@ssmeo/mcp-local-rag 0.18.14 → 0.19.0
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.de.md +248 -215
- package/README.es.md +241 -215
- package/README.fr.md +249 -215
- package/README.md +207 -297
- package/README.pt-BR.md +236 -215
- package/README.zh-CN.md +133 -206
- package/dist/chunker/semantic-chunker.d.ts +23 -5
- package/dist/chunker/semantic-chunker.d.ts.map +1 -1
- package/dist/chunker/semantic-chunker.js +13 -5
- package/dist/chunker/semantic-chunker.js.map +1 -1
- package/dist/cli/common.d.ts +7 -0
- package/dist/cli/common.d.ts.map +1 -1
- package/dist/cli/common.js +26 -2
- package/dist/cli/common.js.map +1 -1
- package/dist/cli/delete.d.ts.map +1 -1
- package/dist/cli/delete.js +2 -0
- package/dist/cli/delete.js.map +1 -1
- package/dist/cli/file-collection.d.ts +2 -1
- package/dist/cli/file-collection.d.ts.map +1 -1
- package/dist/cli/file-collection.js +2 -2
- package/dist/cli/file-collection.js.map +1 -1
- package/dist/cli/ingest.d.ts +1 -0
- package/dist/cli/ingest.d.ts.map +1 -1
- package/dist/cli/ingest.js +45 -25
- package/dist/cli/ingest.js.map +1 -1
- package/dist/cli/list.d.ts.map +1 -1
- package/dist/cli/list.js +9 -13
- package/dist/cli/list.js.map +1 -1
- package/dist/cli/options.d.ts +8 -0
- package/dist/cli/options.d.ts.map +1 -1
- package/dist/cli/options.js +21 -0
- package/dist/cli/options.js.map +1 -1
- package/dist/cli/query.d.ts.map +1 -1
- package/dist/cli/query.js +2 -1
- package/dist/cli/query.js.map +1 -1
- package/dist/cli/read-neighbors.d.ts.map +1 -1
- package/dist/cli/read-neighbors.js +1 -0
- package/dist/cli/read-neighbors.js.map +1 -1
- package/dist/cli/relocate.d.ts +3 -0
- package/dist/cli/relocate.d.ts.map +1 -0
- package/dist/cli/relocate.js +135 -0
- package/dist/cli/relocate.js.map +1 -0
- package/dist/cli/sync.d.ts.map +1 -1
- package/dist/cli/sync.js +17 -8
- package/dist/cli/sync.js.map +1 -1
- package/dist/cli-main.d.ts +1 -1
- package/dist/cli-main.d.ts.map +1 -1
- package/dist/cli-main.js +5 -0
- package/dist/cli-main.js.map +1 -1
- package/dist/embedder/embedding-snapshot.d.ts +52 -0
- package/dist/embedder/embedding-snapshot.d.ts.map +1 -0
- package/dist/embedder/embedding-snapshot.js +413 -0
- package/dist/embedder/embedding-snapshot.js.map +1 -0
- package/dist/embedder/index.d.ts +36 -5
- package/dist/embedder/index.d.ts.map +1 -1
- package/dist/embedder/index.js +346 -23
- package/dist/embedder/index.js.map +1 -1
- package/dist/embedder/remote.d.ts +16 -2
- package/dist/embedder/remote.d.ts.map +1 -1
- package/dist/embedder/remote.js +48 -11
- package/dist/embedder/remote.js.map +1 -1
- package/dist/embedder/sentence-transformers-config.d.ts +24 -0
- package/dist/embedder/sentence-transformers-config.d.ts.map +1 -0
- package/dist/embedder/sentence-transformers-config.js +93 -0
- package/dist/embedder/sentence-transformers-config.js.map +1 -0
- package/dist/features/sync.js +1 -1
- package/dist/features/sync.js.map +1 -1
- package/dist/ingest/compute.d.ts +15 -7
- package/dist/ingest/compute.d.ts.map +1 -1
- package/dist/ingest/compute.js +74 -14
- package/dist/ingest/compute.js.map +1 -1
- package/dist/ingest/file.d.ts +3 -3
- package/dist/ingest/file.d.ts.map +1 -1
- package/dist/ingest/file.js +0 -5
- package/dist/ingest/file.js.map +1 -1
- package/dist/ingest/visual.d.ts +6 -2
- package/dist/ingest/visual.d.ts.map +1 -1
- package/dist/ingest/visual.js +20 -4
- package/dist/ingest/visual.js.map +1 -1
- package/dist/parser/docx-parser.d.ts +2 -0
- package/dist/parser/docx-parser.d.ts.map +1 -1
- package/dist/parser/docx-parser.js +12 -3
- package/dist/parser/docx-parser.js.map +1 -1
- package/dist/parser/index.d.ts +11 -1
- package/dist/parser/index.d.ts.map +1 -1
- package/dist/parser/index.js +27 -11
- package/dist/parser/index.js.map +1 -1
- package/dist/parser/pdf-extract.d.ts +2 -0
- package/dist/parser/pdf-extract.d.ts.map +1 -1
- package/dist/parser/pdf-extract.js +3 -0
- package/dist/parser/pdf-extract.js.map +1 -1
- package/dist/parser/pdf-filter.d.ts +2 -9
- package/dist/parser/pdf-filter.d.ts.map +1 -1
- package/dist/parser/pdf-filter.js +110 -144
- package/dist/parser/pdf-filter.js.map +1 -1
- package/dist/parser/pdf-headings.d.ts +6 -0
- package/dist/parser/pdf-headings.d.ts.map +1 -0
- package/dist/parser/pdf-headings.js +166 -0
- package/dist/parser/pdf-headings.js.map +1 -0
- package/dist/parser/text-source-map.d.ts +3 -0
- package/dist/parser/text-source-map.d.ts.map +1 -0
- package/dist/parser/text-source-map.js +33 -0
- package/dist/parser/text-source-map.js.map +1 -0
- package/dist/pdf-visual/detector.js +3 -3
- package/dist/pdf-visual/detector.js.map +1 -1
- package/dist/rerank/command.d.ts +10 -7
- package/dist/rerank/command.d.ts.map +1 -1
- package/dist/rerank/command.js +51 -10
- package/dist/rerank/command.js.map +1 -1
- package/dist/rerank/index.d.ts +3 -3
- package/dist/rerank/index.d.ts.map +1 -1
- package/dist/rerank/index.js +3 -3
- package/dist/rerank/index.js.map +1 -1
- package/dist/server/index.d.ts +4 -1
- package/dist/server/index.d.ts.map +1 -1
- package/dist/server/index.js +114 -79
- package/dist/server/index.js.map +1 -1
- package/dist/server/list-scanner.d.ts +2 -1
- package/dist/server/list-scanner.d.ts.map +1 -1
- package/dist/server/list-scanner.js +2 -2
- package/dist/server/list-scanner.js.map +1 -1
- package/dist/server/tool-definitions.js +2 -2
- package/dist/server/tool-definitions.js.map +1 -1
- package/dist/server/types.d.ts +9 -0
- package/dist/server/types.d.ts.map +1 -1
- package/dist/server-main.d.ts +2 -2
- package/dist/server-main.d.ts.map +1 -1
- package/dist/server-main.js +22 -20
- package/dist/server-main.js.map +1 -1
- package/dist/utils/base-dirs.d.ts +9 -10
- package/dist/utils/base-dirs.d.ts.map +1 -1
- package/dist/utils/base-dirs.js +28 -30
- package/dist/utils/base-dirs.js.map +1 -1
- package/dist/utils/scan.d.ts +25 -5
- package/dist/utils/scan.d.ts.map +1 -1
- package/dist/utils/scan.js +40 -10
- package/dist/utils/scan.js.map +1 -1
- package/dist/utils/source-context.d.ts +29 -0
- package/dist/utils/source-context.d.ts.map +1 -0
- package/dist/utils/source-context.js +86 -0
- package/dist/utils/source-context.js.map +1 -0
- package/dist/utils/sync-path-key.d.ts +1 -1
- package/dist/utils/sync-path-key.d.ts.map +1 -1
- package/dist/utils/sync-path-key.js +1 -1
- package/dist/utils/sync-path-key.js.map +1 -1
- package/dist/vectordb/index.d.ts +4 -2
- package/dist/vectordb/index.d.ts.map +1 -1
- package/dist/vectordb/index.js +30 -7
- package/dist/vectordb/index.js.map +1 -1
- package/dist/vectordb/relocate.d.ts +18 -0
- package/dist/vectordb/relocate.d.ts.map +1 -0
- package/dist/vectordb/relocate.js +179 -0
- package/dist/vectordb/relocate.js.map +1 -0
- package/dist/vectordb/types.d.ts +10 -23
- package/dist/vectordb/types.d.ts.map +1 -1
- package/dist/vectordb/types.js +14 -23
- package/dist/vectordb/types.js.map +1 -1
- package/package.json +14 -14
package/dist/ingest/file.js
CHANGED
|
@@ -6,7 +6,6 @@ export function buildPreparedFileVectorChunks(prepared) {
|
|
|
6
6
|
filePath: prepared.filePath,
|
|
7
7
|
chunks: prepared.chunks,
|
|
8
8
|
embeddings: prepared.embeddings,
|
|
9
|
-
fileSize: prepared.textLength,
|
|
10
9
|
fileTitle: prepared.title,
|
|
11
10
|
contentHash: prepared.contentHash,
|
|
12
11
|
visualProfile: prepared.visualProfile,
|
|
@@ -31,7 +30,6 @@ export async function prepareFileForIngest(filePath, collaborators, options) {
|
|
|
31
30
|
const { parser, chunker, embedder } = collaborators;
|
|
32
31
|
const contentHash = await readPreParseContentHash(filePath, parser);
|
|
33
32
|
const isPdf = filePath.toLowerCase().endsWith('.pdf');
|
|
34
|
-
let text;
|
|
35
33
|
let title;
|
|
36
34
|
let chunks;
|
|
37
35
|
let embeddings;
|
|
@@ -43,7 +41,6 @@ export async function prepareFileForIngest(filePath, collaborators, options) {
|
|
|
43
41
|
...(options.captioner === undefined ? {} : { captioner: options.captioner }),
|
|
44
42
|
};
|
|
45
43
|
const result = await prepareVisualPdfChunks(filePath, collaborators, visualOptions);
|
|
46
|
-
text = result.text;
|
|
47
44
|
title = result.title;
|
|
48
45
|
chunks = result.chunks;
|
|
49
46
|
embeddings = result.embeddings;
|
|
@@ -54,7 +51,6 @@ export async function prepareFileForIngest(filePath, collaborators, options) {
|
|
|
54
51
|
const result = isPdf
|
|
55
52
|
? await parser.parsePdf(filePath, embedder)
|
|
56
53
|
: await parser.parseFile(filePath, { images: options.images });
|
|
57
|
-
text = result.content;
|
|
58
54
|
title = result.title || null;
|
|
59
55
|
({ chunks, embeddings, visualAttachments, omittedImageCount } =
|
|
60
56
|
await buildChunksFromParseResult(result, chunker, embedder));
|
|
@@ -63,7 +59,6 @@ export async function prepareFileForIngest(filePath, collaborators, options) {
|
|
|
63
59
|
filePath,
|
|
64
60
|
chunks,
|
|
65
61
|
embeddings,
|
|
66
|
-
textLength: text.length,
|
|
67
62
|
title,
|
|
68
63
|
contentHash,
|
|
69
64
|
visualProfile: isPdf ? (options.captioner?.profile ?? null) : null,
|
package/dist/ingest/file.js.map
CHANGED
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"file.js","sourceRoot":"","sources":["../../src/ingest/file.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,QAAQ,EAAE,MAAM,kBAAkB,CAAA;AAO3C,OAAO,EAAE,0BAA0B,EAAE,iBAAiB,EAAE,kBAAkB,EAAE,MAAM,cAAc,CAAA;AAChG,OAAO,EAGL,sBAAsB,GACvB,MAAM,aAAa,CAAA;
|
|
1
|
+
{"version":3,"file":"file.js","sourceRoot":"","sources":["../../src/ingest/file.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,QAAQ,EAAE,MAAM,kBAAkB,CAAA;AAO3C,OAAO,EAAE,0BAA0B,EAAE,iBAAiB,EAAE,kBAAkB,EAAE,MAAM,cAAc,CAAA;AAChG,OAAO,EAGL,sBAAsB,GACvB,MAAM,aAAa,CAAA;AA0BpB,MAAM,UAAU,6BAA6B,CAAC,QAA4B;IACxE,OAAO,iBAAiB,CAAC;QACvB,QAAQ,EAAE,QAAQ,CAAC,QAAQ;QAC3B,MAAM,EAAE,QAAQ,CAAC,MAAM;QACvB,UAAU,EAAE,QAAQ,CAAC,UAAU;QAC/B,SAAS,EAAE,QAAQ,CAAC,KAAK;QACzB,WAAW,EAAE,QAAQ,CAAC,WAAW;QACjC,aAAa,EAAE,QAAQ,CAAC,aAAa;QACrC,iBAAiB,EAAE,QAAQ,CAAC,iBAAiB;KAC9C,CAAC,CAAA;AACJ,CAAC;AAED,KAAK,UAAU,uBAAuB,CAAC,QAAgB,EAAE,MAAsB;IAC7E,MAAM,MAAM,CAAC,gBAAgB,CAAC,QAAQ,CAAC,CAAA;IACvC,MAAM,CAAC,gBAAgB,CAAC,QAAQ,CAAC,CAAA;IACjC,OAAO,kBAAkB,CAAC,MAAM,QAAQ,CAAC,QAAQ,CAAC,CAAC,CAAA;AACrD,CAAC;AAED;;;;;;;;GAQG;AACH,MAAM,CAAC,KAAK,UAAU,oBAAoB,CACxC,QAAgB,EAChB,aAAsC,EACtC,OAAoC;IAEpC,MAAM,EAAE,MAAM,EAAE,OAAO,EAAE,QAAQ,EAAE,GAAG,aAAa,CAAA;IACnD,MAAM,WAAW,GAAG,MAAM,uBAAuB,CAAC,QAAQ,EAAE,MAAM,CAAC,CAAA;IACnE,MAAM,KAAK,GAAG,QAAQ,CAAC,WAAW,EAAE,CAAC,QAAQ,CAAC,MAAM,CAAC,CAAA;IAErD,IAAI,KAAoB,CAAA;IACxB,IAAI,MAAwE,CAAA;IAC5E,IAAI,UAAgF,CAAA;IACpF,IAAI,iBAA8F,CAAA;IAClG,IAAI,iBAAyB,CAAA;IAE7B,IAAI,KAAK,IAAI,CAAC,OAAO,CAAC,SAAS,KAAK,SAAS,IAAI,OAAO,CAAC,MAAM,CAAC,EAAE,CAAC;QACjE,MAAM,aAAa,GAAkC;YACnD,MAAM,EAAE,OAAO,CAAC,MAAM;YACtB,GAAG,CAAC,OAAO,CAAC,SAAS,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,SAAS,EAAE,OAAO,CAAC,SAAS,EAAE,CAAC;SAC7E,CAAA;QACD,MAAM,MAAM,GAAG,MAAM,sBAAsB,CAAC,QAAQ,EAAE,aAAa,EAAE,aAAa,CAAC,CAAA;QACnF,KAAK,GAAG,MAAM,CAAC,KAAK,CAAA;QACpB,MAAM,GAAG,MAAM,CAAC,MAAM,CAAA;QACtB,UAAU,GAAG,MAAM,CAAC,UAAU,CAAA;QAC9B,iBAAiB,GAAG,MAAM,CAAC,iBAAiB,CAAA;QAC5C,iBAAiB,GAAG,MAAM,CAAC,iBAAiB,CAAA;IAC9C,CAAC;SAAM,CAAC;QACN,MAAM,MAAM,GAAG,KAAK;YAClB,CAAC,CAAC,MAAM,MAAM,CAAC,QAAQ,CAAC,QAAQ,EAAE,QAAQ,CAAC;YAC3C,CAAC,CAAC,MAAM,MAAM,CAAC,SAAS,CAAC,QAAQ,EAAE,EAAE,MAAM,EAAE,OAAO,CAAC,MAAM,EAAE,CAAC,CAAA;QAChE,KAAK,GAAG,MAAM,CAAC,KAAK,IAAI,IAAI,CAC3B;QAAA,CAAC,EAAE,MAAM,EAAE,UAAU,EAAE,iBAAiB,EAAE,iBAAiB,EAAE;YAC5D,MAAM,0BAA0B,CAAC,MAAM,EAAE,OAAO,EAAE,QAAQ,CAAC,CAAC,CAAA;IAChE,CAAC;IAED,OAAO;QACL,QAAQ;QACR,MAAM;QACN,UAAU;QACV,KAAK;QACL,WAAW;QACX,aAAa,EAAE,KAAK,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC,SAAS,EAAE,OAAO,IAAI,IAAI,CAAC,CAAC,CAAC,CAAC,IAAI;QAClE,iBAAiB;QACjB,iBAAiB;KAClB,CAAA;AACH,CAAC"}
|
package/dist/ingest/visual.d.ts
CHANGED
|
@@ -1,10 +1,12 @@
|
|
|
1
|
-
import type { AtomicTextRange, SemanticChunker
|
|
1
|
+
import type { AtomicTextRange, SemanticChunker } from '../chunker/index.js';
|
|
2
2
|
import type { EmbedderInterface } from '../chunker/semantic-chunker.js';
|
|
3
3
|
import type { DocumentParser } from '../parser/index.js';
|
|
4
4
|
import type { FilteredTextFragment } from '../parser/pdf-filter.js';
|
|
5
5
|
import type { ProcessedVisualRegion } from '../pdf-visual/types.js';
|
|
6
|
+
import type { HeadingAnchor, SourceMap } from '../utils/source-context.js';
|
|
6
7
|
import type { QualityProfile } from '../utils/visual-profile.js';
|
|
7
8
|
import type { VisualAttachment } from '../vectordb/types.js';
|
|
9
|
+
import { buildChunksAndEmbeddings } from './compute.js';
|
|
8
10
|
export interface VisualPdfParser {
|
|
9
11
|
parsePdfPages: DocumentParser['parsePdfPages'];
|
|
10
12
|
}
|
|
@@ -24,7 +26,7 @@ export interface PrepareVisualPdfChunksOptions {
|
|
|
24
26
|
captioner?: CaptionerConfig;
|
|
25
27
|
}
|
|
26
28
|
export interface PrepareVisualPdfChunksResult {
|
|
27
|
-
chunks:
|
|
29
|
+
chunks: Awaited<ReturnType<typeof buildChunksAndEmbeddings>>['chunks'];
|
|
28
30
|
embeddings: number[][];
|
|
29
31
|
title: string | null;
|
|
30
32
|
text: string;
|
|
@@ -33,6 +35,7 @@ export interface PrepareVisualPdfChunksResult {
|
|
|
33
35
|
omittedImageCount: number;
|
|
34
36
|
}
|
|
35
37
|
export interface OrderedVisualPage {
|
|
38
|
+
headings?: HeadingAnchor[];
|
|
36
39
|
pageNum: number;
|
|
37
40
|
text: string;
|
|
38
41
|
textFragments: readonly FilteredTextFragment[];
|
|
@@ -43,6 +46,7 @@ export interface OrderedVisualRegion extends ProcessedVisualRegion {
|
|
|
43
46
|
captionRange?: AtomicTextRange;
|
|
44
47
|
}
|
|
45
48
|
export interface OrderedVisualDocument {
|
|
49
|
+
sourceMap?: SourceMap;
|
|
46
50
|
text: string;
|
|
47
51
|
atomicRanges: AtomicTextRange[];
|
|
48
52
|
regions: OrderedVisualRegion[];
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"visual.d.ts","sourceRoot":"","sources":["../../src/ingest/visual.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,eAAe,EAAE,eAAe,
|
|
1
|
+
{"version":3,"file":"visual.d.ts","sourceRoot":"","sources":["../../src/ingest/visual.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,eAAe,EAAE,eAAe,EAAa,MAAM,qBAAqB,CAAA;AACtF,OAAO,KAAK,EAAE,iBAAiB,EAAE,MAAM,gCAAgC,CAAA;AACvE,OAAO,KAAK,EAAE,cAAc,EAAE,MAAM,oBAAoB,CAAA;AACxD,OAAO,KAAK,EAAE,oBAAoB,EAAE,MAAM,yBAAyB,CAAA;AACnE,OAAO,KAAK,EAAwB,qBAAqB,EAAE,MAAM,wBAAwB,CAAA;AACzF,OAAO,KAAK,EAAE,aAAa,EAAE,SAAS,EAAE,MAAM,4BAA4B,CAAA;AAC1E,OAAO,KAAK,EAAE,cAAc,EAAE,MAAM,4BAA4B,CAAA;AAChE,OAAO,KAAK,EAAE,gBAAgB,EAAE,MAAM,sBAAsB,CAAA;AAC5D,OAAO,EAAE,wBAAwB,EAA4C,MAAM,cAAc,CAAA;AAEjG,MAAM,WAAW,eAAe;IAC9B,aAAa,EAAE,cAAc,CAAC,eAAe,CAAC,CAAA;CAC/C;AAED,MAAM,WAAW,eAAe;IAC9B,OAAO,EAAE,cAAc,CAAA;IACvB,QAAQ,EAAE,MAAM,CAAA;IAChB,MAAM,CAAC,EAAE,MAAM,GAAG,SAAS,CAAA;CAC5B;AAED,yEAAyE;AACzE,MAAM,WAAW,yBAAyB;IACxC,MAAM,EAAE,eAAe,CAAA;IACvB,OAAO,EAAE,eAAe,CAAA;IACxB,QAAQ,EAAE,iBAAiB,CAAA;CAC5B;AAED,MAAM,WAAW,6BAA6B;IAC5C,MAAM,EAAE,OAAO,CAAA;IACf,SAAS,CAAC,EAAE,eAAe,CAAA;CAC5B;AAED,MAAM,WAAW,4BAA4B;IAC3C,MAAM,EAAE,OAAO,CAAC,UAAU,CAAC,OAAO,wBAAwB,CAAC,CAAC,CAAC,QAAQ,CAAC,CAAA;IACtE,UAAU,EAAE,MAAM,EAAE,EAAE,CAAA;IACtB,KAAK,EAAE,MAAM,GAAG,IAAI,CAAA;IACpB,IAAI,EAAE,MAAM,CAAA;IACZ,YAAY,EAAE,eAAe,EAAE,CAAA;IAC/B,iBAAiB,EAAE,GAAG,CAAC,MAAM,EAAE,gBAAgB,EAAE,CAAC,CAAA;IAClD,iBAAiB,EAAE,MAAM,CAAA;CAC1B;AAED,MAAM,WAAW,iBAAiB;IAChC,QAAQ,CAAC,EAAE,aAAa,EAAE,CAAA;IAC1B,OAAO,EAAE,MAAM,CAAA;IACf,IAAI,EAAE,MAAM,CAAA;IACZ,aAAa,EAAE,SAAS,oBAAoB,EAAE,CAAA;CAC/C;AAED,MAAM,WAAW,mBAAoB,SAAQ,qBAAqB;IAChE,WAAW,EAAE,MAAM,CAAA;IACnB,YAAY,EAAE,MAAM,CAAA;IACpB,YAAY,CAAC,EAAE,eAAe,CAAA;CAC/B;AAED,MAAM,WAAW,qBAAqB;IACpC,SAAS,CAAC,EAAE,SAAS,CAAA;IACrB,IAAI,EAAE,MAAM,CAAA;IACZ,YAAY,EAAE,eAAe,EAAE,CAAA;IAC/B,OAAO,EAAE,mBAAmB,EAAE,CAAA;CAC/B;AA0ND,wBAAgB,0BAA0B,CACxC,KAAK,EAAE,SAAS,iBAAiB,EAAE,EACnC,gBAAgB,EAAE,SAAS,qBAAqB,EAAE,GACjD,qBAAqB,CA0CvB;AAoHD,wBAAsB,sBAAsB,CAC1C,QAAQ,EAAE,MAAM,EAChB,aAAa,EAAE,yBAAyB,EACxC,OAAO,EAAE,6BAA6B,GACrC,OAAO,CAAC,4BAA4B,CAAC,CAwCvC"}
|
package/dist/ingest/visual.js
CHANGED
|
@@ -102,6 +102,7 @@ function appendGroupCaptions(pageText, group, context) {
|
|
|
102
102
|
function renderPageWithCaptions(page, insertionGroups, pageVisualIndices) {
|
|
103
103
|
let pageText = '';
|
|
104
104
|
let pageCursor = 0;
|
|
105
|
+
const shifts = [];
|
|
105
106
|
const captionRanges = new Map();
|
|
106
107
|
const anchorOffsets = new Map();
|
|
107
108
|
for (const [offset, group] of [...insertionGroups].sort((left, right) => left[0] - right[0])) {
|
|
@@ -119,10 +120,16 @@ function renderPageWithCaptions(page, insertionGroups, pageVisualIndices) {
|
|
|
119
120
|
if (hasCaption && offset < page.text.length) {
|
|
120
121
|
pageText += blankLinePadding(page.text.slice(offset));
|
|
121
122
|
}
|
|
123
|
+
shifts.push({ offset, delta: pageText.length - offset });
|
|
122
124
|
pageCursor = offset;
|
|
123
125
|
}
|
|
124
126
|
pageText += page.text.slice(pageCursor);
|
|
125
|
-
|
|
127
|
+
const headings = (page.headings ?? []).map((heading) => ({
|
|
128
|
+
...heading,
|
|
129
|
+
offset: heading.offset +
|
|
130
|
+
(shifts.filter((shift) => shift.offset <= heading.offset).at(-1)?.delta ?? 0),
|
|
131
|
+
}));
|
|
132
|
+
return { pageText, captionRanges, anchorOffsets, headings };
|
|
126
133
|
}
|
|
127
134
|
/** Rebase one page-relative region onto the whole-document text. */
|
|
128
135
|
function toOrderedRegion(region, rendered, pageStart, visualIndex) {
|
|
@@ -143,6 +150,7 @@ function toOrderedRegion(region, rendered, pageStart, visualIndex) {
|
|
|
143
150
|
}
|
|
144
151
|
export function buildOrderedVisualDocument(pages, processedRegions) {
|
|
145
152
|
let text = '';
|
|
153
|
+
const sourceMap = { headings: [], pages: [] };
|
|
146
154
|
let visualIndex = 0;
|
|
147
155
|
const atomicRanges = [];
|
|
148
156
|
const regions = [];
|
|
@@ -156,6 +164,10 @@ export function buildOrderedVisualDocument(pages, processedRegions) {
|
|
|
156
164
|
const pageSeparator = text.length > 0 && rendered.pageText.length > 0 ? '\n\n' : '';
|
|
157
165
|
const pageStart = text.length + pageSeparator.length;
|
|
158
166
|
text += pageSeparator + rendered.pageText;
|
|
167
|
+
if (rendered.pageText) {
|
|
168
|
+
sourceMap.pages?.push({ start: pageStart, end: text.length, page: page.pageNum });
|
|
169
|
+
}
|
|
170
|
+
sourceMap.headings.push(...rendered.headings.map((heading) => ({ ...heading, offset: pageStart + heading.offset })));
|
|
159
171
|
for (const positionedRegion of positioned) {
|
|
160
172
|
const record = toOrderedRegion(positionedRegion.region, rendered, pageStart, visualIndex++);
|
|
161
173
|
if (record.captionRange) {
|
|
@@ -168,7 +180,7 @@ export function buildOrderedVisualDocument(pages, processedRegions) {
|
|
|
168
180
|
if (consumedRegions.size !== processedRegions.length) {
|
|
169
181
|
throw new Error('A processed visual region has no matching PDF page');
|
|
170
182
|
}
|
|
171
|
-
return { text, atomicRanges, regions };
|
|
183
|
+
return { text, atomicRanges, regions, sourceMap };
|
|
172
184
|
}
|
|
173
185
|
function assignVisualAttachments(document, chunks, attachments) {
|
|
174
186
|
const attachmentsByChunkIndex = new Map();
|
|
@@ -257,14 +269,18 @@ function collectAttachments(regions) {
|
|
|
257
269
|
export async function prepareVisualPdfChunks(filePath, collaborators, options) {
|
|
258
270
|
const { parser, chunker, embedder } = collaborators;
|
|
259
271
|
const captionerConfig = options.captioner;
|
|
260
|
-
const { doc, title, pages } = await parser.parsePdfPages(filePath, embedder);
|
|
272
|
+
const { doc, title, embeddingTitle, pages } = await parser.parsePdfPages(filePath, embedder);
|
|
261
273
|
try {
|
|
262
274
|
let omittedImageCount = 0;
|
|
263
275
|
const detected = await detectAndProcessRegions(pages, doc, captionerConfig, options.images);
|
|
264
276
|
const processed = detected.processed;
|
|
265
277
|
omittedImageCount += detected.omittedImageCount;
|
|
266
278
|
const ordered = buildOrderedVisualDocument(pages, processed);
|
|
267
|
-
const { chunks, embeddings } = await buildChunksAndEmbeddings(ordered.text, chunker, embedder,
|
|
279
|
+
const { chunks, embeddings } = await buildChunksAndEmbeddings(ordered.text, chunker, embedder, {
|
|
280
|
+
sourceMap: ordered.sourceMap,
|
|
281
|
+
atomicRanges: ordered.atomicRanges,
|
|
282
|
+
title: embeddingTitle,
|
|
283
|
+
});
|
|
268
284
|
const collected = options.images
|
|
269
285
|
? collectAttachments(ordered.regions)
|
|
270
286
|
: { attachments: [], omittedImageCount: 0 };
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"visual.js","sourceRoot":"","sources":["../../src/ingest/visual.ts"],"names":[],"mappings":"AAOA,OAAO,EAAE,wBAAwB,EAAE,sBAAsB,EAAE,gBAAgB,EAAE,MAAM,cAAc,CAAA;AAoDjG,SAAS,iBAAiB,CAAC,IAAuB,EAAE,KAAwB;IAC1E,OAAO,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC,EAAE,KAAK,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,IAAI,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC,EAAE,KAAK,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAA;AACvF,CAAC;AAED,SAAS,OAAO,CAAC,IAAuB;IACtC,OAAO,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,GAAG,CAAC,CAAA;AAC9C,CAAC;AAED,SAAS,OAAO,CAAC,IAAuB;IACtC,OAAO,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,GAAG,CAAC,CAAA;AAC9C,CAAC;AAED,SAAS,kBAAkB,CACzB,SAA0C,EAC1C,MAA4B;IAE5B,IAAI,SAAS,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QAC3B,OAAO,CAAC,CAAA;IACV,CAAC;IACD,IAAI,UAAU,GAAG,SAAS;SACvB,GAAG,CAAC,CAAC,QAAQ,EAAE,KAAK,EAAE,EAAE,CAAC,CAAC,EAAE,QAAQ,EAAE,KAAK,EAAE,CAAC,CAAC;SAC/C,MAAM,CAAC,CAAC,EAAE,QAAQ,EAAE,EAAE,EAAE,CAAC,iBAAiB,CAAC,QAAQ,CAAC,IAAI,EAAE,MAAM,CAAC,IAAI,CAAC,CAAC,CAAA;IAE1E,IAAI,UAAU,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QAC5B,MAAM,YAAY,GAAG,OAAO,CAAC,MAAM,CAAC,IAAI,CAAC,CAAA;QACzC,MAAM,eAAe,GAAG,IAAI,CAAC,GAAG,CAC9B,GAAG,SAAS,CAAC,GAAG,CAAC,CAAC,QAAQ,EAAE,EAAE,CAAC,IAAI,CAAC,GAAG,CAAC,OAAO,CAAC,QAAQ,CAAC,IAAI,CAAC,GAAG,YAAY,CAAC,CAAC,CAChF,CAAA;QACD,UAAU,GAAG,SAAS;aACnB,GAAG,CAAC,CAAC,QAAQ,EAAE,KAAK,EAAE,EAAE,CAAC,CAAC,EAAE,QAAQ,EAAE,KAAK,EAAE,CAAC,CAAC;aAC/C,MAAM,CAAC,CAAC,EAAE,QAAQ,EAAE,EAAE,EAAE,CAAC,IAAI,CAAC,GAAG,CAAC,OAAO,CAAC,QAAQ,CAAC,IAAI,CAAC,GAAG,YAAY,CAAC,KAAK,eAAe,CAAC,CAAA;IAClG,CAAC;IAED,MAAM,cAAc,GAAG,OAAO,CAAC,MAAM,CAAC,IAAI,CAAC,CAAA;IAC3C,MAAM,SAAS,GAAG,UAAU,CAAC,MAAM,CAAC,CAAC,EAAE,QAAQ,EAAE,EAAE,EAAE,CAAC,OAAO,CAAC,QAAQ,CAAC,IAAI,CAAC,IAAI,cAAc,CAAC,CAAA;IAC/F,IAAI,SAAS,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QACzB,OAAO,IAAI,CAAC,GAAG,CAAC,GAAG,SAAS,CAAC,GAAG,CAAC,CAAC,EAAE,KAAK,EAAE,EAAE,EAAE,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC,CAAA;IAC7D,CAAC;IACD,OAAO,IAAI,CAAC,GAAG,CAAC,GAAG,UAAU,CAAC,GAAG,CAAC,CAAC,EAAE,KAAK,EAAE,EAAE,EAAE,CAAC,KAAK,CAAC,CAAC,CAAA;AAC1D,CAAC;AAED,SAAS,iBAAiB,CAAC,IAAuB;IAChD,IAAI,IAAI,CAAC,aAAa,CAAC,MAAM,GAAG,CAAC,IAAI,IAAI,CAAC,IAAI,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QAC5D,OAAO,IAAI,CAAC,aAAa,CAAA;IAC3B,CAAC;IACD,OAAO;QACL;YACE,OAAO,EAAE,IAAI,CAAC,OAAO;YACrB,YAAY,EAAE,CAAC;YACf,WAAW,EAAE,CAAC;YACd,eAAe,EAAE,CAAC;YAClB,IAAI,EAAE,CAAC,CAAC,EAAE,CAAC,EAAE,CAAC,EAAE,CAAC,CAAC;YAClB,IAAI,EAAE,IAAI,CAAC,IAAI;YACf,aAAa,EAAE,CAAC;YAChB,WAAW,EAAE,IAAI,CAAC,IAAI,CAAC,MAAM;SAC9B;KACF,CAAA;AACH,CAAC;AAQD,2EAA2E;AAC3E,SAAS,eAAe,CACtB,IAAuB,EACvB,aAA8C,EAC9C,gBAAkD;IAElD,OAAO,gBAAgB;SACpB,MAAM,CAAC,CAAC,MAAM,EAAE,EAAE,CAAC,MAAM,CAAC,OAAO,KAAK,IAAI,CAAC,OAAO,CAAC;SACnD,GAAG,CAAC,CAAC,MAAM,EAAE,EAAE,CAAC,CAAC,EAAE,MAAM,EAAE,cAAc,EAAE,kBAAkB,CAAC,aAAa,EAAE,MAAM,CAAC,EAAE,CAAC,CAAC;SACxF,IAAI,CACH,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE,CACd,IAAI,CAAC,cAAc,GAAG,KAAK,CAAC,cAAc;QAC1C,IAAI,CAAC,MAAM,CAAC,cAAc,GAAG,KAAK,CAAC,MAAM,CAAC,cAAc,CAC3D,CAAA;AACL,CAAC;AAED;;;GAGG;AACH,SAAS,sBAAsB,CAC7B,IAAuB,EACvB,aAA8C,EAC9C,UAAuC;IAEvC,MAAM,MAAM,GAAG,IAAI,GAAG,EAAmC,CAAA;IACzD,KAAK,MAAM,IAAI,IAAI,UAAU,EAAE,CAAC;QAC9B,MAAM,SAAS,GAAG,aAAa,CAAC,IAAI,CAAC,cAAc,GAAG,CAAC,CAAC,CAAA;QACxD,MAAM,SAAS,GAAG,aAAa,CAAC,IAAI,CAAC,cAAc,CAAC,CAAA;QACpD,MAAM,MAAM,GAAG,IAAI,CAAC,GAAG,CACrB,CAAC,EACD,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,IAAI,CAAC,MAAM,EAAE,SAAS,EAAE,WAAW,IAAI,SAAS,EAAE,aAAa,IAAI,CAAC,CAAC,CACpF,CAAA;QACD,MAAM,KAAK,GAAG,MAAM,CAAC,GAAG,CAAC,MAAM,CAAC,IAAI,EAAE,CAAA;QACtC,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,MAAM,CAAC,CAAA;QACvB,MAAM,CAAC,GAAG,CAAC,MAAM,EAAE,KAAK,CAAC,CAAA;IAC3B,CAAC;IACD,OAAO,MAAM,CAAA;AACf,CAAC;AASD,kEAAkE;AAClE,SAAS,gBAAgB,CAAC,QAAgB;IACxC,MAAM,OAAO,GAAG,QAAQ,CAAC,KAAK,CAAC,MAAM,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,MAAM,IAAI,CAAC,CAAA;IACvD,OAAO,IAAI,CAAC,MAAM,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,CAAC,GAAG,OAAO,CAAC,CAAC,CAAA;AAC9C,CAAC;AAED;;;GAGG;AACH,SAAS,mBAAmB,CAC1B,QAAgB,EAChB,KAAuC,EACvC,OAKC;IAED,MAAM,EAAE,IAAI,EAAE,iBAAiB,EAAE,GAAG,IAAI,EAAE,GAAG,OAAO,CAAA;IACpD,IAAI,IAAI,GAAG,QAAQ,CAAA;IACnB,IAAI,YAAY,GAAG,CAAC,CAAA;IACpB,KAAK,MAAM,MAAM,IAAI,KAAK,EAAE,CAAC;QAC3B,IAAI,MAAM,CAAC,OAAO,KAAK,IAAI,EAAE,CAAC;YAC5B,IAAI,CAAC,aAAa,CAAC,GAAG,CAAC,MAAM,EAAE,IAAI,CAAC,MAAM,CAAC,CAAA;YAC3C,SAAQ;QACV,CAAC;QACD,IAAI,YAAY,GAAG,CAAC,EAAE,CAAC;YACrB,IAAI,IAAI,MAAM,CAAA;QAChB,CAAC;QACD,MAAM,KAAK,GAAG,IAAI,CAAC,MAAM,CAAA;QACzB,IAAI,IAAI,2BAA2B,IAAI,CAAC,OAAO,YAAY,iBAAiB,CAAC,GAAG,CAAC,MAAM,CAAC,KAAK,MAAM,CAAC,OAAO,GAAG,CAAA;QAC9G,IAAI,CAAC,aAAa,CAAC,GAAG,CAAC,MAAM,EAAE,EAAE,KAAK,EAAE,GAAG,EAAE,IAAI,CAAC,MAAM,EAAE,CAAC,CAAA;QAC3D,IAAI,CAAC,aAAa,CAAC,GAAG,CAAC,MAAM,EAAE,KAAK,CAAC,CAAA;QACrC,YAAY,IAAI,CAAC,CAAA;IACnB,CAAC;IACD,OAAO,IAAI,CAAA;AACb,CAAC;AAED,SAAS,sBAAsB,CAC7B,IAAuB,EACvB,eAAqD,EACrD,iBAAqD;IAErD,IAAI,QAAQ,GAAG,EAAE,CAAA;IACjB,IAAI,UAAU,GAAG,CAAC,CAAA;IAClB,MAAM,aAAa,GAAG,IAAI,GAAG,EAA0C,CAAA;IACvE,MAAM,aAAa,GAAG,IAAI,GAAG,EAAiC,CAAA;IAE9D,KAAK,MAAM,CAAC,MAAM,EAAE,KAAK,CAAC,IAAI,CAAC,GAAG,eAAe,CAAC,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE,CAAC,IAAI,CAAC,CAAC,CAAC,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC;QAC7F,QAAQ,IAAI,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,UAAU,EAAE,MAAM,CAAC,CAAA;QAC/C,MAAM,UAAU,GAAG,KAAK,CAAC,IAAI,CAAC,CAAC,MAAM,EAAE,EAAE,CAAC,MAAM,CAAC,OAAO,KAAK,IAAI,CAAC,CAAA;QAClE,IAAI,UAAU,IAAI,QAAQ,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;YACtC,QAAQ,IAAI,gBAAgB,CAAC,QAAQ,CAAC,KAAK,CAAC,MAAM,CAAC,EAAE,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAA;QACjE,CAAC;QACD,QAAQ,GAAG,mBAAmB,CAAC,QAAQ,EAAE,KAAK,EAAE;YAC9C,IAAI;YACJ,iBAAiB;YACjB,aAAa;YACb,aAAa;SACd,CAAC,CAAA;QACF,IAAI,UAAU,IAAI,MAAM,GAAG,IAAI,CAAC,IAAI,CAAC,MAAM,EAAE,CAAC;YAC5C,QAAQ,IAAI,gBAAgB,CAAC,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC,CAAA;QACvD,CAAC;QACD,UAAU,GAAG,MAAM,CAAA;IACrB,CAAC;IACD,QAAQ,IAAI,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,UAAU,CAAC,CAAA;IAEvC,OAAO,EAAE,QAAQ,EAAE,aAAa,EAAE,aAAa,EAAE,CAAA;AACnD,CAAC;AAED,oEAAoE;AACpE,SAAS,eAAe,CACtB,MAA6B,EAC7B,QAAsB,EACtB,SAAiB,EACjB,WAAmB;IAEnB,MAAM,QAAQ,GAAG,QAAQ,CAAC,aAAa,CAAC,GAAG,CAAC,MAAM,CAAC,CAAA;IACnD,MAAM,YAAY,GAAG,QAAQ;QAC3B,CAAC,CAAC,EAAE,KAAK,EAAE,SAAS,GAAG,QAAQ,CAAC,KAAK,EAAE,GAAG,EAAE,SAAS,GAAG,QAAQ,CAAC,GAAG,EAAE;QACtE,CAAC,CAAC,SAAS,CAAA;IACb,MAAM,YAAY,GAAG,QAAQ,CAAC,aAAa,CAAC,GAAG,CAAC,MAAM,CAAC,CAAA;IACvD,IAAI,YAAY,KAAK,SAAS,EAAE,CAAC;QAC/B,MAAM,IAAI,KAAK,CAAC,oCAAoC,CAAC,CAAA;IACvD,CAAC;IACD,OAAO;QACL,GAAG,MAAM;QACT,WAAW;QACX,YAAY,EAAE,SAAS,GAAG,YAAY;QACtC,GAAG,CAAC,YAAY,CAAC,CAAC,CAAC,EAAE,YAAY,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;KAC1C,CAAA;AACH,CAAC;AAED,MAAM,UAAU,0BAA0B,CACxC,KAAmC,EACnC,gBAAkD;IAElD,IAAI,IAAI,GAAG,EAAE,CAAA;IACb,IAAI,WAAW,GAAG,CAAC,CAAA;IACnB,MAAM,YAAY,GAAsB,EAAE,CAAA;IAC1C,MAAM,OAAO,GAA0B,EAAE,CAAA;IACzC,MAAM,eAAe,GAAG,IAAI,GAAG,EAAyB,CAAA;IAExD,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QACzB,MAAM,aAAa,GAAG,iBAAiB,CAAC,IAAI,CAAC,CAAA;QAC7C,MAAM,UAAU,GAAG,eAAe,CAAC,IAAI,EAAE,aAAa,EAAE,gBAAgB,CAAC,CAAA;QACzE,MAAM,iBAAiB,GAAG,IAAI,GAAG,CAC/B,UAAU,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE,CAAC,CAAC,IAAI,CAAC,MAAM,EAAE,WAAW,GAAG,KAAK,CAAU,CAAC,CAC7E,CAAA;QACD,MAAM,eAAe,GAAG,sBAAsB,CAAC,IAAI,EAAE,aAAa,EAAE,UAAU,CAAC,CAAA;QAC/E,MAAM,QAAQ,GAAG,sBAAsB,CAAC,IAAI,EAAE,eAAe,EAAE,iBAAiB,CAAC,CAAA;QAEjF,MAAM,aAAa,GAAG,IAAI,CAAC,MAAM,GAAG,CAAC,IAAI,QAAQ,CAAC,QAAQ,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,CAAA;QACnF,MAAM,SAAS,GAAG,IAAI,CAAC,MAAM,GAAG,aAAa,CAAC,MAAM,CAAA;QACpD,IAAI,IAAI,aAAa,GAAG,QAAQ,CAAC,QAAQ,CAAA;QAEzC,KAAK,MAAM,gBAAgB,IAAI,UAAU,EAAE,CAAC;YAC1C,MAAM,MAAM,GAAG,eAAe,CAAC,gBAAgB,CAAC,MAAM,EAAE,QAAQ,EAAE,SAAS,EAAE,WAAW,EAAE,CAAC,CAAA;YAC3F,IAAI,MAAM,CAAC,YAAY,EAAE,CAAC;gBACxB,YAAY,CAAC,IAAI,CAAC,MAAM,CAAC,YAAY,CAAC,CAAA;YACxC,CAAC;YACD,OAAO,CAAC,IAAI,CAAC,MAAM,CAAC,CAAA;YACpB,eAAe,CAAC,GAAG,CAAC,gBAAgB,CAAC,MAAM,CAAC,CAAA;QAC9C,CAAC;IACH,CAAC;IAED,IAAI,eAAe,CAAC,IAAI,KAAK,gBAAgB,CAAC,MAAM,EAAE,CAAC;QACrD,MAAM,IAAI,KAAK,CAAC,oDAAoD,CAAC,CAAA;IACvE,CAAC;IAED,OAAO,EAAE,IAAI,EAAE,YAAY,EAAE,OAAO,EAAE,CAAA;AACxC,CAAC;AAED,SAAS,uBAAuB,CAC9B,QAA+B,EAC/B,MAA4B,EAC5B,WAAwC;IAExC,MAAM,uBAAuB,GAAG,IAAI,GAAG,EAA8B,CAAA;IACrE,IAAI,MAAM,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QACxB,OAAO,uBAAuB,CAAA;IAChC,CAAC;IACD,MAAM,mBAAmB,GAAG,IAAI,GAAG,CACjC,QAAQ,CAAC,OAAO,CAAC,GAAG,CAAC,CAAC,MAAM,EAAE,EAAE,CAAC,CAAC,MAAM,CAAC,WAAW,EAAE,MAAM,CAAU,CAAC,CACxE,CAAA;IACD,MAAM,iBAAiB,GAAG,IAAI,GAAG,EAAU,CAAA;IAE3C,KAAK,MAAM,UAAU,IAAI,CAAC,GAAG,WAAW,CAAC,CAAC,IAAI,CAC5C,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE,CAAC,IAAI,CAAC,UAAU,GAAG,KAAK,CAAC,UAAU,CACpD,EAAE,CAAC;QACF,IAAI,iBAAiB,CAAC,GAAG,CAAC,UAAU,CAAC,UAAU,CAAC,EAAE,CAAC;YACjD,MAAM,IAAI,KAAK,CAAC,sCAAsC,UAAU,CAAC,UAAU,EAAE,CAAC,CAAA;QAChF,CAAC;QACD,iBAAiB,CAAC,GAAG,CAAC,UAAU,CAAC,UAAU,CAAC,CAAA;QAC5C,MAAM,MAAM,GAAG,mBAAmB,CAAC,GAAG,CAAC,UAAU,CAAC,UAAU,CAAC,CAAA;QAC7D,IAAI,CAAC,MAAM,EAAE,CAAC;YACZ,MAAM,IAAI,KAAK,CAAC,4CAA4C,UAAU,CAAC,UAAU,EAAE,CAAC,CAAA;QACtF,CAAC;QAED,MAAM,KAAK,GAAG,gBAAgB,CAAC,MAAM,EAAE,MAAM,CAAC,YAAY,CAAC,CAAA;QAC3D,IAAI,CAAC,KAAK,EAAE,CAAC;YACX,MAAM,IAAI,KAAK,CAAC,UAAU,UAAU,CAAC,UAAU,sBAAsB,CAAC,CAAA;QACxE,CAAC;QACD,MAAM,OAAO,GAAG,uBAAuB,CAAC,GAAG,CAAC,KAAK,CAAC,KAAK,CAAC,IAAI,EAAE,CAAA;QAC9D,OAAO,CAAC,IAAI,CAAC,UAAU,CAAC,CAAA;QACxB,uBAAuB,CAAC,GAAG,CAAC,KAAK,CAAC,KAAK,EAAE,OAAO,CAAC,CAAA;IACnD,CAAC;IACD,OAAO,uBAAuB,CAAA;AAChC,CAAC;AAED,KAAK,UAAU,uBAAuB,CACpC,OAA+B,EAC/B,GAAiE;IAKjE,MAAM,QAAQ,GAAG,MAAM,MAAM,CAAC,2BAA2B,CAAC,CAAA;IAC1D,MAAM,SAAS,GAA4B,EAAE,CAAA;IAC7C,IAAI,iBAAiB,GAAG,CAAC,CAAA;IACzB,KAAK,MAAM,MAAM,IAAI,OAAO,EAAE,CAAC;QAC7B,IAAI,CAAC;YACH,MAAM,KAAK,GAAG,MAAM,QAAQ,CAAC,kBAAkB,CAC7C,GAAG,EACH,MAAM,CAAC,OAAO,EACd,MAAM,CAAC,IAAI,EACX,MAAM,CAAC,QAAQ,CAChB,CAAA;YACD,SAAS,CAAC,IAAI,CAAC,EAAE,GAAG,MAAM,EAAE,OAAO,EAAE,IAAI,EAAE,SAAS,EAAE,KAAK,EAAE,CAAC,CAAA;QAChE,CAAC;QAAC,MAAM,CAAC;YACP,iBAAiB,IAAI,CAAC,CAAA;YACtB,SAAS,CAAC,IAAI,CAAC,EAAE,GAAG,MAAM,EAAE,OAAO,EAAE,IAAI,EAAE,CAAC,CAAA;QAC9C,CAAC;IACH,CAAC;IACD,OAAO,EAAE,SAAS,EAAE,iBAAiB,EAAE,CAAA;AACzC,CAAC;AAED;;;;GAIG;AACH,KAAK,UAAU,uBAAuB,CACpC,KAAqE,EACrE,GAAiE,EACjE,eAA4C,EAC5C,aAAsB;IAEtB,MAAM,QAAQ,GAAG,KAAK,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC,EAAE,OAAO,EAAE,IAAI,CAAC,OAAO,EAAE,SAAS,EAAE,IAAI,CAAC,SAAS,EAAE,CAAC,CAAC,CAAA;IAC5F,IAAI,eAAe,KAAK,SAAS,EAAE,CAAC;QAClC,MAAM,QAAQ,GAAG,MAAM,MAAM,CAAC,2BAA2B,CAAC,CAAA;QAC1D,OAAO,uBAAuB,CAAC,QAAQ,CAAC,mBAAmB,CAAC,QAAQ,EAAE,GAAG,CAAC,EAAE,GAAG,CAAC,CAAA;IAClF,CAAC;IACD,MAAM,SAAS,GAAG,MAAM,MAAM,CAAC,wBAAwB,CAAC,CAAA;IACxD,MAAM,OAAO,GAAG,SAAS,CAAC,mBAAmB,CAAC,QAAQ,EAAE,GAAG,CAAC,CAAA;IAC5D,MAAM,SAAS,GAAG,SAAS,CAAC,eAAe,CAAC,eAAe,CAAC,CAAA;IAC5D,IAAI,CAAC;QACH,MAAM,SAAS,GAAG,MAAM,SAAS,CAAC,oBAAoB,CAAC,OAAO,EAAE,GAAG,EAAE;YACnE,SAAS;YACT,aAAa;SACd,CAAC,CAAA;QACF,OAAO,EAAE,SAAS,EAAE,iBAAiB,EAAE,CAAC,EAAE,CAAA;IAC5C,CAAC;YAAS,CAAC;QACT,MAAM,SAAS,CAAC,OAAO,EAAE,CAAA;IAC3B,CAAC;AACH,CAAC;AAED,sFAAsF;AACtF,SAAS,kBAAkB,CAAC,OAAuC;IAIjE,MAAM,WAAW,GAAuB,EAAE,CAAA;IAC1C,IAAI,iBAAiB,GAAG,CAAC,CAAA;IACzB,KAAK,MAAM,MAAM,IAAI,OAAO,EAAE,CAAC;QAC7B,IAAI,CAAC,MAAM,CAAC,SAAS,EAAE,CAAC;YACtB,SAAQ;QACV,CAAC;QACD,IAAI,CAAC;YACH,WAAW,CAAC,IAAI,CAAC,sBAAsB,CAAC,MAAM,CAAC,WAAW,EAAE,MAAM,CAAC,SAAS,CAAC,CAAC,CAAA;QAChF,CAAC;QAAC,MAAM,CAAC;YACP,iBAAiB,IAAI,CAAC,CAAA;QACxB,CAAC;IACH,CAAC;IACD,OAAO,EAAE,WAAW,EAAE,iBAAiB,EAAE,CAAA;AAC3C,CAAC;AAED,MAAM,CAAC,KAAK,UAAU,sBAAsB,CAC1C,QAAgB,EAChB,aAAwC,EACxC,OAAsC;IAEtC,MAAM,EAAE,MAAM,EAAE,OAAO,EAAE,QAAQ,EAAE,GAAG,aAAa,CAAA;IACnD,MAAM,eAAe,GAAG,OAAO,CAAC,SAAS,CAAA;IACzC,MAAM,EAAE,GAAG,EAAE,KAAK,EAAE,KAAK,EAAE,GAAG,MAAM,MAAM,CAAC,aAAa,CAAC,QAAQ,EAAE,QAAQ,CAAC,CAAA;IAC5E,IAAI,CAAC;QACH,IAAI,iBAAiB,GAAG,CAAC,CAAA;QAEzB,MAAM,QAAQ,GAAG,MAAM,uBAAuB,CAAC,KAAK,EAAE,GAAG,EAAE,eAAe,EAAE,OAAO,CAAC,MAAM,CAAC,CAAA;QAC3F,MAAM,SAAS,GAAG,QAAQ,CAAC,SAAS,CAAA;QACpC,iBAAiB,IAAI,QAAQ,CAAC,iBAAiB,CAAA;QAE/C,MAAM,OAAO,GAAG,0BAA0B,CAAC,KAAK,EAAE,SAAS,CAAC,CAAA;QAC5D,MAAM,EAAE,MAAM,EAAE,UAAU,EAAE,GAAG,MAAM,wBAAwB,CAC3D,OAAO,CAAC,IAAI,EACZ,OAAO,EACP,QAAQ,EACR,OAAO,CAAC,YAAY,CACrB,CAAA;QACD,MAAM,SAAS,GAAG,OAAO,CAAC,MAAM;YAC9B,CAAC,CAAC,kBAAkB,CAAC,OAAO,CAAC,OAAO,CAAC;YACrC,CAAC,CAAC,EAAE,WAAW,EAAE,EAAE,EAAE,iBAAiB,EAAE,CAAC,EAAE,CAAA;QAC7C,MAAM,WAAW,GAAG,SAAS,CAAC,WAAW,CAAA;QACzC,iBAAiB,IAAI,SAAS,CAAC,iBAAiB,CAAA;QAChD,MAAM,iBAAiB,GAAG,uBAAuB,CAAC,OAAO,EAAE,MAAM,EAAE,WAAW,CAAC,CAAA;QAC/E,OAAO;YACL,MAAM;YACN,UAAU;YACV,KAAK;YACL,IAAI,EAAE,OAAO,CAAC,IAAI;YAClB,YAAY,EAAE,OAAO,CAAC,YAAY;YAClC,iBAAiB;YACjB,iBAAiB;SAClB,CAAA;IACH,CAAC;YAAS,CAAC;QACT,IAAI,CAAC;YACH,GAAG,CAAC,OAAO,EAAE,CAAA;QACf,CAAC;QAAC,OAAO,YAAY,EAAE,CAAC;YACtB,MAAM,OAAO,GAAG,YAAY,YAAY,KAAK,CAAC,CAAC,CAAC,YAAY,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,YAAY,CAAC,CAAA;YAC3F,OAAO,CAAC,IAAI,CAAC,iDAAiD,OAAO,EAAE,CAAC,CAAA;QAC1E,CAAC;IACH,CAAC;AACH,CAAC"}
|
|
1
|
+
{"version":3,"file":"visual.js","sourceRoot":"","sources":["../../src/ingest/visual.ts"],"names":[],"mappings":"AAQA,OAAO,EAAE,wBAAwB,EAAE,sBAAsB,EAAE,gBAAgB,EAAE,MAAM,cAAc,CAAA;AAsDjG,SAAS,iBAAiB,CAAC,IAAuB,EAAE,KAAwB;IAC1E,OAAO,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC,EAAE,KAAK,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,IAAI,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC,EAAE,KAAK,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAA;AACvF,CAAC;AAED,SAAS,OAAO,CAAC,IAAuB;IACtC,OAAO,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,GAAG,CAAC,CAAA;AAC9C,CAAC;AAED,SAAS,OAAO,CAAC,IAAuB;IACtC,OAAO,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,GAAG,CAAC,CAAA;AAC9C,CAAC;AAED,SAAS,kBAAkB,CACzB,SAA0C,EAC1C,MAA4B;IAE5B,IAAI,SAAS,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QAC3B,OAAO,CAAC,CAAA;IACV,CAAC;IACD,IAAI,UAAU,GAAG,SAAS;SACvB,GAAG,CAAC,CAAC,QAAQ,EAAE,KAAK,EAAE,EAAE,CAAC,CAAC,EAAE,QAAQ,EAAE,KAAK,EAAE,CAAC,CAAC;SAC/C,MAAM,CAAC,CAAC,EAAE,QAAQ,EAAE,EAAE,EAAE,CAAC,iBAAiB,CAAC,QAAQ,CAAC,IAAI,EAAE,MAAM,CAAC,IAAI,CAAC,CAAC,CAAA;IAE1E,IAAI,UAAU,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QAC5B,MAAM,YAAY,GAAG,OAAO,CAAC,MAAM,CAAC,IAAI,CAAC,CAAA;QACzC,MAAM,eAAe,GAAG,IAAI,CAAC,GAAG,CAC9B,GAAG,SAAS,CAAC,GAAG,CAAC,CAAC,QAAQ,EAAE,EAAE,CAAC,IAAI,CAAC,GAAG,CAAC,OAAO,CAAC,QAAQ,CAAC,IAAI,CAAC,GAAG,YAAY,CAAC,CAAC,CAChF,CAAA;QACD,UAAU,GAAG,SAAS;aACnB,GAAG,CAAC,CAAC,QAAQ,EAAE,KAAK,EAAE,EAAE,CAAC,CAAC,EAAE,QAAQ,EAAE,KAAK,EAAE,CAAC,CAAC;aAC/C,MAAM,CAAC,CAAC,EAAE,QAAQ,EAAE,EAAE,EAAE,CAAC,IAAI,CAAC,GAAG,CAAC,OAAO,CAAC,QAAQ,CAAC,IAAI,CAAC,GAAG,YAAY,CAAC,KAAK,eAAe,CAAC,CAAA;IAClG,CAAC;IAED,MAAM,cAAc,GAAG,OAAO,CAAC,MAAM,CAAC,IAAI,CAAC,CAAA;IAC3C,MAAM,SAAS,GAAG,UAAU,CAAC,MAAM,CAAC,CAAC,EAAE,QAAQ,EAAE,EAAE,EAAE,CAAC,OAAO,CAAC,QAAQ,CAAC,IAAI,CAAC,IAAI,cAAc,CAAC,CAAA;IAC/F,IAAI,SAAS,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QACzB,OAAO,IAAI,CAAC,GAAG,CAAC,GAAG,SAAS,CAAC,GAAG,CAAC,CAAC,EAAE,KAAK,EAAE,EAAE,EAAE,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC,CAAA;IAC7D,CAAC;IACD,OAAO,IAAI,CAAC,GAAG,CAAC,GAAG,UAAU,CAAC,GAAG,CAAC,CAAC,EAAE,KAAK,EAAE,EAAE,EAAE,CAAC,KAAK,CAAC,CAAC,CAAA;AAC1D,CAAC;AAED,SAAS,iBAAiB,CAAC,IAAuB;IAChD,IAAI,IAAI,CAAC,aAAa,CAAC,MAAM,GAAG,CAAC,IAAI,IAAI,CAAC,IAAI,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QAC5D,OAAO,IAAI,CAAC,aAAa,CAAA;IAC3B,CAAC;IACD,OAAO;QACL;YACE,OAAO,EAAE,IAAI,CAAC,OAAO;YACrB,YAAY,EAAE,CAAC;YACf,WAAW,EAAE,CAAC;YACd,eAAe,EAAE,CAAC;YAClB,IAAI,EAAE,CAAC,CAAC,EAAE,CAAC,EAAE,CAAC,EAAE,CAAC,CAAC;YAClB,IAAI,EAAE,IAAI,CAAC,IAAI;YACf,aAAa,EAAE,CAAC;YAChB,WAAW,EAAE,IAAI,CAAC,IAAI,CAAC,MAAM;SAC9B;KACF,CAAA;AACH,CAAC;AAQD,2EAA2E;AAC3E,SAAS,eAAe,CACtB,IAAuB,EACvB,aAA8C,EAC9C,gBAAkD;IAElD,OAAO,gBAAgB;SACpB,MAAM,CAAC,CAAC,MAAM,EAAE,EAAE,CAAC,MAAM,CAAC,OAAO,KAAK,IAAI,CAAC,OAAO,CAAC;SACnD,GAAG,CAAC,CAAC,MAAM,EAAE,EAAE,CAAC,CAAC,EAAE,MAAM,EAAE,cAAc,EAAE,kBAAkB,CAAC,aAAa,EAAE,MAAM,CAAC,EAAE,CAAC,CAAC;SACxF,IAAI,CACH,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE,CACd,IAAI,CAAC,cAAc,GAAG,KAAK,CAAC,cAAc;QAC1C,IAAI,CAAC,MAAM,CAAC,cAAc,GAAG,KAAK,CAAC,MAAM,CAAC,cAAc,CAC3D,CAAA;AACL,CAAC;AAED;;;GAGG;AACH,SAAS,sBAAsB,CAC7B,IAAuB,EACvB,aAA8C,EAC9C,UAAuC;IAEvC,MAAM,MAAM,GAAG,IAAI,GAAG,EAAmC,CAAA;IACzD,KAAK,MAAM,IAAI,IAAI,UAAU,EAAE,CAAC;QAC9B,MAAM,SAAS,GAAG,aAAa,CAAC,IAAI,CAAC,cAAc,GAAG,CAAC,CAAC,CAAA;QACxD,MAAM,SAAS,GAAG,aAAa,CAAC,IAAI,CAAC,cAAc,CAAC,CAAA;QACpD,MAAM,MAAM,GAAG,IAAI,CAAC,GAAG,CACrB,CAAC,EACD,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,IAAI,CAAC,MAAM,EAAE,SAAS,EAAE,WAAW,IAAI,SAAS,EAAE,aAAa,IAAI,CAAC,CAAC,CACpF,CAAA;QACD,MAAM,KAAK,GAAG,MAAM,CAAC,GAAG,CAAC,MAAM,CAAC,IAAI,EAAE,CAAA;QACtC,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,MAAM,CAAC,CAAA;QACvB,MAAM,CAAC,GAAG,CAAC,MAAM,EAAE,KAAK,CAAC,CAAA;IAC3B,CAAC;IACD,OAAO,MAAM,CAAA;AACf,CAAC;AAUD,kEAAkE;AAClE,SAAS,gBAAgB,CAAC,QAAgB;IACxC,MAAM,OAAO,GAAG,QAAQ,CAAC,KAAK,CAAC,MAAM,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,MAAM,IAAI,CAAC,CAAA;IACvD,OAAO,IAAI,CAAC,MAAM,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,CAAC,GAAG,OAAO,CAAC,CAAC,CAAA;AAC9C,CAAC;AAED;;;GAGG;AACH,SAAS,mBAAmB,CAC1B,QAAgB,EAChB,KAAuC,EACvC,OAKC;IAED,MAAM,EAAE,IAAI,EAAE,iBAAiB,EAAE,GAAG,IAAI,EAAE,GAAG,OAAO,CAAA;IACpD,IAAI,IAAI,GAAG,QAAQ,CAAA;IACnB,IAAI,YAAY,GAAG,CAAC,CAAA;IACpB,KAAK,MAAM,MAAM,IAAI,KAAK,EAAE,CAAC;QAC3B,IAAI,MAAM,CAAC,OAAO,KAAK,IAAI,EAAE,CAAC;YAC5B,IAAI,CAAC,aAAa,CAAC,GAAG,CAAC,MAAM,EAAE,IAAI,CAAC,MAAM,CAAC,CAAA;YAC3C,SAAQ;QACV,CAAC;QACD,IAAI,YAAY,GAAG,CAAC,EAAE,CAAC;YACrB,IAAI,IAAI,MAAM,CAAA;QAChB,CAAC;QACD,MAAM,KAAK,GAAG,IAAI,CAAC,MAAM,CAAA;QACzB,IAAI,IAAI,2BAA2B,IAAI,CAAC,OAAO,YAAY,iBAAiB,CAAC,GAAG,CAAC,MAAM,CAAC,KAAK,MAAM,CAAC,OAAO,GAAG,CAAA;QAC9G,IAAI,CAAC,aAAa,CAAC,GAAG,CAAC,MAAM,EAAE,EAAE,KAAK,EAAE,GAAG,EAAE,IAAI,CAAC,MAAM,EAAE,CAAC,CAAA;QAC3D,IAAI,CAAC,aAAa,CAAC,GAAG,CAAC,MAAM,EAAE,KAAK,CAAC,CAAA;QACrC,YAAY,IAAI,CAAC,CAAA;IACnB,CAAC;IACD,OAAO,IAAI,CAAA;AACb,CAAC;AAED,SAAS,sBAAsB,CAC7B,IAAuB,EACvB,eAAqD,EACrD,iBAAqD;IAErD,IAAI,QAAQ,GAAG,EAAE,CAAA;IACjB,IAAI,UAAU,GAAG,CAAC,CAAA;IAClB,MAAM,MAAM,GAAwC,EAAE,CAAA;IACtD,MAAM,aAAa,GAAG,IAAI,GAAG,EAA0C,CAAA;IACvE,MAAM,aAAa,GAAG,IAAI,GAAG,EAAiC,CAAA;IAE9D,KAAK,MAAM,CAAC,MAAM,EAAE,KAAK,CAAC,IAAI,CAAC,GAAG,eAAe,CAAC,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE,CAAC,IAAI,CAAC,CAAC,CAAC,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC;QAC7F,QAAQ,IAAI,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,UAAU,EAAE,MAAM,CAAC,CAAA;QAC/C,MAAM,UAAU,GAAG,KAAK,CAAC,IAAI,CAAC,CAAC,MAAM,EAAE,EAAE,CAAC,MAAM,CAAC,OAAO,KAAK,IAAI,CAAC,CAAA;QAClE,IAAI,UAAU,IAAI,QAAQ,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;YACtC,QAAQ,IAAI,gBAAgB,CAAC,QAAQ,CAAC,KAAK,CAAC,MAAM,CAAC,EAAE,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAA;QACjE,CAAC;QACD,QAAQ,GAAG,mBAAmB,CAAC,QAAQ,EAAE,KAAK,EAAE;YAC9C,IAAI;YACJ,iBAAiB;YACjB,aAAa;YACb,aAAa;SACd,CAAC,CAAA;QACF,IAAI,UAAU,IAAI,MAAM,GAAG,IAAI,CAAC,IAAI,CAAC,MAAM,EAAE,CAAC;YAC5C,QAAQ,IAAI,gBAAgB,CAAC,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC,CAAA;QACvD,CAAC;QACD,MAAM,CAAC,IAAI,CAAC,EAAE,MAAM,EAAE,KAAK,EAAE,QAAQ,CAAC,MAAM,GAAG,MAAM,EAAE,CAAC,CAAA;QACxD,UAAU,GAAG,MAAM,CAAA;IACrB,CAAC;IACD,QAAQ,IAAI,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,UAAU,CAAC,CAAA;IAEvC,MAAM,QAAQ,GAAG,CAAC,IAAI,CAAC,QAAQ,IAAI,EAAE,CAAC,CAAC,GAAG,CAAC,CAAC,OAAO,EAAE,EAAE,CAAC,CAAC;QACvD,GAAG,OAAO;QACV,MAAM,EACJ,OAAO,CAAC,MAAM;YACd,CAAC,MAAM,CAAC,MAAM,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,MAAM,IAAI,OAAO,CAAC,MAAM,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,EAAE,KAAK,IAAI,CAAC,CAAC;KAChF,CAAC,CAAC,CAAA;IACH,OAAO,EAAE,QAAQ,EAAE,aAAa,EAAE,aAAa,EAAE,QAAQ,EAAE,CAAA;AAC7D,CAAC;AAED,oEAAoE;AACpE,SAAS,eAAe,CACtB,MAA6B,EAC7B,QAAsB,EACtB,SAAiB,EACjB,WAAmB;IAEnB,MAAM,QAAQ,GAAG,QAAQ,CAAC,aAAa,CAAC,GAAG,CAAC,MAAM,CAAC,CAAA;IACnD,MAAM,YAAY,GAAG,QAAQ;QAC3B,CAAC,CAAC,EAAE,KAAK,EAAE,SAAS,GAAG,QAAQ,CAAC,KAAK,EAAE,GAAG,EAAE,SAAS,GAAG,QAAQ,CAAC,GAAG,EAAE;QACtE,CAAC,CAAC,SAAS,CAAA;IACb,MAAM,YAAY,GAAG,QAAQ,CAAC,aAAa,CAAC,GAAG,CAAC,MAAM,CAAC,CAAA;IACvD,IAAI,YAAY,KAAK,SAAS,EAAE,CAAC;QAC/B,MAAM,IAAI,KAAK,CAAC,oCAAoC,CAAC,CAAA;IACvD,CAAC;IACD,OAAO;QACL,GAAG,MAAM;QACT,WAAW;QACX,YAAY,EAAE,SAAS,GAAG,YAAY;QACtC,GAAG,CAAC,YAAY,CAAC,CAAC,CAAC,EAAE,YAAY,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;KAC1C,CAAA;AACH,CAAC;AAED,MAAM,UAAU,0BAA0B,CACxC,KAAmC,EACnC,gBAAkD;IAElD,IAAI,IAAI,GAAG,EAAE,CAAA;IACb,MAAM,SAAS,GAAc,EAAE,QAAQ,EAAE,EAAE,EAAE,KAAK,EAAE,EAAE,EAAE,CAAA;IACxD,IAAI,WAAW,GAAG,CAAC,CAAA;IACnB,MAAM,YAAY,GAAsB,EAAE,CAAA;IAC1C,MAAM,OAAO,GAA0B,EAAE,CAAA;IACzC,MAAM,eAAe,GAAG,IAAI,GAAG,EAAyB,CAAA;IAExD,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QACzB,MAAM,aAAa,GAAG,iBAAiB,CAAC,IAAI,CAAC,CAAA;QAC7C,MAAM,UAAU,GAAG,eAAe,CAAC,IAAI,EAAE,aAAa,EAAE,gBAAgB,CAAC,CAAA;QACzE,MAAM,iBAAiB,GAAG,IAAI,GAAG,CAC/B,UAAU,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE,CAAC,CAAC,IAAI,CAAC,MAAM,EAAE,WAAW,GAAG,KAAK,CAAU,CAAC,CAC7E,CAAA;QACD,MAAM,eAAe,GAAG,sBAAsB,CAAC,IAAI,EAAE,aAAa,EAAE,UAAU,CAAC,CAAA;QAC/E,MAAM,QAAQ,GAAG,sBAAsB,CAAC,IAAI,EAAE,eAAe,EAAE,iBAAiB,CAAC,CAAA;QAEjF,MAAM,aAAa,GAAG,IAAI,CAAC,MAAM,GAAG,CAAC,IAAI,QAAQ,CAAC,QAAQ,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,CAAA;QACnF,MAAM,SAAS,GAAG,IAAI,CAAC,MAAM,GAAG,aAAa,CAAC,MAAM,CAAA;QACpD,IAAI,IAAI,aAAa,GAAG,QAAQ,CAAC,QAAQ,CAAA;QACzC,IAAI,QAAQ,CAAC,QAAQ,EAAE,CAAC;YACtB,SAAS,CAAC,KAAK,EAAE,IAAI,CAAC,EAAE,KAAK,EAAE,SAAS,EAAE,GAAG,EAAE,IAAI,CAAC,MAAM,EAAE,IAAI,EAAE,IAAI,CAAC,OAAO,EAAE,CAAC,CAAA;QACnF,CAAC;QACD,SAAS,CAAC,QAAQ,CAAC,IAAI,CACrB,GAAG,QAAQ,CAAC,QAAQ,CAAC,GAAG,CAAC,CAAC,OAAO,EAAE,EAAE,CAAC,CAAC,EAAE,GAAG,OAAO,EAAE,MAAM,EAAE,SAAS,GAAG,OAAO,CAAC,MAAM,EAAE,CAAC,CAAC,CAC5F,CAAA;QAED,KAAK,MAAM,gBAAgB,IAAI,UAAU,EAAE,CAAC;YAC1C,MAAM,MAAM,GAAG,eAAe,CAAC,gBAAgB,CAAC,MAAM,EAAE,QAAQ,EAAE,SAAS,EAAE,WAAW,EAAE,CAAC,CAAA;YAC3F,IAAI,MAAM,CAAC,YAAY,EAAE,CAAC;gBACxB,YAAY,CAAC,IAAI,CAAC,MAAM,CAAC,YAAY,CAAC,CAAA;YACxC,CAAC;YACD,OAAO,CAAC,IAAI,CAAC,MAAM,CAAC,CAAA;YACpB,eAAe,CAAC,GAAG,CAAC,gBAAgB,CAAC,MAAM,CAAC,CAAA;QAC9C,CAAC;IACH,CAAC;IAED,IAAI,eAAe,CAAC,IAAI,KAAK,gBAAgB,CAAC,MAAM,EAAE,CAAC;QACrD,MAAM,IAAI,KAAK,CAAC,oDAAoD,CAAC,CAAA;IACvE,CAAC;IAED,OAAO,EAAE,IAAI,EAAE,YAAY,EAAE,OAAO,EAAE,SAAS,EAAE,CAAA;AACnD,CAAC;AAED,SAAS,uBAAuB,CAC9B,QAA+B,EAC/B,MAA4B,EAC5B,WAAwC;IAExC,MAAM,uBAAuB,GAAG,IAAI,GAAG,EAA8B,CAAA;IACrE,IAAI,MAAM,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QACxB,OAAO,uBAAuB,CAAA;IAChC,CAAC;IACD,MAAM,mBAAmB,GAAG,IAAI,GAAG,CACjC,QAAQ,CAAC,OAAO,CAAC,GAAG,CAAC,CAAC,MAAM,EAAE,EAAE,CAAC,CAAC,MAAM,CAAC,WAAW,EAAE,MAAM,CAAU,CAAC,CACxE,CAAA;IACD,MAAM,iBAAiB,GAAG,IAAI,GAAG,EAAU,CAAA;IAE3C,KAAK,MAAM,UAAU,IAAI,CAAC,GAAG,WAAW,CAAC,CAAC,IAAI,CAC5C,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE,CAAC,IAAI,CAAC,UAAU,GAAG,KAAK,CAAC,UAAU,CACpD,EAAE,CAAC;QACF,IAAI,iBAAiB,CAAC,GAAG,CAAC,UAAU,CAAC,UAAU,CAAC,EAAE,CAAC;YACjD,MAAM,IAAI,KAAK,CAAC,sCAAsC,UAAU,CAAC,UAAU,EAAE,CAAC,CAAA;QAChF,CAAC;QACD,iBAAiB,CAAC,GAAG,CAAC,UAAU,CAAC,UAAU,CAAC,CAAA;QAC5C,MAAM,MAAM,GAAG,mBAAmB,CAAC,GAAG,CAAC,UAAU,CAAC,UAAU,CAAC,CAAA;QAC7D,IAAI,CAAC,MAAM,EAAE,CAAC;YACZ,MAAM,IAAI,KAAK,CAAC,4CAA4C,UAAU,CAAC,UAAU,EAAE,CAAC,CAAA;QACtF,CAAC;QAED,MAAM,KAAK,GAAG,gBAAgB,CAAC,MAAM,EAAE,MAAM,CAAC,YAAY,CAAC,CAAA;QAC3D,IAAI,CAAC,KAAK,EAAE,CAAC;YACX,MAAM,IAAI,KAAK,CAAC,UAAU,UAAU,CAAC,UAAU,sBAAsB,CAAC,CAAA;QACxE,CAAC;QACD,MAAM,OAAO,GAAG,uBAAuB,CAAC,GAAG,CAAC,KAAK,CAAC,KAAK,CAAC,IAAI,EAAE,CAAA;QAC9D,OAAO,CAAC,IAAI,CAAC,UAAU,CAAC,CAAA;QACxB,uBAAuB,CAAC,GAAG,CAAC,KAAK,CAAC,KAAK,EAAE,OAAO,CAAC,CAAA;IACnD,CAAC;IACD,OAAO,uBAAuB,CAAA;AAChC,CAAC;AAED,KAAK,UAAU,uBAAuB,CACpC,OAA+B,EAC/B,GAAiE;IAKjE,MAAM,QAAQ,GAAG,MAAM,MAAM,CAAC,2BAA2B,CAAC,CAAA;IAC1D,MAAM,SAAS,GAA4B,EAAE,CAAA;IAC7C,IAAI,iBAAiB,GAAG,CAAC,CAAA;IACzB,KAAK,MAAM,MAAM,IAAI,OAAO,EAAE,CAAC;QAC7B,IAAI,CAAC;YACH,MAAM,KAAK,GAAG,MAAM,QAAQ,CAAC,kBAAkB,CAC7C,GAAG,EACH,MAAM,CAAC,OAAO,EACd,MAAM,CAAC,IAAI,EACX,MAAM,CAAC,QAAQ,CAChB,CAAA;YACD,SAAS,CAAC,IAAI,CAAC,EAAE,GAAG,MAAM,EAAE,OAAO,EAAE,IAAI,EAAE,SAAS,EAAE,KAAK,EAAE,CAAC,CAAA;QAChE,CAAC;QAAC,MAAM,CAAC;YACP,iBAAiB,IAAI,CAAC,CAAA;YACtB,SAAS,CAAC,IAAI,CAAC,EAAE,GAAG,MAAM,EAAE,OAAO,EAAE,IAAI,EAAE,CAAC,CAAA;QAC9C,CAAC;IACH,CAAC;IACD,OAAO,EAAE,SAAS,EAAE,iBAAiB,EAAE,CAAA;AACzC,CAAC;AAED;;;;GAIG;AACH,KAAK,UAAU,uBAAuB,CACpC,KAAqE,EACrE,GAAiE,EACjE,eAA4C,EAC5C,aAAsB;IAEtB,MAAM,QAAQ,GAAG,KAAK,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC,EAAE,OAAO,EAAE,IAAI,CAAC,OAAO,EAAE,SAAS,EAAE,IAAI,CAAC,SAAS,EAAE,CAAC,CAAC,CAAA;IAC5F,IAAI,eAAe,KAAK,SAAS,EAAE,CAAC;QAClC,MAAM,QAAQ,GAAG,MAAM,MAAM,CAAC,2BAA2B,CAAC,CAAA;QAC1D,OAAO,uBAAuB,CAAC,QAAQ,CAAC,mBAAmB,CAAC,QAAQ,EAAE,GAAG,CAAC,EAAE,GAAG,CAAC,CAAA;IAClF,CAAC;IACD,MAAM,SAAS,GAAG,MAAM,MAAM,CAAC,wBAAwB,CAAC,CAAA;IACxD,MAAM,OAAO,GAAG,SAAS,CAAC,mBAAmB,CAAC,QAAQ,EAAE,GAAG,CAAC,CAAA;IAC5D,MAAM,SAAS,GAAG,SAAS,CAAC,eAAe,CAAC,eAAe,CAAC,CAAA;IAC5D,IAAI,CAAC;QACH,MAAM,SAAS,GAAG,MAAM,SAAS,CAAC,oBAAoB,CAAC,OAAO,EAAE,GAAG,EAAE;YACnE,SAAS;YACT,aAAa;SACd,CAAC,CAAA;QACF,OAAO,EAAE,SAAS,EAAE,iBAAiB,EAAE,CAAC,EAAE,CAAA;IAC5C,CAAC;YAAS,CAAC;QACT,MAAM,SAAS,CAAC,OAAO,EAAE,CAAA;IAC3B,CAAC;AACH,CAAC;AAED,sFAAsF;AACtF,SAAS,kBAAkB,CAAC,OAAuC;IAIjE,MAAM,WAAW,GAAuB,EAAE,CAAA;IAC1C,IAAI,iBAAiB,GAAG,CAAC,CAAA;IACzB,KAAK,MAAM,MAAM,IAAI,OAAO,EAAE,CAAC;QAC7B,IAAI,CAAC,MAAM,CAAC,SAAS,EAAE,CAAC;YACtB,SAAQ;QACV,CAAC;QACD,IAAI,CAAC;YACH,WAAW,CAAC,IAAI,CAAC,sBAAsB,CAAC,MAAM,CAAC,WAAW,EAAE,MAAM,CAAC,SAAS,CAAC,CAAC,CAAA;QAChF,CAAC;QAAC,MAAM,CAAC;YACP,iBAAiB,IAAI,CAAC,CAAA;QACxB,CAAC;IACH,CAAC;IACD,OAAO,EAAE,WAAW,EAAE,iBAAiB,EAAE,CAAA;AAC3C,CAAC;AAED,MAAM,CAAC,KAAK,UAAU,sBAAsB,CAC1C,QAAgB,EAChB,aAAwC,EACxC,OAAsC;IAEtC,MAAM,EAAE,MAAM,EAAE,OAAO,EAAE,QAAQ,EAAE,GAAG,aAAa,CAAA;IACnD,MAAM,eAAe,GAAG,OAAO,CAAC,SAAS,CAAA;IACzC,MAAM,EAAE,GAAG,EAAE,KAAK,EAAE,cAAc,EAAE,KAAK,EAAE,GAAG,MAAM,MAAM,CAAC,aAAa,CAAC,QAAQ,EAAE,QAAQ,CAAC,CAAA;IAC5F,IAAI,CAAC;QACH,IAAI,iBAAiB,GAAG,CAAC,CAAA;QAEzB,MAAM,QAAQ,GAAG,MAAM,uBAAuB,CAAC,KAAK,EAAE,GAAG,EAAE,eAAe,EAAE,OAAO,CAAC,MAAM,CAAC,CAAA;QAC3F,MAAM,SAAS,GAAG,QAAQ,CAAC,SAAS,CAAA;QACpC,iBAAiB,IAAI,QAAQ,CAAC,iBAAiB,CAAA;QAE/C,MAAM,OAAO,GAAG,0BAA0B,CAAC,KAAK,EAAE,SAAS,CAAC,CAAA;QAC5D,MAAM,EAAE,MAAM,EAAE,UAAU,EAAE,GAAG,MAAM,wBAAwB,CAAC,OAAO,CAAC,IAAI,EAAE,OAAO,EAAE,QAAQ,EAAE;YAC7F,SAAS,EAAE,OAAO,CAAC,SAAS;YAC5B,YAAY,EAAE,OAAO,CAAC,YAAY;YAClC,KAAK,EAAE,cAAc;SACtB,CAAC,CAAA;QACF,MAAM,SAAS,GAAG,OAAO,CAAC,MAAM;YAC9B,CAAC,CAAC,kBAAkB,CAAC,OAAO,CAAC,OAAO,CAAC;YACrC,CAAC,CAAC,EAAE,WAAW,EAAE,EAAE,EAAE,iBAAiB,EAAE,CAAC,EAAE,CAAA;QAC7C,MAAM,WAAW,GAAG,SAAS,CAAC,WAAW,CAAA;QACzC,iBAAiB,IAAI,SAAS,CAAC,iBAAiB,CAAA;QAChD,MAAM,iBAAiB,GAAG,uBAAuB,CAAC,OAAO,EAAE,MAAM,EAAE,WAAW,CAAC,CAAA;QAC/E,OAAO;YACL,MAAM;YACN,UAAU;YACV,KAAK;YACL,IAAI,EAAE,OAAO,CAAC,IAAI;YAClB,YAAY,EAAE,OAAO,CAAC,YAAY;YAClC,iBAAiB;YACjB,iBAAiB;SAClB,CAAA;IACH,CAAC;YAAS,CAAC;QACT,IAAI,CAAC;YACH,GAAG,CAAC,OAAO,EAAE,CAAA;QACf,CAAC;QAAC,OAAO,YAAY,EAAE,CAAC;YACtB,MAAM,OAAO,GAAG,YAAY,YAAY,KAAK,CAAC,CAAC,CAAC,YAAY,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,YAAY,CAAC,CAAA;YAC3F,OAAO,CAAC,IAAI,CAAC,iDAAiD,OAAO,EAAE,CAAC,CAAA;QAC1E,CAAC;IACH,CAAC;AACH,CAAC"}
|
|
@@ -1,5 +1,7 @@
|
|
|
1
1
|
import type { AtomicTextRange } from '../chunker/index.js';
|
|
2
|
+
import type { SourceMap } from '../utils/source-context.js';
|
|
2
3
|
export interface DocxBodyResult {
|
|
4
|
+
sourceMap?: SourceMap;
|
|
3
5
|
content: string;
|
|
4
6
|
atomicRanges: readonly AtomicTextRange[];
|
|
5
7
|
imageAnchors?: readonly {
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"docx-parser.d.ts","sourceRoot":"","sources":["../../src/parser/docx-parser.ts"],"names":[],"mappings":"AAEA,OAAO,KAAK,EAAE,eAAe,EAAE,MAAM,qBAAqB,CAAA;
|
|
1
|
+
{"version":3,"file":"docx-parser.d.ts","sourceRoot":"","sources":["../../src/parser/docx-parser.ts"],"names":[],"mappings":"AAEA,OAAO,KAAK,EAAE,eAAe,EAAE,MAAM,qBAAqB,CAAA;AAC1D,OAAO,KAAK,EAAiB,SAAS,EAAE,MAAM,4BAA4B,CAAA;AAU1E,MAAM,WAAW,cAAc;IAC7B,SAAS,CAAC,EAAE,SAAS,CAAA;IACrB,OAAO,EAAE,MAAM,CAAA;IACf,YAAY,EAAE,SAAS,eAAe,EAAE,CAAA;IACxC,YAAY,CAAC,EAAE,SAAS;QAAE,MAAM,EAAE,MAAM,CAAC;QAAC,UAAU,EAAE,MAAM,CAAA;KAAE,EAAE,CAAA;CACjE;AAkWD,wBAAsB,oBAAoB,CAAC,MAAM,EAAE,MAAM,GAAG,OAAO,CAAC,MAAM,GAAG,SAAS,CAAC,CAsBtF;AAsCD,wBAAgB,yBAAyB,CAAC,QAAQ,EAAE,QAAQ,GAAG,cAAc,CAwC5E"}
|
|
@@ -256,9 +256,13 @@ function blockSerializerFor(element) {
|
|
|
256
256
|
function emitDocumentBlocks(document) {
|
|
257
257
|
const blocks = [];
|
|
258
258
|
/** Push one non-empty prose block. */
|
|
259
|
-
const pushText = (text) => {
|
|
259
|
+
const pushText = (text, tag = '') => {
|
|
260
260
|
if (text) {
|
|
261
|
-
blocks.push({
|
|
261
|
+
blocks.push({
|
|
262
|
+
text,
|
|
263
|
+
atomic: false,
|
|
264
|
+
...(/^H[1-6]$/.test(tag) ? { headingLevel: Number(tag[1]) } : {}),
|
|
265
|
+
});
|
|
262
266
|
}
|
|
263
267
|
};
|
|
264
268
|
const emitElement = (element) => {
|
|
@@ -267,7 +271,7 @@ function emitDocumentBlocks(document) {
|
|
|
267
271
|
}
|
|
268
272
|
const serialize = blockSerializerFor(element);
|
|
269
273
|
if (serialize !== null) {
|
|
270
|
-
pushText(serialize(element));
|
|
274
|
+
pushText(serialize(element), element.tagName);
|
|
271
275
|
return;
|
|
272
276
|
}
|
|
273
277
|
if (element.tagName === 'TABLE') {
|
|
@@ -344,6 +348,7 @@ function dedupeAnchors(imageAnchors) {
|
|
|
344
348
|
}
|
|
345
349
|
export function convertDocxDocumentToText(document) {
|
|
346
350
|
let content = '';
|
|
351
|
+
const headings = [];
|
|
347
352
|
const atomicRanges = [];
|
|
348
353
|
const imageAnchors = [];
|
|
349
354
|
for (const block of emitDocumentBlocks(document)) {
|
|
@@ -359,6 +364,9 @@ export function convertDocxDocumentToText(document) {
|
|
|
359
364
|
content += '\n\n';
|
|
360
365
|
}
|
|
361
366
|
const start = content.length;
|
|
367
|
+
if (block.headingLevel) {
|
|
368
|
+
headings.push({ offset: start, level: block.headingLevel, text: blockText });
|
|
369
|
+
}
|
|
362
370
|
imageAnchors.push(...anchorsWithinBlock(block.text, matches, start));
|
|
363
371
|
content += blockText;
|
|
364
372
|
if (block.atomic) {
|
|
@@ -368,6 +376,7 @@ export function convertDocxDocumentToText(document) {
|
|
|
368
376
|
const uniqueImageAnchors = dedupeAnchors(imageAnchors);
|
|
369
377
|
return {
|
|
370
378
|
content,
|
|
379
|
+
...(headings.length ? { sourceMap: { headings } } : {}),
|
|
371
380
|
atomicRanges,
|
|
372
381
|
...(uniqueImageAnchors.length === 0 ? {} : { imageAnchors: uniqueImageAnchors }),
|
|
373
382
|
};
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"docx-parser.js","sourceRoot":"","sources":["../../src/parser/docx-parser.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,KAAK,EAAE,MAAM,OAAO,CAAA;AAC7B,OAAO,KAAK,MAAM,OAAO,CAAA;AAGzB,MAAM,oBAAoB,GAAG,kCAAkC,CAAA;AAC/D,MAAM,gBAAgB,GAAG,IAAI,GAAG,CAAC,CAAC,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,GAAG,EAAE,YAAY,EAAE,KAAK,CAAC,CAAC,CAAA;AAChG,MAAM,SAAS,GAAG,IAAI,GAAG,CAAC,CAAC,IAAI,EAAE,IAAI,CAAC,CAAC,CAAA;AACvC,MAAM,YAAY,GAAG,IAAI,GAAG,CAAC,CAAC,QAAQ,EAAE,OAAO,EAAE,UAAU,CAAC,CAAC,CAAA;AAC7D,MAAM,mBAAmB,GAAG,kBAAkB,CAAA;AAC9C,MAAM,mBAAmB,GAAG,QAAQ,CAAA;AACpC,MAAM,oBAAoB,GAAG,IAAI,MAAM,CAAC,GAAG,mBAAmB,SAAS,mBAAmB,EAAE,EAAE,IAAI,CAAC,CAAA;AAanG,iFAAiF;AACjF,SAAS,SAAS,CAAC,IAAU;IAC3B,OAAO,IAAI,CAAC,QAAQ,KAAK,CAAC,CAAA;AAC5B,CAAC;AAED,SAAS,WAAW,CAAC,OAAgB;IACnC,IAAI,OAAO,CAAC,OAAO,KAAK,KAAK,EAAE,CAAC;QAC9B,OAAO,EAAE,CAAA;IACX,CAAC;IACD,MAAM,KAAK,GAAG,OAAO,CAAC,YAAY,CAAC,sBAAsB,CAAC,CAAA;IAC1D,OAAO,KAAK,KAAK,IAAI,IAAI,OAAO,CAAC,IAAI,CAAC,KAAK,CAAC;QAC1C,CAAC,CAAC,GAAG,mBAAmB,GAAG,KAAK,GAAG,mBAAmB,EAAE;QACxD,CAAC,CAAC,EAAE,CAAA;AACR,CAAC;AAED,SAAS,mBAAmB,CAAC,IAAY;IACvC,OAAO,IAAI,CAAC,OAAO,CAAC,MAAM,EAAE,GAAG,CAAC,CAAC,IAAI,EAAE,CAAA;AACzC,CAAC;AASD;;;;GAIG;AACH,SAAS,YAAY,CAAC,IAAU;IAC9B,IAAI,IAAI,CAAC,QAAQ,KAAK,CAAC,EAAE,CAAC;QACxB,OAAO,EAAE,IAAI,EAAE,MAAM,EAAE,IAAI,EAAE,IAAI,CAAC,SAAS,IAAI,EAAE,EAAE,CAAA;IACrD,CAAC;IACD,IAAI,CAAC,SAAS,CAAC,IAAI,CAAC,IAAI,YAAY,CAAC,GAAG,CAAC,IAAI,CAAC,OAAO,CAAC,EAAE,CAAC;QACvD,OAAO,EAAE,IAAI,EAAE,MAAM,EAAE,CAAA;IACzB,CAAC;IACD,MAAM,MAAM,GAAG,WAAW,CAAC,IAAI,CAAC,CAAA;IAChC,IAAI,MAAM,EAAE,CAAC;QACX,OAAO,EAAE,IAAI,EAAE,MAAM,EAAE,IAAI,EAAE,MAAM,EAAE,CAAA;IACvC,CAAC;IACD,IAAI,IAAI,CAAC,OAAO,KAAK,IAAI,EAAE,CAAC;QAC1B,OAAO,EAAE,IAAI,EAAE,OAAO,EAAE,CAAA;IAC1B,CAAC;IACD,OAAO,EAAE,IAAI,EAAE,SAAS,EAAE,OAAO,EAAE,IAAI,EAAE,CAAA;AAC3C,CAAC;AAED,SAAS,oBAAoB,CAAC,OAAgB;IAC5C,MAAM,KAAK,GAAa,EAAE,CAAA;IAC1B,MAAM,KAAK,GAAG,CAAC,IAAU,EAAQ,EAAE;QACjC,MAAM,MAAM,GAAG,YAAY,CAAC,IAAI,CAAC,CAAA;QACjC,IAAI,MAAM,CAAC,IAAI,KAAK,MAAM,EAAE,CAAC;YAC3B,OAAM;QACR,CAAC;QACD,IAAI,MAAM,CAAC,IAAI,KAAK,MAAM,EAAE,CAAC;YAC3B,KAAK,CAAC,IAAI,CAAC,MAAM,CAAC,IAAI,CAAC,CAAA;YACvB,OAAM;QACR,CAAC;QACD,IAAI,MAAM,CAAC,IAAI,KAAK,OAAO,EAAE,CAAC;YAC5B,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,CAAA;YACf,OAAM;QACR,CAAC;QAED,MAAM,KAAK,GAAG,MAAM,CAAC,OAAO,CAAA;QAC5B,MAAM,gBAAgB,GACpB,gBAAgB,CAAC,GAAG,CAAC,KAAK,CAAC,OAAO,CAAC,IAAI,SAAS,CAAC,GAAG,CAAC,KAAK,CAAC,OAAO,CAAC,IAAI,KAAK,CAAC,OAAO,KAAK,IAAI,CAAA;QAC/F,IAAI,gBAAgB,EAAE,CAAC;YACrB,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,CAAA;QACjB,CAAC;QACD,KAAK,MAAM,UAAU,IAAI,KAAK,CAAC,UAAU,EAAE,CAAC;YAC1C,KAAK,CAAC,UAAU,CAAC,CAAA;QACnB,CAAC;QACD,IAAI,gBAAgB,EAAE,CAAC;YACrB,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,CAAA;QACjB,CAAC;IACH,CAAC,CAAA;IAED,KAAK,MAAM,KAAK,IAAI,OAAO,CAAC,UAAU,EAAE,CAAC;QACvC,KAAK,CAAC,KAAK,CAAC,CAAA;IACd,CAAC;IACD,OAAO,mBAAmB,CAAC,KAAK,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC,CAAA;AAC5C,CAAC;AAED,SAAS,mBAAmB,CAAC,OAAgB;IAC3C,MAAM,KAAK,GAAa,EAAE,CAAA;IAE1B,MAAM,KAAK,GAAG,CAAC,IAAU,EAAQ,EAAE;QACjC,MAAM,MAAM,GAAG,YAAY,CAAC,IAAI,CAAC,CAAA;QACjC,IAAI,MAAM,CAAC,IAAI,KAAK,MAAM,EAAE,CAAC;YAC3B,OAAM;QACR,CAAC;QACD,IAAI,MAAM,CAAC,IAAI,KAAK,MAAM,EAAE,CAAC;YAC3B,KAAK,CAAC,IAAI,CAAC,MAAM,CAAC,IAAI,CAAC,CAAA;YACvB,OAAM;QACR,CAAC;QACD,IAAI,MAAM,CAAC,IAAI,KAAK,OAAO,EAAE,CAAC;YAC5B,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAA;YAChB,OAAM;QACR,CAAC;QACD,KAAK,MAAM,UAAU,IAAI,MAAM,CAAC,OAAO,CAAC,UAAU,EAAE,CAAC;YACnD,KAAK,CAAC,UAAU,CAAC,CAAA;QACnB,CAAC;IACH,CAAC,CAAA;IAED,KAAK,MAAM,KAAK,IAAI,OAAO,CAAC,UAAU,EAAE,CAAC;QACvC,KAAK,CAAC,KAAK,CAAC,CAAA;IACd,CAAC;IAED,MAAM,KAAK,GAAG,KAAK,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,GAAG,CAAC,mBAAmB,CAAC,CAAA;IACjE,OAAO,KAAK,CAAC,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC;QACvB,KAAK,CAAC,KAAK,EAAE,CAAA;IACf,CAAC;IACD,OAAO,KAAK,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC;QAC3B,KAAK,CAAC,GAAG,EAAE,CAAA;IACb,CAAC;IACD,OAAO,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAA;AACzB,CAAC;AAED,SAAS,aAAa,CAAC,IAAa;IAClC,MAAM,KAAK,GAAa,EAAE,CAAA;IAE1B,MAAM,oBAAoB,GAAG,CAAC,WAAoB,EAAQ,EAAE;QAC1D,KAAK,MAAM,KAAK,IAAI,WAAW,CAAC,QAAQ,EAAE,CAAC;YACzC,IAAI,KAAK,CAAC,OAAO,KAAK,IAAI,EAAE,CAAC;gBAC3B,iBAAiB,CAAC,KAAK,EAAE,KAAK,EAAE,oBAAoB,CAAC,CAAA;YACvD,CAAC;QACH,CAAC;IACH,CAAC,CAAA;IAED,oBAAoB,CAAC,IAAI,CAAC,CAAA;IAC1B,OAAO,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAA;AACzB,CAAC;AAED;;;GAGG;AACH,SAAS,iBAAiB,CACxB,IAAa,EACb,KAAe,EACf,eAAwC;IAExC,IAAI,WAAW,GAAG,EAAE,CAAA;IACpB,MAAM,KAAK,GAAG,GAAS,EAAE;QACvB,MAAM,UAAU,GAAG,mBAAmB,CAAC,WAAW,CAAC,CAAA;QACnD,IAAI,UAAU,EAAE,CAAC;YACf,KAAK,CAAC,IAAI,CAAC,UAAU,CAAC,CAAA;QACxB,CAAC;QACD,WAAW,GAAG,EAAE,CAAA;IAClB,CAAC,CAAA;IAED,MAAM,aAAa,GAAG,CAAC,IAAU,EAAQ,EAAE;QACzC,MAAM,MAAM,GAAG,YAAY,CAAC,IAAI,CAAC,CAAA;QACjC,IAAI,MAAM,CAAC,IAAI,KAAK,MAAM,EAAE,CAAC;YAC3B,OAAM;QACR,CAAC;QACD,IAAI,MAAM,CAAC,IAAI,KAAK,MAAM,EAAE,CAAC;YAC3B,WAAW,IAAI,MAAM,CAAC,IAAI,CAAA;YAC1B,OAAM;QACR,CAAC;QACD,IAAI,MAAM,CAAC,IAAI,KAAK,OAAO,EAAE,CAAC;YAC5B,KAAK,EAAE,CAAA;YACP,OAAM;QACR,CAAC;QACD,IAAI,SAAS,CAAC,GAAG,CAAC,MAAM,CAAC,OAAO,CAAC,OAAO,CAAC,EAAE,CAAC;YAC1C,KAAK,EAAE,CAAA;YACP,eAAe,CAAC,MAAM,CAAC,OAAO,CAAC,CAAA;YAC/B,OAAM;QACR,CAAC;QACD,KAAK,MAAM,UAAU,IAAI,MAAM,CAAC,OAAO,CAAC,UAAU,EAAE,CAAC;YACnD,aAAa,CAAC,UAAU,CAAC,CAAA;QAC3B,CAAC;IACH,CAAC,CAAA;IAED,KAAK,MAAM,SAAS,IAAI,IAAI,CAAC,UAAU,EAAE,CAAC;QACxC,aAAa,CAAC,SAAS,CAAC,CAAA;IAC1B,CAAC;IACD,KAAK,EAAE,CAAA;AACT,CAAC;AAED,SAAS,eAAe,CAAC,KAAc;IACrC,OAAO,KAAK,CAAC,IAAI,CAAC,KAAK,CAAC,gBAAgB,CAAC,IAAI,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,GAAG,CAAC,OAAO,CAAC,OAAO,CAAC,KAAK,KAAK,CAAC,CAAA;AACjG,CAAC;AAED,SAAS,cAAc,CAAC,GAAY;IAClC,OAAO,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,QAAQ,CAAC,CAAC,MAAM,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,OAAO,KAAK,IAAI,IAAI,IAAI,CAAC,OAAO,KAAK,IAAI,CAAC,CAAA;AAClG,CAAC;AAED,SAAS,eAAe,CAAC,KAAyB;IAChD,OAAO,KAAK,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,EAAE;QACzB,MAAM,OAAO,GAAG,MAAM,CAAC,QAAQ,CAAC,IAAI,CAAC,YAAY,CAAC,SAAS,CAAC,IAAI,GAAG,EAAE,EAAE,CAAC,CAAA;QACxE,MAAM,UAAU,GAAG,MAAM,CAAC,QAAQ,CAAC,IAAI,CAAC,YAAY,CAAC,SAAS,CAAC,IAAI,GAAG,EAAE,EAAE,CAAC,CAAA;QAC3E,OAAO,OAAO,GAAG,CAAC,IAAI,UAAU,GAAG,CAAC,CAAA;IACtC,CAAC,CAAC,CAAA;AACJ,CAAC;AAED,SAAS,gBAAgB,CAAC,KAAc,EAAE,IAAwB;IAChE,IAAI,KAAK,CAAC,aAAa,CAAC,OAAO,CAAC,KAAK,IAAI,IAAI,IAAI,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QAC7D,OAAO,KAAK,CAAA;IACd,CAAC;IAED,MAAM,QAAQ,GAAG,IAAI,CAAC,GAAG,CAAC,cAAc,CAAC,CAAA;IACzC,MAAM,WAAW,GAAG,QAAQ,CAAC,CAAC,CAAC,EAAE,MAAM,IAAI,CAAC,CAAA;IAC5C,OAAO,CACL,WAAW,GAAG,CAAC;QACf,QAAQ,CAAC,KAAK,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,MAAM,KAAK,WAAW,IAAI,CAAC,eAAe,CAAC,KAAK,CAAC,CAAC,CACnF,CAAA;AACH,CAAC;AAED,SAAS,iBAAiB,CAAC,OAAgB;IACzC,MAAM,MAAM,GAAa,EAAE,CAAA;IAC3B,MAAM,KAAK,GAAG,CAAC,IAAU,EAAQ,EAAE;QACjC,IAAI,IAAI,CAAC,QAAQ,KAAK,CAAC,EAAE,CAAC;YACxB,MAAM,KAAK,GAAG,mBAAmB,CAAC,IAAI,CAAC,SAAS,IAAI,EAAE,CAAC,CAAA;YACvD,IAAI,KAAK,EAAE,CAAC;gBACV,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,CAAA;YACpB,CAAC;YACD,OAAM;QACR,CAAC;QACD,IAAI,CAAC,SAAS,CAAC,IAAI,CAAC,EAAE,CAAC;YACrB,OAAM;QACR,CAAC;QAED,MAAM,KAAK,GAAG,IAAI,CAAA;QAClB,IAAI,YAAY,CAAC,GAAG,CAAC,KAAK,CAAC,OAAO,CAAC,EAAE,CAAC;YACpC,OAAM;QACR,CAAC;QACD,MAAM,MAAM,GAAG,WAAW,CAAC,KAAK,CAAC,CAAA;QACjC,IAAI,MAAM,EAAE,CAAC;YACX,MAAM,CAAC,IAAI,CAAC,MAAM,CAAC,CAAA;YACnB,OAAM;QACR,CAAC;QACD,KAAK,MAAM,UAAU,IAAI,KAAK,CAAC,UAAU,EAAE,CAAC;YAC1C,KAAK,CAAC,UAAU,CAAC,CAAA;QACnB,CAAC;IACH,CAAC,CAAA;IACD,KAAK,CAAC,OAAO,CAAC,CAAA;IACd,OAAO,MAAM,CAAC,IAAI,CAAC,GAAG,CAAC,CAAA;AACzB,CAAC;AAED,SAAS,SAAS,CAAC,KAAc;IAC/B,MAAM,IAAI,GAAG,eAAe,CAAC,KAAK,CAAC,CAAA;IACnC,MAAM,SAAS,GAAG,IAAI,CAAC,CAAC,CAAC,CAAA;IACzB,IAAI,CAAC,gBAAgB,CAAC,KAAK,EAAE,IAAI,CAAC,IAAI,SAAS,KAAK,SAAS,EAAE,CAAC;QAC9D,MAAM,IAAI,GAAG,iBAAiB,CAAC,KAAK,CAAC,CAAA;QACrC,OAAO,IAAI,CAAC,CAAC,CAAC,CAAC,EAAE,IAAI,EAAE,MAAM,EAAE,KAAK,EAAE,CAAC,CAAC,CAAC,CAAC,EAAE,CAAA;IAC9C,CAAC;IAED,6EAA6E;IAC7E,+EAA+E;IAC/E,MAAM,aAAa,GAAa,EAAE,CAAA;IAClC,MAAM,OAAO,GAAG,cAAc,CAAC,SAAS,CAAC,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE;QAC5D,MAAM,IAAI,GAAG,oBAAoB,CAAC,IAAI,CAAC,CAAA;QACvC,aAAa,CAAC,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,KAAK,CAAC,oBAAoB,CAAC,IAAI,EAAE,CAAC,CAAC,CAAA;QAC/D,OAAO,IAAI,CAAC,OAAO,CAAC,oBAAoB,EAAE,EAAE,CAAC,IAAI,UAAU,KAAK,GAAG,CAAC,EAAE,CAAA;IACxE,CAAC,CAAC,CAAA;IAEF,MAAM,SAAS,GAAG,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,GAAG,EAAE,EAAE;QAC9C,MAAM,MAAM,GAAG,cAAc,CAAC,GAAG,CAAC,CAAA;QAClC,MAAM,IAAI,GAAG,MAAM;aAChB,GAAG,CACF,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE,CAAC,GAAG,OAAO,CAAC,KAAK,CAAC,IAAI,UAAU,KAAK,GAAG,CAAC,EAAE,KAAK,oBAAoB,CAAC,IAAI,CAAC,EAAE,CAC7F;aACA,IAAI,CAAC,IAAI,CAAC,CAAA;QACb,OAAO,IAAI,CAAC,CAAC,CAAC,CAAC,EAAE,IAAI,EAAE,MAAM,EAAE,IAAI,EAAE,CAAC,CAAC,CAAC,CAAC,EAAE,CAAA;IAC7C,CAAC,CAAC,CAAA;IACF,MAAM,KAAK,GAAG,SAAS,CAAC,CAAC,CAAC,CAAA;IAC1B,IAAI,KAAK,IAAI,aAAa,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QACtC,KAAK,CAAC,IAAI,GAAG,aAAa,CAAC,IAAI,CAAC,EAAE,CAAC,GAAG,KAAK,CAAC,IAAI,CAAA;IAClD,CAAC;IACD,OAAO,SAAS,CAAA;AAClB,CAAC;AAED,SAAS,4BAA4B,CAAC,OAAgB;IACpD,OAAO,KAAK,CAAC,IAAI,CAAC,OAAO,CAAC,gBAAgB,CAAC,GAAG,CAAC,CAAC,CAAC,IAAI,CACnD,CAAC,UAAU,EAAE,EAAE,CACb,gBAAgB,CAAC,GAAG,CAAC,UAAU,CAAC,OAAO,CAAC;QACxC,SAAS,CAAC,GAAG,CAAC,UAAU,CAAC,OAAO,CAAC;QACjC,UAAU,CAAC,OAAO,KAAK,OAAO,CACjC,CAAA;AACH,CAAC;AAED,oFAAoF;AACpF,SAAS,kBAAkB,CAAC,OAAgB;IAC1C,IAAI,gBAAgB,CAAC,GAAG,CAAC,OAAO,CAAC,OAAO,CAAC,EAAE,CAAC;QAC1C,OAAO,mBAAmB,CAAA;IAC5B,CAAC;IACD,IAAI,SAAS,CAAC,GAAG,CAAC,OAAO,CAAC,OAAO,CAAC,EAAE,CAAC;QACnC,OAAO,aAAa,CAAA;IACtB,CAAC;IACD,OAAO,IAAI,CAAA;AACb,CAAC;AAED,SAAS,kBAAkB,CAAC,QAAkB;IAC5C,MAAM,MAAM,GAAmB,EAAE,CAAA;IAEjC,sCAAsC;IACtC,MAAM,QAAQ,GAAG,CAAC,IAAY,EAAQ,EAAE;QACtC,IAAI,IAAI,EAAE,CAAC;YACT,MAAM,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,MAAM,EAAE,KAAK,EAAE,CAAC,CAAA;QACtC,CAAC;IACH,CAAC,CAAA;IAED,MAAM,WAAW,GAAG,CAAC,OAAgB,EAAQ,EAAE;QAC7C,IAAI,YAAY,CAAC,GAAG,CAAC,OAAO,CAAC,OAAO,CAAC,EAAE,CAAC;YACtC,OAAM;QACR,CAAC;QACD,MAAM,SAAS,GAAG,kBAAkB,CAAC,OAAO,CAAC,CAAA;QAC7C,IAAI,SAAS,KAAK,IAAI,EAAE,CAAC;YACvB,QAAQ,CAAC,SAAS,CAAC,OAAO,CAAC,CAAC,CAAA;YAC5B,OAAM;QACR,CAAC;QACD,IAAI,OAAO,CAAC,OAAO,KAAK,OAAO,EAAE,CAAC;YAChC,MAAM,CAAC,IAAI,CAAC,GAAG,SAAS,CAAC,OAAO,CAAC,CAAC,CAAA;YAClC,OAAM;QACR,CAAC;QACD,IAAI,CAAC,4BAA4B,CAAC,OAAO,CAAC,EAAE,CAAC;YAC3C,QAAQ,CAAC,oBAAoB,CAAC,OAAO,CAAC,CAAC,CAAA;YACvC,OAAM;QACR,CAAC;QACD,KAAK,MAAM,KAAK,IAAI,OAAO,CAAC,UAAU,EAAE,CAAC;YACvC,QAAQ,CAAC,KAAK,CAAC,CAAA;QACjB,CAAC;IACH,CAAC,CAAA;IAED,MAAM,QAAQ,GAAG,CAAC,IAAU,EAAQ,EAAE;QACpC,IAAI,IAAI,CAAC,QAAQ,KAAK,CAAC,EAAE,CAAC;YACxB,QAAQ,CAAC,mBAAmB,CAAC,IAAI,CAAC,SAAS,IAAI,EAAE,CAAC,CAAC,CAAA;YACnD,OAAM;QACR,CAAC;QACD,IAAI,SAAS,CAAC,IAAI,CAAC,EAAE,CAAC;YACpB,WAAW,CAAC,IAAI,CAAC,CAAA;QACnB,CAAC;IACH,CAAC,CAAA;IAED,KAAK,MAAM,IAAI,IAAI,QAAQ,CAAC,IAAI,CAAC,UAAU,EAAE,CAAC;QAC5C,QAAQ,CAAC,IAAI,CAAC,CAAA;IAChB,CAAC;IACD,OAAO,MAAM,CAAA;AACf,CAAC;AAED,MAAM,CAAC,KAAK,UAAU,oBAAoB,CAAC,MAAc;IACvD,MAAM,GAAG,GAAG,MAAM,KAAK,CAAC,SAAS,CAAC,MAAM,CAAC,CAAA;IACzC,MAAM,cAAc,GAAG,GAAG,CAAC,IAAI,CAAC,mBAAmB,CAAC,CAAA;IACpD,IAAI,CAAC,cAAc,EAAE,CAAC;QACpB,OAAO,SAAS,CAAA;IAClB,CAAC;IAED,MAAM,GAAG,GAAG,MAAM,cAAc,CAAC,KAAK,CAAC,QAAQ,CAAC,CAAA;IAChD,IAAI,CAAC;QACH,MAAM,QAAQ,GAAG,IAAI,KAAK,CAAC,GAAG,EAAE,EAAE,WAAW,EAAE,UAAU,EAAE,CAAC,CAAC,MAAM,CAAC,QAAQ,CAAA;QAC5E,IAAI,QAAQ,CAAC,aAAa,CAAC,aAAa,CAAC,EAAE,CAAC;YAC1C,OAAO,SAAS,CAAA;QAClB,CAAC;QAED,MAAM,KAAK,GAAG,QAAQ,CAAC,sBAAsB,CAAC,oBAAoB,EAAE,OAAO,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAA;QACpF,MAAM,UAAU,GAAG,mBAAmB,CAAC,KAAK,EAAE,WAAW,IAAI,EAAE,CAAC,CAAA;QAChE,OAAO,UAAU,IAAI,SAAS,CAAA;IAChC,CAAC;IAAC,MAAM,CAAC;QACP,qEAAqE;QACrE,8EAA8E;QAC9E,OAAO,SAAS,CAAA;IAClB,CAAC;AACH,CAAC;AAED,uEAAuE;AACvE,SAAS,iBAAiB,CAAC,IAAY;IACrC,OAAO,IAAI,CAAC,OAAO,CAAC,oBAAoB,EAAE,EAAE,CAAC,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,GAAG,CAAC,mBAAmB,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,CAAA;AAC/F,CAAC;AAED,6EAA6E;AAC7E,SAAS,kBAAkB,CACzB,SAAiB,EACjB,OAAmC,EACnC,KAAa;IAEb,OAAO,OAAO,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC;QAC7B,MAAM,EAAE,KAAK,GAAG,iBAAiB,CAAC,SAAS,CAAC,KAAK,CAAC,CAAC,EAAE,KAAK,CAAC,KAAK,IAAI,CAAC,CAAC,CAAC,CAAC,MAAM;QAC9E,UAAU,EAAE,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC;KAC7B,CAAC,CAAC,CAAA;AACL,CAAC;AAED,6DAA6D;AAC7D,SAAS,aAAa,CAAC,YAAoC;IACzD,MAAM,IAAI,GAAG,IAAI,GAAG,EAAU,CAAA;IAC9B,OAAO,CAAC,GAAG,YAAY,CAAC;SACrB,IAAI,CAAC,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE,CAAC,IAAI,CAAC,UAAU,GAAG,KAAK,CAAC,UAAU,CAAC;SACzD,MAAM,CAAC,CAAC,MAAM,EAAE,EAAE;QACjB,IAAI,IAAI,CAAC,GAAG,CAAC,MAAM,CAAC,UAAU,CAAC,EAAE,CAAC;YAChC,OAAO,KAAK,CAAA;QACd,CAAC;QACD,IAAI,CAAC,GAAG,CAAC,MAAM,CAAC,UAAU,CAAC,CAAA;QAC3B,OAAO,IAAI,CAAA;IACb,CAAC,CAAC,CAAA;AACN,CAAC;AAOD,MAAM,UAAU,yBAAyB,CAAC,QAAkB;IAC1D,IAAI,OAAO,GAAG,EAAE,CAAA;IAChB,MAAM,YAAY,GAAsB,EAAE,CAAA;IAC1C,MAAM,YAAY,GAAkB,EAAE,CAAA;IAEtC,KAAK,MAAM,KAAK,IAAI,kBAAkB,CAAC,QAAQ,CAAC,EAAE,CAAC;QACjD,MAAM,OAAO,GAAG,CAAC,GAAG,KAAK,CAAC,IAAI,CAAC,QAAQ,CAAC,oBAAoB,CAAC,CAAC,CAAA;QAC9D,MAAM,SAAS,GAAG,OAAO,CAAC,MAAM,KAAK,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC,iBAAiB,CAAC,KAAK,CAAC,IAAI,CAAC,CAAA;QAEnF,sEAAsE;QACtE,qDAAqD;QACrD,IAAI,OAAO,CAAC,MAAM,GAAG,CAAC,IAAI,CAAC,SAAS,EAAE,CAAC;YACrC,YAAY,CAAC,IAAI,CACf,GAAG,OAAO,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC,EAAE,MAAM,EAAE,OAAO,CAAC,MAAM,EAAE,UAAU,EAAE,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CACtF,CAAA;YACD,SAAQ;QACV,CAAC;QAED,IAAI,OAAO,EAAE,CAAC;YACZ,OAAO,IAAI,MAAM,CAAA;QACnB,CAAC;QACD,MAAM,KAAK,GAAG,OAAO,CAAC,MAAM,CAAA;QAC5B,YAAY,CAAC,IAAI,CAAC,GAAG,kBAAkB,CAAC,KAAK,CAAC,IAAI,EAAE,OAAO,EAAE,KAAK,CAAC,CAAC,CAAA;QACpE,OAAO,IAAI,SAAS,CAAA;QACpB,IAAI,KAAK,CAAC,MAAM,EAAE,CAAC;YACjB,YAAY,CAAC,IAAI,CAAC,EAAE,KAAK,EAAE,GAAG,EAAE,OAAO,CAAC,MAAM,EAAE,CAAC,CAAA;QACnD,CAAC;IACH,CAAC;IAED,MAAM,kBAAkB,GAAG,aAAa,CAAC,YAAY,CAAC,CAAA;IACtD,OAAO;QACL,OAAO;QACP,YAAY;QACZ,GAAG,CAAC,kBAAkB,CAAC,MAAM,KAAK,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,YAAY,EAAE,kBAAkB,EAAE,CAAC;KACjF,CAAA;AACH,CAAC"}
|
|
1
|
+
{"version":3,"file":"docx-parser.js","sourceRoot":"","sources":["../../src/parser/docx-parser.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,KAAK,EAAE,MAAM,OAAO,CAAA;AAC7B,OAAO,KAAK,MAAM,OAAO,CAAA;AAIzB,MAAM,oBAAoB,GAAG,kCAAkC,CAAA;AAC/D,MAAM,gBAAgB,GAAG,IAAI,GAAG,CAAC,CAAC,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,GAAG,EAAE,YAAY,EAAE,KAAK,CAAC,CAAC,CAAA;AAChG,MAAM,SAAS,GAAG,IAAI,GAAG,CAAC,CAAC,IAAI,EAAE,IAAI,CAAC,CAAC,CAAA;AACvC,MAAM,YAAY,GAAG,IAAI,GAAG,CAAC,CAAC,QAAQ,EAAE,OAAO,EAAE,UAAU,CAAC,CAAC,CAAA;AAC7D,MAAM,mBAAmB,GAAG,kBAAkB,CAAA;AAC9C,MAAM,mBAAmB,GAAG,QAAQ,CAAA;AACpC,MAAM,oBAAoB,GAAG,IAAI,MAAM,CAAC,GAAG,mBAAmB,SAAS,mBAAmB,EAAE,EAAE,IAAI,CAAC,CAAA;AAenG,iFAAiF;AACjF,SAAS,SAAS,CAAC,IAAU;IAC3B,OAAO,IAAI,CAAC,QAAQ,KAAK,CAAC,CAAA;AAC5B,CAAC;AAED,SAAS,WAAW,CAAC,OAAgB;IACnC,IAAI,OAAO,CAAC,OAAO,KAAK,KAAK,EAAE,CAAC;QAC9B,OAAO,EAAE,CAAA;IACX,CAAC;IACD,MAAM,KAAK,GAAG,OAAO,CAAC,YAAY,CAAC,sBAAsB,CAAC,CAAA;IAC1D,OAAO,KAAK,KAAK,IAAI,IAAI,OAAO,CAAC,IAAI,CAAC,KAAK,CAAC;QAC1C,CAAC,CAAC,GAAG,mBAAmB,GAAG,KAAK,GAAG,mBAAmB,EAAE;QACxD,CAAC,CAAC,EAAE,CAAA;AACR,CAAC;AAED,SAAS,mBAAmB,CAAC,IAAY;IACvC,OAAO,IAAI,CAAC,OAAO,CAAC,MAAM,EAAE,GAAG,CAAC,CAAC,IAAI,EAAE,CAAA;AACzC,CAAC;AASD;;;;GAIG;AACH,SAAS,YAAY,CAAC,IAAU;IAC9B,IAAI,IAAI,CAAC,QAAQ,KAAK,CAAC,EAAE,CAAC;QACxB,OAAO,EAAE,IAAI,EAAE,MAAM,EAAE,IAAI,EAAE,IAAI,CAAC,SAAS,IAAI,EAAE,EAAE,CAAA;IACrD,CAAC;IACD,IAAI,CAAC,SAAS,CAAC,IAAI,CAAC,IAAI,YAAY,CAAC,GAAG,CAAC,IAAI,CAAC,OAAO,CAAC,EAAE,CAAC;QACvD,OAAO,EAAE,IAAI,EAAE,MAAM,EAAE,CAAA;IACzB,CAAC;IACD,MAAM,MAAM,GAAG,WAAW,CAAC,IAAI,CAAC,CAAA;IAChC,IAAI,MAAM,EAAE,CAAC;QACX,OAAO,EAAE,IAAI,EAAE,MAAM,EAAE,IAAI,EAAE,MAAM,EAAE,CAAA;IACvC,CAAC;IACD,IAAI,IAAI,CAAC,OAAO,KAAK,IAAI,EAAE,CAAC;QAC1B,OAAO,EAAE,IAAI,EAAE,OAAO,EAAE,CAAA;IAC1B,CAAC;IACD,OAAO,EAAE,IAAI,EAAE,SAAS,EAAE,OAAO,EAAE,IAAI,EAAE,CAAA;AAC3C,CAAC;AAED,SAAS,oBAAoB,CAAC,OAAgB;IAC5C,MAAM,KAAK,GAAa,EAAE,CAAA;IAC1B,MAAM,KAAK,GAAG,CAAC,IAAU,EAAQ,EAAE;QACjC,MAAM,MAAM,GAAG,YAAY,CAAC,IAAI,CAAC,CAAA;QACjC,IAAI,MAAM,CAAC,IAAI,KAAK,MAAM,EAAE,CAAC;YAC3B,OAAM;QACR,CAAC;QACD,IAAI,MAAM,CAAC,IAAI,KAAK,MAAM,EAAE,CAAC;YAC3B,KAAK,CAAC,IAAI,CAAC,MAAM,CAAC,IAAI,CAAC,CAAA;YACvB,OAAM;QACR,CAAC;QACD,IAAI,MAAM,CAAC,IAAI,KAAK,OAAO,EAAE,CAAC;YAC5B,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,CAAA;YACf,OAAM;QACR,CAAC;QAED,MAAM,KAAK,GAAG,MAAM,CAAC,OAAO,CAAA;QAC5B,MAAM,gBAAgB,GACpB,gBAAgB,CAAC,GAAG,CAAC,KAAK,CAAC,OAAO,CAAC,IAAI,SAAS,CAAC,GAAG,CAAC,KAAK,CAAC,OAAO,CAAC,IAAI,KAAK,CAAC,OAAO,KAAK,IAAI,CAAA;QAC/F,IAAI,gBAAgB,EAAE,CAAC;YACrB,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,CAAA;QACjB,CAAC;QACD,KAAK,MAAM,UAAU,IAAI,KAAK,CAAC,UAAU,EAAE,CAAC;YAC1C,KAAK,CAAC,UAAU,CAAC,CAAA;QACnB,CAAC;QACD,IAAI,gBAAgB,EAAE,CAAC;YACrB,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,CAAA;QACjB,CAAC;IACH,CAAC,CAAA;IAED,KAAK,MAAM,KAAK,IAAI,OAAO,CAAC,UAAU,EAAE,CAAC;QACvC,KAAK,CAAC,KAAK,CAAC,CAAA;IACd,CAAC;IACD,OAAO,mBAAmB,CAAC,KAAK,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC,CAAA;AAC5C,CAAC;AAED,SAAS,mBAAmB,CAAC,OAAgB;IAC3C,MAAM,KAAK,GAAa,EAAE,CAAA;IAE1B,MAAM,KAAK,GAAG,CAAC,IAAU,EAAQ,EAAE;QACjC,MAAM,MAAM,GAAG,YAAY,CAAC,IAAI,CAAC,CAAA;QACjC,IAAI,MAAM,CAAC,IAAI,KAAK,MAAM,EAAE,CAAC;YAC3B,OAAM;QACR,CAAC;QACD,IAAI,MAAM,CAAC,IAAI,KAAK,MAAM,EAAE,CAAC;YAC3B,KAAK,CAAC,IAAI,CAAC,MAAM,CAAC,IAAI,CAAC,CAAA;YACvB,OAAM;QACR,CAAC;QACD,IAAI,MAAM,CAAC,IAAI,KAAK,OAAO,EAAE,CAAC;YAC5B,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAA;YAChB,OAAM;QACR,CAAC;QACD,KAAK,MAAM,UAAU,IAAI,MAAM,CAAC,OAAO,CAAC,UAAU,EAAE,CAAC;YACnD,KAAK,CAAC,UAAU,CAAC,CAAA;QACnB,CAAC;IACH,CAAC,CAAA;IAED,KAAK,MAAM,KAAK,IAAI,OAAO,CAAC,UAAU,EAAE,CAAC;QACvC,KAAK,CAAC,KAAK,CAAC,CAAA;IACd,CAAC;IAED,MAAM,KAAK,GAAG,KAAK,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,GAAG,CAAC,mBAAmB,CAAC,CAAA;IACjE,OAAO,KAAK,CAAC,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC;QACvB,KAAK,CAAC,KAAK,EAAE,CAAA;IACf,CAAC;IACD,OAAO,KAAK,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC;QAC3B,KAAK,CAAC,GAAG,EAAE,CAAA;IACb,CAAC;IACD,OAAO,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAA;AACzB,CAAC;AAED,SAAS,aAAa,CAAC,IAAa;IAClC,MAAM,KAAK,GAAa,EAAE,CAAA;IAE1B,MAAM,oBAAoB,GAAG,CAAC,WAAoB,EAAQ,EAAE;QAC1D,KAAK,MAAM,KAAK,IAAI,WAAW,CAAC,QAAQ,EAAE,CAAC;YACzC,IAAI,KAAK,CAAC,OAAO,KAAK,IAAI,EAAE,CAAC;gBAC3B,iBAAiB,CAAC,KAAK,EAAE,KAAK,EAAE,oBAAoB,CAAC,CAAA;YACvD,CAAC;QACH,CAAC;IACH,CAAC,CAAA;IAED,oBAAoB,CAAC,IAAI,CAAC,CAAA;IAC1B,OAAO,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAA;AACzB,CAAC;AAED;;;GAGG;AACH,SAAS,iBAAiB,CACxB,IAAa,EACb,KAAe,EACf,eAAwC;IAExC,IAAI,WAAW,GAAG,EAAE,CAAA;IACpB,MAAM,KAAK,GAAG,GAAS,EAAE;QACvB,MAAM,UAAU,GAAG,mBAAmB,CAAC,WAAW,CAAC,CAAA;QACnD,IAAI,UAAU,EAAE,CAAC;YACf,KAAK,CAAC,IAAI,CAAC,UAAU,CAAC,CAAA;QACxB,CAAC;QACD,WAAW,GAAG,EAAE,CAAA;IAClB,CAAC,CAAA;IAED,MAAM,aAAa,GAAG,CAAC,IAAU,EAAQ,EAAE;QACzC,MAAM,MAAM,GAAG,YAAY,CAAC,IAAI,CAAC,CAAA;QACjC,IAAI,MAAM,CAAC,IAAI,KAAK,MAAM,EAAE,CAAC;YAC3B,OAAM;QACR,CAAC;QACD,IAAI,MAAM,CAAC,IAAI,KAAK,MAAM,EAAE,CAAC;YAC3B,WAAW,IAAI,MAAM,CAAC,IAAI,CAAA;YAC1B,OAAM;QACR,CAAC;QACD,IAAI,MAAM,CAAC,IAAI,KAAK,OAAO,EAAE,CAAC;YAC5B,KAAK,EAAE,CAAA;YACP,OAAM;QACR,CAAC;QACD,IAAI,SAAS,CAAC,GAAG,CAAC,MAAM,CAAC,OAAO,CAAC,OAAO,CAAC,EAAE,CAAC;YAC1C,KAAK,EAAE,CAAA;YACP,eAAe,CAAC,MAAM,CAAC,OAAO,CAAC,CAAA;YAC/B,OAAM;QACR,CAAC;QACD,KAAK,MAAM,UAAU,IAAI,MAAM,CAAC,OAAO,CAAC,UAAU,EAAE,CAAC;YACnD,aAAa,CAAC,UAAU,CAAC,CAAA;QAC3B,CAAC;IACH,CAAC,CAAA;IAED,KAAK,MAAM,SAAS,IAAI,IAAI,CAAC,UAAU,EAAE,CAAC;QACxC,aAAa,CAAC,SAAS,CAAC,CAAA;IAC1B,CAAC;IACD,KAAK,EAAE,CAAA;AACT,CAAC;AAED,SAAS,eAAe,CAAC,KAAc;IACrC,OAAO,KAAK,CAAC,IAAI,CAAC,KAAK,CAAC,gBAAgB,CAAC,IAAI,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,GAAG,CAAC,OAAO,CAAC,OAAO,CAAC,KAAK,KAAK,CAAC,CAAA;AACjG,CAAC;AAED,SAAS,cAAc,CAAC,GAAY;IAClC,OAAO,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,QAAQ,CAAC,CAAC,MAAM,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,OAAO,KAAK,IAAI,IAAI,IAAI,CAAC,OAAO,KAAK,IAAI,CAAC,CAAA;AAClG,CAAC;AAED,SAAS,eAAe,CAAC,KAAyB;IAChD,OAAO,KAAK,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,EAAE;QACzB,MAAM,OAAO,GAAG,MAAM,CAAC,QAAQ,CAAC,IAAI,CAAC,YAAY,CAAC,SAAS,CAAC,IAAI,GAAG,EAAE,EAAE,CAAC,CAAA;QACxE,MAAM,UAAU,GAAG,MAAM,CAAC,QAAQ,CAAC,IAAI,CAAC,YAAY,CAAC,SAAS,CAAC,IAAI,GAAG,EAAE,EAAE,CAAC,CAAA;QAC3E,OAAO,OAAO,GAAG,CAAC,IAAI,UAAU,GAAG,CAAC,CAAA;IACtC,CAAC,CAAC,CAAA;AACJ,CAAC;AAED,SAAS,gBAAgB,CAAC,KAAc,EAAE,IAAwB;IAChE,IAAI,KAAK,CAAC,aAAa,CAAC,OAAO,CAAC,KAAK,IAAI,IAAI,IAAI,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QAC7D,OAAO,KAAK,CAAA;IACd,CAAC;IAED,MAAM,QAAQ,GAAG,IAAI,CAAC,GAAG,CAAC,cAAc,CAAC,CAAA;IACzC,MAAM,WAAW,GAAG,QAAQ,CAAC,CAAC,CAAC,EAAE,MAAM,IAAI,CAAC,CAAA;IAC5C,OAAO,CACL,WAAW,GAAG,CAAC;QACf,QAAQ,CAAC,KAAK,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,MAAM,KAAK,WAAW,IAAI,CAAC,eAAe,CAAC,KAAK,CAAC,CAAC,CACnF,CAAA;AACH,CAAC;AAED,SAAS,iBAAiB,CAAC,OAAgB;IACzC,MAAM,MAAM,GAAa,EAAE,CAAA;IAC3B,MAAM,KAAK,GAAG,CAAC,IAAU,EAAQ,EAAE;QACjC,IAAI,IAAI,CAAC,QAAQ,KAAK,CAAC,EAAE,CAAC;YACxB,MAAM,KAAK,GAAG,mBAAmB,CAAC,IAAI,CAAC,SAAS,IAAI,EAAE,CAAC,CAAA;YACvD,IAAI,KAAK,EAAE,CAAC;gBACV,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,CAAA;YACpB,CAAC;YACD,OAAM;QACR,CAAC;QACD,IAAI,CAAC,SAAS,CAAC,IAAI,CAAC,EAAE,CAAC;YACrB,OAAM;QACR,CAAC;QAED,MAAM,KAAK,GAAG,IAAI,CAAA;QAClB,IAAI,YAAY,CAAC,GAAG,CAAC,KAAK,CAAC,OAAO,CAAC,EAAE,CAAC;YACpC,OAAM;QACR,CAAC;QACD,MAAM,MAAM,GAAG,WAAW,CAAC,KAAK,CAAC,CAAA;QACjC,IAAI,MAAM,EAAE,CAAC;YACX,MAAM,CAAC,IAAI,CAAC,MAAM,CAAC,CAAA;YACnB,OAAM;QACR,CAAC;QACD,KAAK,MAAM,UAAU,IAAI,KAAK,CAAC,UAAU,EAAE,CAAC;YAC1C,KAAK,CAAC,UAAU,CAAC,CAAA;QACnB,CAAC;IACH,CAAC,CAAA;IACD,KAAK,CAAC,OAAO,CAAC,CAAA;IACd,OAAO,MAAM,CAAC,IAAI,CAAC,GAAG,CAAC,CAAA;AACzB,CAAC;AAED,SAAS,SAAS,CAAC,KAAc;IAC/B,MAAM,IAAI,GAAG,eAAe,CAAC,KAAK,CAAC,CAAA;IACnC,MAAM,SAAS,GAAG,IAAI,CAAC,CAAC,CAAC,CAAA;IACzB,IAAI,CAAC,gBAAgB,CAAC,KAAK,EAAE,IAAI,CAAC,IAAI,SAAS,KAAK,SAAS,EAAE,CAAC;QAC9D,MAAM,IAAI,GAAG,iBAAiB,CAAC,KAAK,CAAC,CAAA;QACrC,OAAO,IAAI,CAAC,CAAC,CAAC,CAAC,EAAE,IAAI,EAAE,MAAM,EAAE,KAAK,EAAE,CAAC,CAAC,CAAC,CAAC,EAAE,CAAA;IAC9C,CAAC;IAED,6EAA6E;IAC7E,+EAA+E;IAC/E,MAAM,aAAa,GAAa,EAAE,CAAA;IAClC,MAAM,OAAO,GAAG,cAAc,CAAC,SAAS,CAAC,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE;QAC5D,MAAM,IAAI,GAAG,oBAAoB,CAAC,IAAI,CAAC,CAAA;QACvC,aAAa,CAAC,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,KAAK,CAAC,oBAAoB,CAAC,IAAI,EAAE,CAAC,CAAC,CAAA;QAC/D,OAAO,IAAI,CAAC,OAAO,CAAC,oBAAoB,EAAE,EAAE,CAAC,IAAI,UAAU,KAAK,GAAG,CAAC,EAAE,CAAA;IACxE,CAAC,CAAC,CAAA;IAEF,MAAM,SAAS,GAAG,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,GAAG,EAAE,EAAE;QAC9C,MAAM,MAAM,GAAG,cAAc,CAAC,GAAG,CAAC,CAAA;QAClC,MAAM,IAAI,GAAG,MAAM;aAChB,GAAG,CACF,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE,CAAC,GAAG,OAAO,CAAC,KAAK,CAAC,IAAI,UAAU,KAAK,GAAG,CAAC,EAAE,KAAK,oBAAoB,CAAC,IAAI,CAAC,EAAE,CAC7F;aACA,IAAI,CAAC,IAAI,CAAC,CAAA;QACb,OAAO,IAAI,CAAC,CAAC,CAAC,CAAC,EAAE,IAAI,EAAE,MAAM,EAAE,IAAI,EAAE,CAAC,CAAC,CAAC,CAAC,EAAE,CAAA;IAC7C,CAAC,CAAC,CAAA;IACF,MAAM,KAAK,GAAG,SAAS,CAAC,CAAC,CAAC,CAAA;IAC1B,IAAI,KAAK,IAAI,aAAa,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QACtC,KAAK,CAAC,IAAI,GAAG,aAAa,CAAC,IAAI,CAAC,EAAE,CAAC,GAAG,KAAK,CAAC,IAAI,CAAA;IAClD,CAAC;IACD,OAAO,SAAS,CAAA;AAClB,CAAC;AAED,SAAS,4BAA4B,CAAC,OAAgB;IACpD,OAAO,KAAK,CAAC,IAAI,CAAC,OAAO,CAAC,gBAAgB,CAAC,GAAG,CAAC,CAAC,CAAC,IAAI,CACnD,CAAC,UAAU,EAAE,EAAE,CACb,gBAAgB,CAAC,GAAG,CAAC,UAAU,CAAC,OAAO,CAAC;QACxC,SAAS,CAAC,GAAG,CAAC,UAAU,CAAC,OAAO,CAAC;QACjC,UAAU,CAAC,OAAO,KAAK,OAAO,CACjC,CAAA;AACH,CAAC;AAED,oFAAoF;AACpF,SAAS,kBAAkB,CAAC,OAAgB;IAC1C,IAAI,gBAAgB,CAAC,GAAG,CAAC,OAAO,CAAC,OAAO,CAAC,EAAE,CAAC;QAC1C,OAAO,mBAAmB,CAAA;IAC5B,CAAC;IACD,IAAI,SAAS,CAAC,GAAG,CAAC,OAAO,CAAC,OAAO,CAAC,EAAE,CAAC;QACnC,OAAO,aAAa,CAAA;IACtB,CAAC;IACD,OAAO,IAAI,CAAA;AACb,CAAC;AAED,SAAS,kBAAkB,CAAC,QAAkB;IAC5C,MAAM,MAAM,GAAmB,EAAE,CAAA;IAEjC,sCAAsC;IACtC,MAAM,QAAQ,GAAG,CAAC,IAAY,EAAE,GAAG,GAAG,EAAE,EAAQ,EAAE;QAChD,IAAI,IAAI,EAAE,CAAC;YACT,MAAM,CAAC,IAAI,CAAC;gBACV,IAAI;gBACJ,MAAM,EAAE,KAAK;gBACb,GAAG,CAAC,UAAU,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,EAAE,YAAY,EAAE,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;aAClE,CAAC,CAAA;QACJ,CAAC;IACH,CAAC,CAAA;IAED,MAAM,WAAW,GAAG,CAAC,OAAgB,EAAQ,EAAE;QAC7C,IAAI,YAAY,CAAC,GAAG,CAAC,OAAO,CAAC,OAAO,CAAC,EAAE,CAAC;YACtC,OAAM;QACR,CAAC;QACD,MAAM,SAAS,GAAG,kBAAkB,CAAC,OAAO,CAAC,CAAA;QAC7C,IAAI,SAAS,KAAK,IAAI,EAAE,CAAC;YACvB,QAAQ,CAAC,SAAS,CAAC,OAAO,CAAC,EAAE,OAAO,CAAC,OAAO,CAAC,CAAA;YAC7C,OAAM;QACR,CAAC;QACD,IAAI,OAAO,CAAC,OAAO,KAAK,OAAO,EAAE,CAAC;YAChC,MAAM,CAAC,IAAI,CAAC,GAAG,SAAS,CAAC,OAAO,CAAC,CAAC,CAAA;YAClC,OAAM;QACR,CAAC;QACD,IAAI,CAAC,4BAA4B,CAAC,OAAO,CAAC,EAAE,CAAC;YAC3C,QAAQ,CAAC,oBAAoB,CAAC,OAAO,CAAC,CAAC,CAAA;YACvC,OAAM;QACR,CAAC;QACD,KAAK,MAAM,KAAK,IAAI,OAAO,CAAC,UAAU,EAAE,CAAC;YACvC,QAAQ,CAAC,KAAK,CAAC,CAAA;QACjB,CAAC;IACH,CAAC,CAAA;IAED,MAAM,QAAQ,GAAG,CAAC,IAAU,EAAQ,EAAE;QACpC,IAAI,IAAI,CAAC,QAAQ,KAAK,CAAC,EAAE,CAAC;YACxB,QAAQ,CAAC,mBAAmB,CAAC,IAAI,CAAC,SAAS,IAAI,EAAE,CAAC,CAAC,CAAA;YACnD,OAAM;QACR,CAAC;QACD,IAAI,SAAS,CAAC,IAAI,CAAC,EAAE,CAAC;YACpB,WAAW,CAAC,IAAI,CAAC,CAAA;QACnB,CAAC;IACH,CAAC,CAAA;IAED,KAAK,MAAM,IAAI,IAAI,QAAQ,CAAC,IAAI,CAAC,UAAU,EAAE,CAAC;QAC5C,QAAQ,CAAC,IAAI,CAAC,CAAA;IAChB,CAAC;IACD,OAAO,MAAM,CAAA;AACf,CAAC;AAED,MAAM,CAAC,KAAK,UAAU,oBAAoB,CAAC,MAAc;IACvD,MAAM,GAAG,GAAG,MAAM,KAAK,CAAC,SAAS,CAAC,MAAM,CAAC,CAAA;IACzC,MAAM,cAAc,GAAG,GAAG,CAAC,IAAI,CAAC,mBAAmB,CAAC,CAAA;IACpD,IAAI,CAAC,cAAc,EAAE,CAAC;QACpB,OAAO,SAAS,CAAA;IAClB,CAAC;IAED,MAAM,GAAG,GAAG,MAAM,cAAc,CAAC,KAAK,CAAC,QAAQ,CAAC,CAAA;IAChD,IAAI,CAAC;QACH,MAAM,QAAQ,GAAG,IAAI,KAAK,CAAC,GAAG,EAAE,EAAE,WAAW,EAAE,UAAU,EAAE,CAAC,CAAC,MAAM,CAAC,QAAQ,CAAA;QAC5E,IAAI,QAAQ,CAAC,aAAa,CAAC,aAAa,CAAC,EAAE,CAAC;YAC1C,OAAO,SAAS,CAAA;QAClB,CAAC;QAED,MAAM,KAAK,GAAG,QAAQ,CAAC,sBAAsB,CAAC,oBAAoB,EAAE,OAAO,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAA;QACpF,MAAM,UAAU,GAAG,mBAAmB,CAAC,KAAK,EAAE,WAAW,IAAI,EAAE,CAAC,CAAA;QAChE,OAAO,UAAU,IAAI,SAAS,CAAA;IAChC,CAAC;IAAC,MAAM,CAAC;QACP,qEAAqE;QACrE,8EAA8E;QAC9E,OAAO,SAAS,CAAA;IAClB,CAAC;AACH,CAAC;AAED,uEAAuE;AACvE,SAAS,iBAAiB,CAAC,IAAY;IACrC,OAAO,IAAI,CAAC,OAAO,CAAC,oBAAoB,EAAE,EAAE,CAAC,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,GAAG,CAAC,mBAAmB,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,CAAA;AAC/F,CAAC;AAED,6EAA6E;AAC7E,SAAS,kBAAkB,CACzB,SAAiB,EACjB,OAAmC,EACnC,KAAa;IAEb,OAAO,OAAO,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC;QAC7B,MAAM,EAAE,KAAK,GAAG,iBAAiB,CAAC,SAAS,CAAC,KAAK,CAAC,CAAC,EAAE,KAAK,CAAC,KAAK,IAAI,CAAC,CAAC,CAAC,CAAC,MAAM;QAC9E,UAAU,EAAE,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC;KAC7B,CAAC,CAAC,CAAA;AACL,CAAC;AAED,6DAA6D;AAC7D,SAAS,aAAa,CAAC,YAAoC;IACzD,MAAM,IAAI,GAAG,IAAI,GAAG,EAAU,CAAA;IAC9B,OAAO,CAAC,GAAG,YAAY,CAAC;SACrB,IAAI,CAAC,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE,CAAC,IAAI,CAAC,UAAU,GAAG,KAAK,CAAC,UAAU,CAAC;SACzD,MAAM,CAAC,CAAC,MAAM,EAAE,EAAE;QACjB,IAAI,IAAI,CAAC,GAAG,CAAC,MAAM,CAAC,UAAU,CAAC,EAAE,CAAC;YAChC,OAAO,KAAK,CAAA;QACd,CAAC;QACD,IAAI,CAAC,GAAG,CAAC,MAAM,CAAC,UAAU,CAAC,CAAA;QAC3B,OAAO,IAAI,CAAA;IACb,CAAC,CAAC,CAAA;AACN,CAAC;AAOD,MAAM,UAAU,yBAAyB,CAAC,QAAkB;IAC1D,IAAI,OAAO,GAAG,EAAE,CAAA;IAChB,MAAM,QAAQ,GAAoB,EAAE,CAAA;IACpC,MAAM,YAAY,GAAsB,EAAE,CAAA;IAC1C,MAAM,YAAY,GAAkB,EAAE,CAAA;IAEtC,KAAK,MAAM,KAAK,IAAI,kBAAkB,CAAC,QAAQ,CAAC,EAAE,CAAC;QACjD,MAAM,OAAO,GAAG,CAAC,GAAG,KAAK,CAAC,IAAI,CAAC,QAAQ,CAAC,oBAAoB,CAAC,CAAC,CAAA;QAC9D,MAAM,SAAS,GAAG,OAAO,CAAC,MAAM,KAAK,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC,iBAAiB,CAAC,KAAK,CAAC,IAAI,CAAC,CAAA;QAEnF,sEAAsE;QACtE,qDAAqD;QACrD,IAAI,OAAO,CAAC,MAAM,GAAG,CAAC,IAAI,CAAC,SAAS,EAAE,CAAC;YACrC,YAAY,CAAC,IAAI,CACf,GAAG,OAAO,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC,EAAE,MAAM,EAAE,OAAO,CAAC,MAAM,EAAE,UAAU,EAAE,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CACtF,CAAA;YACD,SAAQ;QACV,CAAC;QAED,IAAI,OAAO,EAAE,CAAC;YACZ,OAAO,IAAI,MAAM,CAAA;QACnB,CAAC;QACD,MAAM,KAAK,GAAG,OAAO,CAAC,MAAM,CAAA;QAC5B,IAAI,KAAK,CAAC,YAAY,EAAE,CAAC;YACvB,QAAQ,CAAC,IAAI,CAAC,EAAE,MAAM,EAAE,KAAK,EAAE,KAAK,EAAE,KAAK,CAAC,YAAY,EAAE,IAAI,EAAE,SAAS,EAAE,CAAC,CAAA;QAC9E,CAAC;QACD,YAAY,CAAC,IAAI,CAAC,GAAG,kBAAkB,CAAC,KAAK,CAAC,IAAI,EAAE,OAAO,EAAE,KAAK,CAAC,CAAC,CAAA;QACpE,OAAO,IAAI,SAAS,CAAA;QACpB,IAAI,KAAK,CAAC,MAAM,EAAE,CAAC;YACjB,YAAY,CAAC,IAAI,CAAC,EAAE,KAAK,EAAE,GAAG,EAAE,OAAO,CAAC,MAAM,EAAE,CAAC,CAAA;QACnD,CAAC;IACH,CAAC;IAED,MAAM,kBAAkB,GAAG,aAAa,CAAC,YAAY,CAAC,CAAA;IACtD,OAAO;QACL,OAAO;QACP,GAAG,CAAC,QAAQ,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,SAAS,EAAE,EAAE,QAAQ,EAAE,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;QACvD,YAAY;QACZ,GAAG,CAAC,kBAAkB,CAAC,MAAM,KAAK,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,YAAY,EAAE,kBAAkB,EAAE,CAAC;KACjF,CAAA;AACH,CAAC"}
|
package/dist/parser/index.d.ts
CHANGED
|
@@ -1,6 +1,7 @@
|
|
|
1
1
|
import type { Document as MupdfDocument } from 'mupdf';
|
|
2
2
|
import { type AtomicTextRange } from '../chunker/index.js';
|
|
3
3
|
import { AppError } from '../utils/errors.js';
|
|
4
|
+
import type { HeadingAnchor, SourceMap } from '../utils/source-context.js';
|
|
4
5
|
import type { EmbedderInterface, FilteredTextFragment } from './pdf-filter.js';
|
|
5
6
|
/**
|
|
6
7
|
* File extensions supported by the parser module (parseFile + parsePdf).
|
|
@@ -10,11 +11,18 @@ import type { EmbedderInterface, FilteredTextFragment } from './pdf-filter.js';
|
|
|
10
11
|
export declare const SUPPORTED_EXTENSIONS: Set<string>;
|
|
11
12
|
/**
|
|
12
13
|
* Result from parsing a document, containing both content and extracted title.
|
|
13
|
-
* Title is display
|
|
14
|
+
* Title is display metadata; it reaches search only through the opt-in
|
|
15
|
+
* embedding title prefix.
|
|
14
16
|
*/
|
|
15
17
|
export interface ParseResult {
|
|
18
|
+
sourceMap?: SourceMap;
|
|
16
19
|
content: string;
|
|
17
20
|
title: string;
|
|
21
|
+
/**
|
|
22
|
+
* Title for the embedding prefix; `title` is used when absent. PDFs set it
|
|
23
|
+
* so a display title taken from page-1 body text embeds as the file name.
|
|
24
|
+
*/
|
|
25
|
+
embeddingTitle?: string;
|
|
18
26
|
atomicRanges?: readonly AtomicTextRange[];
|
|
19
27
|
imageAnchors?: readonly ParsedImageAnchor[];
|
|
20
28
|
}
|
|
@@ -105,8 +113,10 @@ export declare class DocumentParser {
|
|
|
105
113
|
parsePdfPages(filePath: string, embedder: EmbedderInterface): Promise<{
|
|
106
114
|
doc: MupdfDocument;
|
|
107
115
|
title: string;
|
|
116
|
+
embeddingTitle: string;
|
|
108
117
|
pages: Array<{
|
|
109
118
|
pageNum: number;
|
|
119
|
+
headings?: HeadingAnchor[];
|
|
110
120
|
text: string;
|
|
111
121
|
textFragments: FilteredTextFragment[];
|
|
112
122
|
stextJson: unknown;
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../../src/parser/index.ts"],"names":[],"mappings":"AAOA,OAAO,KAAK,EAAE,QAAQ,IAAI,aAAa,EAAE,MAAM,OAAO,CAAA;AACtD,OAAO,EAAE,KAAK,eAAe,EAAmB,MAAM,qBAAqB,CAAA;AAE3E,OAAO,EAAE,QAAQ,EAAuB,MAAM,oBAAoB,CAAA;
|
|
1
|
+
{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../../src/parser/index.ts"],"names":[],"mappings":"AAOA,OAAO,KAAK,EAAE,QAAQ,IAAI,aAAa,EAAE,MAAM,OAAO,CAAA;AACtD,OAAO,EAAE,KAAK,eAAe,EAAmB,MAAM,qBAAqB,CAAA;AAE3E,OAAO,EAAE,QAAQ,EAAuB,MAAM,oBAAoB,CAAA;AAClE,OAAO,KAAK,EAAE,aAAa,EAAE,SAAS,EAAE,MAAM,4BAA4B,CAAA;AAI1E,OAAO,KAAK,EAAE,iBAAiB,EAAE,oBAAoB,EAAE,MAAM,iBAAiB,CAAA;AAa9E;;;;GAIG;AACH,eAAO,MAAM,oBAAoB,aAA4C,CAAA;AAM7E;;;;GAIG;AACH,MAAM,WAAW,WAAW;IAC1B,SAAS,CAAC,EAAE,SAAS,CAAA;IACrB,OAAO,EAAE,MAAM,CAAA;IACf,KAAK,EAAE,MAAM,CAAA;IACb;;;OAGG;IACH,cAAc,CAAC,EAAE,MAAM,CAAA;IACvB,YAAY,CAAC,EAAE,SAAS,eAAe,EAAE,CAAA;IACzC,YAAY,CAAC,EAAE,SAAS,iBAAiB,EAAE,CAAA;CAC5C;AAED,MAAM,WAAW,iBAAiB;IAChC,MAAM,EAAE,MAAM,CAAA;IACd,UAAU,EAAE,MAAM,CAAA;IAClB,QAAQ,EAAE,WAAW,GAAG,YAAY,CAAA;IACpC,KAAK,EAAE,UAAU,CAAA;CAClB;AAED,MAAM,WAAW,gBAAgB;IAC/B,MAAM,CAAC,EAAE,OAAO,CAAA;CACjB;AAyCD;;;;GAIG;AACH,MAAM,MAAM,YAAY,GACpB;IACE,8DAA8D;IAC9D,OAAO,EAAE,MAAM,CAAA;IACf,QAAQ,CAAC,EAAE,SAAS,CAAA;IACpB,iCAAiC;IACjC,WAAW,EAAE,MAAM,CAAA;CACpB,GACD;IACE,yEAAyE;IACzE,QAAQ,EAAE,SAAS,MAAM,EAAE,CAAA;IAC3B,OAAO,CAAC,EAAE,SAAS,CAAA;IACnB,iCAAiC;IACjC,WAAW,EAAE,MAAM,CAAA;CACpB,CAAA;AAEL;;GAEG;AACH,qBAAa,eAAgB,SAAQ,QAAQ;IAC3C,YAAY,OAAO,EAAE,MAAM,EAAE,OAAO,CAAC,EAAE;QAAE,KAAK,CAAC,EAAE,KAAK,CAAA;KAAE,EAGvD;CACF;AAED;;GAEG;AACH,qBAAa,kBAAmB,SAAQ,QAAQ;IAC9C,YAAY,OAAO,EAAE,MAAM,EAAE,OAAO,CAAC,EAAE;QAAE,KAAK,CAAC,EAAE,KAAK,CAAA;KAAE,EAGvD;CACF;AAMD,6DAA6D;AAC7D,qBAAa,cAAc;IACzB,OAAO,CAAC,QAAQ,CAAC,MAAM,CAAc;IACrC,yEAAyE;IACzE,OAAO,CAAC,QAAQ,CAAC,WAAW,CAAmB;IAC/C;;;;OAIG;IACH,OAAO,CAAC,gBAAgB,CAAwB;IAEhD,YAAY,MAAM,EAAE,YAAY,EAW/B;IAED;;;;;;;;OAQG;IACG,gBAAgB,CAAC,QAAQ,EAAE,MAAM,GAAG,OAAO,CAAC,IAAI,CAAC,CAkEtD;IAED,+EAA+E;IAC/E,gBAAgB,CAAC,QAAQ,EAAE,MAAM,GAAG,IAAI,CAoBvC;IAED,6DAA6D;IACvD,SAAS,CAAC,QAAQ,EAAE,MAAM,EAAE,OAAO,GAAE,gBAAqB,GAAG,OAAO,CAAC,WAAW,CAAC,CAiBtF;IAED;;;;OAIG;IACG,QAAQ,CAAC,QAAQ,EAAE,MAAM,EAAE,QAAQ,EAAE,iBAAiB,GAAG,OAAO,CAAC,WAAW,CAAC,CA4DlF;IAED;;;;;;;;OAQG;IACG,aAAa,CACjB,QAAQ,EAAE,MAAM,EAChB,QAAQ,EAAE,iBAAiB,GAC1B,OAAO,CAAC;QACT,GAAG,EAAE,aAAa,CAAA;QAClB,KAAK,EAAE,MAAM,CAAA;QACb,cAAc,EAAE,MAAM,CAAA;QACtB,KAAK,EAAE,KAAK,CAAC;YACX,OAAO,EAAE,MAAM,CAAA;YACf,QAAQ,CAAC,EAAE,aAAa,EAAE,CAAA;YAC1B,IAAI,EAAE,MAAM,CAAA;YACZ,aAAa,EAAE,oBAAoB,EAAE,CAAA;YACrC,SAAS,EAAE,OAAO,CAAA;SACnB,CAAC,CAAA;KACH,CAAC,CAsDD;IAED;;;OAGG;YACW,SAAS;IA6DvB,mBAAmB;YACL,QAAQ;IAYtB,kBAAkB;YACJ,OAAO;CAWtB"}
|
package/dist/parser/index.js
CHANGED
|
@@ -10,6 +10,7 @@ import { AppError, isAppError, toError } from '../utils/errors.js';
|
|
|
10
10
|
import { errorCode } from '../utils/type-guards.js';
|
|
11
11
|
import { convertDocxDocumentToText, extractDocxCoreTitle } from './docx-parser.js';
|
|
12
12
|
import { extractPdfPages } from './pdf-extract.js';
|
|
13
|
+
import { textSourceMap } from './text-source-map.js';
|
|
13
14
|
import { extractDocxTitle, extractMarkdownTitle, extractPdfTitle, extractTxtTitle, } from './title-extractor.js';
|
|
14
15
|
// ============================================
|
|
15
16
|
// Supported Extensions
|
|
@@ -37,7 +38,10 @@ async function resolvePdfTitle(filePath, pages, hints, embedder) {
|
|
|
37
38
|
}
|
|
38
39
|
console.error(`Title extraction failed, falling back to filename: ${titleError}`);
|
|
39
40
|
}
|
|
40
|
-
return
|
|
41
|
+
return {
|
|
42
|
+
title: extractPdfTitle(metadataTitle, firstPageChunkText, fileName, page1FontHint).title,
|
|
43
|
+
embeddingTitle: extractPdfTitle(metadataTitle, undefined, fileName, page1FontHint).title,
|
|
44
|
+
};
|
|
41
45
|
}
|
|
42
46
|
/**
|
|
43
47
|
* Validation error (equivalent to 400)
|
|
@@ -203,13 +207,23 @@ export class DocumentParser {
|
|
|
203
207
|
const mupdf = await import('mupdf');
|
|
204
208
|
doc = mupdf.Document.openDocument(buffer, 'application/pdf');
|
|
205
209
|
const { pages, metadataTitle, page1FontHint } = await extractPdfPages(doc, embedder, 'preserve-whitespace');
|
|
206
|
-
|
|
207
|
-
|
|
208
|
-
|
|
209
|
-
.
|
|
210
|
-
|
|
210
|
+
let text = '';
|
|
211
|
+
const sourceMap = { headings: [], pages: [] };
|
|
212
|
+
for (const page of pages) {
|
|
213
|
+
if (!page.text) {
|
|
214
|
+
continue;
|
|
215
|
+
}
|
|
216
|
+
if (text) {
|
|
217
|
+
text += '\n\n';
|
|
218
|
+
}
|
|
219
|
+
const start = text.length;
|
|
220
|
+
text += page.text;
|
|
221
|
+
sourceMap.pages?.push({ start, end: text.length, page: page.pageNum });
|
|
222
|
+
sourceMap.headings.push(...page.headings.map((heading) => ({ ...heading, offset: start + heading.offset })));
|
|
223
|
+
}
|
|
224
|
+
const { title, embeddingTitle } = await resolvePdfTitle(filePath, pages, { metadataTitle, page1FontHint }, embedder);
|
|
211
225
|
console.error(`Parsed PDF: ${filePath} (${text.length} characters, ${pages.length} pages)`);
|
|
212
|
-
return { content: text, title };
|
|
226
|
+
return { content: text, title, embeddingTitle, sourceMap };
|
|
213
227
|
}
|
|
214
228
|
catch (error) {
|
|
215
229
|
// A foreign domain error (an `EmbeddingError` raised while the parser
|
|
@@ -254,13 +268,14 @@ export class DocumentParser {
|
|
|
254
268
|
const { pages: helperPages, metadataTitle, page1FontHint } = extracted;
|
|
255
269
|
const pages = helperPages.map((page) => ({
|
|
256
270
|
pageNum: page.pageNum,
|
|
271
|
+
headings: page.headings,
|
|
257
272
|
text: page.text,
|
|
258
273
|
textFragments: page.textFragments,
|
|
259
274
|
stextJson: page.stextJson,
|
|
260
275
|
}));
|
|
261
|
-
const title = await resolvePdfTitle(filePath, helperPages, { metadataTitle, page1FontHint }, embedder);
|
|
276
|
+
const { title, embeddingTitle } = await resolvePdfTitle(filePath, helperPages, { metadataTitle, page1FontHint }, embedder);
|
|
262
277
|
console.error(`Parsed PDF pages: ${filePath} (${pages.length} pages; caller owns doc disposal)`);
|
|
263
|
-
return { doc, title, pages };
|
|
278
|
+
return { doc, title, embeddingTitle, pages };
|
|
264
279
|
}
|
|
265
280
|
catch (error) {
|
|
266
281
|
// `doc` is undefined when `openDocument` itself threw — nothing to free.
|
|
@@ -325,6 +340,7 @@ export class DocumentParser {
|
|
|
325
340
|
});
|
|
326
341
|
return {
|
|
327
342
|
content: body.content,
|
|
343
|
+
...(body.sourceMap ? { sourceMap: body.sourceMap } : {}),
|
|
328
344
|
title: titleResult.title,
|
|
329
345
|
...(body.atomicRanges.length === 0 ? {} : { atomicRanges: body.atomicRanges }),
|
|
330
346
|
...(imageAnchors.length === 0 ? {} : { imageAnchors }),
|
|
@@ -341,7 +357,7 @@ export class DocumentParser {
|
|
|
341
357
|
const fileName = basename(filePath);
|
|
342
358
|
const titleResult = extractTxtTitle(text, fileName);
|
|
343
359
|
console.error(`Parsed TXT: ${filePath} (${text.length} characters)`);
|
|
344
|
-
return { content: text, title: titleResult.title };
|
|
360
|
+
return { content: text, title: titleResult.title, sourceMap: textSourceMap(text, false) };
|
|
345
361
|
}
|
|
346
362
|
catch (error) {
|
|
347
363
|
throw new FileOperationError(`Failed to parse TXT: ${filePath}`, { cause: toError(error) });
|
|
@@ -354,7 +370,7 @@ export class DocumentParser {
|
|
|
354
370
|
const fileName = basename(filePath);
|
|
355
371
|
const titleResult = extractMarkdownTitle(text, fileName);
|
|
356
372
|
console.error(`Parsed MD: ${filePath} (${text.length} characters)`);
|
|
357
|
-
return { content: text, title: titleResult.title };
|
|
373
|
+
return { content: text, title: titleResult.title, sourceMap: textSourceMap(text, true) };
|
|
358
374
|
}
|
|
359
375
|
catch (error) {
|
|
360
376
|
throw new FileOperationError(`Failed to parse MD: ${filePath}`, { cause: toError(error) });
|
package/dist/parser/index.js.map
CHANGED
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"index.js","sourceRoot":"","sources":["../../src/parser/index.ts"],"names":[],"mappings":"AAAA,6DAA6D;AAE7D,OAAO,EAAE,QAAQ,EAAE,MAAM,SAAS,CAAA;AAClC,OAAO,EAAE,KAAK,EAAE,QAAQ,EAAE,QAAQ,EAAE,MAAM,kBAAkB,CAAA;AAC5D,OAAO,EAAE,QAAQ,EAAE,OAAO,EAAE,UAAU,EAAE,OAAO,EAAE,MAAM,WAAW,CAAA;AAClE,OAAO,EAAE,KAAK,EAAE,MAAM,OAAO,CAAA;AAC7B,OAAO,OAAO,MAAM,SAAS,CAAA;AAE7B,OAAO,EAAwB,eAAe,EAAE,MAAM,qBAAqB,CAAA;AAC3E,OAAO,EAAE,qBAAqB,EAAE,MAAM,uBAAuB,CAAA;AAC7D,OAAO,EAAE,QAAQ,EAAE,UAAU,EAAE,OAAO,EAAE,MAAM,oBAAoB,CAAA;AAClE,OAAO,EAAE,SAAS,EAAE,MAAM,yBAAyB,CAAA;AACnD,OAAO,EAAE,yBAAyB,EAAE,oBAAoB,EAAE,MAAM,kBAAkB,CAAA;AAClF,OAAO,EAAE,eAAe,EAAE,MAAM,kBAAkB,CAAA;AAElD,OAAO,EACL,gBAAgB,EAChB,oBAAoB,EACpB,eAAe,EACf,eAAe,GAChB,MAAM,sBAAsB,CAAA;AAE7B,+CAA+C;AAC/C,uBAAuB;AACvB,+CAA+C;AAE/C;;;;GAIG;AACH,MAAM,CAAC,MAAM,oBAAoB,GAAG,IAAI,GAAG,CAAC,CAAC,MAAM,EAAE,OAAO,EAAE,MAAM,EAAE,KAAK,CAAC,CAAC,CAAA;AAkC7E,KAAK,UAAU,eAAe,CAC5B,QAAgB,EAChB,KAAkC,EAClC,KAAoB,EACpB,QAA2B;IAE3B,MAAM,EAAE,aAAa,EAAE,aAAa,EAAE,GAAG,KAAK,CAAA;IAC9C,MAAM,QAAQ,GAAG,QAAQ,CAAC,QAAQ,CAAC,CAAA;IACnC,IAAI,kBAAsC,CAAA;IAC1C,IAAI,CAAC;QACH,MAAM,aAAa,GAAG,KAAK,CAAC,CAAC,CAAC,EAAE,IAAI,CAAA;QACpC,IAAI,aAAa,IAAI,aAAa,CAAC,IAAI,EAAE,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;YACrD,MAAM,WAAW,GAAG,MAAM,IAAI,eAAe,EAAE,CAAC,SAAS,CAAC,aAAa,EAAE,QAAQ,CAAC,CAAA;YAClF,kBAAkB,GAAG,WAAW,CAAC,CAAC,CAAC,EAAE,IAAI,CAAA;QAC3C,CAAC;IACH,CAAC;IAAC,OAAO,UAAU,EAAE,CAAC;QACpB,IAAI,UAAU,CAAC,UAAU,CAAC,EAAE,CAAC;YAC3B,MAAM,UAAU,CAAA;QAClB,CAAC;QACD,OAAO,CAAC,KAAK,CAAC,sDAAsD,UAAU,EAAE,CAAC,CAAA;IACnF,CAAC;IACD,OAAO,eAAe,CAAC,aAAa,EAAE,kBAAkB,EAAE,QAAQ,EAAE,aAAa,CAAC,CAAC,KAAK,CAAA;AAC1F,CAAC;AAuBD;;GAEG;AACH,MAAM,OAAO,eAAgB,SAAQ,QAAQ;IAC3C,YAAY,OAAe,EAAE,OAA2B;QACtD,KAAK,CAAC,OAAO,EAAE,QAAQ,EAAE,YAAY,EAAE,OAAO,CAAC,CAAA;QAC/C,IAAI,CAAC,IAAI,GAAG,iBAAiB,CAAA;IAC/B,CAAC;CACF;AAED;;GAEG;AACH,MAAM,OAAO,kBAAmB,SAAQ,QAAQ;IAC9C,YAAY,OAAe,EAAE,OAA2B;QACtD,KAAK,CAAC,OAAO,EAAE,QAAQ,EAAE,IAAI,EAAE,OAAO,CAAC,CAAA;QACvC,IAAI,CAAC,IAAI,GAAG,oBAAoB,CAAA;IAClC,CAAC;CACF;AAED,+CAA+C;AAC/C,uBAAuB;AACvB,+CAA+C;AAE/C,6DAA6D;AAC7D,MAAM,OAAO,cAAc;IACR,MAAM,CAAc;IACrC,yEAAyE;IACxD,WAAW,CAAmB;IAC/C;;;;OAIG;IACK,gBAAgB,GAAoB,IAAI,CAAA;IAEhD,YAAY,MAAoB;QAC9B,IAAI,CAAC,MAAM,GAAG,MAAM,CAAA;QACpB,yEAAyE;QACzE,oEAAoE;QACpE,uEAAuE;QACvE,mCAAmC;QACnC,IAAI,MAAM,CAAC,QAAQ,KAAK,SAAS,EAAE,CAAC;YAClC,IAAI,CAAC,WAAW,GAAG,MAAM,CAAC,QAAQ,CAAA;QACpC,CAAC;aAAM,CAAC;YACN,IAAI,CAAC,WAAW,GAAG,CAAC,MAAM,CAAC,OAAO,CAAC,CAAA;QACrC,CAAC;IACH,CAAC;IAED;;;;;;;;OAQG;IACH,KAAK,CAAC,gBAAgB,CAAC,QAAgB;QACrC,sEAAsE;QACtE,yEAAyE;QACzE,gDAAgD;QAChD,IAAI,IAAI,CAAC,WAAW,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;YAClC,MAAM,IAAI,eAAe,CACvB,4JAA4J,CAC7J,CAAA;QACH,CAAC;QAED,wDAAwD;QACxD,IAAI,CAAC,UAAU,CAAC,QAAQ,CAAC,EAAE,CAAC;YAC1B,MAAM,IAAI,eAAe,CACvB,8CAA8C,QAAQ,wGAAwG,CAC/J,CAAA;QACH,CAAC;QAED,uEAAuE;QACvE,gEAAgE;QAChE,+CAA+C;QAC/C,IAAI,CAAC,IAAI,CAAC,gBAAgB,EAAE,CAAC;YAC3B,MAAM,YAAY,GAAa,EAAE,CAAA;YACjC,KAAK,MAAM,GAAG,IAAI,IAAI,CAAC,WAAW,EAAE,CAAC;gBACnC,MAAM,QAAQ,GAAG,MAAM,QAAQ,CAAC,OAAO,CAAC,GAAG,CAAC,CAAC,CAAA;gBAC7C,YAAY,CAAC,IAAI,CAAC,qBAAqB,CAAC,QAAQ,CAAC,CAAC,CAAA;YACpD,CAAC;YACD,IAAI,CAAC,gBAAgB,GAAG,YAAY,CAAA;QACtC,CAAC;QAED,uDAAuD;QACvD,IAAI,YAAoB,CAAA;QACxB,IAAI,CAAC;YACH,YAAY,GAAG,MAAM,QAAQ,CAAC,QAAQ,CAAC,CAAA;QACzC,CAAC;QAAC,OAAO,KAAK,EAAE,CAAC;YACf,sDAAsD;YACtD,iEAAiE;YACjE,oEAAoE;YACpE,4EAA4E;YAC5E,MAAM,SAAS,GAAG,MAAM,KAAK,CAAC,QAAQ,CAAC;iBACpC,IAAI,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,cAAc,EAAE,CAAC;iBACvC,KAAK,CAAC,GAAG,EAAE,CAAC,KAAK,CAAC,CAAA;YAErB,IAAI,SAAS,EAAE,CAAC;gBACd,MAAM,IAAI,eAAe,CACvB,6BAA6B,QAAQ,kDAAkD,EACvF,EAAE,KAAK,EAAE,OAAO,CAAC,KAAK,CAAC,EAAE,CAC1B,CAAA;YACH,CAAC;YAED,0EAA0E;YAC1E,kFAAkF;YAClF,2FAA2F;YAC3F,YAAY,GAAG,OAAO,CAAC,QAAQ,CAAC,CAAA;QAClC,CAAC;QAED,qDAAqD;QACrD,MAAM,OAAO,GAAG,IAAI,CAAC,gBAAgB,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,YAAY,CAAC,UAAU,CAAC,IAAI,CAAC,CAAC,CAAA;QACnF,IAAI,CAAC,OAAO,EAAE,CAAC;YACb,MAAM,YAAY,GAChB,IAAI,CAAC,gBAAgB,CAAC,MAAM,KAAK,CAAC;gBAChC,CAAC,CAAC,IAAI,CAAC,gBAAgB,CAAC,CAAC,CAAC;gBAC1B,CAAC,CAAC,IAAI,CAAC,gBAAgB,CAAC,IAAI,CAAC,IAAI,CAAC,CAAA;YACtC,MAAM,IAAI,eAAe,CACvB,wGAAwG,YAAY,iDAAiD,QAAQ,EAAE,CAChL,CAAA;QACH,CAAC;IACH,CAAC;IAED,+EAA+E;IAC/E,gBAAgB,CAAC,QAAgB;QAC/B,IAAI,CAAC;YACH,MAAM,KAAK,GAAG,QAAQ,CAAC,QAAQ,CAAC,CAAA;YAChC,IAAI,KAAK,CAAC,IAAI,GAAG,IAAI,CAAC,MAAM,CAAC,WAAW,EAAE,CAAC;gBACzC,MAAM,IAAI,eAAe,CACvB,4BAA4B,KAAK,CAAC,IAAI,MAAM,IAAI,CAAC,MAAM,CAAC,WAAW,EAAE,CACtE,CAAA;YACH,CAAC;QACH,CAAC;QAAC,OAAO,KAAK,EAAE,CAAC;YACf,IAAI,KAAK,YAAY,eAAe,EAAE,CAAC;gBACrC,MAAM,KAAK,CAAA;YACb,CAAC;YACD,oDAAoD;YACpD,IAAI,SAAS,CAAC,KAAK,CAAC,KAAK,QAAQ,EAAE,CAAC;gBAClC,MAAM,IAAI,eAAe,CAAC,mBAAmB,QAAQ,EAAE,CAAC,CAAA;YAC1D,CAAC;YACD,MAAM,IAAI,kBAAkB,CAAC,8BAA8B,QAAQ,EAAE,EAAE;gBACrE,KAAK,EAAE,OAAO,CAAC,KAAK,CAAC;aACtB,CAAC,CAAA;QACJ,CAAC;IACH,CAAC;IAED,6DAA6D;IAC7D,KAAK,CAAC,SAAS,CAAC,QAAgB,EAAE,OAAO,GAAqB,EAAE;QAC9D,aAAa;QACb,MAAM,IAAI,CAAC,gBAAgB,CAAC,QAAQ,CAAC,CAAA;QACrC,IAAI,CAAC,gBAAgB,CAAC,QAAQ,CAAC,CAAA;QAE/B,gDAAgD;QAChD,MAAM,GAAG,GAAG,OAAO,CAAC,QAAQ,CAAC,CAAC,WAAW,EAAE,CAAA;QAC3C,QAAQ,GAAG,EAAE,CAAC;YACZ,KAAK,OAAO;gBACV,OAAO,MAAM,IAAI,CAAC,SAAS,CAAC,QAAQ,EAAE,OAAO,CAAC,MAAM,KAAK,IAAI,CAAC,CAAA;YAChE,KAAK,MAAM;gBACT,OAAO,MAAM,IAAI,CAAC,QAAQ,CAAC,QAAQ,CAAC,CAAA;YACtC,KAAK,KAAK;gBACR,OAAO,MAAM,IAAI,CAAC,OAAO,CAAC,QAAQ,CAAC,CAAA;YACrC;gBACE,MAAM,IAAI,eAAe,CAAC,4BAA4B,GAAG,EAAE,CAAC,CAAA;QAChE,CAAC;IACH,CAAC;IAED;;;;OAIG;IACH,KAAK,CAAC,QAAQ,CAAC,QAAgB,EAAE,QAA2B;QAC1D,aAAa;QACb,MAAM,IAAI,CAAC,gBAAgB,CAAC,QAAQ,CAAC,CAAA;QACrC,IAAI,CAAC,gBAAgB,CAAC,QAAQ,CAAC,CAAA;QAE/B,yEAAyE;QACzE,sEAAsE;QACtE,0EAA0E;QAC1E,uDAAuD;QACvD,IAAI,GAA8B,CAAA;QAClC,IAAI,CAAC;YACH,MAAM,MAAM,GAAG,MAAM,QAAQ,CAAC,QAAQ,CAAC,CAAA;YACvC,MAAM,KAAK,GAAG,MAAM,MAAM,CAAC,OAAO,CAAC,CAAA;YACnC,GAAG,GAAG,KAAK,CAAC,QAAQ,CAAC,YAAY,CAAC,MAAM,EAAE,iBAAiB,CAAC,CAAA;YAE5D,MAAM,EAAE,KAAK,EAAE,aAAa,EAAE,aAAa,EAAE,GAAG,MAAM,eAAe,CACnE,GAAG,EACH,QAAQ,EACR,qBAAqB,CACtB,CAAA;YACD,MAAM,IAAI,GAAG,KAAK;iBACf,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC;iBAClB,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,MAAM,GAAG,CAAC,CAAC;iBAC3B,IAAI,CAAC,MAAM,CAAC,CAAA;YAEf,MAAM,KAAK,GAAG,MAAM,eAAe,CACjC,QAAQ,EACR,KAAK,EACL,EAAE,aAAa,EAAE,aAAa,EAAE,EAChC,QAAQ,CACT,CAAA;YAED,OAAO,CAAC,KAAK,CAAC,eAAe,QAAQ,KAAK,IAAI,CAAC,MAAM,gBAAgB,KAAK,CAAC,MAAM,SAAS,CAAC,CAAA;YAE3F,OAAO,EAAE,OAAO,EAAE,IAAI,EAAE,KAAK,EAAE,CAAA;QACjC,CAAC;QAAC,OAAO,KAAK,EAAE,CAAC;YACf,sEAAsE;YACtE,uEAAuE;YACvE,4EAA4E;YAC5E,IAAI,UAAU,CAAC,KAAK,CAAC,EAAE,CAAC;gBACtB,MAAM,KAAK,CAAA;YACb,CAAC;YACD,MAAM,IAAI,kBAAkB,CAAC,wBAAwB,QAAQ,EAAE,EAAE,EAAE,KAAK,EAAE,OAAO,CAAC,KAAK,CAAC,EAAE,CAAC,CAAA;QAC7F,CAAC;gBAAS,CAAC;YACT,sEAAsE;YACtE,2BAA2B;YAC3B,GAAG,EAAE,OAAO,EAAE,CAAA;QAChB,CAAC;IACH,CAAC;IAED;;;;;;;;OAQG;IACH,KAAK,CAAC,aAAa,CACjB,QAAgB,EAChB,QAA2B;QAW3B,yEAAyE;QACzE,2CAA2C;QAC3C,MAAM,IAAI,CAAC,gBAAgB,CAAC,QAAQ,CAAC,CAAA;QACrC,IAAI,CAAC,gBAAgB,CAAC,QAAQ,CAAC,CAAA;QAE/B,qEAAqE;QACrE,+BAA+B;QAC/B,yEAAyE;QACzE,0EAA0E;QAC1E,uEAAuE;QACvE,IAAI,GAA8B,CAAA;QAClC,IAAI,CAAC;YACH,MAAM,MAAM,GAAG,MAAM,QAAQ,CAAC,QAAQ,CAAC,CAAA;YACvC,MAAM,KAAK,GAAG,MAAM,MAAM,CAAC,OAAO,CAAC,CAAA;YACnC,GAAG,GAAG,KAAK,CAAC,QAAQ,CAAC,YAAY,CAAC,MAAM,EAAE,iBAAiB,CAAC,CAAA;YAC5D,MAAM,SAAS,GAAG,MAAM,eAAe,CAAC,GAAG,EAAE,QAAQ,EAAE,qCAAqC,CAAC,CAAA;YAE7F,MAAM,EAAE,KAAK,EAAE,WAAW,EAAE,aAAa,EAAE,aAAa,EAAE,GAAG,SAAS,CAAA;YACtE,MAAM,KAAK,GAAG,WAAW,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC;gBACvC,OAAO,EAAE,IAAI,CAAC,OAAO;gBACrB,IAAI,EAAE,IAAI,CAAC,IAAI;gBACf,aAAa,EAAE,IAAI,CAAC,aAAa;gBACjC,SAAS,EAAE,IAAI,CAAC,SAAS;aAC1B,CAAC,CAAC,CAAA;YACH,MAAM,KAAK,GAAG,MAAM,eAAe,CACjC,QAAQ,EACR,WAAW,EACX,EAAE,aAAa,EAAE,aAAa,EAAE,EAChC,QAAQ,CACT,CAAA;YAED,OAAO,CAAC,KAAK,CACX,qBAAqB,QAAQ,KAAK,KAAK,CAAC,MAAM,mCAAmC,CAClF,CAAA;YAED,OAAO,EAAE,GAAG,EAAE,KAAK,EAAE,KAAK,EAAE,CAAA;QAC9B,CAAC;QAAC,OAAO,KAAK,EAAE,CAAC;YACf,yEAAyE;YACzE,0EAA0E;YAC1E,uEAAuE;YACvE,qCAAqC;YACrC,GAAG,EAAE,OAAO,EAAE,CAAA;YACd,sEAAsE;YACtE,uEAAuE;YACvE,uDAAuD;YACvD,IAAI,UAAU,CAAC,KAAK,CAAC,EAAE,CAAC;gBACtB,MAAM,KAAK,CAAA;YACb,CAAC;YACD,MAAM,IAAI,kBAAkB,CAAC,8BAA8B,QAAQ,EAAE,EAAE;gBACrE,KAAK,EAAE,OAAO,CAAC,KAAK,CAAC;aACtB,CAAC,CAAA;QACJ,CAAC;IACH,CAAC;IAED;;;OAGG;IACK,KAAK,CAAC,SAAS,CAAC,QAAgB,EAAE,aAAsB;QAC9D,IAAI,CAAC;YACH,MAAM,MAAM,GAAG,MAAM,QAAQ,CAAC,QAAQ,CAAC,CAAA;YACvC,MAAM,cAAc,GAAG,IAAI,GAAG,EAG3B,CAAA;YACH,IAAI,cAAc,GAAG,CAAC,CAAA;YACtB,IAAI,iBAAiB,GAAG,CAAC,CAAA;YACzB,MAAM,UAAU,GAAG,aAAa;gBAC9B,CAAC,CAAC,MAAM,OAAO,CAAC,aAAa,CACzB,EAAE,MAAM,EAAE,EACV;oBACE,YAAY,EAAE,OAAO,CAAC,MAAM,CAAC,UAAU,CAAC,KAAK,EAAE,KAAK,EAAE,EAAE;wBACtD,MAAM,UAAU,GAAG,cAAc,EAAE,CAAA;wBACnC,IAAI,KAAK,CAAC,WAAW,KAAK,WAAW,IAAI,KAAK,CAAC,WAAW,KAAK,YAAY,EAAE,CAAC;4BAC5E,iBAAiB,IAAI,CAAC,CAAA;4BACtB,OAAO,EAAE,GAAG,EAAE,QAAQ,KAAK,CAAC,WAAW,UAAU,EAAE,CAAA;wBACrD,CAAC;wBACD,MAAM,KAAK,GAAG,MAAM,KAAK,CAAC,YAAY,EAAE,CAAC,KAAK,CAAC,GAAG,EAAE,CAAC,IAAI,CAAC,CAAA;wBAC1D,IAAI,KAAK,KAAK,IAAI,EAAE,CAAC;4BACnB,iBAAiB,IAAI,CAAC,CAAA;4BACtB,OAAO,EAAE,GAAG,EAAE,QAAQ,KAAK,CAAC,WAAW,UAAU,EAAE,CAAA;wBACrD,CAAC;wBACD,cAAc,CAAC,GAAG,CAAC,UAAU,EAAE,EAAE,QAAQ,EAAE,KAAK,CAAC,WAAW,EAAE,KAAK,EAAE,CAAC,CAAA;wBACtE,OAAO;4BACL,GAAG,EAAE,QAAQ,KAAK,CAAC,WAAW,UAAU;4BACxC,sBAAsB,EAAE,MAAM,CAAC,UAAU,CAAC;yBAC3C,CAAA;oBACH,CAAC,CAAC;iBACH,CACF;gBACH,CAAC,CAAC,MAAM,OAAO,CAAC,aAAa,CAAC,EAAE,MAAM,EAAE,CAAC,CAAA;YAC3C,MAAM,YAAY,GAAG,IAAI,KAAK,CAAC,UAAU,CAAC,KAAK,CAAC,CAAC,MAAM,CAAC,QAAQ,CAAA;YAChE,MAAM,SAAS,GAAG,MAAM,oBAAoB,CAAC,MAAM,CAAC,CAAA;YACpD,MAAM,IAAI,GAAG,yBAAyB,CAAC,YAAY,CAAC,CAAA;YACpD,MAAM,QAAQ,GAAG,QAAQ,CAAC,QAAQ,CAAC,CAAA;YACnC,MAAM,WAAW,GAAG,gBAAgB,CAAC,YAAY,EAAE,QAAQ,EAAE,SAAS,CAAC,CAAA;YAEvE,OAAO,CAAC,KAAK,CAAC,gBAAgB,QAAQ,KAAK,IAAI,CAAC,OAAO,CAAC,MAAM,cAAc,CAAC,CAAA;YAC7E,IAAI,iBAAiB,GAAG,CAAC,EAAE,CAAC;gBAC1B,OAAO,CAAC,IAAI,CACV,WAAW,iBAAiB,+CAA+C,QAAQ,4CAA4C,CAChI,CAAA;YACH,CAAC;YACD,MAAM,YAAY,GAAG,CAAC,IAAI,CAAC,YAAY,IAAI,EAAE,CAAC,CAAC,OAAO,CAAC,CAAC,MAAM,EAAE,EAAE;gBAChE,MAAM,QAAQ,GAAG,cAAc,CAAC,GAAG,CAAC,MAAM,CAAC,UAAU,CAAC,CAAA;gBACtD,OAAO,QAAQ,CAAC,CAAC,CAAC,CAAC,EAAE,GAAG,MAAM,EAAE,GAAG,QAAQ,EAAE,CAAC,CAAC,CAAC,CAAC,EAAE,CAAA;YACrD,CAAC,CAAC,CAAA;YACF,OAAO;gBACL,OAAO,EAAE,IAAI,CAAC,OAAO;gBACrB,KAAK,EAAE,WAAW,CAAC,KAAK;gBACxB,GAAG,CAAC,IAAI,CAAC,YAAY,CAAC,MAAM,KAAK,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,YAAY,EAAE,IAAI,CAAC,YAAY,EAAE,CAAC;gBAC9E,GAAG,CAAC,YAAY,CAAC,MAAM,KAAK,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,YAAY,EAAE,CAAC;aACvD,CAAA;QACH,CAAC;QAAC,OAAO,KAAK,EAAE,CAAC;YACf,MAAM,IAAI,kBAAkB,CAAC,yBAAyB,QAAQ,EAAE,EAAE,EAAE,KAAK,EAAE,OAAO,CAAC,KAAK,CAAC,EAAE,CAAC,CAAA;QAC9F,CAAC;IACH,CAAC;IAED,mBAAmB;IACX,KAAK,CAAC,QAAQ,CAAC,QAAgB;QACrC,IAAI,CAAC;YACH,MAAM,IAAI,GAAG,MAAM,QAAQ,CAAC,QAAQ,EAAE,OAAO,CAAC,CAAA;YAC9C,MAAM,QAAQ,GAAG,QAAQ,CAAC,QAAQ,CAAC,CAAA;YACnC,MAAM,WAAW,GAAG,eAAe,CAAC,IAAI,EAAE,QAAQ,CAAC,CAAA;YACnD,OAAO,CAAC,KAAK,CAAC,eAAe,QAAQ,KAAK,IAAI,CAAC,MAAM,cAAc,CAAC,CAAA;YACpE,OAAO,EAAE,OAAO,EAAE,IAAI,EAAE,KAAK,EAAE,WAAW,CAAC,KAAK,EAAE,CAAA;QACpD,CAAC;QAAC,OAAO,KAAK,EAAE,CAAC;YACf,MAAM,IAAI,kBAAkB,CAAC,wBAAwB,QAAQ,EAAE,EAAE,EAAE,KAAK,EAAE,OAAO,CAAC,KAAK,CAAC,EAAE,CAAC,CAAA;QAC7F,CAAC;IACH,CAAC;IAED,kBAAkB;IACV,KAAK,CAAC,OAAO,CAAC,QAAgB;QACpC,IAAI,CAAC;YACH,MAAM,IAAI,GAAG,MAAM,QAAQ,CAAC,QAAQ,EAAE,OAAO,CAAC,CAAA;YAC9C,MAAM,QAAQ,GAAG,QAAQ,CAAC,QAAQ,CAAC,CAAA;YACnC,MAAM,WAAW,GAAG,oBAAoB,CAAC,IAAI,EAAE,QAAQ,CAAC,CAAA;YACxD,OAAO,CAAC,KAAK,CAAC,cAAc,QAAQ,KAAK,IAAI,CAAC,MAAM,cAAc,CAAC,CAAA;YACnE,OAAO,EAAE,OAAO,EAAE,IAAI,EAAE,KAAK,EAAE,WAAW,CAAC,KAAK,EAAE,CAAA;QACpD,CAAC;QAAC,OAAO,KAAK,EAAE,CAAC;YACf,MAAM,IAAI,kBAAkB,CAAC,uBAAuB,QAAQ,EAAE,EAAE,EAAE,KAAK,EAAE,OAAO,CAAC,KAAK,CAAC,EAAE,CAAC,CAAA;QAC5F,CAAC;IACH,CAAC;CACF"}
|
|
1
|
+
{"version":3,"file":"index.js","sourceRoot":"","sources":["../../src/parser/index.ts"],"names":[],"mappings":"AAAA,6DAA6D;AAE7D,OAAO,EAAE,QAAQ,EAAE,MAAM,SAAS,CAAA;AAClC,OAAO,EAAE,KAAK,EAAE,QAAQ,EAAE,QAAQ,EAAE,MAAM,kBAAkB,CAAA;AAC5D,OAAO,EAAE,QAAQ,EAAE,OAAO,EAAE,UAAU,EAAE,OAAO,EAAE,MAAM,WAAW,CAAA;AAClE,OAAO,EAAE,KAAK,EAAE,MAAM,OAAO,CAAA;AAC7B,OAAO,OAAO,MAAM,SAAS,CAAA;AAE7B,OAAO,EAAwB,eAAe,EAAE,MAAM,qBAAqB,CAAA;AAC3E,OAAO,EAAE,qBAAqB,EAAE,MAAM,uBAAuB,CAAA;AAC7D,OAAO,EAAE,QAAQ,EAAE,UAAU,EAAE,OAAO,EAAE,MAAM,oBAAoB,CAAA;AAElE,OAAO,EAAE,SAAS,EAAE,MAAM,yBAAyB,CAAA;AACnD,OAAO,EAAE,yBAAyB,EAAE,oBAAoB,EAAE,MAAM,kBAAkB,CAAA;AAClF,OAAO,EAAE,eAAe,EAAE,MAAM,kBAAkB,CAAA;AAElD,OAAO,EAAE,aAAa,EAAE,MAAM,sBAAsB,CAAA;AACpD,OAAO,EACL,gBAAgB,EAChB,oBAAoB,EACpB,eAAe,EACf,eAAe,GAChB,MAAM,sBAAsB,CAAA;AAE7B,+CAA+C;AAC/C,uBAAuB;AACvB,+CAA+C;AAE/C;;;;GAIG;AACH,MAAM,CAAC,MAAM,oBAAoB,GAAG,IAAI,GAAG,CAAC,CAAC,MAAM,EAAE,OAAO,EAAE,MAAM,EAAE,KAAK,CAAC,CAAC,CAAA;AA+C7E,KAAK,UAAU,eAAe,CAC5B,QAAgB,EAChB,KAAkC,EAClC,KAAoB,EACpB,QAA2B;IAE3B,MAAM,EAAE,aAAa,EAAE,aAAa,EAAE,GAAG,KAAK,CAAA;IAC9C,MAAM,QAAQ,GAAG,QAAQ,CAAC,QAAQ,CAAC,CAAA;IACnC,IAAI,kBAAsC,CAAA;IAC1C,IAAI,CAAC;QACH,MAAM,aAAa,GAAG,KAAK,CAAC,CAAC,CAAC,EAAE,IAAI,CAAA;QACpC,IAAI,aAAa,IAAI,aAAa,CAAC,IAAI,EAAE,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;YACrD,MAAM,WAAW,GAAG,MAAM,IAAI,eAAe,EAAE,CAAC,SAAS,CAAC,aAAa,EAAE,QAAQ,CAAC,CAAA;YAClF,kBAAkB,GAAG,WAAW,CAAC,CAAC,CAAC,EAAE,IAAI,CAAA;QAC3C,CAAC;IACH,CAAC;IAAC,OAAO,UAAU,EAAE,CAAC;QACpB,IAAI,UAAU,CAAC,UAAU,CAAC,EAAE,CAAC;YAC3B,MAAM,UAAU,CAAA;QAClB,CAAC;QACD,OAAO,CAAC,KAAK,CAAC,sDAAsD,UAAU,EAAE,CAAC,CAAA;IACnF,CAAC;IACD,OAAO;QACL,KAAK,EAAE,eAAe,CAAC,aAAa,EAAE,kBAAkB,EAAE,QAAQ,EAAE,aAAa,CAAC,CAAC,KAAK;QACxF,cAAc,EAAE,eAAe,CAAC,aAAa,EAAE,SAAS,EAAE,QAAQ,EAAE,aAAa,CAAC,CAAC,KAAK;KACzF,CAAA;AACH,CAAC;AAuBD;;GAEG;AACH,MAAM,OAAO,eAAgB,SAAQ,QAAQ;IAC3C,YAAY,OAAe,EAAE,OAA2B;QACtD,KAAK,CAAC,OAAO,EAAE,QAAQ,EAAE,YAAY,EAAE,OAAO,CAAC,CAAA;QAC/C,IAAI,CAAC,IAAI,GAAG,iBAAiB,CAAA;IAC/B,CAAC;CACF;AAED;;GAEG;AACH,MAAM,OAAO,kBAAmB,SAAQ,QAAQ;IAC9C,YAAY,OAAe,EAAE,OAA2B;QACtD,KAAK,CAAC,OAAO,EAAE,QAAQ,EAAE,IAAI,EAAE,OAAO,CAAC,CAAA;QACvC,IAAI,CAAC,IAAI,GAAG,oBAAoB,CAAA;IAClC,CAAC;CACF;AAED,+CAA+C;AAC/C,uBAAuB;AACvB,+CAA+C;AAE/C,6DAA6D;AAC7D,MAAM,OAAO,cAAc;IACR,MAAM,CAAc;IACrC,yEAAyE;IACxD,WAAW,CAAmB;IAC/C;;;;OAIG;IACK,gBAAgB,GAAoB,IAAI,CAAA;IAEhD,YAAY,MAAoB;QAC9B,IAAI,CAAC,MAAM,GAAG,MAAM,CAAA;QACpB,yEAAyE;QACzE,oEAAoE;QACpE,uEAAuE;QACvE,mCAAmC;QACnC,IAAI,MAAM,CAAC,QAAQ,KAAK,SAAS,EAAE,CAAC;YAClC,IAAI,CAAC,WAAW,GAAG,MAAM,CAAC,QAAQ,CAAA;QACpC,CAAC;aAAM,CAAC;YACN,IAAI,CAAC,WAAW,GAAG,CAAC,MAAM,CAAC,OAAO,CAAC,CAAA;QACrC,CAAC;IACH,CAAC;IAED;;;;;;;;OAQG;IACH,KAAK,CAAC,gBAAgB,CAAC,QAAgB;QACrC,sEAAsE;QACtE,yEAAyE;QACzE,gDAAgD;QAChD,IAAI,IAAI,CAAC,WAAW,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;YAClC,MAAM,IAAI,eAAe,CACvB,4JAA4J,CAC7J,CAAA;QACH,CAAC;QAED,wDAAwD;QACxD,IAAI,CAAC,UAAU,CAAC,QAAQ,CAAC,EAAE,CAAC;YAC1B,MAAM,IAAI,eAAe,CACvB,8CAA8C,QAAQ,wGAAwG,CAC/J,CAAA;QACH,CAAC;QAED,uEAAuE;QACvE,gEAAgE;QAChE,+CAA+C;QAC/C,IAAI,CAAC,IAAI,CAAC,gBAAgB,EAAE,CAAC;YAC3B,MAAM,YAAY,GAAa,EAAE,CAAA;YACjC,KAAK,MAAM,GAAG,IAAI,IAAI,CAAC,WAAW,EAAE,CAAC;gBACnC,MAAM,QAAQ,GAAG,MAAM,QAAQ,CAAC,OAAO,CAAC,GAAG,CAAC,CAAC,CAAA;gBAC7C,YAAY,CAAC,IAAI,CAAC,qBAAqB,CAAC,QAAQ,CAAC,CAAC,CAAA;YACpD,CAAC;YACD,IAAI,CAAC,gBAAgB,GAAG,YAAY,CAAA;QACtC,CAAC;QAED,uDAAuD;QACvD,IAAI,YAAoB,CAAA;QACxB,IAAI,CAAC;YACH,YAAY,GAAG,MAAM,QAAQ,CAAC,QAAQ,CAAC,CAAA;QACzC,CAAC;QAAC,OAAO,KAAK,EAAE,CAAC;YACf,sDAAsD;YACtD,iEAAiE;YACjE,oEAAoE;YACpE,4EAA4E;YAC5E,MAAM,SAAS,GAAG,MAAM,KAAK,CAAC,QAAQ,CAAC;iBACpC,IAAI,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,cAAc,EAAE,CAAC;iBACvC,KAAK,CAAC,GAAG,EAAE,CAAC,KAAK,CAAC,CAAA;YAErB,IAAI,SAAS,EAAE,CAAC;gBACd,MAAM,IAAI,eAAe,CACvB,6BAA6B,QAAQ,kDAAkD,EACvF,EAAE,KAAK,EAAE,OAAO,CAAC,KAAK,CAAC,EAAE,CAC1B,CAAA;YACH,CAAC;YAED,0EAA0E;YAC1E,kFAAkF;YAClF,2FAA2F;YAC3F,YAAY,GAAG,OAAO,CAAC,QAAQ,CAAC,CAAA;QAClC,CAAC;QAED,qDAAqD;QACrD,MAAM,OAAO,GAAG,IAAI,CAAC,gBAAgB,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,YAAY,CAAC,UAAU,CAAC,IAAI,CAAC,CAAC,CAAA;QACnF,IAAI,CAAC,OAAO,EAAE,CAAC;YACb,MAAM,YAAY,GAChB,IAAI,CAAC,gBAAgB,CAAC,MAAM,KAAK,CAAC;gBAChC,CAAC,CAAC,IAAI,CAAC,gBAAgB,CAAC,CAAC,CAAC;gBAC1B,CAAC,CAAC,IAAI,CAAC,gBAAgB,CAAC,IAAI,CAAC,IAAI,CAAC,CAAA;YACtC,MAAM,IAAI,eAAe,CACvB,wGAAwG,YAAY,iDAAiD,QAAQ,EAAE,CAChL,CAAA;QACH,CAAC;IACH,CAAC;IAED,+EAA+E;IAC/E,gBAAgB,CAAC,QAAgB;QAC/B,IAAI,CAAC;YACH,MAAM,KAAK,GAAG,QAAQ,CAAC,QAAQ,CAAC,CAAA;YAChC,IAAI,KAAK,CAAC,IAAI,GAAG,IAAI,CAAC,MAAM,CAAC,WAAW,EAAE,CAAC;gBACzC,MAAM,IAAI,eAAe,CACvB,4BAA4B,KAAK,CAAC,IAAI,MAAM,IAAI,CAAC,MAAM,CAAC,WAAW,EAAE,CACtE,CAAA;YACH,CAAC;QACH,CAAC;QAAC,OAAO,KAAK,EAAE,CAAC;YACf,IAAI,KAAK,YAAY,eAAe,EAAE,CAAC;gBACrC,MAAM,KAAK,CAAA;YACb,CAAC;YACD,oDAAoD;YACpD,IAAI,SAAS,CAAC,KAAK,CAAC,KAAK,QAAQ,EAAE,CAAC;gBAClC,MAAM,IAAI,eAAe,CAAC,mBAAmB,QAAQ,EAAE,CAAC,CAAA;YAC1D,CAAC;YACD,MAAM,IAAI,kBAAkB,CAAC,8BAA8B,QAAQ,EAAE,EAAE;gBACrE,KAAK,EAAE,OAAO,CAAC,KAAK,CAAC;aACtB,CAAC,CAAA;QACJ,CAAC;IACH,CAAC;IAED,6DAA6D;IAC7D,KAAK,CAAC,SAAS,CAAC,QAAgB,EAAE,OAAO,GAAqB,EAAE;QAC9D,aAAa;QACb,MAAM,IAAI,CAAC,gBAAgB,CAAC,QAAQ,CAAC,CAAA;QACrC,IAAI,CAAC,gBAAgB,CAAC,QAAQ,CAAC,CAAA;QAE/B,gDAAgD;QAChD,MAAM,GAAG,GAAG,OAAO,CAAC,QAAQ,CAAC,CAAC,WAAW,EAAE,CAAA;QAC3C,QAAQ,GAAG,EAAE,CAAC;YACZ,KAAK,OAAO;gBACV,OAAO,MAAM,IAAI,CAAC,SAAS,CAAC,QAAQ,EAAE,OAAO,CAAC,MAAM,KAAK,IAAI,CAAC,CAAA;YAChE,KAAK,MAAM;gBACT,OAAO,MAAM,IAAI,CAAC,QAAQ,CAAC,QAAQ,CAAC,CAAA;YACtC,KAAK,KAAK;gBACR,OAAO,MAAM,IAAI,CAAC,OAAO,CAAC,QAAQ,CAAC,CAAA;YACrC;gBACE,MAAM,IAAI,eAAe,CAAC,4BAA4B,GAAG,EAAE,CAAC,CAAA;QAChE,CAAC;IACH,CAAC;IAED;;;;OAIG;IACH,KAAK,CAAC,QAAQ,CAAC,QAAgB,EAAE,QAA2B;QAC1D,aAAa;QACb,MAAM,IAAI,CAAC,gBAAgB,CAAC,QAAQ,CAAC,CAAA;QACrC,IAAI,CAAC,gBAAgB,CAAC,QAAQ,CAAC,CAAA;QAE/B,yEAAyE;QACzE,sEAAsE;QACtE,0EAA0E;QAC1E,uDAAuD;QACvD,IAAI,GAA8B,CAAA;QAClC,IAAI,CAAC;YACH,MAAM,MAAM,GAAG,MAAM,QAAQ,CAAC,QAAQ,CAAC,CAAA;YACvC,MAAM,KAAK,GAAG,MAAM,MAAM,CAAC,OAAO,CAAC,CAAA;YACnC,GAAG,GAAG,KAAK,CAAC,QAAQ,CAAC,YAAY,CAAC,MAAM,EAAE,iBAAiB,CAAC,CAAA;YAE5D,MAAM,EAAE,KAAK,EAAE,aAAa,EAAE,aAAa,EAAE,GAAG,MAAM,eAAe,CACnE,GAAG,EACH,QAAQ,EACR,qBAAqB,CACtB,CAAA;YACD,IAAI,IAAI,GAAG,EAAE,CAAA;YACb,MAAM,SAAS,GAAc,EAAE,QAAQ,EAAE,EAAE,EAAE,KAAK,EAAE,EAAE,EAAE,CAAA;YACxD,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;gBACzB,IAAI,CAAC,IAAI,CAAC,IAAI,EAAE,CAAC;oBACf,SAAQ;gBACV,CAAC;gBACD,IAAI,IAAI,EAAE,CAAC;oBACT,IAAI,IAAI,MAAM,CAAA;gBAChB,CAAC;gBACD,MAAM,KAAK,GAAG,IAAI,CAAC,MAAM,CAAA;gBACzB,IAAI,IAAI,IAAI,CAAC,IAAI,CAAA;gBACjB,SAAS,CAAC,KAAK,EAAE,IAAI,CAAC,EAAE,KAAK,EAAE,GAAG,EAAE,IAAI,CAAC,MAAM,EAAE,IAAI,EAAE,IAAI,CAAC,OAAO,EAAE,CAAC,CAAA;gBACtE,SAAS,CAAC,QAAQ,CAAC,IAAI,CACrB,GAAG,IAAI,CAAC,QAAQ,CAAC,GAAG,CAAC,CAAC,OAAO,EAAE,EAAE,CAAC,CAAC,EAAE,GAAG,OAAO,EAAE,MAAM,EAAE,KAAK,GAAG,OAAO,CAAC,MAAM,EAAE,CAAC,CAAC,CACpF,CAAA;YACH,CAAC;YAED,MAAM,EAAE,KAAK,EAAE,cAAc,EAAE,GAAG,MAAM,eAAe,CACrD,QAAQ,EACR,KAAK,EACL,EAAE,aAAa,EAAE,aAAa,EAAE,EAChC,QAAQ,CACT,CAAA;YAED,OAAO,CAAC,KAAK,CAAC,eAAe,QAAQ,KAAK,IAAI,CAAC,MAAM,gBAAgB,KAAK,CAAC,MAAM,SAAS,CAAC,CAAA;YAE3F,OAAO,EAAE,OAAO,EAAE,IAAI,EAAE,KAAK,EAAE,cAAc,EAAE,SAAS,EAAE,CAAA;QAC5D,CAAC;QAAC,OAAO,KAAK,EAAE,CAAC;YACf,sEAAsE;YACtE,uEAAuE;YACvE,4EAA4E;YAC5E,IAAI,UAAU,CAAC,KAAK,CAAC,EAAE,CAAC;gBACtB,MAAM,KAAK,CAAA;YACb,CAAC;YACD,MAAM,IAAI,kBAAkB,CAAC,wBAAwB,QAAQ,EAAE,EAAE,EAAE,KAAK,EAAE,OAAO,CAAC,KAAK,CAAC,EAAE,CAAC,CAAA;QAC7F,CAAC;gBAAS,CAAC;YACT,sEAAsE;YACtE,2BAA2B;YAC3B,GAAG,EAAE,OAAO,EAAE,CAAA;QAChB,CAAC;IACH,CAAC;IAED;;;;;;;;OAQG;IACH,KAAK,CAAC,aAAa,CACjB,QAAgB,EAChB,QAA2B;QAa3B,yEAAyE;QACzE,2CAA2C;QAC3C,MAAM,IAAI,CAAC,gBAAgB,CAAC,QAAQ,CAAC,CAAA;QACrC,IAAI,CAAC,gBAAgB,CAAC,QAAQ,CAAC,CAAA;QAE/B,qEAAqE;QACrE,+BAA+B;QAC/B,yEAAyE;QACzE,0EAA0E;QAC1E,uEAAuE;QACvE,IAAI,GAA8B,CAAA;QAClC,IAAI,CAAC;YACH,MAAM,MAAM,GAAG,MAAM,QAAQ,CAAC,QAAQ,CAAC,CAAA;YACvC,MAAM,KAAK,GAAG,MAAM,MAAM,CAAC,OAAO,CAAC,CAAA;YACnC,GAAG,GAAG,KAAK,CAAC,QAAQ,CAAC,YAAY,CAAC,MAAM,EAAE,iBAAiB,CAAC,CAAA;YAC5D,MAAM,SAAS,GAAG,MAAM,eAAe,CAAC,GAAG,EAAE,QAAQ,EAAE,qCAAqC,CAAC,CAAA;YAE7F,MAAM,EAAE,KAAK,EAAE,WAAW,EAAE,aAAa,EAAE,aAAa,EAAE,GAAG,SAAS,CAAA;YACtE,MAAM,KAAK,GAAG,WAAW,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC;gBACvC,OAAO,EAAE,IAAI,CAAC,OAAO;gBACrB,QAAQ,EAAE,IAAI,CAAC,QAAQ;gBACvB,IAAI,EAAE,IAAI,CAAC,IAAI;gBACf,aAAa,EAAE,IAAI,CAAC,aAAa;gBACjC,SAAS,EAAE,IAAI,CAAC,SAAS;aAC1B,CAAC,CAAC,CAAA;YACH,MAAM,EAAE,KAAK,EAAE,cAAc,EAAE,GAAG,MAAM,eAAe,CACrD,QAAQ,EACR,WAAW,EACX,EAAE,aAAa,EAAE,aAAa,EAAE,EAChC,QAAQ,CACT,CAAA;YAED,OAAO,CAAC,KAAK,CACX,qBAAqB,QAAQ,KAAK,KAAK,CAAC,MAAM,mCAAmC,CAClF,CAAA;YAED,OAAO,EAAE,GAAG,EAAE,KAAK,EAAE,cAAc,EAAE,KAAK,EAAE,CAAA;QAC9C,CAAC;QAAC,OAAO,KAAK,EAAE,CAAC;YACf,yEAAyE;YACzE,0EAA0E;YAC1E,uEAAuE;YACvE,qCAAqC;YACrC,GAAG,EAAE,OAAO,EAAE,CAAA;YACd,sEAAsE;YACtE,uEAAuE;YACvE,uDAAuD;YACvD,IAAI,UAAU,CAAC,KAAK,CAAC,EAAE,CAAC;gBACtB,MAAM,KAAK,CAAA;YACb,CAAC;YACD,MAAM,IAAI,kBAAkB,CAAC,8BAA8B,QAAQ,EAAE,EAAE;gBACrE,KAAK,EAAE,OAAO,CAAC,KAAK,CAAC;aACtB,CAAC,CAAA;QACJ,CAAC;IACH,CAAC;IAED;;;OAGG;IACK,KAAK,CAAC,SAAS,CAAC,QAAgB,EAAE,aAAsB;QAC9D,IAAI,CAAC;YACH,MAAM,MAAM,GAAG,MAAM,QAAQ,CAAC,QAAQ,CAAC,CAAA;YACvC,MAAM,cAAc,GAAG,IAAI,GAAG,EAG3B,CAAA;YACH,IAAI,cAAc,GAAG,CAAC,CAAA;YACtB,IAAI,iBAAiB,GAAG,CAAC,CAAA;YACzB,MAAM,UAAU,GAAG,aAAa;gBAC9B,CAAC,CAAC,MAAM,OAAO,CAAC,aAAa,CACzB,EAAE,MAAM,EAAE,EACV;oBACE,YAAY,EAAE,OAAO,CAAC,MAAM,CAAC,UAAU,CAAC,KAAK,EAAE,KAAK,EAAE,EAAE;wBACtD,MAAM,UAAU,GAAG,cAAc,EAAE,CAAA;wBACnC,IAAI,KAAK,CAAC,WAAW,KAAK,WAAW,IAAI,KAAK,CAAC,WAAW,KAAK,YAAY,EAAE,CAAC;4BAC5E,iBAAiB,IAAI,CAAC,CAAA;4BACtB,OAAO,EAAE,GAAG,EAAE,QAAQ,KAAK,CAAC,WAAW,UAAU,EAAE,CAAA;wBACrD,CAAC;wBACD,MAAM,KAAK,GAAG,MAAM,KAAK,CAAC,YAAY,EAAE,CAAC,KAAK,CAAC,GAAG,EAAE,CAAC,IAAI,CAAC,CAAA;wBAC1D,IAAI,KAAK,KAAK,IAAI,EAAE,CAAC;4BACnB,iBAAiB,IAAI,CAAC,CAAA;4BACtB,OAAO,EAAE,GAAG,EAAE,QAAQ,KAAK,CAAC,WAAW,UAAU,EAAE,CAAA;wBACrD,CAAC;wBACD,cAAc,CAAC,GAAG,CAAC,UAAU,EAAE,EAAE,QAAQ,EAAE,KAAK,CAAC,WAAW,EAAE,KAAK,EAAE,CAAC,CAAA;wBACtE,OAAO;4BACL,GAAG,EAAE,QAAQ,KAAK,CAAC,WAAW,UAAU;4BACxC,sBAAsB,EAAE,MAAM,CAAC,UAAU,CAAC;yBAC3C,CAAA;oBACH,CAAC,CAAC;iBACH,CACF;gBACH,CAAC,CAAC,MAAM,OAAO,CAAC,aAAa,CAAC,EAAE,MAAM,EAAE,CAAC,CAAA;YAC3C,MAAM,YAAY,GAAG,IAAI,KAAK,CAAC,UAAU,CAAC,KAAK,CAAC,CAAC,MAAM,CAAC,QAAQ,CAAA;YAChE,MAAM,SAAS,GAAG,MAAM,oBAAoB,CAAC,MAAM,CAAC,CAAA;YACpD,MAAM,IAAI,GAAG,yBAAyB,CAAC,YAAY,CAAC,CAAA;YACpD,MAAM,QAAQ,GAAG,QAAQ,CAAC,QAAQ,CAAC,CAAA;YACnC,MAAM,WAAW,GAAG,gBAAgB,CAAC,YAAY,EAAE,QAAQ,EAAE,SAAS,CAAC,CAAA;YAEvE,OAAO,CAAC,KAAK,CAAC,gBAAgB,QAAQ,KAAK,IAAI,CAAC,OAAO,CAAC,MAAM,cAAc,CAAC,CAAA;YAC7E,IAAI,iBAAiB,GAAG,CAAC,EAAE,CAAC;gBAC1B,OAAO,CAAC,IAAI,CACV,WAAW,iBAAiB,+CAA+C,QAAQ,4CAA4C,CAChI,CAAA;YACH,CAAC;YACD,MAAM,YAAY,GAAG,CAAC,IAAI,CAAC,YAAY,IAAI,EAAE,CAAC,CAAC,OAAO,CAAC,CAAC,MAAM,EAAE,EAAE;gBAChE,MAAM,QAAQ,GAAG,cAAc,CAAC,GAAG,CAAC,MAAM,CAAC,UAAU,CAAC,CAAA;gBACtD,OAAO,QAAQ,CAAC,CAAC,CAAC,CAAC,EAAE,GAAG,MAAM,EAAE,GAAG,QAAQ,EAAE,CAAC,CAAC,CAAC,CAAC,EAAE,CAAA;YACrD,CAAC,CAAC,CAAA;YACF,OAAO;gBACL,OAAO,EAAE,IAAI,CAAC,OAAO;gBACrB,GAAG,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC,CAAC,EAAE,SAAS,EAAE,IAAI,CAAC,SAAS,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;gBACxD,KAAK,EAAE,WAAW,CAAC,KAAK;gBACxB,GAAG,CAAC,IAAI,CAAC,YAAY,CAAC,MAAM,KAAK,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,YAAY,EAAE,IAAI,CAAC,YAAY,EAAE,CAAC;gBAC9E,GAAG,CAAC,YAAY,CAAC,MAAM,KAAK,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,YAAY,EAAE,CAAC;aACvD,CAAA;QACH,CAAC;QAAC,OAAO,KAAK,EAAE,CAAC;YACf,MAAM,IAAI,kBAAkB,CAAC,yBAAyB,QAAQ,EAAE,EAAE,EAAE,KAAK,EAAE,OAAO,CAAC,KAAK,CAAC,EAAE,CAAC,CAAA;QAC9F,CAAC;IACH,CAAC;IAED,mBAAmB;IACX,KAAK,CAAC,QAAQ,CAAC,QAAgB;QACrC,IAAI,CAAC;YACH,MAAM,IAAI,GAAG,MAAM,QAAQ,CAAC,QAAQ,EAAE,OAAO,CAAC,CAAA;YAC9C,MAAM,QAAQ,GAAG,QAAQ,CAAC,QAAQ,CAAC,CAAA;YACnC,MAAM,WAAW,GAAG,eAAe,CAAC,IAAI,EAAE,QAAQ,CAAC,CAAA;YACnD,OAAO,CAAC,KAAK,CAAC,eAAe,QAAQ,KAAK,IAAI,CAAC,MAAM,cAAc,CAAC,CAAA;YACpE,OAAO,EAAE,OAAO,EAAE,IAAI,EAAE,KAAK,EAAE,WAAW,CAAC,KAAK,EAAE,SAAS,EAAE,aAAa,CAAC,IAAI,EAAE,KAAK,CAAC,EAAE,CAAA;QAC3F,CAAC;QAAC,OAAO,KAAK,EAAE,CAAC;YACf,MAAM,IAAI,kBAAkB,CAAC,wBAAwB,QAAQ,EAAE,EAAE,EAAE,KAAK,EAAE,OAAO,CAAC,KAAK,CAAC,EAAE,CAAC,CAAA;QAC7F,CAAC;IACH,CAAC;IAED,kBAAkB;IACV,KAAK,CAAC,OAAO,CAAC,QAAgB;QACpC,IAAI,CAAC;YACH,MAAM,IAAI,GAAG,MAAM,QAAQ,CAAC,QAAQ,EAAE,OAAO,CAAC,CAAA;YAC9C,MAAM,QAAQ,GAAG,QAAQ,CAAC,QAAQ,CAAC,CAAA;YACnC,MAAM,WAAW,GAAG,oBAAoB,CAAC,IAAI,EAAE,QAAQ,CAAC,CAAA;YACxD,OAAO,CAAC,KAAK,CAAC,cAAc,QAAQ,KAAK,IAAI,CAAC,MAAM,cAAc,CAAC,CAAA;YACnE,OAAO,EAAE,OAAO,EAAE,IAAI,EAAE,KAAK,EAAE,WAAW,CAAC,KAAK,EAAE,SAAS,EAAE,aAAa,CAAC,IAAI,EAAE,IAAI,CAAC,EAAE,CAAA;QAC1F,CAAC;QAAC,OAAO,KAAK,EAAE,CAAC;YACf,MAAM,IAAI,kBAAkB,CAAC,uBAAuB,QAAQ,EAAE,EAAE,EAAE,KAAK,EAAE,OAAO,CAAC,KAAK,CAAC,EAAE,CAAC,CAAA;QAC5F,CAAC;IACH,CAAC;CACF"}
|
|
@@ -1,4 +1,5 @@
|
|
|
1
1
|
import type { Document as MupdfDocument } from 'mupdf';
|
|
2
|
+
import type { HeadingAnchor } from '../utils/source-context.js';
|
|
2
3
|
import { type EmbedderInterface, type FilteredTextFragment } from './pdf-filter.js';
|
|
3
4
|
interface StextBbox {
|
|
4
5
|
x: number;
|
|
@@ -30,6 +31,7 @@ interface StextJson {
|
|
|
30
31
|
* kept for the downstream visual detector.
|
|
31
32
|
*/
|
|
32
33
|
interface ExtractedPage {
|
|
34
|
+
headings: HeadingAnchor[];
|
|
33
35
|
pageNum: number;
|
|
34
36
|
text: string;
|
|
35
37
|
textFragments: FilteredTextFragment[];
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"pdf-extract.d.ts","sourceRoot":"","sources":["../../src/parser/pdf-extract.ts"],"names":[],"mappings":"AAUA,OAAO,KAAK,EAAE,QAAQ,IAAI,aAAa,EAAqB,MAAM,OAAO,CAAA;
|
|
1
|
+
{"version":3,"file":"pdf-extract.d.ts","sourceRoot":"","sources":["../../src/parser/pdf-extract.ts"],"names":[],"mappings":"AAUA,OAAO,KAAK,EAAE,QAAQ,IAAI,aAAa,EAAqB,MAAM,OAAO,CAAA;AACzE,OAAO,KAAK,EAAE,aAAa,EAAE,MAAM,4BAA4B,CAAA;AAC/D,OAAO,EACL,KAAK,iBAAiB,EACtB,KAAK,oBAAoB,EAG1B,MAAM,iBAAiB,CAAA;AAGxB,UAAU,SAAS;IACjB,CAAC,EAAE,MAAM,CAAA;IACT,CAAC,EAAE,MAAM,CAAA;IACT,CAAC,EAAE,MAAM,CAAA;IACT,CAAC,EAAE,MAAM,CAAA;CACV;AAED,kEAAkE;AAClE,UAAU,SAAS;IACjB,MAAM,EAAE,KAAK,CAAC;QACZ,IAAI,EAAE,MAAM,CAAA;QACZ,+EAA+E;QAC/E,IAAI,CAAC,EAAE,SAAS,CAAA;QAChB,KAAK,CAAC,EAAE,KAAK,CAAC;YACZ,IAAI,EAAE,MAAM,CAAA;YACZ,CAAC,EAAE,MAAM,CAAA;YACT,CAAC,EAAE,MAAM,CAAA;YACT,IAAI,CAAC,EAAE,SAAS,CAAA;YAChB,IAAI,EAAE;gBAAE,IAAI,EAAE,MAAM,CAAC;gBAAC,IAAI,CAAC,EAAE,MAAM,CAAC;gBAAC,MAAM,CAAC,EAAE,MAAM,CAAA;aAAE,CAAA;SACvD,CAAC,CAAA;KACH,CAAC,CAAA;CACH;AAED;;;GAGG;AACH,UAAU,aAAa;IACrB,QAAQ,EAAE,aAAa,EAAE,CAAA;IACzB,OAAO,EAAE,MAAM,CAAA;IACf,IAAI,EAAE,MAAM,CAAA;IACZ,aAAa,EAAE,oBAAoB,EAAE,CAAA;IACrC,SAAS,EAAE,SAAS,CAAA;CACrB;AAED,4CAA4C;AAC5C,UAAU,YAAY;IACpB,KAAK,EAAE,aAAa,EAAE,CAAA;IACtB,aAAa,EAAE,MAAM,GAAG,SAAS,CAAA;IACjC,aAAa,EAAE;QAAE,IAAI,EAAE,MAAM,CAAC;QAAC,QAAQ,EAAE,MAAM,CAAA;KAAE,GAAG,SAAS,CAAA;CAC9D;AAwFD,wBAAsB,eAAe,CACnC,GAAG,EAAE,aAAa,EAClB,QAAQ,EAAE,iBAAiB,EAC3B,YAAY,EAAE,MAAM,GACnB,OAAO,CAAC,YAAY,CAAC,CAkCvB"}
|
|
@@ -8,6 +8,7 @@
|
|
|
8
8
|
// emits `block.type === 'image'` entries for the downstream
|
|
9
9
|
// visual-candidate detector.
|
|
10
10
|
import { filterPageBoundaryLayouts, } from './pdf-filter.js';
|
|
11
|
+
import { pdfHeadings } from './pdf-headings.js';
|
|
11
12
|
/**
|
|
12
13
|
* Per-page extraction shared by `parsePdf` and `parsePdfPages`, which differ
|
|
13
14
|
* only in `stextOptions`: `parsePdfPages` adds `preserve-images` so mupdf
|
|
@@ -102,9 +103,11 @@ export async function extractPdfPages(doc, embedder, stextOptions) {
|
|
|
102
103
|
}
|
|
103
104
|
// Apply sentence-level header/footer filtering while retaining each survivor's layout data.
|
|
104
105
|
const filteredPages = await filterPageBoundaryLayouts(pageDataList, embedder);
|
|
106
|
+
const headings = pdfHeadings(doc, pageDataList, filteredPages);
|
|
105
107
|
const page1FontHint = largestFontTitleHint(pageDataList[0]?.items ?? []);
|
|
106
108
|
const pages = pageDataList.map((p, idx) => ({
|
|
107
109
|
pageNum: p.pageNum,
|
|
110
|
+
headings: headings[idx] ?? [],
|
|
108
111
|
text: filteredPages[idx]?.text ?? '',
|
|
109
112
|
textFragments: filteredPages[idx]?.textFragments ?? [],
|
|
110
113
|
stextJson: stextJsonList[idx] ?? { blocks: [] },
|