实测:一个 3 段画面的测试视频,阈值降到 0.1 仍然只有 1 帧)。只靠均匀采样又会漏掉
|
|
54
|
+
* 真正的画面切换。两路都跑,按时间排序、近邻合并、内容去重。
|
|
55
|
+
*/
|
|
56
|
+
export async function extractKeyframes(file, durationSeconds, scratch, options = {}, signal) {
|
|
57
|
+
const samples = Math.max(1, options.samples ?? 12);
|
|
58
|
+
const maxFrames = Math.max(1, options.maxFrames ?? 40);
|
|
59
|
+
const scene = clamp01(options.sceneThreshold ?? 0.3);
|
|
60
|
+
const raw = join(scratch, 'raw');
|
|
61
|
+
await mkdir(join(raw, 'scene'), { recursive: true });
|
|
62
|
+
const count = Math.max(1, Math.min(samples, maxFrames));
|
|
63
|
+
const step = durationSeconds > 0 ? Math.max(durationSeconds / count, 0.5) : 1;
|
|
64
|
+
await runFfmpeg([
|
|
65
|
+
'-hide_banner', '-loglevel', 'error', '-i', file,
|
|
66
|
+
'-vf', `fps=1/${step.toFixed(3)}`, '-frames:v', String(count), '-q:v', '3',
|
|
67
|
+
join(raw, 'even-%04d.jpg'),
|
|
68
|
+
], { allowFailure: true });
|
|
69
|
+
const sceneRun = await runFfmpeg([
|
|
70
|
+
'-hide_banner', '-loglevel', 'info', '-i', file,
|
|
71
|
+
'-vf', `select='gt(scene\\,${scene})',showinfo`, '-fps_mode', 'vfr',
|
|
72
|
+
'-frames:v', String(maxFrames), '-q:v', '3',
|
|
73
|
+
join(raw, 'scene', 'scene-%04d.jpg'),
|
|
74
|
+
], { allowFailure: true });
|
|
75
|
+
const sceneTimes = [...(sceneRun.stderr ?? '').matchAll(/pts_time:([0-9.]+)/g)].map(match => Number(match[1]));
|
|
76
|
+
const evenFrames = (await readdirSafe(raw)).filter(name => name.startsWith('even-')).sort()
|
|
77
|
+
.map((name, index) => ({ seconds: Number((index * step).toFixed(3)), path: join(raw, name) }));
|
|
78
|
+
const sceneFrames = (await readdirSafe(join(raw, 'scene'))).filter(name => name.startsWith('scene-')).sort()
|
|
79
|
+
.map((name, index) => ({
|
|
80
|
+
seconds: Number.isFinite(sceneTimes[index]) ? Number(sceneTimes[index]) : 0,
|
|
81
|
+
path: join(raw, 'scene', name),
|
|
82
|
+
}));
|
|
83
|
+
const picked = [];
|
|
84
|
+
for (const candidate of [...sceneFrames, ...evenFrames].sort((a, b) => a.seconds - b.seconds)) {
|
|
85
|
+
if (picked.length >= maxFrames)
|
|
86
|
+
break;
|
|
87
|
+
if (picked.some(kept => Math.abs(kept.seconds - candidate.seconds) < 0.75))
|
|
88
|
+
continue;
|
|
89
|
+
picked.push(candidate);
|
|
90
|
+
}
|
|
91
|
+
const { createHash } = await import('node:crypto');
|
|
92
|
+
const frames = [];
|
|
93
|
+
const seen = new Set();
|
|
94
|
+
for (const item of picked) {
|
|
95
|
+
if (signal?.aborted)
|
|
96
|
+
throw new MediaToMarkdownError('ABORTED', '已取消。');
|
|
97
|
+
const data = await readFile(item.path);
|
|
98
|
+
// 静止画面会被采样多次,按内容去重,别让 assets/ 里塞满一模一样的帧。
|
|
99
|
+
const digest = createHash('sha1').update(data).digest('hex');
|
|
100
|
+
if (seen.has(digest))
|
|
101
|
+
continue;
|
|
102
|
+
seen.add(digest);
|
|
103
|
+
frames.push({
|
|
104
|
+
rel: `assets/frame-${String(frames.length + 1).padStart(4, '0')}.jpg`,
|
|
105
|
+
sourcePath: item.path,
|
|
106
|
+
data,
|
|
107
|
+
seconds: item.seconds,
|
|
108
|
+
});
|
|
109
|
+
}
|
|
110
|
+
return frames;
|
|
111
|
+
}
|
|
112
|
+
/** 内嵌字幕轨优先,没有再看同名旁挂 `.srt`。 */
|
|
113
|
+
export async function extractSubtitles(file, probe, scratch, signal) {
|
|
114
|
+
if (probe.hasSubtitleTrack) {
|
|
115
|
+
const target = join(scratch, 'subs.srt');
|
|
116
|
+
try {
|
|
117
|
+
await execFileAsync('ffmpeg', ['-hide_banner', '-loglevel', 'error', '-y', '-i', file, '-map', '0:s:0', target], { signal });
|
|
118
|
+
const text = await readFile(target, 'utf8');
|
|
119
|
+
if (text.trim())
|
|
120
|
+
return text;
|
|
121
|
+
}
|
|
122
|
+
catch {
|
|
123
|
+
/* 抽轨失败就退到旁挂字幕 */
|
|
124
|
+
}
|
|
125
|
+
}
|
|
126
|
+
const sidecar = join(dirname(file), `${basename(file, extname(file))}.srt`);
|
|
127
|
+
try {
|
|
128
|
+
const text = await readFile(sidecar, 'utf8');
|
|
129
|
+
if (text.trim())
|
|
130
|
+
return text;
|
|
131
|
+
}
|
|
132
|
+
catch {
|
|
133
|
+
/* 没有旁挂字幕 */
|
|
134
|
+
}
|
|
135
|
+
return undefined;
|
|
136
|
+
}
|
|
137
|
+
/** 极简 SRT 解析:取起止时间与文本,顺带去掉自动字幕常见的滚动重复行。 */
|
|
138
|
+
export function parseSrt(source) {
|
|
139
|
+
const cues = [];
|
|
140
|
+
const range = /(\d{1,2}):(\d{2}):(\d{2})[,.](\d{1,3})\s*-->\s*(\d{1,2}):(\d{2}):(\d{2})[,.](\d{1,3})/;
|
|
141
|
+
const toSeconds = (h, m, s, ms) => Number(h) * 3600 + Number(m) * 60 + Number(s) + Number(ms.padEnd(3, '0')) / 1000;
|
|
142
|
+
for (const block of source.replace(/\r/g, '').split(/\n{2,}/)) {
|
|
143
|
+
const lines = block.split('\n').map(line => line.trim()).filter(Boolean);
|
|
144
|
+
if (lines.length === 0)
|
|
145
|
+
continue;
|
|
146
|
+
const timeLine = lines.find(line => line.includes('-->'));
|
|
147
|
+
if (!timeLine)
|
|
148
|
+
continue;
|
|
149
|
+
const match = range.exec(timeLine);
|
|
150
|
+
const text = lines.filter(line => line !== timeLine && !/^\d+$/.test(line)).join(' ').trim();
|
|
151
|
+
if (!match || !text)
|
|
152
|
+
continue;
|
|
153
|
+
const seconds = toSeconds(match[1], match[2], match[3], match[4]);
|
|
154
|
+
const endSeconds = toSeconds(match[5], match[6], match[7], match[8]);
|
|
155
|
+
const previous = cues[cues.length - 1];
|
|
156
|
+
if (previous && previous.text === text)
|
|
157
|
+
continue;
|
|
158
|
+
cues.push({ seconds, endSeconds, text });
|
|
159
|
+
}
|
|
160
|
+
return cues;
|
|
161
|
+
}
|
|
162
|
+
function clamp01(value) {
|
|
163
|
+
if (!Number.isFinite(value))
|
|
164
|
+
return 0.3;
|
|
165
|
+
return Math.min(1, Math.max(0, value));
|
|
166
|
+
}
|
|
167
|
+
//# sourceMappingURL=media.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"media.js","sourceRoot":"","sources":["../../src/media/media.ts"],"names":[],"mappings":"AAAA;;GAEG;AAEH,OAAO,EAAE,QAAQ,EAAE,MAAM,oBAAoB,CAAA;AAC7C,OAAO,EAAE,KAAK,EAAE,OAAO,EAAE,QAAQ,EAAE,MAAM,kBAAkB,CAAA;AAC3D,OAAO,EAAE,QAAQ,EAAE,OAAO,EAAE,OAAO,EAAE,IAAI,EAAE,MAAM,WAAW,CAAA;AAC5D,OAAO,EAAE,SAAS,EAAE,MAAM,WAAW,CAAA;AACrC,OAAO,EAAE,oBAAoB,EAAyC,MAAM,YAAY,CAAA;AACxF,OAAO,EAAE,SAAS,EAAE,MAAM,eAAe,CAAA;AAEzC,MAAM,aAAa,GAAG,SAAS,CAAC,QAAQ,CAAC,CAAA;AAuBzC,MAAM,CAAC,KAAK,UAAU,UAAU,CAAC,IAAY,EAAE,MAAoB;IACjE,IAAI,MAAc,CAAA;IAClB,IAAI,CAAC;QACH,CAAC,EAAE,MAAM,EAAE,GAAG,MAAM,aAAa,CAAC,SAAS,EAAE;YAC3C,IAAI,EAAE,OAAO,EAAE,eAAe,EAAE,MAAM,EAAE,cAAc,EAAE,eAAe,EAAE,IAAI;SAC9E,EAAE,EAAE,SAAS,EAAE,EAAE,GAAG,IAAI,GAAG,IAAI,EAAE,GAAG,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,MAAM,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,CAAC,CAAC,CAAA;IACrE,CAAC;IAAC,OAAO,KAAK,EAAE,CAAC;QACf,MAAM,IAAI,oBAAoB,CAC5B,kBAAkB,EAClB,mBAAmB,KAAK,YAAY,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,EAAE,CAC5E,CAAA;IACH,CAAC;IACD,MAAM,MAAM,GAAG,IAAI,CAAC,KAAK,CAAC,MAAM,CAA+E,CAAA;IAC/G,MAAM,OAAO,GAAG,MAAM,CAAC,OAAO,IAAI,EAAE,CAAA;IACpC,MAAM,KAAK,GAAG,OAAO,CAAC,IAAI,CAAC,MAAM,CAAC,EAAE,CAAC,MAAM,CAAC,UAAU,KAAK,OAAO,CAAC,CAAA;IACnE,MAAM,KAAK,GAAG,OAAO,CAAC,IAAI,CAAC,MAAM,CAAC,EAAE,CAAC,MAAM,CAAC,UAAU,KAAK,OAAO,CAAC,CAAA;IACnE,MAAM,QAAQ,GAAG,OAAO,CAAC,IAAI,CAAC,MAAM,CAAC,EAAE,CAAC,MAAM,CAAC,UAAU,KAAK,UAAU,CAAC,CAAA;IACzE,MAAM,IAAI,GAAG,KAAK,EAAE,YAAY,EAAE,KAAK,CAAC,GAAG,CAAC,CAAA;IAC5C,MAAM,GAAG,GAAG,IAAI,IAAI,MAAM,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,GAAG,MAAM,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,SAAS,CAAA;IACvF,OAAO;QACL,eAAe,EAAE,MAAM,CAAC,MAAM,CAAC,MAAM,EAAE,QAAQ,IAAI,CAAC,CAAC;QACrD,SAAS,EAAE,MAAM,CAAC,MAAM,CAAC,MAAM,EAAE,IAAI,IAAI,CAAC,CAAC;QAC3C,QAAQ,EAAE,KAAK,KAAK,SAAS;QAC7B,QAAQ,EAAE,KAAK,KAAK,SAAS;QAC7B,gBAAgB,EAAE,QAAQ,KAAK,SAAS;QACxC,GAAG,CAAC,KAAK,EAAE,KAAK,CAAC,CAAC,CAAC,EAAE,KAAK,EAAE,KAAK,CAAC,KAAK,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;QAC/C,GAAG,CAAC,KAAK,EAAE,MAAM,CAAC,CAAC,CAAC,EAAE,MAAM,EAAE,KAAK,CAAC,MAAM,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;QAClD,GAAG,CAAC,GAAG,KAAK,SAAS,IAAI,MAAM,CAAC,QAAQ,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,EAAE,GAAG,EAAE,MAAM,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;QACrF,GAAG,CAAC,KAAK,EAAE,UAAU,CAAC,CAAC,CAAC,EAAE,UAAU,EAAE,KAAK,CAAC,UAAU,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;QAC9D,GAAG,CAAC,KAAK,EAAE,UAAU,CAAC,CAAC,CAAC,EAAE,UAAU,EAAE,KAAK,CAAC,UAAU,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;KAC/D,CAAA;AACH,CAAC;AAWD,KAAK,UAAU,WAAW,CAAC,GAAW;IACpC,IAAI,CAAC;QACH,OAAO,MAAM,OAAO,CAAC,GAAG,CAAC,CAAA;IAC3B,CAAC;IAAC,MAAM,CAAC;QACP,OAAO,EAAE,CAAA;IACX,CAAC;AACH,CAAC;AAED;;;;;;GAMG;AACH,MAAM,CAAC,KAAK,UAAU,gBAAgB,CACpC,IAAY,EACZ,eAAuB,EACvB,OAAe,EACf,UAA6B,EAAE,EAC/B,MAAoB;IAEpB,MAAM,OAAO,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,OAAO,CAAC,OAAO,IAAI,EAAE,CAAC,CAAA;IAClD,MAAM,SAAS,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,OAAO,CAAC,SAAS,IAAI,EAAE,CAAC,CAAA;IACtD,MAAM,KAAK,GAAG,OAAO,CAAC,OAAO,CAAC,cAAc,IAAI,GAAG,CAAC,CAAA;IACpD,MAAM,GAAG,GAAG,IAAI,CAAC,OAAO,EAAE,KAAK,CAAC,CAAA;IAChC,MAAM,KAAK,CAAC,IAAI,CAAC,GAAG,EAAE,OAAO,CAAC,EAAE,EAAE,SAAS,EAAE,IAAI,EAAE,CAAC,CAAA;IAEpD,MAAM,KAAK,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,IAAI,CAAC,GAAG,CAAC,OAAO,EAAE,SAAS,CAAC,CAAC,CAAA;IACvD,MAAM,IAAI,GAAG,eAAe,GAAG,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,eAAe,GAAG,KAAK,EAAE,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,CAAA;IAC7E,MAAM,SAAS,CAAC;QACd,cAAc,EAAE,WAAW,EAAE,OAAO,EAAE,IAAI,EAAE,IAAI;QAChD,KAAK,EAAE,SAAS,IAAI,CAAC,OAAO,CAAC,CAAC,CAAC,EAAE,EAAE,WAAW,EAAE,MAAM,CAAC,KAAK,CAAC,EAAE,MAAM,EAAE,GAAG;QAC1E,IAAI,CAAC,GAAG,EAAE,eAAe,CAAC;KAC3B,EAAE,EAAE,YAAY,EAAE,IAAI,EAAE,CAAC,CAAA;IAE1B,MAAM,QAAQ,GAAG,MAAM,SAAS,CAAC;QAC/B,cAAc,EAAE,WAAW,EAAE,MAAM,EAAE,IAAI,EAAE,IAAI;QAC/C,KAAK,EAAE,sBAAsB,KAAK,aAAa,EAAE,WAAW,EAAE,KAAK;QACnE,WAAW,EAAE,MAAM,CAAC,SAAS,CAAC,EAAE,MAAM,EAAE,GAAG;QAC3C,IAAI,CAAC,GAAG,EAAE,OAAO,EAAE,gBAAgB,CAAC;KACrC,EAAE,EAAE,YAAY,EAAE,IAAI,EAAE,CAAC,CAAA;IAC1B,MAAM,UAAU,GAAG,CAAC,GAAG,CAAC,QAAQ,CAAC,MAAM,IAAI,EAAE,CAAC,CAAC,QAAQ,CAAC,qBAAqB,CAAC,CAAC,CAAC,GAAG,CAAC,KAAK,CAAC,EAAE,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAA;IAE9G,MAAM,UAAU,GAAG,CAAC,MAAM,WAAW,CAAC,GAAG,CAAC,CAAC,CAAC,MAAM,CAAC,IAAI,CAAC,EAAE,CAAC,IAAI,CAAC,UAAU,CAAC,OAAO,CAAC,CAAC,CAAC,IAAI,EAAE;SACxF,GAAG,CAAC,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE,CAAC,CAAC,EAAE,OAAO,EAAE,MAAM,CAAC,CAAC,KAAK,GAAG,IAAI,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC,EAAE,IAAI,EAAE,IAAI,CAAC,GAAG,EAAE,IAAI,CAAC,EAAE,CAAC,CAAC,CAAA;IAChG,MAAM,WAAW,GAAG,CAAC,MAAM,WAAW,CAAC,IAAI,CAAC,GAAG,EAAE,OAAO,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,IAAI,CAAC,EAAE,CAAC,IAAI,CAAC,UAAU,CAAC,QAAQ,CAAC,CAAC,CAAC,IAAI,EAAE;SACzG,GAAG,CAAC,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE,CAAC,CAAC;QACrB,OAAO,EAAE,MAAM,CAAC,QAAQ,CAAC,UAAU,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,UAAU,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;QAC3E,IAAI,EAAE,IAAI,CAAC,GAAG,EAAE,OAAO,EAAE,IAAI,CAAC;KAC/B,CAAC,CAAC,CAAA;IAEL,MAAM,MAAM,GAA6C,EAAE,CAAA;IAC3D,KAAK,MAAM,SAAS,IAAI,CAAC,GAAG,WAAW,EAAE,GAAG,UAAU,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,OAAO,GAAG,CAAC,CAAC,OAAO,CAAC,EAAE,CAAC;QAC9F,IAAI,MAAM,CAAC,MAAM,IAAI,SAAS;YAAE,MAAK;QACrC,IAAI,MAAM,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,OAAO,GAAG,SAAS,CAAC,OAAO,CAAC,GAAG,IAAI,CAAC;YAAE,SAAQ;QACpF,MAAM,CAAC,IAAI,CAAC,SAAS,CAAC,CAAA;IACxB,CAAC;IAED,MAAM,EAAE,UAAU,EAAE,GAAG,MAAM,MAAM,CAAC,aAAa,CAAC,CAAA;IAClD,MAAM,MAAM,GAAqB,EAAE,CAAA;IACnC,MAAM,IAAI,GAAG,IAAI,GAAG,EAAU,CAAA;IAC9B,KAAK,MAAM,IAAI,IAAI,MAAM,EAAE,CAAC;QAC1B,IAAI,MAAM,EAAE,OAAO;YAAE,MAAM,IAAI,oBAAoB,CAAC,SAAS,EAAE,MAAM,CAAC,CAAA;QACtE,MAAM,IAAI,GAAG,MAAM,QAAQ,CAAC,IAAI,CAAC,IAAI,CAAC,CAAA;QACtC,yCAAyC;QACzC,MAAM,MAAM,GAAG,UAAU,CAAC,MAAM,CAAC,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,CAAA;QAC5D,IAAI,IAAI,CAAC,GAAG,CAAC,MAAM,CAAC;YAAE,SAAQ;QAC9B,IAAI,CAAC,GAAG,CAAC,MAAM,CAAC,CAAA;QAChB,MAAM,CAAC,IAAI,CAAC;YACV,GAAG,EAAE,gBAAgB,MAAM,CAAC,MAAM,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,QAAQ,CAAC,CAAC,EAAE,GAAG,CAAC,MAAM;YACrE,UAAU,EAAE,IAAI,CAAC,IAAI;YACrB,IAAI;YACJ,OAAO,EAAE,IAAI,CAAC,OAAO;SACtB,CAAC,CAAA;IACJ,CAAC;IACD,OAAO,MAAM,CAAA;AACf,CAAC;AAED,+BAA+B;AAC/B,MAAM,CAAC,KAAK,UAAU,gBAAgB,CAAC,IAAY,EAAE,KAAiB,EAAE,OAAe,EAAE,MAAoB;IAC3G,IAAI,KAAK,CAAC,gBAAgB,EAAE,CAAC;QAC3B,MAAM,MAAM,GAAG,IAAI,CAAC,OAAO,EAAE,UAAU,CAAC,CAAA;QACxC,IAAI,CAAC;YACH,MAAM,aAAa,CAAC,QAAQ,EAAE,CAAC,cAAc,EAAE,WAAW,EAAE,OAAO,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,MAAM,EAAE,OAAO,EAAE,MAAM,CAAC,EAAE,EAAE,MAAM,EAAE,CAAC,CAAA;YAC5H,MAAM,IAAI,GAAG,MAAM,QAAQ,CAAC,MAAM,EAAE,MAAM,CAAC,CAAA;YAC3C,IAAI,IAAI,CAAC,IAAI,EAAE;gBAAE,OAAO,IAAI,CAAA;QAC9B,CAAC;QAAC,MAAM,CAAC;YACP,iBAAiB;QACnB,CAAC;IACH,CAAC;IACD,MAAM,OAAO,GAAG,IAAI,CAAC,OAAO,CAAC,IAAI,CAAC,EAAE,GAAG,QAAQ,CAAC,IAAI,EAAE,OAAO,CAAC,IAAI,CAAC,CAAC,MAAM,CAAC,CAAA;IAC3E,IAAI,CAAC;QACH,MAAM,IAAI,GAAG,MAAM,QAAQ,CAAC,OAAO,EAAE,MAAM,CAAC,CAAA;QAC5C,IAAI,IAAI,CAAC,IAAI,EAAE;YAAE,OAAO,IAAI,CAAA;IAC9B,CAAC;IAAC,MAAM,CAAC;QACP,YAAY;IACd,CAAC;IACD,OAAO,SAAS,CAAA;AAClB,CAAC;AAED,2CAA2C;AAC3C,MAAM,UAAU,QAAQ,CAAC,MAAc;IACrC,MAAM,IAAI,GAAe,EAAE,CAAA;IAC3B,MAAM,KAAK,GAAG,uFAAuF,CAAA;IACrG,MAAM,SAAS,GAAG,CAAC,CAAS,EAAE,CAAS,EAAE,CAAS,EAAE,EAAU,EAAU,EAAE,CACxE,MAAM,CAAC,CAAC,CAAC,GAAG,IAAI,GAAG,MAAM,CAAC,CAAC,CAAC,GAAG,EAAE,GAAG,MAAM,CAAC,CAAC,CAAC,GAAG,MAAM,CAAC,EAAE,CAAC,MAAM,CAAC,CAAC,EAAE,GAAG,CAAC,CAAC,GAAG,IAAI,CAAA;IAClF,KAAK,MAAM,KAAK,IAAI,MAAM,CAAC,OAAO,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC,KAAK,CAAC,QAAQ,CAAC,EAAE,CAAC;QAC9D,MAAM,KAAK,GAAG,KAAK,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,GAAG,CAAC,IAAI,CAAC,EAAE,CAAC,IAAI,CAAC,IAAI,EAAE,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,CAAA;QACxE,IAAI,KAAK,CAAC,MAAM,KAAK,CAAC;YAAE,SAAQ;QAChC,MAAM,QAAQ,GAAG,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC,IAAI,CAAC,QAAQ,CAAC,KAAK,CAAC,CAAC,CAAA;QACzD,IAAI,CAAC,QAAQ;YAAE,SAAQ;QACvB,MAAM,KAAK,GAAG,KAAK,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAA;QAClC,MAAM,IAAI,GAAG,KAAK,CAAC,MAAM,CAAC,IAAI,CAAC,EAAE,CAAC,IAAI,KAAK,QAAQ,IAAI,CAAC,OAAO,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,CAAA;QAC5F,IAAI,CAAC,KAAK,IAAI,CAAC,IAAI;YAAE,SAAQ;QAC7B,MAAM,OAAO,GAAG,SAAS,CAAC,KAAK,CAAC,CAAC,CAAE,EAAE,KAAK,CAAC,CAAC,CAAE,EAAE,KAAK,CAAC,CAAC,CAAE,EAAE,KAAK,CAAC,CAAC,CAAE,CAAC,CAAA;QACrE,MAAM,UAAU,GAAG,SAAS,CAAC,KAAK,CAAC,CAAC,CAAE,EAAE,KAAK,CAAC,CAAC,CAAE,EAAE,KAAK,CAAC,CAAC,CAAE,EAAE,KAAK,CAAC,CAAC,CAAE,CAAC,CAAA;QACxE,MAAM,QAAQ,GAAG,IAAI,CAAC,IAAI,CAAC,MAAM,GAAG,CAAC,CAAC,CAAA;QACtC,IAAI,QAAQ,IAAI,QAAQ,CAAC,IAAI,KAAK,IAAI;YAAE,SAAQ;QAChD,IAAI,CAAC,IAAI,CAAC,EAAE,OAAO,EAAE,UAAU,EAAE,IAAI,EAAE,CAAC,CAAA;IAC1C,CAAC;IACD,OAAO,IAAI,CAAA;AACb,CAAC;AAED,SAAS,OAAO,CAAC,KAAa;IAC5B,IAAI,CAAC,MAAM,CAAC,QAAQ,CAAC,KAAK,CAAC;QAAE,OAAO,GAAG,CAAA;IACvC,OAAO,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,KAAK,CAAC,CAAC,CAAA;AACxC,CAAC"}
|
|
@@ -0,0 +1,31 @@
|
|
|
1
|
+
/**
|
|
2
|
+
* 两个**专用小模型**的调用点:语音识别(whisper)与画面文字 OCR。
|
|
3
|
+
*
|
|
4
|
+
* 都不是大模型,也都不联网推理:模型下载一次后完全离线。helper 脚本随包发布
|
|
5
|
+
* (`scripts/*.py`,由 `uvx --with …` 按需拉起),所以这个包不把 Python/模型塞进 npm tarball,
|
|
6
|
+
* 但会把「要不要下载模型」这个决定**显式**交回调用方。
|
|
7
|
+
*/
|
|
8
|
+
import { type MediaAsrOptions, type MediaCue, type MediaToolOptions } from './types.js';
|
|
9
|
+
export declare const ASR_MODEL_SIZES: Record<string, string>;
|
|
10
|
+
export declare function describeModel(model: string): string;
|
|
11
|
+
export interface TranscriptionResult {
|
|
12
|
+
cues: MediaCue[];
|
|
13
|
+
model: string;
|
|
14
|
+
language?: string;
|
|
15
|
+
}
|
|
16
|
+
/** 音频 → 带时间戳的转写。`allowModelDownload` 为 false 且本地无模型时抛 `ASR_MODEL_REQUIRED`。 */
|
|
17
|
+
export declare function transcribe(audioFile: string, scratch: string, asr?: MediaAsrOptions, tools?: MediaToolOptions, signal?: AbortSignal): Promise<TranscriptionResult>;
|
|
18
|
+
export interface OcrEntry {
|
|
19
|
+
path: string;
|
|
20
|
+
lines: string[];
|
|
21
|
+
text: string;
|
|
22
|
+
}
|
|
23
|
+
/** 图片(或关键帧) → 画面文字。仅在调用方显式开启时才该被调到。 */
|
|
24
|
+
export declare function ocrImages(files: string[], tools?: MediaToolOptions, signal?: AbortSignal): Promise<{
|
|
25
|
+
entries: OcrEntry[];
|
|
26
|
+
warnings: Array<{
|
|
27
|
+
code: string;
|
|
28
|
+
message: string;
|
|
29
|
+
}>;
|
|
30
|
+
}>;
|
|
31
|
+
//# sourceMappingURL=models.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"models.d.ts","sourceRoot":"","sources":["../../src/media/models.ts"],"names":[],"mappings":"AAAA;;;;;;GAMG;AAKH,OAAO,EAAwB,KAAK,eAAe,EAAE,KAAK,QAAQ,EAAE,KAAK,gBAAgB,EAAE,MAAM,YAAY,CAAA;AAU7G,eAAO,MAAM,eAAe,EAAE,MAAM,CAAC,MAAM,EAAE,MAAM,CAOlD,CAAA;AAED,wBAAgB,aAAa,CAAC,KAAK,EAAE,MAAM,GAAG,MAAM,CAEnD;AAQD,MAAM,WAAW,mBAAmB;IAClC,IAAI,EAAE,QAAQ,EAAE,CAAA;IAChB,KAAK,EAAE,MAAM,CAAA;IACb,QAAQ,CAAC,EAAE,MAAM,CAAA;CAClB;AAED,+EAA+E;AAC/E,wBAAsB,UAAU,CAC9B,SAAS,EAAE,MAAM,EACjB,OAAO,EAAE,MAAM,EACf,GAAG,GAAE,eAAoB,EACzB,KAAK,CAAC,EAAE,gBAAgB,EACxB,MAAM,CAAC,EAAE,WAAW,GACnB,OAAO,CAAC,mBAAmB,CAAC,CA2C9B;AAED,MAAM,WAAW,QAAQ;IACvB,IAAI,EAAE,MAAM,CAAA;IACZ,KAAK,EAAE,MAAM,EAAE,CAAA;IACf,IAAI,EAAE,MAAM,CAAA;CACb;AAED,uCAAuC;AACvC,wBAAsB,SAAS,CAC7B,KAAK,EAAE,MAAM,EAAE,EACf,KAAK,CAAC,EAAE,gBAAgB,EACxB,MAAM,CAAC,EAAE,WAAW,GACnB,OAAO,CAAC;IAAE,OAAO,EAAE,QAAQ,EAAE,CAAC;IAAC,QAAQ,EAAE,KAAK,CAAC;QAAE,IAAI,EAAE,MAAM,CAAC;QAAC,OAAO,EAAE,MAAM,CAAA;KAAE,CAAC,CAAA;CAAE,CAAC,CActF"}
|
|
@@ -0,0 +1,87 @@
|
|
|
1
|
+
/**
|
|
2
|
+
* 两个**专用小模型**的调用点:语音识别(whisper)与画面文字 OCR。
|
|
3
|
+
*
|
|
4
|
+
* 都不是大模型,也都不联网推理:模型下载一次后完全离线。helper 脚本随包发布
|
|
5
|
+
* (`scripts/*.py`,由 `uvx --with …` 按需拉起),所以这个包不把 Python/模型塞进 npm tarball,
|
|
6
|
+
* 但会把「要不要下载模型」这个决定**显式**交回调用方。
|
|
7
|
+
*/
|
|
8
|
+
import { execFile } from 'node:child_process';
|
|
9
|
+
import { fileURLToPath } from 'node:url';
|
|
10
|
+
import { promisify } from 'node:util';
|
|
11
|
+
import { MediaToMarkdownError } from './types.js';
|
|
12
|
+
import { UVX_HINT } from './binaries.js';
|
|
13
|
+
const execFileAsync = promisify(execFile);
|
|
14
|
+
/** helper 脚本路径:`lib/*.js` → `scripts/*.py`。 */
|
|
15
|
+
function helperPath(name) {
|
|
16
|
+
return fileURLToPath(new URL(`../../scripts/${name}`, import.meta.url));
|
|
17
|
+
}
|
|
18
|
+
export const ASR_MODEL_SIZES = {
|
|
19
|
+
tiny: '约 75 MB',
|
|
20
|
+
base: '约 145 MB',
|
|
21
|
+
small: '约 460 MB',
|
|
22
|
+
medium: '约 1.5 GB',
|
|
23
|
+
'large-v3': '约 3 GB',
|
|
24
|
+
'large-v3-turbo': '约 1.5 GB',
|
|
25
|
+
};
|
|
26
|
+
export function describeModel(model) {
|
|
27
|
+
return ASR_MODEL_SIZES[model] ?? '未知大小';
|
|
28
|
+
}
|
|
29
|
+
function uvxArgs(tools, packages, script, args) {
|
|
30
|
+
const uvx = tools?.uvxPath?.trim() || process.env.MEDIA_TO_MD_UVX?.trim() || 'uvx';
|
|
31
|
+
const python = tools?.pythonVersion?.trim() || process.env.MEDIA_TO_MD_PYTHON?.trim() || '3.12';
|
|
32
|
+
return [uvx, ['--python', python, ...packages.flatMap(pkg => ['--with', pkg]), 'python', script, ...args]];
|
|
33
|
+
}
|
|
34
|
+
/** 音频 → 带时间戳的转写。`allowModelDownload` 为 false 且本地无模型时抛 `ASR_MODEL_REQUIRED`。 */
|
|
35
|
+
export async function transcribe(audioFile, scratch, asr = {}, tools, signal) {
|
|
36
|
+
const model = asr.model?.trim() || 'small';
|
|
37
|
+
const [cmd, args] = uvxArgs(tools, ['faster-whisper'], helperPath('asr.py'), [audioFile, model, asr.language?.trim() ?? '']);
|
|
38
|
+
let stdout;
|
|
39
|
+
try {
|
|
40
|
+
({ stdout } = await execFileAsync(cmd, args, {
|
|
41
|
+
maxBuffer: 64 * 1024 * 1024,
|
|
42
|
+
env: {
|
|
43
|
+
...process.env,
|
|
44
|
+
...(asr.device ? { MEDIA_TO_MD_ASR_DEVICE: asr.device } : {}),
|
|
45
|
+
...(asr.computeType ? { MEDIA_TO_MD_ASR_COMPUTE: asr.computeType } : {}),
|
|
46
|
+
...(asr.beamSize ? { MEDIA_TO_MD_ASR_BEAM: String(asr.beamSize) } : {}),
|
|
47
|
+
},
|
|
48
|
+
...(signal ? { signal } : {}),
|
|
49
|
+
}));
|
|
50
|
+
}
|
|
51
|
+
catch (cause) {
|
|
52
|
+
const detail = cause instanceof Error ? cause.message : String(cause);
|
|
53
|
+
const looksLikeDownload = /LocalEntryNotFoundError|ConnectError|SSL|Connection|snapshot_download/i.test(detail);
|
|
54
|
+
if (looksLikeDownload && asr.allowModelDownload === false) {
|
|
55
|
+
throw new MediaToMarkdownError('ASR_MODEL_REQUIRED', `无字幕,转写需要下载本地 whisper 模型(${model},${describeModel(model)});当前未允许下载,已放弃转换。`, `先取得模型再重试:设 allowModelDownload=true,或把模型放到本地目录后用 asr.model 指向它。中文建议 large-v3-turbo。`);
|
|
56
|
+
}
|
|
57
|
+
throw new MediaToMarkdownError('ASR_FAILED', `本地转写失败:${detail}`, '检查网络(首次需下载模型,国内可设 HF_ENDPOINT=https://hf-mirror.com)或预先下载模型目录并用 asr.model 指向它。');
|
|
58
|
+
}
|
|
59
|
+
const payload = JSON.parse(stdout.trim().split('\n').filter(Boolean).pop() ?? '{}');
|
|
60
|
+
const cues = (payload.segments ?? [])
|
|
61
|
+
.map(segment => ({
|
|
62
|
+
seconds: Number(segment.start) || 0,
|
|
63
|
+
...(Number(segment.end) ? { endSeconds: Number(segment.end) } : {}),
|
|
64
|
+
text: (segment.text ?? '').trim(),
|
|
65
|
+
}))
|
|
66
|
+
.filter(cue => cue.text.length > 0);
|
|
67
|
+
return { cues, model, ...(payload.language ? { language: payload.language } : {}) };
|
|
68
|
+
}
|
|
69
|
+
/** 图片(或关键帧) → 画面文字。仅在调用方显式开启时才该被调到。 */
|
|
70
|
+
export async function ocrImages(files, tools, signal) {
|
|
71
|
+
if (files.length === 0)
|
|
72
|
+
return { entries: [], warnings: [] };
|
|
73
|
+
const [cmd, args] = uvxArgs(tools, ['rapidocr-onnxruntime'], helperPath('ocr.py'), files);
|
|
74
|
+
try {
|
|
75
|
+
const { stdout } = await execFileAsync(cmd, args, { maxBuffer: 32 * 1024 * 1024, ...(signal ? { signal } : {}) });
|
|
76
|
+
const entries = JSON.parse(stdout.trim().split('\n').filter(Boolean).pop() ?? '[]');
|
|
77
|
+
return { entries, warnings: [] };
|
|
78
|
+
}
|
|
79
|
+
catch (cause) {
|
|
80
|
+
const detail = cause instanceof Error ? cause.message : String(cause);
|
|
81
|
+
if (/ENOENT/.test(detail)) {
|
|
82
|
+
return { entries: [], warnings: [{ code: 'OCR_UNAVAILABLE', message: `未找到 uvx,跳过 OCR。${UVX_HINT}` }] };
|
|
83
|
+
}
|
|
84
|
+
return { entries: [], warnings: [{ code: 'OCR_FAILED', message: `OCR 失败:${detail}` }] };
|
|
85
|
+
}
|
|
86
|
+
}
|
|
87
|
+
//# sourceMappingURL=models.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"models.js","sourceRoot":"","sources":["../../src/media/models.ts"],"names":[],"mappings":"AAAA;;;;;;GAMG;AAEH,OAAO,EAAE,QAAQ,EAAE,MAAM,oBAAoB,CAAA;AAC7C,OAAO,EAAE,aAAa,EAAE,MAAM,UAAU,CAAA;AACxC,OAAO,EAAE,SAAS,EAAE,MAAM,WAAW,CAAA;AACrC,OAAO,EAAE,oBAAoB,EAA8D,MAAM,YAAY,CAAA;AAC7G,OAAO,EAAE,QAAQ,EAAE,MAAM,eAAe,CAAA;AAExC,MAAM,aAAa,GAAG,SAAS,CAAC,QAAQ,CAAC,CAAA;AAEzC,+CAA+C;AAC/C,SAAS,UAAU,CAAC,IAAY;IAC9B,OAAO,aAAa,CAAC,IAAI,GAAG,CAAC,iBAAiB,IAAI,EAAE,EAAE,MAAM,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,CAAA;AACzE,CAAC;AAED,MAAM,CAAC,MAAM,eAAe,GAA2B;IACrD,IAAI,EAAE,SAAS;IACf,IAAI,EAAE,UAAU;IAChB,KAAK,EAAE,UAAU;IACjB,MAAM,EAAE,UAAU;IAClB,UAAU,EAAE,QAAQ;IACpB,gBAAgB,EAAE,UAAU;CAC7B,CAAA;AAED,MAAM,UAAU,aAAa,CAAC,KAAa;IACzC,OAAO,eAAe,CAAC,KAAK,CAAC,IAAI,MAAM,CAAA;AACzC,CAAC;AAED,SAAS,OAAO,CAAC,KAAmC,EAAE,QAAkB,EAAE,MAAc,EAAE,IAAc;IACtG,MAAM,GAAG,GAAG,KAAK,EAAE,OAAO,EAAE,IAAI,EAAE,IAAI,OAAO,CAAC,GAAG,CAAC,eAAe,EAAE,IAAI,EAAE,IAAI,KAAK,CAAA;IAClF,MAAM,MAAM,GAAG,KAAK,EAAE,aAAa,EAAE,IAAI,EAAE,IAAI,OAAO,CAAC,GAAG,CAAC,kBAAkB,EAAE,IAAI,EAAE,IAAI,MAAM,CAAA;IAC/F,OAAO,CAAC,GAAG,EAAE,CAAC,UAAU,EAAE,MAAM,EAAE,GAAG,QAAQ,CAAC,OAAO,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC,QAAQ,EAAE,GAAG,CAAC,CAAC,EAAE,QAAQ,EAAE,MAAM,EAAE,GAAG,IAAI,CAAC,CAAC,CAAA;AAC5G,CAAC;AAQD,+EAA+E;AAC/E,MAAM,CAAC,KAAK,UAAU,UAAU,CAC9B,SAAiB,EACjB,OAAe,EACf,MAAuB,EAAE,EACzB,KAAwB,EACxB,MAAoB;IAEpB,MAAM,KAAK,GAAG,GAAG,CAAC,KAAK,EAAE,IAAI,EAAE,IAAI,OAAO,CAAA;IAC1C,MAAM,CAAC,GAAG,EAAE,IAAI,CAAC,GAAG,OAAO,CAAC,KAAK,EAAE,CAAC,gBAAgB,CAAC,EAAE,UAAU,CAAC,QAAQ,CAAC,EAAE,CAAC,SAAS,EAAE,KAAK,EAAE,GAAG,CAAC,QAAQ,EAAE,IAAI,EAAE,IAAI,EAAE,CAAC,CAAC,CAAA;IAC5H,IAAI,MAAc,CAAA;IAClB,IAAI,CAAC;QACH,CAAC,EAAE,MAAM,EAAE,GAAG,MAAM,aAAa,CAAC,GAAG,EAAE,IAAI,EAAE;YAC3C,SAAS,EAAE,EAAE,GAAG,IAAI,GAAG,IAAI;YAC3B,GAAG,EAAE;gBACH,GAAG,OAAO,CAAC,GAAG;gBACd,GAAG,CAAC,GAAG,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,sBAAsB,EAAE,GAAG,CAAC,MAAM,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;gBAC7D,GAAG,CAAC,GAAG,CAAC,WAAW,CAAC,CAAC,CAAC,EAAE,uBAAuB,EAAE,GAAG,CAAC,WAAW,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;gBACxE,GAAG,CAAC,GAAG,CAAC,QAAQ,CAAC,CAAC,CAAC,EAAE,oBAAoB,EAAE,MAAM,CAAC,GAAG,CAAC,QAAQ,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;aACxE;YACD,GAAG,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,MAAM,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;SAC9B,CAAC,CAAC,CAAA;IACL,CAAC;IAAC,OAAO,KAAK,EAAE,CAAC;QACf,MAAM,MAAM,GAAG,KAAK,YAAY,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,CAAA;QACrE,MAAM,iBAAiB,GAAG,wEAAwE,CAAC,IAAI,CAAC,MAAM,CAAC,CAAA;QAC/G,IAAI,iBAAiB,IAAI,GAAG,CAAC,kBAAkB,KAAK,KAAK,EAAE,CAAC;YAC1D,MAAM,IAAI,oBAAoB,CAC5B,oBAAoB,EACpB,2BAA2B,KAAK,IAAI,aAAa,CAAC,KAAK,CAAC,kBAAkB,EAC1E,oFAAoF,CACrF,CAAA;QACH,CAAC;QACD,MAAM,IAAI,oBAAoB,CAC5B,YAAY,EACZ,UAAU,MAAM,EAAE,EAClB,gFAAgF,CACjF,CAAA;IACH,CAAC;IACD,MAAM,OAAO,GAAG,IAAI,CAAC,KAAK,CAAC,MAAM,CAAC,IAAI,EAAE,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC,GAAG,EAAE,IAAI,IAAI,CAGjF,CAAA;IACD,MAAM,IAAI,GAAG,CAAC,OAAO,CAAC,QAAQ,IAAI,EAAE,CAAC;SAClC,GAAG,CAAC,OAAO,CAAC,EAAE,CAAC,CAAC;QACf,OAAO,EAAE,MAAM,CAAC,OAAO,CAAC,KAAK,CAAC,IAAI,CAAC;QACnC,GAAG,CAAC,MAAM,CAAC,OAAO,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,EAAE,UAAU,EAAE,MAAM,CAAC,OAAO,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;QACnE,IAAI,EAAE,CAAC,OAAO,CAAC,IAAI,IAAI,EAAE,CAAC,CAAC,IAAI,EAAE;KAClC,CAAC,CAAC;SACF,MAAM,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,CAAC,IAAI,CAAC,MAAM,GAAG,CAAC,CAAC,CAAA;IACrC,OAAO,EAAE,IAAI,EAAE,KAAK,EAAE,GAAG,CAAC,OAAO,CAAC,QAAQ,CAAC,CAAC,CAAC,EAAE,QAAQ,EAAE,OAAO,CAAC,QAAQ,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,CAAA;AACrF,CAAC;AAQD,uCAAuC;AACvC,MAAM,CAAC,KAAK,UAAU,SAAS,CAC7B,KAAe,EACf,KAAwB,EACxB,MAAoB;IAEpB,IAAI,KAAK,CAAC,MAAM,KAAK,CAAC;QAAE,OAAO,EAAE,OAAO,EAAE,EAAE,EAAE,QAAQ,EAAE,EAAE,EAAE,CAAA;IAC5D,MAAM,CAAC,GAAG,EAAE,IAAI,CAAC,GAAG,OAAO,CAAC,KAAK,EAAE,CAAC,sBAAsB,CAAC,EAAE,UAAU,CAAC,QAAQ,CAAC,EAAE,KAAK,CAAC,CAAA;IACzF,IAAI,CAAC;QACH,MAAM,EAAE,MAAM,EAAE,GAAG,MAAM,aAAa,CAAC,GAAG,EAAE,IAAI,EAAE,EAAE,SAAS,EAAE,EAAE,GAAG,IAAI,GAAG,IAAI,EAAE,GAAG,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,MAAM,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,CAAC,CAAA;QACjH,MAAM,OAAO,GAAG,IAAI,CAAC,KAAK,CAAC,MAAM,CAAC,IAAI,EAAE,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC,GAAG,EAAE,IAAI,IAAI,CAAe,CAAA;QACjG,OAAO,EAAE,OAAO,EAAE,QAAQ,EAAE,EAAE,EAAE,CAAA;IAClC,CAAC;IAAC,OAAO,KAAK,EAAE,CAAC;QACf,MAAM,MAAM,GAAG,KAAK,YAAY,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,CAAA;QACrE,IAAI,QAAQ,CAAC,IAAI,CAAC,MAAM,CAAC,EAAE,CAAC;YAC1B,OAAO,EAAE,OAAO,EAAE,EAAE,EAAE,QAAQ,EAAE,CAAC,EAAE,IAAI,EAAE,iBAAiB,EAAE,OAAO,EAAE,kBAAkB,QAAQ,EAAE,EAAE,CAAC,EAAE,CAAA;QACxG,CAAC;QACD,OAAO,EAAE,OAAO,EAAE,EAAE,EAAE,QAAQ,EAAE,CAAC,EAAE,IAAI,EAAE,YAAY,EAAE,OAAO,EAAE,UAAU,MAAM,EAAE,EAAE,CAAC,EAAE,CAAA;IACzF,CAAC;AACH,CAAC"}
|
|
@@ -0,0 +1,18 @@
|
|
|
1
|
+
/**
|
|
2
|
+
* 轻量扫描档:纯图片页 PDF → 逐页本地 OCR → Markdown。
|
|
3
|
+
*
|
|
4
|
+
* 解决的**唯一**问题:扫描件里的字读不出来(anydoc 只吃文本层 PDF)。
|
|
5
|
+
* 刻意不做哪些事:
|
|
6
|
+
*
|
|
7
|
+
* - **不引入 docling / PaddlePaddle / MinerU**:那是 issues/0017 的"可选高质量档",
|
|
8
|
+
* 用来做版面顺序、表格结构、公式还原;代价是几百 MB ~ 20 GB 的依赖。
|
|
9
|
+
* - **不承诺版面**:输出就是"第 N 页 + 该页图片 + 该页文字",顺序 = 页序。
|
|
10
|
+
* - **不联网**:整条链只有两个本地组件 —— pypdfium2(渲染页图)与 RapidOCR(识别文字)。
|
|
11
|
+
*
|
|
12
|
+
* 渲染用 pypdfium2 而不是 poppler,是因为 `pdftoppm` 是外部二进制,不是每台机器都有
|
|
13
|
+
* (本机就没有);pypdfium2 是 pip wheel,自带 pdfium,由 uvx 按需拉起。
|
|
14
|
+
*/
|
|
15
|
+
import { type MediaToMarkdownResult, type ScannedPdfOptions } from './types.js';
|
|
16
|
+
/** 纯图片页 PDF → 逐页 OCR 的 Markdown(轻量档)。 */
|
|
17
|
+
export declare function convertScannedDocument(options: ScannedPdfOptions): Promise<MediaToMarkdownResult>;
|
|
18
|
+
//# sourceMappingURL=scanned.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"scanned.d.ts","sourceRoot":"","sources":["../../src/media/scanned.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;GAaG;AAaH,OAAO,EAEL,KAAK,qBAAqB,EAI1B,KAAK,iBAAiB,EACvB,MAAM,YAAY,CAAA;AAgEnB,yCAAyC;AACzC,wBAAsB,sBAAsB,CAAC,OAAO,EAAE,iBAAiB,GAAG,OAAO,CAAC,qBAAqB,CAAC,CAqGvG"}
|
|
@@ -0,0 +1,179 @@
|
|
|
1
|
+
/**
|
|
2
|
+
* 轻量扫描档:纯图片页 PDF → 逐页本地 OCR → Markdown。
|
|
3
|
+
*
|
|
4
|
+
* 解决的**唯一**问题:扫描件里的字读不出来(anydoc 只吃文本层 PDF)。
|
|
5
|
+
* 刻意不做哪些事:
|
|
6
|
+
*
|
|
7
|
+
* - **不引入 docling / PaddlePaddle / MinerU**:那是 issues/0017 的"可选高质量档",
|
|
8
|
+
* 用来做版面顺序、表格结构、公式还原;代价是几百 MB ~ 20 GB 的依赖。
|
|
9
|
+
* - **不承诺版面**:输出就是"第 N 页 + 该页图片 + 该页文字",顺序 = 页序。
|
|
10
|
+
* - **不联网**:整条链只有两个本地组件 —— pypdfium2(渲染页图)与 RapidOCR(识别文字)。
|
|
11
|
+
*
|
|
12
|
+
* 渲染用 pypdfium2 而不是 poppler,是因为 `pdftoppm` 是外部二进制,不是每台机器都有
|
|
13
|
+
* (本机就没有);pypdfium2 是 pip wheel,自带 pdfium,由 uvx 按需拉起。
|
|
14
|
+
*/
|
|
15
|
+
import { createHash } from 'node:crypto';
|
|
16
|
+
import { createReadStream } from 'node:fs';
|
|
17
|
+
import { execFile } from 'node:child_process';
|
|
18
|
+
import { mkdir, mkdtemp, readFile, rm, stat } from 'node:fs/promises';
|
|
19
|
+
import { basename, extname, join } from 'node:path';
|
|
20
|
+
import { tmpdir } from 'node:os';
|
|
21
|
+
import { fileURLToPath } from 'node:url';
|
|
22
|
+
import { promisify } from 'node:util';
|
|
23
|
+
import { UVX_HINT } from './binaries.js';
|
|
24
|
+
import { buildScannedMarkdown } from './markdown.js';
|
|
25
|
+
import { ocrImages } from './models.js';
|
|
26
|
+
import { MediaToMarkdownError, } from './types.js';
|
|
27
|
+
const execFileAsync = promisify(execFile);
|
|
28
|
+
function helperPath(name) {
|
|
29
|
+
return fileURLToPath(new URL(`../../scripts/${name}`, import.meta.url));
|
|
30
|
+
}
|
|
31
|
+
function uvxCommand(tools) {
|
|
32
|
+
return {
|
|
33
|
+
cmd: tools?.uvxPath?.trim() || process.env.MEDIA_TO_MD_UVX?.trim() || 'uvx',
|
|
34
|
+
python: tools?.pythonVersion?.trim() || process.env.MEDIA_TO_MD_PYTHON?.trim() || '3.12',
|
|
35
|
+
};
|
|
36
|
+
}
|
|
37
|
+
async function sha256OfFile(file) {
|
|
38
|
+
const hash = createHash('sha256');
|
|
39
|
+
await new Promise((resolve, reject) => {
|
|
40
|
+
createReadStream(file)
|
|
41
|
+
.on('data', chunk => hash.update(chunk))
|
|
42
|
+
.on('error', reject)
|
|
43
|
+
.on('end', () => resolve());
|
|
44
|
+
});
|
|
45
|
+
return `sha256:${hash.digest('hex')}`;
|
|
46
|
+
}
|
|
47
|
+
async function renderPages(file, scratch, dpi, maxPages, tools, signal) {
|
|
48
|
+
const { cmd, python } = uvxCommand(tools);
|
|
49
|
+
const outDir = join(scratch, 'pages');
|
|
50
|
+
await mkdir(outDir, { recursive: true });
|
|
51
|
+
let stdout;
|
|
52
|
+
try {
|
|
53
|
+
({ stdout } = await execFileAsync(cmd, [
|
|
54
|
+
'--python', python, '--with', 'pypdfium2', '--with', 'pillow', 'python', helperPath('pdf_pages.py'),
|
|
55
|
+
file, outDir, String(dpi), String(maxPages),
|
|
56
|
+
], { maxBuffer: 32 * 1024 * 1024, ...(signal ? { signal } : {}) }));
|
|
57
|
+
}
|
|
58
|
+
catch (cause) {
|
|
59
|
+
const detail = cause instanceof Error ? cause.message : String(cause);
|
|
60
|
+
if (/ENOENT/.test(detail)) {
|
|
61
|
+
throw new MediaToMarkdownError('DEPENDENCY_MISSING', '扫描档需要 uvx(用于拉起 pypdfium2 渲染页图与 RapidOCR 识别),本机没有找到。', UVX_HINT);
|
|
62
|
+
}
|
|
63
|
+
throw new MediaToMarkdownError('PDF_RENDER_FAILED', `页图渲染失败:${detail}`);
|
|
64
|
+
}
|
|
65
|
+
const payload = JSON.parse(stdout.trim().split('\n').filter(Boolean).pop() ?? '{}');
|
|
66
|
+
// uvx 版本只在能力上报里用;这里顺手探测一次,失败不影响主流程
|
|
67
|
+
let uvxVersion;
|
|
68
|
+
try {
|
|
69
|
+
const version = await execFileAsync(cmd, ['--version']);
|
|
70
|
+
uvxVersion = /(\d+\.\d+\.\d+)/.exec(version.stdout)?.[1];
|
|
71
|
+
}
|
|
72
|
+
catch {
|
|
73
|
+
/* 可选 */
|
|
74
|
+
}
|
|
75
|
+
return { ...payload, ...(uvxVersion ? { uvxVersion } : {}) };
|
|
76
|
+
}
|
|
77
|
+
/** 纯图片页 PDF → 逐页 OCR 的 Markdown(轻量档)。 */
|
|
78
|
+
export async function convertScannedDocument(options) {
|
|
79
|
+
const file = options.file;
|
|
80
|
+
const fileInfo = await stat(file).catch(() => undefined);
|
|
81
|
+
if (!fileInfo?.isFile())
|
|
82
|
+
throw new MediaToMarkdownError('INPUT_NOT_A_FILE', `输入不是文件:${file}`);
|
|
83
|
+
if (extname(file).toLowerCase() !== '.pdf') {
|
|
84
|
+
throw new MediaToMarkdownError('INPUT_NOT_PDF', `扫描档只处理 PDF,收到:${extname(file) || '(无扩展名)'}`);
|
|
85
|
+
}
|
|
86
|
+
const title = options.title?.trim() || basename(file, extname(file));
|
|
87
|
+
const dpi = Math.min(300, Math.max(72, options.dpi ?? 150));
|
|
88
|
+
const maxPages = Math.max(1, options.maxPages ?? 50);
|
|
89
|
+
const pageImages = options.pageImages !== false;
|
|
90
|
+
const report = (stage, progress, message) => {
|
|
91
|
+
options.onProgress?.({ stage, progress, ...(message ? { message } : {}) });
|
|
92
|
+
};
|
|
93
|
+
const warnings = [];
|
|
94
|
+
const scratch = await mkdtemp(join(tmpdir(), 'media-to-md-scan-'));
|
|
95
|
+
try {
|
|
96
|
+
report('probing', 0.05, '渲染 PDF 页图');
|
|
97
|
+
const rendered = await renderPages(file, scratch, dpi, maxPages, options.tools, options.signal);
|
|
98
|
+
if (rendered.totalPages > rendered.pages.length) {
|
|
99
|
+
warnings.push({
|
|
100
|
+
code: 'PAGE_LIMIT_REACHED',
|
|
101
|
+
message: `PDF 共 ${rendered.totalPages} 页,本次只渲染并识别前 ${rendered.pages.length} 页(上限 ${maxPages});调大 maxPages 可覆盖更多。`,
|
|
102
|
+
});
|
|
103
|
+
}
|
|
104
|
+
if (rendered.pages.length === 0) {
|
|
105
|
+
warnings.push({ code: 'NO_PAGES', message: '该 PDF 没有可渲染的页面。' });
|
|
106
|
+
}
|
|
107
|
+
report('ocr', 0.4, `本地 OCR(${rendered.pages.length} 页)`);
|
|
108
|
+
const ocr = await ocrImages(rendered.pages.map(page => page.path), options.tools, options.signal);
|
|
109
|
+
warnings.push(...ocr.warnings);
|
|
110
|
+
const byPath = new Map(ocr.entries.map(entry => [entry.path, entry]));
|
|
111
|
+
const pages = [];
|
|
112
|
+
const assets = [];
|
|
113
|
+
let emptyPages = 0;
|
|
114
|
+
for (const page of rendered.pages) {
|
|
115
|
+
const rel = `assets/page-${String(page.index).padStart(4, '0')}.png`;
|
|
116
|
+
const entry = byPath.get(page.path);
|
|
117
|
+
const text = (entry?.text ?? '').trim();
|
|
118
|
+
if (!text)
|
|
119
|
+
emptyPages += 1;
|
|
120
|
+
pages.push({
|
|
121
|
+
index: page.index,
|
|
122
|
+
rel,
|
|
123
|
+
...(page.width ? { width: page.width } : {}),
|
|
124
|
+
...(page.height ? { height: page.height } : {}),
|
|
125
|
+
text,
|
|
126
|
+
ocrLines: entry?.lines.length ?? 0,
|
|
127
|
+
});
|
|
128
|
+
if (pageImages) {
|
|
129
|
+
assets.push({ rel, data: await readFile(page.path), contentType: 'image/png' });
|
|
130
|
+
}
|
|
131
|
+
}
|
|
132
|
+
const identified = pages.filter(page => page.text).length;
|
|
133
|
+
warnings.push({
|
|
134
|
+
code: 'SCANNED_OCR_LOCAL',
|
|
135
|
+
message: `扫描档走本地 OCR:${identified}/${pages.length} 页识别到文字(${dpi} DPI);不联网、不做版面还原。`,
|
|
136
|
+
});
|
|
137
|
+
if (emptyPages > 0) {
|
|
138
|
+
warnings.push({ code: 'OCR_EMPTY_PAGES', message: `有 ${emptyPages} 页没有识别到文字(可能是空白页、手写体或图片型内容)。` });
|
|
139
|
+
}
|
|
140
|
+
if (!pageImages) {
|
|
141
|
+
warnings.push({ code: 'PAGE_IMAGES_SKIPPED', message: 'pageImages=false:只保留文字,未落地页图。' });
|
|
142
|
+
}
|
|
143
|
+
report('assembling', 0.9, '装配 Markdown');
|
|
144
|
+
const meta = {
|
|
145
|
+
sourceFormat: 'pdf',
|
|
146
|
+
containerExtension: 'pdf',
|
|
147
|
+
durationSeconds: 0,
|
|
148
|
+
sizeBytes: fileInfo.size,
|
|
149
|
+
hasVideo: false,
|
|
150
|
+
hasAudio: false,
|
|
151
|
+
hasSubtitleTrack: false,
|
|
152
|
+
transcriptSource: 'none',
|
|
153
|
+
};
|
|
154
|
+
const dependencyVersions = { pypdfium2: 'via uvx', rapidocr: 'via uvx' };
|
|
155
|
+
if (rendered.uvxVersion)
|
|
156
|
+
dependencyVersions.uvx = rendered.uvxVersion;
|
|
157
|
+
return {
|
|
158
|
+
title: title.slice(0, 300),
|
|
159
|
+
markdown: buildScannedMarkdown({
|
|
160
|
+
title,
|
|
161
|
+
totalPages: rendered.totalPages,
|
|
162
|
+
pageImages,
|
|
163
|
+
pages: pages.map(page => ({ index: page.index, rel: page.rel, text: page.text })),
|
|
164
|
+
}),
|
|
165
|
+
assets,
|
|
166
|
+
meta,
|
|
167
|
+
transcript: { source: 'none', cues: [] },
|
|
168
|
+
warnings,
|
|
169
|
+
dependencyVersions,
|
|
170
|
+
externalId: await sha256OfFile(file),
|
|
171
|
+
pages,
|
|
172
|
+
};
|
|
173
|
+
}
|
|
174
|
+
finally {
|
|
175
|
+
if (!options.keepScratch)
|
|
176
|
+
await rm(scratch, { recursive: true, force: true }).catch(() => undefined);
|
|
177
|
+
}
|
|
178
|
+
}
|
|
179
|
+
//# sourceMappingURL=scanned.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"scanned.js","sourceRoot":"","sources":["../../src/media/scanned.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;GAaG;AAEH,OAAO,EAAE,UAAU,EAAE,MAAM,aAAa,CAAA;AACxC,OAAO,EAAE,gBAAgB,EAAE,MAAM,SAAS,CAAA;AAC1C,OAAO,EAAE,QAAQ,EAAE,MAAM,oBAAoB,CAAA;AAC7C,OAAO,EAAE,KAAK,EAAE,OAAO,EAAE,QAAQ,EAAE,EAAE,EAAE,IAAI,EAAE,MAAM,kBAAkB,CAAA;AACrE,OAAO,EAAE,QAAQ,EAAE,OAAO,EAAE,IAAI,EAAE,MAAM,WAAW,CAAA;AACnD,OAAO,EAAE,MAAM,EAAE,MAAM,SAAS,CAAA;AAChC,OAAO,EAAE,aAAa,EAAE,MAAM,UAAU,CAAA;AACxC,OAAO,EAAE,SAAS,EAAE,MAAM,WAAW,CAAA;AACrC,OAAO,EAAE,QAAQ,EAAE,MAAM,eAAe,CAAA;AACxC,OAAO,EAAE,oBAAoB,EAAE,MAAM,eAAe,CAAA;AACpD,OAAO,EAAE,SAAS,EAAE,MAAM,aAAa,CAAA;AACvC,OAAO,EACL,oBAAoB,GAMrB,MAAM,YAAY,CAAA;AAEnB,MAAM,aAAa,GAAG,SAAS,CAAC,QAAQ,CAAC,CAAA;AAEzC,SAAS,UAAU,CAAC,IAAY;IAC9B,OAAO,aAAa,CAAC,IAAI,GAAG,CAAC,iBAAiB,IAAI,EAAE,EAAE,MAAM,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,CAAA;AACzE,CAAC;AAED,SAAS,UAAU,CAAC,KAAmC;IACrD,OAAO;QACL,GAAG,EAAE,KAAK,EAAE,OAAO,EAAE,IAAI,EAAE,IAAI,OAAO,CAAC,GAAG,CAAC,eAAe,EAAE,IAAI,EAAE,IAAI,KAAK;QAC3E,MAAM,EAAE,KAAK,EAAE,aAAa,EAAE,IAAI,EAAE,IAAI,OAAO,CAAC,GAAG,CAAC,kBAAkB,EAAE,IAAI,EAAE,IAAI,MAAM;KACzF,CAAA;AACH,CAAC;AAED,KAAK,UAAU,YAAY,CAAC,IAAY;IACtC,MAAM,IAAI,GAAG,UAAU,CAAC,QAAQ,CAAC,CAAA;IACjC,MAAM,IAAI,OAAO,CAAO,CAAC,OAAO,EAAE,MAAM,EAAE,EAAE;QAC1C,gBAAgB,CAAC,IAAI,CAAC;aACnB,EAAE,CAAC,MAAM,EAAE,KAAK,CAAC,EAAE,CAAC,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC;aACvC,EAAE,CAAC,OAAO,EAAE,MAAM,CAAC;aACnB,EAAE,CAAC,KAAK,EAAE,GAAG,EAAE,CAAC,OAAO,EAAE,CAAC,CAAA;IAC/B,CAAC,CAAC,CAAA;IACF,OAAO,UAAU,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,EAAE,CAAA;AACvC,CAAC;AAOD,KAAK,UAAU,WAAW,CAAC,IAAY,EAAE,OAAe,EAAE,GAAW,EAAE,QAAgB,EAAE,KAAwB,EAAE,MAAoB;IACrI,MAAM,EAAE,GAAG,EAAE,MAAM,EAAE,GAAG,UAAU,CAAC,KAAK,CAAC,CAAA;IACzC,MAAM,MAAM,GAAG,IAAI,CAAC,OAAO,EAAE,OAAO,CAAC,CAAA;IACrC,MAAM,KAAK,CAAC,MAAM,EAAE,EAAE,SAAS,EAAE,IAAI,EAAE,CAAC,CAAA;IACxC,IAAI,MAAc,CAAA;IAClB,IAAI,CAAC;QACH,CAAC,EAAE,MAAM,EAAE,GAAG,MAAM,aAAa,CAAC,GAAG,EAAE;YACrC,UAAU,EAAE,MAAM,EAAE,QAAQ,EAAE,WAAW,EAAE,QAAQ,EAAE,QAAQ,EAAE,QAAQ,EAAE,UAAU,CAAC,cAAc,CAAC;YACnG,IAAI,EAAE,MAAM,EAAE,MAAM,CAAC,GAAG,CAAC,EAAE,MAAM,CAAC,QAAQ,CAAC;SAC5C,EAAE,EAAE,SAAS,EAAE,EAAE,GAAG,IAAI,GAAG,IAAI,EAAE,GAAG,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,MAAM,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,CAAC,CAAC,CAAA;IACrE,CAAC;IAAC,OAAO,KAAK,EAAE,CAAC;QACf,MAAM,MAAM,GAAG,KAAK,YAAY,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,CAAA;QACrE,IAAI,QAAQ,CAAC,IAAI,CAAC,MAAM,CAAC,EAAE,CAAC;YAC1B,MAAM,IAAI,oBAAoB,CAC5B,oBAAoB,EACpB,qDAAqD,EACrD,QAAQ,CACT,CAAA;QACH,CAAC;QACD,MAAM,IAAI,oBAAoB,CAAC,mBAAmB,EAAE,UAAU,MAAM,EAAE,CAAC,CAAA;IACzE,CAAC;IACD,MAAM,OAAO,GAAG,IAAI,CAAC,KAAK,CAAC,MAAM,CAAC,IAAI,EAAE,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC,GAAG,EAAE,IAAI,IAAI,CAAkB,CAAA;IACpG,mCAAmC;IACnC,IAAI,UAA8B,CAAA;IAClC,IAAI,CAAC;QACH,MAAM,OAAO,GAAG,MAAM,aAAa,CAAC,GAAG,EAAE,CAAC,WAAW,CAAC,CAAC,CAAA;QACvD,UAAU,GAAG,iBAAiB,CAAC,IAAI,CAAC,OAAO,CAAC,MAAM,CAAC,EAAE,CAAC,CAAC,CAAC,CAAA;IAC1D,CAAC;IAAC,MAAM,CAAC;QACP,QAAQ;IACV,CAAC;IACD,OAAO,EAAE,GAAG,OAAO,EAAE,GAAG,CAAC,UAAU,CAAC,CAAC,CAAC,EAAE,UAAU,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,CAAA;AAC9D,CAAC;AAED,yCAAyC;AACzC,MAAM,CAAC,KAAK,UAAU,sBAAsB,CAAC,OAA0B;IACrE,MAAM,IAAI,GAAG,OAAO,CAAC,IAAI,CAAA;IACzB,MAAM,QAAQ,GAAG,MAAM,IAAI,CAAC,IAAI,CAAC,CAAC,KAAK,CAAC,GAAG,EAAE,CAAC,SAAS,CAAC,CAAA;IACxD,IAAI,CAAC,QAAQ,EAAE,MAAM,EAAE;QAAE,MAAM,IAAI,oBAAoB,CAAC,kBAAkB,EAAE,UAAU,IAAI,EAAE,CAAC,CAAA;IAC7F,IAAI,OAAO,CAAC,IAAI,CAAC,CAAC,WAAW,EAAE,KAAK,MAAM,EAAE,CAAC;QAC3C,MAAM,IAAI,oBAAoB,CAAC,eAAe,EAAE,iBAAiB,OAAO,CAAC,IAAI,CAAC,IAAI,QAAQ,EAAE,CAAC,CAAA;IAC/F,CAAC;IAED,MAAM,KAAK,GAAG,OAAO,CAAC,KAAK,EAAE,IAAI,EAAE,IAAI,QAAQ,CAAC,IAAI,EAAE,OAAO,CAAC,IAAI,CAAC,CAAC,CAAA;IACpE,MAAM,GAAG,GAAG,IAAI,CAAC,GAAG,CAAC,GAAG,EAAE,IAAI,CAAC,GAAG,CAAC,EAAE,EAAE,OAAO,CAAC,GAAG,IAAI,GAAG,CAAC,CAAC,CAAA;IAC3D,MAAM,QAAQ,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,OAAO,CAAC,QAAQ,IAAI,EAAE,CAAC,CAAA;IACpD,MAAM,UAAU,GAAG,OAAO,CAAC,UAAU,KAAK,KAAK,CAAA;IAC/C,MAAM,MAAM,GAAG,CAAC,KAA2E,EAAE,QAAuB,EAAE,OAAgB,EAAQ,EAAE;QAC9I,OAAO,CAAC,UAAU,EAAE,CAAC,EAAE,KAAK,EAAE,QAAQ,EAAE,GAAG,CAAC,OAAO,CAAC,CAAC,CAAC,EAAE,OAAO,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,CAAC,CAAA;IAC5E,CAAC,CAAA;IACD,MAAM,QAAQ,GAAmB,EAAE,CAAA;IACnC,MAAM,OAAO,GAAG,MAAM,OAAO,CAAC,IAAI,CAAC,MAAM,EAAE,EAAE,mBAAmB,CAAC,CAAC,CAAA;IAElE,IAAI,CAAC;QACH,MAAM,CAAC,SAAS,EAAE,IAAI,EAAE,WAAW,CAAC,CAAA;QACpC,MAAM,QAAQ,GAAG,MAAM,WAAW,CAAC,IAAI,EAAE,OAAO,EAAE,GAAG,EAAE,QAAQ,EAAE,OAAO,CAAC,KAAK,EAAE,OAAO,CAAC,MAAM,CAAC,CAAA;QAC/F,IAAI,QAAQ,CAAC,UAAU,GAAG,QAAQ,CAAC,KAAK,CAAC,MAAM,EAAE,CAAC;YAChD,QAAQ,CAAC,IAAI,CAAC;gBACZ,IAAI,EAAE,oBAAoB;gBAC1B,OAAO,EAAE,SAAS,QAAQ,CAAC,UAAU,gBAAgB,QAAQ,CAAC,KAAK,CAAC,MAAM,SAAS,QAAQ,sBAAsB;aAClH,CAAC,CAAA;QACJ,CAAC;QACD,IAAI,QAAQ,CAAC,KAAK,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;YAChC,QAAQ,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,UAAU,EAAE,OAAO,EAAE,iBAAiB,EAAE,CAAC,CAAA;QACjE,CAAC;QAED,MAAM,CAAC,KAAK,EAAE,GAAG,EAAE,UAAU,QAAQ,CAAC,KAAK,CAAC,MAAM,KAAK,CAAC,CAAA;QACxD,MAAM,GAAG,GAAG,MAAM,SAAS,CAAC,QAAQ,CAAC,KAAK,CAAC,GAAG,CAAC,IAAI,CAAC,EAAE,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,OAAO,CAAC,KAAK,EAAE,OAAO,CAAC,MAAM,CAAC,CAAA;QACjG,QAAQ,CAAC,IAAI,CAAC,GAAG,GAAG,CAAC,QAAQ,CAAC,CAAA;QAC9B,MAAM,MAAM,GAAG,IAAI,GAAG,CAAC,GAAG,CAAC,OAAO,CAAC,GAAG,CAAC,KAAK,CAAC,EAAE,CAAC,CAAC,KAAK,CAAC,IAAI,EAAE,KAAK,CAAC,CAAC,CAAC,CAAA;QAErE,MAAM,KAAK,GAAkB,EAAE,CAAA;QAC/B,MAAM,MAAM,GAAoC,EAAE,CAAA;QAClD,IAAI,UAAU,GAAG,CAAC,CAAA;QAClB,KAAK,MAAM,IAAI,IAAI,QAAQ,CAAC,KAAK,EAAE,CAAC;YAClC,MAAM,GAAG,GAAG,eAAe,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,QAAQ,CAAC,CAAC,EAAE,GAAG,CAAC,MAAM,CAAA;YACpE,MAAM,KAAK,GAAG,MAAM,CAAC,GAAG,CAAC,IAAI,CAAC,IAAI,CAAC,CAAA;YACnC,MAAM,IAAI,GAAG,CAAC,KAAK,EAAE,IAAI,IAAI,EAAE,CAAC,CAAC,IAAI,EAAE,CAAA;YACvC,IAAI,CAAC,IAAI;gBAAE,UAAU,IAAI,CAAC,CAAA;YAC1B,KAAK,CAAC,IAAI,CAAC;gBACT,KAAK,EAAE,IAAI,CAAC,KAAK;gBACjB,GAAG;gBACH,GAAG,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE,KAAK,EAAE,IAAI,CAAC,KAAK,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;gBAC5C,GAAG,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,MAAM,EAAE,IAAI,CAAC,MAAM,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;gBAC/C,IAAI;gBACJ,QAAQ,EAAE,KAAK,EAAE,KAAK,CAAC,MAAM,IAAI,CAAC;aACnC,CAAC,CAAA;YACF,IAAI,UAAU,EAAE,CAAC;gBACf,MAAM,CAAC,IAAI,CAAC,EAAE,GAAG,EAAE,IAAI,EAAE,MAAM,QAAQ,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,WAAW,EAAE,WAAW,EAAE,CAAC,CAAA;YACjF,CAAC;QACH,CAAC;QACD,MAAM,UAAU,GAAG,KAAK,CAAC,MAAM,CAAC,IAAI,CAAC,EAAE,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,MAAM,CAAA;QACzD,QAAQ,CAAC,IAAI,CAAC;YACZ,IAAI,EAAE,mBAAmB;YACzB,OAAO,EAAE,cAAc,UAAU,IAAI,KAAK,CAAC,MAAM,WAAW,GAAG,mBAAmB;SACnF,CAAC,CAAA;QACF,IAAI,UAAU,GAAG,CAAC,EAAE,CAAC;YACnB,QAAQ,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,iBAAiB,EAAE,OAAO,EAAE,KAAK,UAAU,8BAA8B,EAAE,CAAC,CAAA;QACpG,CAAC;QACD,IAAI,CAAC,UAAU,EAAE,CAAC;YAChB,QAAQ,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,qBAAqB,EAAE,OAAO,EAAE,+BAA+B,EAAE,CAAC,CAAA;QAC1F,CAAC;QAED,MAAM,CAAC,YAAY,EAAE,GAAG,EAAE,aAAa,CAAC,CAAA;QACxC,MAAM,IAAI,GAAG;YACX,YAAY,EAAE,KAAc;YAC5B,kBAAkB,EAAE,KAAK;YACzB,eAAe,EAAE,CAAC;YAClB,SAAS,EAAE,QAAQ,CAAC,IAAI;YACxB,QAAQ,EAAE,KAAK;YACf,QAAQ,EAAE,KAAK;YACf,gBAAgB,EAAE,KAAK;YACvB,gBAAgB,EAAE,MAAe;SAClC,CAAA;QACD,MAAM,kBAAkB,GAA2B,EAAE,SAAS,EAAE,SAAS,EAAE,QAAQ,EAAE,SAAS,EAAE,CAAA;QAChG,IAAI,QAAQ,CAAC,UAAU;YAAE,kBAAkB,CAAC,GAAG,GAAG,QAAQ,CAAC,UAAU,CAAA;QAErE,OAAO;YACL,KAAK,EAAE,KAAK,CAAC,KAAK,CAAC,CAAC,EAAE,GAAG,CAAC;YAC1B,QAAQ,EAAE,oBAAoB,CAAC;gBAC7B,KAAK;gBACL,UAAU,EAAE,QAAQ,CAAC,UAAU;gBAC/B,UAAU;gBACV,KAAK,EAAE,KAAK,CAAC,GAAG,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC,EAAE,KAAK,EAAE,IAAI,CAAC,KAAK,EAAE,GAAG,EAAE,IAAI,CAAC,GAAG,EAAE,IAAI,EAAE,IAAI,CAAC,IAAI,EAAE,CAAC,CAAC;aAClF,CAAC;YACF,MAAM;YACN,IAAI;YACJ,UAAU,EAAE,EAAE,MAAM,EAAE,MAAM,EAAE,IAAI,EAAE,EAAE,EAAE;YACxC,QAAQ;YACR,kBAAkB;YAClB,UAAU,EAAE,MAAM,YAAY,CAAC,IAAI,CAAC;YACpC,KAAK;SACN,CAAA;IACH,CAAC;YAAS,CAAC;QACT,IAAI,CAAC,OAAO,CAAC,WAAW;YAAE,MAAM,EAAE,CAAC,OAAO,EAAE,EAAE,SAAS,EAAE,IAAI,EAAE,KAAK,EAAE,IAAI,EAAE,CAAC,CAAC,KAAK,CAAC,GAAG,EAAE,CAAC,SAAS,CAAC,CAAA;IACtG,CAAC;AACH,CAAC"}
|