@cogitator-ai/rag 0.1.10 → 0.3.0
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/README.md +66 -31
- package/dist/chunkers/create-chunker.js.map +1 -1
- package/dist/chunkers/index.d.ts.map +1 -1
- package/dist/chunkers/index.js.map +1 -1
- package/dist/chunkers/recursive-chunker.d.ts +1 -2
- package/dist/chunkers/recursive-chunker.d.ts.map +1 -1
- package/dist/chunkers/recursive-chunker.js +68 -58
- package/dist/chunkers/recursive-chunker.js.map +1 -1
- package/dist/chunkers/semantic-chunker.d.ts.map +1 -1
- package/dist/chunkers/semantic-chunker.js +5 -2
- package/dist/chunkers/semantic-chunker.js.map +1 -1
- package/dist/index.d.ts +2 -2
- package/dist/index.d.ts.map +1 -1
- package/dist/index.js +4 -2
- package/dist/index.js.map +1 -1
- package/dist/loaders/csv-loader.d.ts.map +1 -1
- package/dist/loaders/csv-loader.js +9 -4
- package/dist/loaders/csv-loader.js.map +1 -1
- package/dist/loaders/html-loader.d.ts +4 -0
- package/dist/loaders/html-loader.d.ts.map +1 -1
- package/dist/loaders/html-loader.js +26 -2
- package/dist/loaders/html-loader.js.map +1 -1
- package/dist/loaders/index.d.ts.map +1 -1
- package/dist/loaders/index.js.map +1 -1
- package/dist/loaders/json-loader.js +1 -1
- package/dist/loaders/json-loader.js.map +1 -1
- package/dist/loaders/markdown-loader.d.ts.map +1 -1
- package/dist/loaders/markdown-loader.js +7 -3
- package/dist/loaders/markdown-loader.js.map +1 -1
- package/dist/loaders/pdf-loader.d.ts +3 -3
- package/dist/loaders/pdf-loader.d.ts.map +1 -1
- package/dist/loaders/pdf-loader.js +45 -42
- package/dist/loaders/pdf-loader.js.map +1 -1
- package/dist/loaders/text-loader.js +1 -1
- package/dist/loaders/text-loader.js.map +1 -1
- package/dist/loaders/web-loader.d.ts +29 -1
- package/dist/loaders/web-loader.d.ts.map +1 -1
- package/dist/loaders/web-loader.js +227 -92
- package/dist/loaders/web-loader.js.map +1 -1
- package/dist/rag-builder.d.ts +10 -1
- package/dist/rag-builder.d.ts.map +1 -1
- package/dist/rag-builder.js +70 -7
- package/dist/rag-builder.js.map +1 -1
- package/dist/rag-pipeline.d.ts +1 -0
- package/dist/rag-pipeline.d.ts.map +1 -1
- package/dist/rag-pipeline.js +55 -19
- package/dist/rag-pipeline.js.map +1 -1
- package/dist/rerankers/index.d.ts.map +1 -1
- package/dist/rerankers/index.js.map +1 -1
- package/dist/retrievers/chunk-indexer.d.ts +19 -0
- package/dist/retrievers/chunk-indexer.d.ts.map +1 -0
- package/dist/retrievers/chunk-indexer.js +8 -0
- package/dist/retrievers/chunk-indexer.js.map +1 -0
- package/dist/retrievers/create-retriever.d.ts +1 -0
- package/dist/retrievers/create-retriever.d.ts.map +1 -1
- package/dist/retrievers/create-retriever.js +1 -0
- package/dist/retrievers/create-retriever.js.map +1 -1
- package/dist/retrievers/hybrid-retriever.d.ts +8 -1
- package/dist/retrievers/hybrid-retriever.d.ts.map +1 -1
- package/dist/retrievers/hybrid-retriever.js +18 -5
- package/dist/retrievers/hybrid-retriever.js.map +1 -1
- package/dist/retrievers/index.d.ts +2 -0
- package/dist/retrievers/index.d.ts.map +1 -1
- package/dist/retrievers/index.js +1 -0
- package/dist/retrievers/index.js.map +1 -1
- package/dist/retrievers/mmr-retriever.d.ts +5 -0
- package/dist/retrievers/mmr-retriever.d.ts.map +1 -1
- package/dist/retrievers/mmr-retriever.js +24 -2
- package/dist/retrievers/mmr-retriever.js.map +1 -1
- package/dist/retrievers/multi-query-retriever.d.ts +12 -1
- package/dist/retrievers/multi-query-retriever.d.ts.map +1 -1
- package/dist/retrievers/multi-query-retriever.js +46 -7
- package/dist/retrievers/multi-query-retriever.js.map +1 -1
- package/dist/retrievers/similarity-retriever.d.ts.map +1 -1
- package/dist/retrievers/similarity-retriever.js +2 -1
- package/dist/retrievers/similarity-retriever.js.map +1 -1
- package/dist/tools.d.ts +11 -2
- package/dist/tools.d.ts.map +1 -1
- package/dist/tools.js +39 -7
- package/dist/tools.js.map +1 -1
- package/package.json +15 -14
|
@@ -9,6 +9,21 @@ async function loadCheerio() {
|
|
|
9
9
|
throw new Error('cheerio is required for HTMLLoader. Install it: pnpm add cheerio');
|
|
10
10
|
}
|
|
11
11
|
}
|
|
12
|
+
const NON_CONTENT_SELECTOR = 'script, style, noscript, template, svg, iframe, object, head';
|
|
13
|
+
const BLOCK_SELECTOR = 'address, article, aside, blockquote, dd, div, dl, dt, fieldset, figcaption, figure, footer, ' +
|
|
14
|
+
'form, h1, h2, h3, h4, h5, h6, header, hr, li, main, nav, ol, p, pre, section, table, tr, ul';
|
|
15
|
+
function normalizeWhitespace(text) {
|
|
16
|
+
return text
|
|
17
|
+
.replace(/\r\n?/g, '\n')
|
|
18
|
+
.replace(/[^\S\n]+/g, ' ')
|
|
19
|
+
.replace(/ *\n */g, '\n')
|
|
20
|
+
.replace(/\n{3,}/g, '\n\n')
|
|
21
|
+
.trim();
|
|
22
|
+
}
|
|
23
|
+
/**
|
|
24
|
+
* Extracts readable text from HTML: scripts, styles and other non-content elements are
|
|
25
|
+
* dropped, block elements become line breaks and whitespace is normalized.
|
|
26
|
+
*/
|
|
12
27
|
export class HTMLLoader {
|
|
13
28
|
supportedTypes = ['html', 'htm'];
|
|
14
29
|
selector;
|
|
@@ -23,8 +38,17 @@ export class HTMLLoader {
|
|
|
23
38
|
async parseHTML(html, source, sourceType) {
|
|
24
39
|
const cheerio = await loadCheerio();
|
|
25
40
|
const $ = cheerio.load(html);
|
|
26
|
-
const
|
|
27
|
-
|
|
41
|
+
const title = $('title').first().text().trim() || undefined;
|
|
42
|
+
$(NON_CONTENT_SELECTOR).remove();
|
|
43
|
+
$('br').replaceWith('\n');
|
|
44
|
+
$(BLOCK_SELECTOR).each((_, element) => {
|
|
45
|
+
$(element).before('\n').after('\n');
|
|
46
|
+
});
|
|
47
|
+
$('td, th').after(' ');
|
|
48
|
+
const content = normalizeWhitespace($(this.selector)
|
|
49
|
+
.map((_, element) => $(element).text())
|
|
50
|
+
.get()
|
|
51
|
+
.join('\n\n'));
|
|
28
52
|
const metadata = {};
|
|
29
53
|
if (title)
|
|
30
54
|
metadata.title = title;
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"html-loader.js","sourceRoot":"","sources":["../../src/loaders/html-loader.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,QAAQ,EAAE,MAAM,kBAAkB,CAAC;AAC5C,OAAO,EAAE,OAAO,EAAE,MAAM,WAAW,CAAC;AACpC,OAAO,EAAE,MAAM,EAAE,MAAM,QAAQ,CAAC;AAGhC,KAAK,UAAU,WAAW;IACxB,IAAI,CAAC;QACH,OAAO,MAAM,MAAM,CAAC,SAAS,CAAC,CAAC;IACjC,CAAC;IAAC,MAAM,CAAC;QACP,MAAM,IAAI,KAAK,CAAC,kEAAkE,CAAC,CAAC;IACtF,CAAC;AACH,CAAC;AAMD,MAAM,OAAO,UAAU;IACZ,cAAc,GAAG,CAAC,MAAM,EAAE,KAAK,CAAC,CAAC;IACjC,QAAQ,CAAS;IAE1B,YAAY,OAA2B;QACrC,IAAI,CAAC,QAAQ,GAAG,OAAO,EAAE,QAAQ,IAAI,MAAM,CAAC;IAC9C,CAAC;IAED,KAAK,CAAC,IAAI,CAAC,MAAc;QACvB,MAAM,QAAQ,GAAG,OAAO,CAAC,MAAM,CAAC,CAAC;QACjC,MAAM,IAAI,GAAG,MAAM,QAAQ,CAAC,QAAQ,EAAE,OAAO,CAAC,CAAC;QAC/C,OAAO,CAAC,MAAM,IAAI,CAAC,SAAS,CAAC,IAAI,EAAE,QAAQ,EAAE,MAAM,CAAC,CAAC,CAAC;IACxD,CAAC;IAED,KAAK,CAAC,SAAS,CAAC,IAAY,EAAE,MAAc,EAAE,UAA0B;QACtE,MAAM,OAAO,GAAG,MAAM,WAAW,EAAE,CAAC;QACpC,MAAM,CAAC,GAAG,OAAO,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;QAE7B,MAAM,
|
|
1
|
+
{"version":3,"file":"html-loader.js","sourceRoot":"","sources":["../../src/loaders/html-loader.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,QAAQ,EAAE,MAAM,kBAAkB,CAAC;AAC5C,OAAO,EAAE,OAAO,EAAE,MAAM,WAAW,CAAC;AACpC,OAAO,EAAE,MAAM,EAAE,MAAM,QAAQ,CAAC;AAGhC,KAAK,UAAU,WAAW;IACxB,IAAI,CAAC;QACH,OAAO,MAAM,MAAM,CAAC,SAAS,CAAC,CAAC;IACjC,CAAC;IAAC,MAAM,CAAC;QACP,MAAM,IAAI,KAAK,CAAC,kEAAkE,CAAC,CAAC;IACtF,CAAC;AACH,CAAC;AAMD,MAAM,oBAAoB,GAAG,8DAA8D,CAAC;AAC5F,MAAM,cAAc,GAClB,8FAA8F;IAC9F,6FAA6F,CAAC;AAEhG,SAAS,mBAAmB,CAAC,IAAY;IACvC,OAAO,IAAI;SACR,OAAO,CAAC,QAAQ,EAAE,IAAI,CAAC;SACvB,OAAO,CAAC,WAAW,EAAE,GAAG,CAAC;SACzB,OAAO,CAAC,SAAS,EAAE,IAAI,CAAC;SACxB,OAAO,CAAC,SAAS,EAAE,MAAM,CAAC;SAC1B,IAAI,EAAE,CAAC;AACZ,CAAC;AAED;;;GAGG;AACH,MAAM,OAAO,UAAU;IACZ,cAAc,GAAG,CAAC,MAAM,EAAE,KAAK,CAAC,CAAC;IACjC,QAAQ,CAAS;IAE1B,YAAY,OAA2B;QACrC,IAAI,CAAC,QAAQ,GAAG,OAAO,EAAE,QAAQ,IAAI,MAAM,CAAC;IAC9C,CAAC;IAED,KAAK,CAAC,IAAI,CAAC,MAAc;QACvB,MAAM,QAAQ,GAAG,OAAO,CAAC,MAAM,CAAC,CAAC;QACjC,MAAM,IAAI,GAAG,MAAM,QAAQ,CAAC,QAAQ,EAAE,OAAO,CAAC,CAAC;QAC/C,OAAO,CAAC,MAAM,IAAI,CAAC,SAAS,CAAC,IAAI,EAAE,QAAQ,EAAE,MAAM,CAAC,CAAC,CAAC;IACxD,CAAC;IAED,KAAK,CAAC,SAAS,CAAC,IAAY,EAAE,MAAc,EAAE,UAA0B;QACtE,MAAM,OAAO,GAAG,MAAM,WAAW,EAAE,CAAC;QACpC,MAAM,CAAC,GAAG,OAAO,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;QAE7B,MAAM,KAAK,GAAG,CAAC,CAAC,OAAO,CAAC,CAAC,KAAK,EAAE,CAAC,IAAI,EAAE,CAAC,IAAI,EAAE,IAAI,SAAS,CAAC;QAE5D,CAAC,CAAC,oBAAoB,CAAC,CAAC,MAAM,EAAE,CAAC;QACjC,CAAC,CAAC,IAAI,CAAC,CAAC,WAAW,CAAC,IAAI,CAAC,CAAC;QAC1B,CAAC,CAAC,cAAc,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,OAAO,EAAE,EAAE;YACpC,CAAC,CAAC,OAAO,CAAC,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;QACtC,CAAC,CAAC,CAAC;QACH,CAAC,CAAC,QAAQ,CAAC,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC;QAEvB,MAAM,OAAO,GAAG,mBAAmB,CACjC,CAAC,CAAC,IAAI,CAAC,QAAQ,CAAC;aACb,GAAG,CAAC,CAAC,CAAC,EAAE,OAAO,EAAE,EAAE,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,IAAI,EAAE,CAAC;aACtC,GAAG,EAAE;aACL,IAAI,CAAC,MAAM,CAAC,CAChB,CAAC;QAEF,MAAM,QAAQ,GAA4B,EAAE,CAAC;QAC7C,IAAI,KAAK;YAAE,QAAQ,CAAC,KAAK,GAAG,KAAK,CAAC;QAElC,OAAO;YACL,EAAE,EAAE,MAAM,EAAE;YACZ,OAAO;YACP,MAAM;YACN,UAAU;YACV,GAAG,CAAC,MAAM,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC,MAAM,GAAG,CAAC,IAAI,EAAE,QAAQ,EAAE,CAAC;SACtD,CAAC;IACJ,CAAC;CACF"}
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../../src/loaders/index.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,UAAU,EAAE,MAAM,
|
|
1
|
+
{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../../src/loaders/index.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,UAAU,EAAE,MAAM,kBAAkB,CAAC;AAC9C,OAAO,EAAE,cAAc,EAAE,MAAM,sBAAsB,CAAC;AACtD,YAAY,EAAE,qBAAqB,EAAE,MAAM,sBAAsB,CAAC;AAClE,OAAO,EAAE,UAAU,EAAE,MAAM,kBAAkB,CAAC;AAC9C,YAAY,EAAE,iBAAiB,EAAE,MAAM,kBAAkB,CAAC;AAC1D,OAAO,EAAE,SAAS,EAAE,MAAM,iBAAiB,CAAC;AAC5C,YAAY,EAAE,gBAAgB,EAAE,MAAM,iBAAiB,CAAC;AACxD,OAAO,EAAE,UAAU,EAAE,MAAM,kBAAkB,CAAC;AAC9C,YAAY,EAAE,iBAAiB,EAAE,MAAM,kBAAkB,CAAC;AAC1D,OAAO,EAAE,SAAS,EAAE,MAAM,iBAAiB,CAAC;AAC5C,YAAY,EAAE,gBAAgB,EAAE,MAAM,iBAAiB,CAAC;AACxD,OAAO,EAAE,SAAS,EAAE,MAAM,iBAAiB,CAAC;AAC5C,YAAY,EAAE,gBAAgB,EAAE,MAAM,iBAAiB,CAAC"}
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"index.js","sourceRoot":"","sources":["../../src/loaders/index.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,UAAU,EAAE,MAAM,
|
|
1
|
+
{"version":3,"file":"index.js","sourceRoot":"","sources":["../../src/loaders/index.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,UAAU,EAAE,MAAM,kBAAkB,CAAC;AAC9C,OAAO,EAAE,cAAc,EAAE,MAAM,sBAAsB,CAAC;AAEtD,OAAO,EAAE,UAAU,EAAE,MAAM,kBAAkB,CAAC;AAE9C,OAAO,EAAE,SAAS,EAAE,MAAM,iBAAiB,CAAC;AAE5C,OAAO,EAAE,UAAU,EAAE,MAAM,kBAAkB,CAAC;AAE9C,OAAO,EAAE,SAAS,EAAE,MAAM,iBAAiB,CAAC;AAE5C,OAAO,EAAE,SAAS,EAAE,MAAM,iBAAiB,CAAC"}
|
|
@@ -19,7 +19,7 @@ export class JSONLoader {
|
|
|
19
19
|
}
|
|
20
20
|
catch (err) {
|
|
21
21
|
const message = err instanceof Error ? err.message : String(err);
|
|
22
|
-
throw new Error(`JSONLoader: failed to parse "${filePath}": ${message}
|
|
22
|
+
throw new Error(`JSONLoader: failed to parse "${filePath}": ${message}`, { cause: err });
|
|
23
23
|
}
|
|
24
24
|
const items = Array.isArray(data) ? data : [data];
|
|
25
25
|
return items.map((item) => this.itemToDocument(item, filePath));
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"json-loader.js","sourceRoot":"","sources":["../../src/loaders/json-loader.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,QAAQ,EAAE,MAAM,kBAAkB,CAAC;AAC5C,OAAO,EAAE,OAAO,EAAE,MAAM,WAAW,CAAC;AACpC,OAAO,EAAE,MAAM,EAAE,MAAM,QAAQ,CAAC;AAGhC,MAAM,cAAc,GAAG,CAAC,SAAS,EAAE,MAAM,EAAE,MAAM,CAAC,CAAC;AAOnD,MAAM,OAAO,UAAU;IACZ,cAAc,GAAG,CAAC,MAAM,CAAC,CAAC;IAClB,YAAY,CAAU;IACtB,cAAc,CAAY;IAE3C,YAAY,OAA2B;QACrC,IAAI,CAAC,YAAY,GAAG,OAAO,EAAE,YAAY,CAAC;QAC1C,IAAI,CAAC,cAAc,GAAG,OAAO,EAAE,cAAc,CAAC;IAChD,CAAC;IAED,KAAK,CAAC,IAAI,CAAC,MAAc;QACvB,MAAM,QAAQ,GAAG,OAAO,CAAC,MAAM,CAAC,CAAC;QACjC,MAAM,GAAG,GAAG,MAAM,QAAQ,CAAC,QAAQ,EAAE,OAAO,CAAC,CAAC;QAE9C,IAAI,IAAa,CAAC;QAClB,IAAI,CAAC;YACH,IAAI,GAAG,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC;QACzB,CAAC;QAAC,OAAO,GAAG,EAAE,CAAC;YACb,MAAM,OAAO,GAAG,GAAG,YAAY,KAAK,CAAC,CAAC,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC;YACjE,MAAM,IAAI,KAAK,CAAC,gCAAgC,QAAQ,MAAM,OAAO,EAAE,CAAC,CAAC;
|
|
1
|
+
{"version":3,"file":"json-loader.js","sourceRoot":"","sources":["../../src/loaders/json-loader.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,QAAQ,EAAE,MAAM,kBAAkB,CAAC;AAC5C,OAAO,EAAE,OAAO,EAAE,MAAM,WAAW,CAAC;AACpC,OAAO,EAAE,MAAM,EAAE,MAAM,QAAQ,CAAC;AAGhC,MAAM,cAAc,GAAG,CAAC,SAAS,EAAE,MAAM,EAAE,MAAM,CAAC,CAAC;AAOnD,MAAM,OAAO,UAAU;IACZ,cAAc,GAAG,CAAC,MAAM,CAAC,CAAC;IAClB,YAAY,CAAU;IACtB,cAAc,CAAY;IAE3C,YAAY,OAA2B;QACrC,IAAI,CAAC,YAAY,GAAG,OAAO,EAAE,YAAY,CAAC;QAC1C,IAAI,CAAC,cAAc,GAAG,OAAO,EAAE,cAAc,CAAC;IAChD,CAAC;IAED,KAAK,CAAC,IAAI,CAAC,MAAc;QACvB,MAAM,QAAQ,GAAG,OAAO,CAAC,MAAM,CAAC,CAAC;QACjC,MAAM,GAAG,GAAG,MAAM,QAAQ,CAAC,QAAQ,EAAE,OAAO,CAAC,CAAC;QAE9C,IAAI,IAAa,CAAC;QAClB,IAAI,CAAC;YACH,IAAI,GAAG,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC;QACzB,CAAC;QAAC,OAAO,GAAG,EAAE,CAAC;YACb,MAAM,OAAO,GAAG,GAAG,YAAY,KAAK,CAAC,CAAC,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC;YACjE,MAAM,IAAI,KAAK,CAAC,gCAAgC,QAAQ,MAAM,OAAO,EAAE,EAAE,EAAE,KAAK,EAAE,GAAG,EAAE,CAAC,CAAC;QAC3F,CAAC;QAED,MAAM,KAAK,GAAG,KAAK,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC;QAClD,OAAO,KAAK,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,cAAc,CAAC,IAAI,EAAE,QAAQ,CAAC,CAAC,CAAC;IAClE,CAAC;IAEO,cAAc,CAAC,IAAa,EAAE,MAAc;QAClD,IAAI,OAAO,IAAI,KAAK,QAAQ,IAAI,IAAI,KAAK,IAAI,IAAI,KAAK,CAAC,OAAO,CAAC,IAAI,CAAC,EAAE,CAAC;YACrE,OAAO;gBACL,EAAE,EAAE,MAAM,EAAE;gBACZ,OAAO,EAAE,MAAM,CAAC,IAAI,CAAC;gBACrB,MAAM;gBACN,UAAU,EAAE,MAAM;aACnB,CAAC;QACJ,CAAC;QACD,MAAM,GAAG,GAAG,IAA+B,CAAC;QAC5C,MAAM,OAAO,GAAG,IAAI,CAAC,cAAc,CAAC,GAAG,CAAC,CAAC;QACzC,MAAM,QAAQ,GAAG,IAAI,CAAC,eAAe,CAAC,GAAG,CAAC,CAAC;QAE3C,OAAO;YACL,EAAE,EAAE,MAAM,EAAE;YACZ,OAAO;YACP,MAAM;YACN,UAAU,EAAE,MAAM;YAClB,GAAG,CAAC,QAAQ,IAAI,MAAM,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC,MAAM,GAAG,CAAC,IAAI,EAAE,QAAQ,EAAE,CAAC;SAClE,CAAC;IACJ,CAAC;IAEO,cAAc,CAAC,GAA4B;QACjD,IAAI,IAAI,CAAC,YAAY,IAAI,IAAI,CAAC,YAAY,IAAI,GAAG,EAAE,CAAC;YAClD,OAAO,IAAI,CAAC,aAAa,CAAC,GAAG,CAAC,IAAI,CAAC,YAAY,CAAC,CAAC,CAAC;QACpD,CAAC;QAED,KAAK,MAAM,KAAK,IAAI,cAAc,EAAE,CAAC;YACnC,IAAI,KAAK,IAAI,GAAG,EAAE,CAAC;gBACjB,OAAO,IAAI,CAAC,aAAa,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC,CAAC;YACxC,CAAC;QACH,CAAC;QAED,OAAO,IAAI,CAAC,SAAS,CAAC,GAAG,EAAE,IAAI,EAAE,CAAC,CAAC,CAAC;IACtC,CAAC;IAEO,aAAa,CAAC,KAAc;QAClC,IAAI,OAAO,KAAK,KAAK,QAAQ;YAAE,OAAO,KAAK,CAAC;QAC5C,OAAO,IAAI,CAAC,SAAS,CAAC,KAAK,EAAE,IAAI,EAAE,CAAC,CAAC,CAAC;IACxC,CAAC;IAEO,eAAe,CAAC,GAA4B;QAClD,IAAI,CAAC,IAAI,CAAC,cAAc;YAAE,OAAO,SAAS,CAAC;QAE3C,MAAM,QAAQ,GAA4B,EAAE,CAAC;QAC7C,KAAK,MAAM,KAAK,IAAI,IAAI,CAAC,cAAc,EAAE,CAAC;YACxC,IAAI,KAAK,IAAI,GAAG,EAAE,CAAC;gBACjB,QAAQ,CAAC,KAAK,CAAC,GAAG,GAAG,CAAC,KAAK,CAAC,CAAC;YAC/B,CAAC;QACH,CAAC;QACD,OAAO,QAAQ,CAAC;IAClB,CAAC;CACF"}
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"markdown-loader.d.ts","sourceRoot":"","sources":["../../src/loaders/markdown-loader.ts"],"names":[],"mappings":"AAGA,OAAO,KAAK,EAAE,cAAc,EAAE,WAAW,EAAE,MAAM,qBAAqB,CAAC;
|
|
1
|
+
{"version":3,"file":"markdown-loader.d.ts","sourceRoot":"","sources":["../../src/loaders/markdown-loader.ts"],"names":[],"mappings":"AAGA,OAAO,KAAK,EAAE,cAAc,EAAE,WAAW,EAAE,MAAM,qBAAqB,CAAC;AAWvE,MAAM,WAAW,qBAAqB;IACpC,gBAAgB,CAAC,EAAE,OAAO,CAAC;CAC5B;AAED,qBAAa,cAAe,YAAW,cAAc;IACnD,QAAQ,CAAC,cAAc,WAAiB;IACxC,OAAO,CAAC,QAAQ,CAAC,gBAAgB,CAAU;gBAE/B,OAAO,CAAC,EAAE,qBAAqB;IAIrC,IAAI,CAAC,MAAM,EAAE,MAAM,GAAG,OAAO,CAAC,WAAW,EAAE,CAAC;YAWpC,QAAQ;IAsBtB,OAAO,CAAC,gBAAgB;YAWV,aAAa;CAiB5B"}
|
|
@@ -3,6 +3,10 @@ import { join, resolve } from 'node:path';
|
|
|
3
3
|
import { nanoid } from 'nanoid';
|
|
4
4
|
const MD_EXTENSIONS = new Set(['md', 'mdx']);
|
|
5
5
|
const FRONTMATTER_RE = /^---\r?\n([\s\S]*?)\r?\n---(?:\r?\n|$)/;
|
|
6
|
+
function unquote(value) {
|
|
7
|
+
const quoted = /^(['"])(.*)\1$/.exec(value);
|
|
8
|
+
return quoted ? quoted[2] : value;
|
|
9
|
+
}
|
|
6
10
|
export class MarkdownLoader {
|
|
7
11
|
supportedTypes = ['md', 'mdx'];
|
|
8
12
|
stripFrontmatter;
|
|
@@ -38,16 +42,16 @@ export class MarkdownLoader {
|
|
|
38
42
|
}
|
|
39
43
|
parseFrontmatter(raw) {
|
|
40
44
|
const result = {};
|
|
41
|
-
for (const line of raw.split(
|
|
45
|
+
for (const line of raw.split(/\r?\n/)) {
|
|
42
46
|
const match = /^(\w[\w\s-]*?):\s*(.+)$/.exec(line);
|
|
43
47
|
if (match) {
|
|
44
|
-
result[match[1].trim()] = match[2].trim();
|
|
48
|
+
result[match[1].trim()] = unquote(match[2].trim());
|
|
45
49
|
}
|
|
46
50
|
}
|
|
47
51
|
return result;
|
|
48
52
|
}
|
|
49
53
|
async loadDirectory(dirPath) {
|
|
50
|
-
const entries = await readdir(dirPath);
|
|
54
|
+
const entries = (await readdir(dirPath)).sort();
|
|
51
55
|
const docs = [];
|
|
52
56
|
for (const entry of entries) {
|
|
53
57
|
const ext = entry.split('.').pop()?.toLowerCase() ?? '';
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"markdown-loader.js","sourceRoot":"","sources":["../../src/loaders/markdown-loader.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,QAAQ,EAAE,OAAO,EAAE,IAAI,EAAE,MAAM,kBAAkB,CAAC;AAC3D,OAAO,EAAE,IAAI,EAAE,OAAO,EAAE,MAAM,WAAW,CAAC;AAC1C,OAAO,EAAE,MAAM,EAAE,MAAM,QAAQ,CAAC;AAGhC,MAAM,aAAa,GAAG,IAAI,GAAG,CAAC,CAAC,IAAI,EAAE,KAAK,CAAC,CAAC,CAAC;AAE7C,MAAM,cAAc,GAAG,wCAAwC,CAAC;
|
|
1
|
+
{"version":3,"file":"markdown-loader.js","sourceRoot":"","sources":["../../src/loaders/markdown-loader.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,QAAQ,EAAE,OAAO,EAAE,IAAI,EAAE,MAAM,kBAAkB,CAAC;AAC3D,OAAO,EAAE,IAAI,EAAE,OAAO,EAAE,MAAM,WAAW,CAAC;AAC1C,OAAO,EAAE,MAAM,EAAE,MAAM,QAAQ,CAAC;AAGhC,MAAM,aAAa,GAAG,IAAI,GAAG,CAAC,CAAC,IAAI,EAAE,KAAK,CAAC,CAAC,CAAC;AAE7C,MAAM,cAAc,GAAG,wCAAwC,CAAC;AAEhE,SAAS,OAAO,CAAC,KAAa;IAC5B,MAAM,MAAM,GAAG,gBAAgB,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;IAC5C,OAAO,MAAM,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAE,CAAC,CAAC,CAAC,KAAK,CAAC;AACrC,CAAC;AAMD,MAAM,OAAO,cAAc;IAChB,cAAc,GAAG,CAAC,IAAI,EAAE,KAAK,CAAC,CAAC;IACvB,gBAAgB,CAAU;IAE3C,YAAY,OAA+B;QACzC,IAAI,CAAC,gBAAgB,GAAG,OAAO,EAAE,gBAAgB,IAAI,KAAK,CAAC;IAC7D,CAAC;IAED,KAAK,CAAC,IAAI,CAAC,MAAc;QACvB,MAAM,QAAQ,GAAG,OAAO,CAAC,MAAM,CAAC,CAAC;QACjC,MAAM,IAAI,GAAG,MAAM,IAAI,CAAC,QAAQ,CAAC,CAAC;QAElC,IAAI,IAAI,CAAC,WAAW,EAAE,EAAE,CAAC;YACvB,OAAO,IAAI,CAAC,aAAa,CAAC,QAAQ,CAAC,CAAC;QACtC,CAAC;QAED,OAAO,CAAC,MAAM,IAAI,CAAC,QAAQ,CAAC,QAAQ,CAAC,CAAC,CAAC;IACzC,CAAC;IAEO,KAAK,CAAC,QAAQ,CAAC,QAAgB;QACrC,MAAM,GAAG,GAAG,MAAM,QAAQ,CAAC,QAAQ,EAAE,OAAO,CAAC,CAAC;QAC9C,IAAI,OAAO,GAAG,GAAG,CAAC;QAClB,IAAI,QAA6C,CAAC;QAElD,IAAI,IAAI,CAAC,gBAAgB,EAAE,CAAC;YAC1B,MAAM,KAAK,GAAG,cAAc,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;YACvC,IAAI,KAAK,EAAE,CAAC;gBACV,QAAQ,GAAG,IAAI,CAAC,gBAAgB,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC;gBAC3C,OAAO,GAAG,GAAG,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC;YACvC,CAAC;QACH,CAAC;QAED,OAAO;YACL,EAAE,EAAE,MAAM,EAAE;YACZ,OAAO;YACP,MAAM,EAAE,QAAQ;YAChB,UAAU,EAAE,UAAU;YACtB,GAAG,CAAC,QAAQ,IAAI,EAAE,QAAQ,EAAE,CAAC;SAC9B,CAAC;IACJ,CAAC;IAEO,gBAAgB,CAAC,GAAW;QAClC,MAAM,MAAM,GAA4B,EAAE,CAAC;QAC3C,KAAK,MAAM,IAAI,IAAI,GAAG,CAAC,KAAK,CAAC,OAAO,CAAC,EAAE,CAAC;YACtC,MAAM,KAAK,GAAG,yBAAyB,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;YACnD,IAAI,KAAK,EAAE,CAAC;gBACV,MAAM,CAAC,KAAK,CAAC,CAAC,CAAE,CAAC,IAAI,EAAE,CAAC,GAAG,OAAO,CAAC,KAAK,CAAC,CAAC,CAAE,CAAC,IAAI,EAAE,CAAC,CAAC;YACvD,CAAC;QACH,CAAC;QACD,OAAO,MAAM,CAAC;IAChB,CAAC;IAEO,KAAK,CAAC,aAAa,CAAC,OAAe;QACzC,MAAM,OAAO,GAAG,CAAC,MAAM,OAAO,CAAC,OAAO,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC;QAChD,MAAM,IAAI,GAAkB,EAAE,CAAC;QAE/B,KAAK,MAAM,KAAK,IAAI,OAAO,EAAE,CAAC;YAC5B,MAAM,GAAG,GAAG,KAAK,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,GAAG,EAAE,EAAE,WAAW,EAAE,IAAI,EAAE,CAAC;YACxD,IAAI,CAAC,aAAa,CAAC,GAAG,CAAC,GAAG,CAAC;gBAAE,SAAS;YAEtC,MAAM,QAAQ,GAAG,IAAI,CAAC,OAAO,EAAE,KAAK,CAAC,CAAC;YACtC,MAAM,IAAI,GAAG,MAAM,IAAI,CAAC,QAAQ,CAAC,CAAC;YAClC,IAAI,IAAI,CAAC,MAAM,EAAE,EAAE,CAAC;gBAClB,IAAI,CAAC,IAAI,CAAC,MAAM,IAAI,CAAC,QAAQ,CAAC,QAAQ,CAAC,CAAC,CAAC;YAC3C,CAAC;QACH,CAAC;QAED,OAAO,IAAI,CAAC;IACd,CAAC;CACF"}
|
|
@@ -7,9 +7,9 @@ export declare class PDFLoader implements DocumentLoader {
|
|
|
7
7
|
private readonly splitPages;
|
|
8
8
|
constructor(options?: PDFLoaderOptions);
|
|
9
9
|
load(source: string): Promise<RAGDocument[]>;
|
|
10
|
-
private
|
|
11
|
-
private
|
|
12
|
-
private
|
|
10
|
+
private parse;
|
|
11
|
+
private buildSingleDoc;
|
|
12
|
+
private buildPageDocs;
|
|
13
13
|
private buildPageDoc;
|
|
14
14
|
}
|
|
15
15
|
//# sourceMappingURL=pdf-loader.d.ts.map
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"pdf-loader.d.ts","sourceRoot":"","sources":["../../src/loaders/pdf-loader.ts"],"names":[],"mappings":"
|
|
1
|
+
{"version":3,"file":"pdf-loader.d.ts","sourceRoot":"","sources":["../../src/loaders/pdf-loader.ts"],"names":[],"mappings":"AAIA,OAAO,KAAK,EAAE,cAAc,EAAE,WAAW,EAAE,MAAM,qBAAqB,CAAC;AA0BvE,MAAM,WAAW,gBAAgB;IAC/B,UAAU,CAAC,EAAE,OAAO,CAAC;CACtB;AAED,qBAAa,SAAU,YAAW,cAAc;IAC9C,QAAQ,CAAC,cAAc,WAAW;IAClC,OAAO,CAAC,QAAQ,CAAC,UAAU,CAAU;gBAEzB,OAAO,CAAC,EAAE,gBAAgB;IAIhC,IAAI,CAAC,MAAM,EAAE,MAAM,GAAG,OAAO,CAAC,WAAW,EAAE,CAAC;YAapC,KAAK;IAwBnB,OAAO,CAAC,cAAc;IAatB,OAAO,CAAC,aAAa;IAYrB,OAAO,CAAC,YAAY;CAkBrB"}
|
|
@@ -4,12 +4,18 @@ import { nanoid } from 'nanoid';
|
|
|
4
4
|
async function loadPdfParse() {
|
|
5
5
|
try {
|
|
6
6
|
const mod = await import('pdf-parse');
|
|
7
|
-
return mod.
|
|
7
|
+
return mod.PDFParse;
|
|
8
8
|
}
|
|
9
9
|
catch {
|
|
10
10
|
throw new Error('pdf-parse is required for PDFLoader. Install it: pnpm add pdf-parse');
|
|
11
11
|
}
|
|
12
12
|
}
|
|
13
|
+
function extractTitle(info) {
|
|
14
|
+
if (typeof info !== 'object' || info === null || !('Title' in info))
|
|
15
|
+
return undefined;
|
|
16
|
+
const title = info.Title;
|
|
17
|
+
return typeof title === 'string' && title.length > 0 ? title : undefined;
|
|
18
|
+
}
|
|
13
19
|
export class PDFLoader {
|
|
14
20
|
supportedTypes = ['pdf'];
|
|
15
21
|
splitPages;
|
|
@@ -17,57 +23,54 @@ export class PDFLoader {
|
|
|
17
23
|
this.splitPages = options?.splitPages ?? false;
|
|
18
24
|
}
|
|
19
25
|
async load(source) {
|
|
20
|
-
const
|
|
26
|
+
const PDFParse = await loadPdfParse();
|
|
21
27
|
const filePath = resolve(source);
|
|
22
28
|
const buffer = await readFile(filePath);
|
|
29
|
+
const parsed = await this.parse(PDFParse, buffer, filePath);
|
|
23
30
|
if (this.splitPages) {
|
|
24
|
-
return this.
|
|
31
|
+
return this.buildPageDocs(parsed, filePath);
|
|
25
32
|
}
|
|
26
|
-
return this.
|
|
33
|
+
return [this.buildSingleDoc(parsed, filePath)];
|
|
27
34
|
}
|
|
28
|
-
async
|
|
29
|
-
|
|
30
|
-
const title = result.info?.Title;
|
|
31
|
-
const metadata = { pages: result.numpages };
|
|
32
|
-
if (title)
|
|
33
|
-
metadata.title = title;
|
|
34
|
-
return [
|
|
35
|
-
{
|
|
36
|
-
id: nanoid(),
|
|
37
|
-
content: result.text,
|
|
38
|
-
source,
|
|
39
|
-
sourceType: 'pdf',
|
|
40
|
-
metadata,
|
|
41
|
-
},
|
|
42
|
-
];
|
|
43
|
-
}
|
|
44
|
-
async loadSplitPages(pdfParse, buffer, source) {
|
|
45
|
-
const pageTexts = [];
|
|
46
|
-
const result = await this.parseWithErrorContext(pdfParse, buffer, source, {
|
|
47
|
-
pagerender: async (pageData) => {
|
|
48
|
-
if (typeof pageData.getTextContent !== 'function')
|
|
49
|
-
return '';
|
|
50
|
-
const textContent = await pageData.getTextContent();
|
|
51
|
-
const text = textContent.items.map((item) => item.str).join('');
|
|
52
|
-
pageTexts.push(text);
|
|
53
|
-
return text;
|
|
54
|
-
},
|
|
55
|
-
});
|
|
56
|
-
const title = result.info?.Title;
|
|
57
|
-
const pages = pageTexts.filter((p) => p.trim().length > 0);
|
|
58
|
-
if (pages.length === 0) {
|
|
59
|
-
return [this.buildPageDoc(result.text, source, 1, result.numpages, title)];
|
|
60
|
-
}
|
|
61
|
-
return pages.map((text, i) => this.buildPageDoc(text.trim(), source, i + 1, result.numpages, title));
|
|
62
|
-
}
|
|
63
|
-
async parseWithErrorContext(pdfParse, buffer, source, options) {
|
|
35
|
+
async parse(PDFParse, buffer, source) {
|
|
36
|
+
let parser;
|
|
64
37
|
try {
|
|
65
|
-
|
|
38
|
+
parser = new PDFParse({ data: new Uint8Array(buffer) });
|
|
39
|
+
const info = await parser.getInfo();
|
|
40
|
+
const result = await parser.getText({ pageJoiner: '' });
|
|
41
|
+
return {
|
|
42
|
+
text: result.text,
|
|
43
|
+
totalPages: result.total,
|
|
44
|
+
pages: result.pages.map(({ num, text }) => ({ num, text })),
|
|
45
|
+
title: extractTitle(info.info),
|
|
46
|
+
};
|
|
66
47
|
}
|
|
67
48
|
catch (err) {
|
|
68
49
|
const message = err instanceof Error ? err.message : String(err);
|
|
69
|
-
throw new Error(`PDFLoader: failed to parse "${source}": ${message}
|
|
50
|
+
throw new Error(`PDFLoader: failed to parse "${source}": ${message}`, { cause: err });
|
|
51
|
+
}
|
|
52
|
+
finally {
|
|
53
|
+
await parser?.destroy();
|
|
54
|
+
}
|
|
55
|
+
}
|
|
56
|
+
buildSingleDoc(parsed, source) {
|
|
57
|
+
const metadata = { pages: parsed.totalPages };
|
|
58
|
+
if (parsed.title)
|
|
59
|
+
metadata.title = parsed.title;
|
|
60
|
+
return {
|
|
61
|
+
id: nanoid(),
|
|
62
|
+
content: parsed.text,
|
|
63
|
+
source,
|
|
64
|
+
sourceType: 'pdf',
|
|
65
|
+
metadata,
|
|
66
|
+
};
|
|
67
|
+
}
|
|
68
|
+
buildPageDocs(parsed, source) {
|
|
69
|
+
const pages = parsed.pages.filter((page) => page.text.trim().length > 0);
|
|
70
|
+
if (pages.length === 0) {
|
|
71
|
+
return [this.buildPageDoc(parsed.text, source, 1, parsed.totalPages, parsed.title)];
|
|
70
72
|
}
|
|
73
|
+
return pages.map((page) => this.buildPageDoc(page.text.trim(), source, page.num, parsed.totalPages, parsed.title));
|
|
71
74
|
}
|
|
72
75
|
buildPageDoc(content, source, pageNumber, totalPages, title) {
|
|
73
76
|
const metadata = { pageNumber, totalPages };
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"pdf-loader.js","sourceRoot":"","sources":["../../src/loaders/pdf-loader.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,QAAQ,EAAE,MAAM,kBAAkB,CAAC;AAC5C,OAAO,EAAE,OAAO,EAAE,MAAM,WAAW,CAAC;AACpC,OAAO,EAAE,MAAM,EAAE,MAAM,QAAQ,CAAC;
|
|
1
|
+
{"version":3,"file":"pdf-loader.js","sourceRoot":"","sources":["../../src/loaders/pdf-loader.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,QAAQ,EAAE,MAAM,kBAAkB,CAAC;AAC5C,OAAO,EAAE,OAAO,EAAE,MAAM,WAAW,CAAC;AACpC,OAAO,EAAE,MAAM,EAAE,MAAM,QAAQ,CAAC;AAahC,KAAK,UAAU,YAAY;IACzB,IAAI,CAAC;QACH,MAAM,GAAG,GAAG,MAAM,MAAM,CAAC,WAAW,CAAC,CAAC;QACtC,OAAO,GAAG,CAAC,QAAQ,CAAC;IACtB,CAAC;IAAC,MAAM,CAAC;QACP,MAAM,IAAI,KAAK,CAAC,qEAAqE,CAAC,CAAC;IACzF,CAAC;AACH,CAAC;AAED,SAAS,YAAY,CAAC,IAAa;IACjC,IAAI,OAAO,IAAI,KAAK,QAAQ,IAAI,IAAI,KAAK,IAAI,IAAI,CAAC,CAAC,OAAO,IAAI,IAAI,CAAC;QAAE,OAAO,SAAS,CAAC;IACtF,MAAM,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC;IACzB,OAAO,OAAO,KAAK,KAAK,QAAQ,IAAI,KAAK,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,SAAS,CAAC;AAC3E,CAAC;AAMD,MAAM,OAAO,SAAS;IACX,cAAc,GAAG,CAAC,KAAK,CAAC,CAAC;IACjB,UAAU,CAAU;IAErC,YAAY,OAA0B;QACpC,IAAI,CAAC,UAAU,GAAG,OAAO,EAAE,UAAU,IAAI,KAAK,CAAC;IACjD,CAAC;IAED,KAAK,CAAC,IAAI,CAAC,MAAc;QACvB,MAAM,QAAQ,GAAG,MAAM,YAAY,EAAE,CAAC;QACtC,MAAM,QAAQ,GAAG,OAAO,CAAC,MAAM,CAAC,CAAC;QACjC,MAAM,MAAM,GAAG,MAAM,QAAQ,CAAC,QAAQ,CAAC,CAAC;QACxC,MAAM,MAAM,GAAG,MAAM,IAAI,CAAC,KAAK,CAAC,QAAQ,EAAE,MAAM,EAAE,QAAQ,CAAC,CAAC;QAE5D,IAAI,IAAI,CAAC,UAAU,EAAE,CAAC;YACpB,OAAO,IAAI,CAAC,aAAa,CAAC,MAAM,EAAE,QAAQ,CAAC,CAAC;QAC9C,CAAC;QAED,OAAO,CAAC,IAAI,CAAC,cAAc,CAAC,MAAM,EAAE,QAAQ,CAAC,CAAC,CAAC;IACjD,CAAC;IAEO,KAAK,CAAC,KAAK,CACjB,QAA6B,EAC7B,MAAc,EACd,MAAc;QAEd,IAAI,MAAiC,CAAC;QACtC,IAAI,CAAC;YACH,MAAM,GAAG,IAAI,QAAQ,CAAC,EAAE,IAAI,EAAE,IAAI,UAAU,CAAC,MAAM,CAAC,EAAE,CAAC,CAAC;YACxD,MAAM,IAAI,GAAG,MAAM,MAAM,CAAC,OAAO,EAAE,CAAC;YACpC,MAAM,MAAM,GAAG,MAAM,MAAM,CAAC,OAAO,CAAC,EAAE,UAAU,EAAE,EAAE,EAAE,CAAC,CAAC;YACxD,OAAO;gBACL,IAAI,EAAE,MAAM,CAAC,IAAI;gBACjB,UAAU,EAAE,MAAM,CAAC,KAAK;gBACxB,KAAK,EAAE,MAAM,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,EAAE,GAAG,EAAE,IAAI,EAAE,EAAE,EAAE,CAAC,CAAC,EAAE,GAAG,EAAE,IAAI,EAAE,CAAC,CAAC;gBAC3D,KAAK,EAAE,YAAY,CAAC,IAAI,CAAC,IAAI,CAAC;aAC/B,CAAC;QACJ,CAAC;QAAC,OAAO,GAAG,EAAE,CAAC;YACb,MAAM,OAAO,GAAG,GAAG,YAAY,KAAK,CAAC,CAAC,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC;YACjE,MAAM,IAAI,KAAK,CAAC,+BAA+B,MAAM,MAAM,OAAO,EAAE,EAAE,EAAE,KAAK,EAAE,GAAG,EAAE,CAAC,CAAC;QACxF,CAAC;gBAAS,CAAC;YACT,MAAM,MAAM,EAAE,OAAO,EAAE,CAAC;QAC1B,CAAC;IACH,CAAC;IAEO,cAAc,CAAC,MAAiB,EAAE,MAAc;QACtD,MAAM,QAAQ,GAA4B,EAAE,KAAK,EAAE,MAAM,CAAC,UAAU,EAAE,CAAC;QACvE,IAAI,MAAM,CAAC,KAAK;YAAE,QAAQ,CAAC,KAAK,GAAG,MAAM,CAAC,KAAK,CAAC;QAEhD,OAAO;YACL,EAAE,EAAE,MAAM,EAAE;YACZ,OAAO,EAAE,MAAM,CAAC,IAAI;YACpB,MAAM;YACN,UAAU,EAAE,KAAK;YACjB,QAAQ;SACT,CAAC;IACJ,CAAC;IAEO,aAAa,CAAC,MAAiB,EAAE,MAAc;QACrD,MAAM,KAAK,GAAG,MAAM,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI,EAAE,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;QAEzE,IAAI,KAAK,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;YACvB,OAAO,CAAC,IAAI,CAAC,YAAY,CAAC,MAAM,CAAC,IAAI,EAAE,MAAM,EAAE,CAAC,EAAE,MAAM,CAAC,UAAU,EAAE,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC;QACtF,CAAC;QAED,OAAO,KAAK,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CACxB,IAAI,CAAC,YAAY,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI,EAAE,EAAE,MAAM,EAAE,IAAI,CAAC,GAAG,EAAE,MAAM,CAAC,UAAU,EAAE,MAAM,CAAC,KAAK,CAAC,CACvF,CAAC;IACJ,CAAC;IAEO,YAAY,CAClB,OAAe,EACf,MAAc,EACd,UAAkB,EAClB,UAAkB,EAClB,KAAyB;QAEzB,MAAM,QAAQ,GAA4B,EAAE,UAAU,EAAE,UAAU,EAAE,CAAC;QACrE,IAAI,KAAK;YAAE,QAAQ,CAAC,KAAK,GAAG,KAAK,CAAC;QAElC,OAAO;YACL,EAAE,EAAE,MAAM,EAAE;YACZ,OAAO;YACP,MAAM;YACN,UAAU,EAAE,KAAK;YACjB,QAAQ;SACT,CAAC;IACJ,CAAC;CACF"}
|
|
@@ -22,7 +22,7 @@ export class TextLoader {
|
|
|
22
22
|
};
|
|
23
23
|
}
|
|
24
24
|
async loadDirectory(dirPath) {
|
|
25
|
-
const entries = await readdir(dirPath);
|
|
25
|
+
const entries = (await readdir(dirPath)).sort();
|
|
26
26
|
const docs = [];
|
|
27
27
|
for (const entry of entries) {
|
|
28
28
|
const ext = entry.split('.').pop()?.toLowerCase() ?? '';
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"text-loader.js","sourceRoot":"","sources":["../../src/loaders/text-loader.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,QAAQ,EAAE,OAAO,EAAE,IAAI,EAAE,MAAM,kBAAkB,CAAC;AAC3D,OAAO,EAAE,IAAI,EAAE,OAAO,EAAE,MAAM,WAAW,CAAC;AAC1C,OAAO,EAAE,MAAM,EAAE,MAAM,QAAQ,CAAC;AAGhC,MAAM,eAAe,GAAG,IAAI,GAAG,CAAC,CAAC,KAAK,EAAE,MAAM,CAAC,CAAC,CAAC;AAEjD,MAAM,OAAO,UAAU;IACZ,cAAc,GAAG,CAAC,KAAK,EAAE,MAAM,CAAC,CAAC;IAE1C,KAAK,CAAC,IAAI,CAAC,MAAc;QACvB,MAAM,QAAQ,GAAG,OAAO,CAAC,MAAM,CAAC,CAAC;QACjC,MAAM,IAAI,GAAG,MAAM,IAAI,CAAC,QAAQ,CAAC,CAAC;QAElC,IAAI,IAAI,CAAC,WAAW,EAAE,EAAE,CAAC;YACvB,OAAO,IAAI,CAAC,aAAa,CAAC,QAAQ,CAAC,CAAC;QACtC,CAAC;QAED,OAAO,CAAC,MAAM,IAAI,CAAC,QAAQ,CAAC,QAAQ,CAAC,CAAC,CAAC;IACzC,CAAC;IAEO,KAAK,CAAC,QAAQ,CAAC,QAAgB;QACrC,MAAM,OAAO,GAAG,MAAM,QAAQ,CAAC,QAAQ,EAAE,OAAO,CAAC,CAAC;QAClD,OAAO;YACL,EAAE,EAAE,MAAM,EAAE;YACZ,OAAO;YACP,MAAM,EAAE,QAAQ;YAChB,UAAU,EAAE,MAAM;SACnB,CAAC;IACJ,CAAC;IAEO,KAAK,CAAC,aAAa,CAAC,OAAe;QACzC,MAAM,OAAO,GAAG,MAAM,OAAO,CAAC,OAAO,CAAC,CAAC;
|
|
1
|
+
{"version":3,"file":"text-loader.js","sourceRoot":"","sources":["../../src/loaders/text-loader.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,QAAQ,EAAE,OAAO,EAAE,IAAI,EAAE,MAAM,kBAAkB,CAAC;AAC3D,OAAO,EAAE,IAAI,EAAE,OAAO,EAAE,MAAM,WAAW,CAAC;AAC1C,OAAO,EAAE,MAAM,EAAE,MAAM,QAAQ,CAAC;AAGhC,MAAM,eAAe,GAAG,IAAI,GAAG,CAAC,CAAC,KAAK,EAAE,MAAM,CAAC,CAAC,CAAC;AAEjD,MAAM,OAAO,UAAU;IACZ,cAAc,GAAG,CAAC,KAAK,EAAE,MAAM,CAAC,CAAC;IAE1C,KAAK,CAAC,IAAI,CAAC,MAAc;QACvB,MAAM,QAAQ,GAAG,OAAO,CAAC,MAAM,CAAC,CAAC;QACjC,MAAM,IAAI,GAAG,MAAM,IAAI,CAAC,QAAQ,CAAC,CAAC;QAElC,IAAI,IAAI,CAAC,WAAW,EAAE,EAAE,CAAC;YACvB,OAAO,IAAI,CAAC,aAAa,CAAC,QAAQ,CAAC,CAAC;QACtC,CAAC;QAED,OAAO,CAAC,MAAM,IAAI,CAAC,QAAQ,CAAC,QAAQ,CAAC,CAAC,CAAC;IACzC,CAAC;IAEO,KAAK,CAAC,QAAQ,CAAC,QAAgB;QACrC,MAAM,OAAO,GAAG,MAAM,QAAQ,CAAC,QAAQ,EAAE,OAAO,CAAC,CAAC;QAClD,OAAO;YACL,EAAE,EAAE,MAAM,EAAE;YACZ,OAAO;YACP,MAAM,EAAE,QAAQ;YAChB,UAAU,EAAE,MAAM;SACnB,CAAC;IACJ,CAAC;IAEO,KAAK,CAAC,aAAa,CAAC,OAAe;QACzC,MAAM,OAAO,GAAG,CAAC,MAAM,OAAO,CAAC,OAAO,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC;QAChD,MAAM,IAAI,GAAkB,EAAE,CAAC;QAE/B,KAAK,MAAM,KAAK,IAAI,OAAO,EAAE,CAAC;YAC5B,MAAM,GAAG,GAAG,KAAK,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,GAAG,EAAE,EAAE,WAAW,EAAE,IAAI,EAAE,CAAC;YACxD,IAAI,CAAC,eAAe,CAAC,GAAG,CAAC,GAAG,CAAC;gBAAE,SAAS;YAExC,MAAM,QAAQ,GAAG,IAAI,CAAC,OAAO,EAAE,KAAK,CAAC,CAAC;YACtC,MAAM,IAAI,GAAG,MAAM,IAAI,CAAC,QAAQ,CAAC,CAAC;YAClC,IAAI,IAAI,CAAC,MAAM,EAAE,EAAE,CAAC;gBAClB,IAAI,CAAC,IAAI,CAAC,MAAM,IAAI,CAAC,QAAQ,CAAC,QAAQ,CAAC,CAAC,CAAC;YAC3C,CAAC;QACH,CAAC;QAED,OAAO,IAAI,CAAC;IACd,CAAC;CACF"}
|
|
@@ -1,18 +1,46 @@
|
|
|
1
|
+
import { type LookupAddress, type LookupOptions } from 'node:dns';
|
|
1
2
|
import type { DocumentLoader, RAGDocument } from '@cogitator-ai/types';
|
|
3
|
+
/**
|
|
4
|
+
* Whether an IP address points to a loopback, private, link-local, CGNAT, multicast or
|
|
5
|
+
* otherwise non-public network (including IPv4-mapped / NAT64 IPv6 forms).
|
|
6
|
+
*/
|
|
7
|
+
export declare function isBlockedAddress(address: string): boolean;
|
|
8
|
+
type LookupCallback = (err: NodeJS.ErrnoException | null, address: string | LookupAddress[], family?: number) => void;
|
|
9
|
+
type Resolver = (hostname: string, callback: (err: NodeJS.ErrnoException | null, addresses: LookupAddress[]) => void) => void;
|
|
10
|
+
/**
|
|
11
|
+
* DNS lookup for `http.request` that rejects private addresses at connect time,
|
|
12
|
+
* which also defeats DNS-rebinding between validation and connection.
|
|
13
|
+
*/
|
|
14
|
+
export declare function createGuardedLookup(resolver?: Resolver): (hostname: string, options: LookupOptions, callback: LookupCallback) => void;
|
|
2
15
|
export interface WebLoaderOptions {
|
|
3
16
|
selector?: string;
|
|
4
17
|
headers?: Record<string, string>;
|
|
5
18
|
timeoutMs?: number;
|
|
6
19
|
maxResponseBytes?: number;
|
|
20
|
+
/**
|
|
21
|
+
* Allow fetching loopback / private-network hosts (e.g. an intranet wiki).
|
|
22
|
+
* Disabled by default to prevent SSRF when URLs come from untrusted input.
|
|
23
|
+
*/
|
|
24
|
+
allowPrivateNetwork?: boolean;
|
|
7
25
|
}
|
|
26
|
+
/**
|
|
27
|
+
* Loads a web page over http(s). HTML is converted to text (scripts and styles removed);
|
|
28
|
+
* plain-text and JSON responses are kept as-is. Redirects are followed (max 5) and every
|
|
29
|
+
* hop — including DNS resolution at connect time — is checked against private networks.
|
|
30
|
+
*/
|
|
8
31
|
export declare class WebLoader implements DocumentLoader {
|
|
9
32
|
readonly supportedTypes: string[];
|
|
10
33
|
private readonly htmlLoader;
|
|
11
34
|
private readonly headers?;
|
|
12
35
|
private readonly timeoutMs;
|
|
13
36
|
private readonly maxResponseBytes;
|
|
37
|
+
private readonly allowPrivateNetwork;
|
|
38
|
+
private readonly lookup;
|
|
14
39
|
constructor(options?: WebLoaderOptions);
|
|
15
40
|
load(source: string): Promise<RAGDocument[]>;
|
|
16
|
-
private
|
|
41
|
+
private fetchPage;
|
|
42
|
+
private request;
|
|
43
|
+
private readBody;
|
|
17
44
|
}
|
|
45
|
+
export {};
|
|
18
46
|
//# sourceMappingURL=web-loader.d.ts.map
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"web-loader.d.ts","sourceRoot":"","sources":["../../src/loaders/web-loader.ts"],"names":[],"mappings":"
|
|
1
|
+
{"version":3,"file":"web-loader.d.ts","sourceRoot":"","sources":["../../src/loaders/web-loader.ts"],"names":[],"mappings":"AAAA,OAAO,EAAuB,KAAK,aAAa,EAAE,KAAK,aAAa,EAAE,MAAM,UAAU,CAAC;AAMvF,OAAO,KAAK,EAAE,cAAc,EAAE,WAAW,EAAE,MAAM,qBAAqB,CAAC;AAqDvE;;;GAGG;AACH,wBAAgB,gBAAgB,CAAC,OAAO,EAAE,MAAM,GAAG,OAAO,CAUzD;AAED,KAAK,cAAc,GAAG,CACpB,GAAG,EAAE,MAAM,CAAC,cAAc,GAAG,IAAI,EACjC,OAAO,EAAE,MAAM,GAAG,aAAa,EAAE,EACjC,MAAM,CAAC,EAAE,MAAM,KACZ,IAAI,CAAC;AAEV,KAAK,QAAQ,GAAG,CACd,QAAQ,EAAE,MAAM,EAChB,QAAQ,EAAE,CAAC,GAAG,EAAE,MAAM,CAAC,cAAc,GAAG,IAAI,EAAE,SAAS,EAAE,aAAa,EAAE,KAAK,IAAI,KAC9E,IAAI,CAAC;AAMV;;;GAGG;AACH,wBAAgB,mBAAmB,CAAC,QAAQ,GAAE,QAAyB,IAC7D,UAAU,MAAM,EAAE,SAAS,aAAa,EAAE,UAAU,cAAc,KAAG,IAAI,CAsClF;AA2DD,MAAM,WAAW,gBAAgB;IAC/B,QAAQ,CAAC,EAAE,MAAM,CAAC;IAClB,OAAO,CAAC,EAAE,MAAM,CAAC,MAAM,EAAE,MAAM,CAAC,CAAC;IACjC,SAAS,CAAC,EAAE,MAAM,CAAC;IACnB,gBAAgB,CAAC,EAAE,MAAM,CAAC;IAC1B;;;OAGG;IACH,mBAAmB,CAAC,EAAE,OAAO,CAAC;CAC/B;AAQD;;;;GAIG;AACH,qBAAa,SAAU,YAAW,cAAc;IAC9C,QAAQ,CAAC,cAAc,WAAqB;IAC5C,OAAO,CAAC,QAAQ,CAAC,UAAU,CAAa;IACxC,OAAO,CAAC,QAAQ,CAAC,OAAO,CAAC,CAAyB;IAClD,OAAO,CAAC,QAAQ,CAAC,SAAS,CAAS;IACnC,OAAO,CAAC,QAAQ,CAAC,gBAAgB,CAAS;IAC1C,OAAO,CAAC,QAAQ,CAAC,mBAAmB,CAAU;IAC9C,OAAO,CAAC,QAAQ,CAAC,MAAM,CAAyB;gBAEpC,OAAO,CAAC,EAAE,gBAAgB;IAQhC,IAAI,CAAC,MAAM,EAAE,MAAM,GAAG,OAAO,CAAC,WAAW,EAAE,CAAC;YA0BpC,SAAS;IAkCvB,OAAO,CAAC,OAAO;IA4Bf,OAAO,CAAC,QAAQ;CA4BjB"}
|