@customize-agent/knowledge 4.0.21 → 4.0.22
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.
|
@@ -11,6 +11,13 @@ import { fileURLToPath } from 'node:url';
|
|
|
11
11
|
import { resolveAndImport, resolvePackage } from './module-resolver.js';
|
|
12
12
|
// ─── 路径工具 ───────────────────────────────────────────────────
|
|
13
13
|
const knowledgeDir = path.dirname(fileURLToPath(import.meta.url));
|
|
14
|
+
const OCR_NATIVE_NOISE_PATTERNS = [/^Image too small to scale!!/u, /^Line cannot be recognized!!$/u];
|
|
15
|
+
const OCR_NOISE_SUPPRESSION_KEY = Symbol.for('customize-agent.ocr-noise-suppression');
|
|
16
|
+
function ocrNoiseSuppressionState() {
|
|
17
|
+
const globalState = globalThis;
|
|
18
|
+
globalState[OCR_NOISE_SUPPRESSION_KEY] ??= { depth: 0 };
|
|
19
|
+
return globalState[OCR_NOISE_SUPPRESSION_KEY];
|
|
20
|
+
}
|
|
14
21
|
function tessdataDir() {
|
|
15
22
|
if (process.env.TESSDATA_PREFIX)
|
|
16
23
|
return process.env.TESSDATA_PREFIX;
|
|
@@ -19,6 +26,44 @@ function tessdataDir() {
|
|
|
19
26
|
return pkg;
|
|
20
27
|
return pkg;
|
|
21
28
|
}
|
|
29
|
+
function isNativeOcrNoise(chunk) {
|
|
30
|
+
const text = Buffer.isBuffer(chunk) ? chunk.toString('utf8') : chunk instanceof Uint8Array ? Buffer.from(chunk).toString('utf8') : typeof chunk === 'string' ? chunk : '';
|
|
31
|
+
if (!text)
|
|
32
|
+
return false;
|
|
33
|
+
const lines = text.split(/\r?\n/u).map(line => line.trim()).filter(Boolean);
|
|
34
|
+
return lines.length > 0 && lines.every(line => OCR_NATIVE_NOISE_PATTERNS.some(pattern => pattern.test(line)));
|
|
35
|
+
}
|
|
36
|
+
async function suppressNativeOcrNoise(operation) {
|
|
37
|
+
const state = ocrNoiseSuppressionState();
|
|
38
|
+
if (state.depth === 0) {
|
|
39
|
+
state.stdout = process.stdout.write;
|
|
40
|
+
state.stderr = process.stderr.write;
|
|
41
|
+
const filter = (original) => function write(chunk, ...args) {
|
|
42
|
+
if (isNativeOcrNoise(chunk)) {
|
|
43
|
+
const callback = args.find((arg) => typeof arg === 'function');
|
|
44
|
+
if (callback)
|
|
45
|
+
process.nextTick(callback);
|
|
46
|
+
return true;
|
|
47
|
+
}
|
|
48
|
+
return original.call(this, chunk, ...args);
|
|
49
|
+
};
|
|
50
|
+
process.stdout.write = filter(state.stdout);
|
|
51
|
+
process.stderr.write = filter(state.stderr);
|
|
52
|
+
}
|
|
53
|
+
state.depth += 1;
|
|
54
|
+
try {
|
|
55
|
+
return await operation();
|
|
56
|
+
}
|
|
57
|
+
finally {
|
|
58
|
+
state.depth -= 1;
|
|
59
|
+
if (state.depth === 0 && state.stdout && state.stderr) {
|
|
60
|
+
process.stdout.write = state.stdout;
|
|
61
|
+
process.stderr.write = state.stderr;
|
|
62
|
+
state.stdout = undefined;
|
|
63
|
+
state.stderr = undefined;
|
|
64
|
+
}
|
|
65
|
+
}
|
|
66
|
+
}
|
|
22
67
|
// ─── Tesseract.js Provider（主 OCR 引擎，跨平台） ──────────────
|
|
23
68
|
export class TesseractJsProvider {
|
|
24
69
|
id = 'tesseract.js';
|
|
@@ -71,7 +116,7 @@ export class TesseractJsProvider {
|
|
|
71
116
|
}
|
|
72
117
|
try {
|
|
73
118
|
const worker = await this.getWorker();
|
|
74
|
-
const result = await worker.recognize(pngPath);
|
|
119
|
+
const result = await suppressNativeOcrNoise(() => worker.recognize(pngPath));
|
|
75
120
|
const text = (result.data.text ?? '').trim();
|
|
76
121
|
const lines = (result.data.lines ?? []);
|
|
77
122
|
const regions = lines
|