@shuji-bonji/pdf-spec-mcp 0.4.1 → 0.4.2
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/dist/services/content-extractor.d.ts +40 -0
- package/dist/services/content-extractor.d.ts.map +1 -1
- package/dist/services/content-extractor.js +92 -0
- package/dist/services/content-extractor.js.map +1 -1
- package/dist/services/search-index.d.ts +7 -0
- package/dist/services/search-index.d.ts.map +1 -1
- package/dist/services/search-index.js +42 -12
- package/dist/services/search-index.js.map +1 -1
- package/dist/services/table-collector.d.ts.map +1 -1
- package/dist/services/table-collector.js +52 -13
- package/dist/services/table-collector.js.map +1 -1
- package/package.json +1 -1
|
@@ -35,4 +35,44 @@ export declare function extractSectionContent(doc: PDFDocumentProxy, startPage:
|
|
|
35
35
|
* trimToSectionStart. Both are pinned by tests.
|
|
36
36
|
*/
|
|
37
37
|
export declare function extractOrphanedStrip(doc: PDFDocumentProxy, seamPage: number, nextSectionNumber: string): Promise<ContentElement[]>;
|
|
38
|
+
/**
|
|
39
|
+
* A page's text, cut at section headings, for the search index (S-8).
|
|
40
|
+
*/
|
|
41
|
+
export interface PageSegment {
|
|
42
|
+
/**
|
|
43
|
+
* Section that owns this text, or null for text above the first heading — the strip,
|
|
44
|
+
* which belongs to whichever section flows in from the previous page. The caller knows
|
|
45
|
+
* that section; this module only knows where the cut is.
|
|
46
|
+
*/
|
|
47
|
+
section: string | null;
|
|
48
|
+
text: string;
|
|
49
|
+
}
|
|
50
|
+
/**
|
|
51
|
+
* Split a page's text at the headings of the sections that start on it (S-8).
|
|
52
|
+
*
|
|
53
|
+
* The search index used to hand each whole page to the last section starting on or before
|
|
54
|
+
* it, so the strip at the top of a page — the tail of the *previous* section — was always
|
|
55
|
+
* misattributed to the next one: search_spec("QuadPoints") reported 12.5.6.11, but
|
|
56
|
+
* QuadPoints is 12.5.6.10's Table 182, and get_tables("12.5.6.11") has no QuadPoints.
|
|
57
|
+
* Search led straight to a section where the term cannot be found.
|
|
58
|
+
*
|
|
59
|
+
* Cutting here uses the same findSectionHeadingIndex as trimToSectionStart /
|
|
60
|
+
* extractOrphanedStrip, so search attributes text to exactly the section whose
|
|
61
|
+
* get_section shows it (the S-5 lesson: one rule, one place). The arms mirror it too:
|
|
62
|
+
* - heading found at index > 0 → text before it goes to the pending owner (the strip)
|
|
63
|
+
* - heading found at index 0 → no strip
|
|
64
|
+
* - heading not found → the section keeps everything pending, and the previous
|
|
65
|
+
* owner adopts nothing (69 sections of ISO 32000-2)
|
|
66
|
+
*
|
|
67
|
+
* When NO heading is found at all, the page is returned as one raw-text segment for the
|
|
68
|
+
* last starting section — exactly what the whole-page index did. This is not just a
|
|
69
|
+
* shortcut: walkStructTree drops text whose only home is an unhandled container (TOC /
|
|
70
|
+
* TOCI / Link / Span chains on the Contents pages, the errata "Issue #NNN" pages, the
|
|
71
|
+
* INTERNATIONAL STANDARD title page), so element text on those pages is near-empty and
|
|
72
|
+
* cutting by elements would silently drop them from the index. Headingless pages cannot
|
|
73
|
+
* be cut anyway — falling back to the raw page keeps them whole.
|
|
74
|
+
*
|
|
75
|
+
* @param startingSections - section numbers starting on this page, in document order
|
|
76
|
+
*/
|
|
77
|
+
export declare function extractPageSegments(doc: PDFDocumentProxy, pageNum: number, startingSections: string[]): Promise<PageSegment[]>;
|
|
38
78
|
//# sourceMappingURL=content-extractor.d.ts.map
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"content-extractor.d.ts","sourceRoot":"","sources":["../../src/services/content-extractor.ts"],"names":[],"mappings":"AAAA;;;GAGG;AAGH,OAAO,KAAK,EACV,cAAc,EAMf,MAAM,mBAAmB,CAAC;AAE3B,OAAO,KAAK,EACV,gBAAgB,EAMjB,MAAM,iBAAiB,CAAC;AAIzB;;;;GAIG;AACH,wBAAsB,qBAAqB,CACzC,GAAG,EAAE,gBAAgB,EACrB,SAAS,EAAE,MAAM,EACjB,OAAO,EAAE,MAAM,EACf,aAAa,CAAC,EAAE,MAAM,GACrB,OAAO,CAAC,cAAc,EAAE,CAAC,CAuB3B;AA8BD;;;;;;;;;;;;;;;;;;;;;;;GAuBG;AACH,wBAAsB,oBAAoB,CACxC,GAAG,EAAE,gBAAgB,EACrB,QAAQ,EAAE,MAAM,EAChB,iBAAiB,EAAE,MAAM,GACxB,OAAO,CAAC,cAAc,EAAE,CAAC,CAM3B"}
|
|
1
|
+
{"version":3,"file":"content-extractor.d.ts","sourceRoot":"","sources":["../../src/services/content-extractor.ts"],"names":[],"mappings":"AAAA;;;GAGG;AAGH,OAAO,KAAK,EACV,cAAc,EAMf,MAAM,mBAAmB,CAAC;AAE3B,OAAO,KAAK,EACV,gBAAgB,EAMjB,MAAM,iBAAiB,CAAC;AAIzB;;;;GAIG;AACH,wBAAsB,qBAAqB,CACzC,GAAG,EAAE,gBAAgB,EACrB,SAAS,EAAE,MAAM,EACjB,OAAO,EAAE,MAAM,EACf,aAAa,CAAC,EAAE,MAAM,GACrB,OAAO,CAAC,cAAc,EAAE,CAAC,CAuB3B;AA8BD;;;;;;;;;;;;;;;;;;;;;;;GAuBG;AACH,wBAAsB,oBAAoB,CACxC,GAAG,EAAE,gBAAgB,EACrB,QAAQ,EAAE,MAAM,EAChB,iBAAiB,EAAE,MAAM,GACxB,OAAO,CAAC,cAAc,EAAE,CAAC,CAM3B;AAED;;GAEG;AACH,MAAM,WAAW,WAAW;IAC1B;;;;OAIG;IACH,OAAO,EAAE,MAAM,GAAG,IAAI,CAAC;IACvB,IAAI,EAAE,MAAM,CAAC;CACd;AAED;;;;;;;;;;;;;;;;;;;;;;;;;;GA0BG;AACH,wBAAsB,mBAAmB,CACvC,GAAG,EAAE,gBAAgB,EACrB,OAAO,EAAE,MAAM,EACf,gBAAgB,EAAE,MAAM,EAAE,GACzB,OAAO,CAAC,WAAW,EAAE,CAAC,CAmCxB"}
|
|
@@ -78,6 +78,98 @@ export async function extractOrphanedStrip(doc, seamPage, nextSectionNumber) {
|
|
|
78
78
|
const headingIdx = findSectionHeadingIndex(elements, nextSectionNumber);
|
|
79
79
|
return headingIdx > 0 ? elements.slice(0, headingIdx) : [];
|
|
80
80
|
}
|
|
81
|
+
/**
|
|
82
|
+
* Split a page's text at the headings of the sections that start on it (S-8).
|
|
83
|
+
*
|
|
84
|
+
* The search index used to hand each whole page to the last section starting on or before
|
|
85
|
+
* it, so the strip at the top of a page — the tail of the *previous* section — was always
|
|
86
|
+
* misattributed to the next one: search_spec("QuadPoints") reported 12.5.6.11, but
|
|
87
|
+
* QuadPoints is 12.5.6.10's Table 182, and get_tables("12.5.6.11") has no QuadPoints.
|
|
88
|
+
* Search led straight to a section where the term cannot be found.
|
|
89
|
+
*
|
|
90
|
+
* Cutting here uses the same findSectionHeadingIndex as trimToSectionStart /
|
|
91
|
+
* extractOrphanedStrip, so search attributes text to exactly the section whose
|
|
92
|
+
* get_section shows it (the S-5 lesson: one rule, one place). The arms mirror it too:
|
|
93
|
+
* - heading found at index > 0 → text before it goes to the pending owner (the strip)
|
|
94
|
+
* - heading found at index 0 → no strip
|
|
95
|
+
* - heading not found → the section keeps everything pending, and the previous
|
|
96
|
+
* owner adopts nothing (69 sections of ISO 32000-2)
|
|
97
|
+
*
|
|
98
|
+
* When NO heading is found at all, the page is returned as one raw-text segment for the
|
|
99
|
+
* last starting section — exactly what the whole-page index did. This is not just a
|
|
100
|
+
* shortcut: walkStructTree drops text whose only home is an unhandled container (TOC /
|
|
101
|
+
* TOCI / Link / Span chains on the Contents pages, the errata "Issue #NNN" pages, the
|
|
102
|
+
* INTERNATIONAL STANDARD title page), so element text on those pages is near-empty and
|
|
103
|
+
* cutting by elements would silently drop them from the index. Headingless pages cannot
|
|
104
|
+
* be cut anyway — falling back to the raw page keeps them whole.
|
|
105
|
+
*
|
|
106
|
+
* @param startingSections - section numbers starting on this page, in document order
|
|
107
|
+
*/
|
|
108
|
+
export async function extractPageSegments(doc, pageNum, startingSections) {
|
|
109
|
+
const elements = await extractPageContent(doc, pageNum);
|
|
110
|
+
const segments = [];
|
|
111
|
+
let owner = null;
|
|
112
|
+
let start = 0;
|
|
113
|
+
let anyHeadingFound = false;
|
|
114
|
+
for (const sec of startingSections) {
|
|
115
|
+
const idx = findSectionHeadingIndex(elements.slice(start), sec);
|
|
116
|
+
if (idx === -1) {
|
|
117
|
+
// Mirrors trimToSectionStart keeping the whole page when the heading is missing:
|
|
118
|
+
// the section takes everything still pending (including any unemitted strip —
|
|
119
|
+
// exactly what its get_section returns), and the previous owner adopts nothing.
|
|
120
|
+
owner = sec;
|
|
121
|
+
continue;
|
|
122
|
+
}
|
|
123
|
+
anyHeadingFound = true;
|
|
124
|
+
const abs = start + idx;
|
|
125
|
+
if (abs > start) {
|
|
126
|
+
segments.push({ section: owner, text: elementsToText(elements.slice(start, abs)) });
|
|
127
|
+
}
|
|
128
|
+
owner = sec;
|
|
129
|
+
start = abs;
|
|
130
|
+
}
|
|
131
|
+
if (!anyHeadingFound) {
|
|
132
|
+
return [{ section: owner, text: await extractRawPageText(doc, pageNum) }].filter((s) => s.text.trim().length > 0);
|
|
133
|
+
}
|
|
134
|
+
segments.push({ section: owner, text: elementsToText(elements.slice(start)) });
|
|
135
|
+
return segments.filter((s) => s.text.trim().length > 0);
|
|
136
|
+
}
|
|
137
|
+
/**
|
|
138
|
+
* Raw page text, independent of tagging (see extractPageSegments' headingless arm).
|
|
139
|
+
*/
|
|
140
|
+
async function extractRawPageText(doc, pageNum) {
|
|
141
|
+
const page = await doc.getPage(pageNum);
|
|
142
|
+
const textContent = await page.getTextContent();
|
|
143
|
+
return textContent.items
|
|
144
|
+
.filter((item) => 'str' in item)
|
|
145
|
+
.map((item) => item.str)
|
|
146
|
+
.join(' ');
|
|
147
|
+
}
|
|
148
|
+
/**
|
|
149
|
+
* Flatten content elements to searchable text.
|
|
150
|
+
*/
|
|
151
|
+
function elementsToText(elements) {
|
|
152
|
+
const parts = [];
|
|
153
|
+
for (const el of elements) {
|
|
154
|
+
switch (el.type) {
|
|
155
|
+
case 'heading':
|
|
156
|
+
case 'paragraph':
|
|
157
|
+
case 'code':
|
|
158
|
+
parts.push(el.text);
|
|
159
|
+
break;
|
|
160
|
+
case 'note':
|
|
161
|
+
parts.push(`${el.label} ${el.text}`);
|
|
162
|
+
break;
|
|
163
|
+
case 'list':
|
|
164
|
+
parts.push(el.items.join('\n'));
|
|
165
|
+
break;
|
|
166
|
+
case 'table':
|
|
167
|
+
parts.push([el.headers.join(' '), ...el.rows.map((r) => r.join(' '))].join('\n'));
|
|
168
|
+
break;
|
|
169
|
+
}
|
|
170
|
+
}
|
|
171
|
+
return parts.join('\n');
|
|
172
|
+
}
|
|
81
173
|
/**
|
|
82
174
|
* Extract structured content from a single page
|
|
83
175
|
*/
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"content-extractor.js","sourceRoot":"","sources":["../../src/services/content-extractor.ts"],"names":[],"mappings":"AAAA;;;GAGG;AAEH,OAAO,EAAE,WAAW,EAAE,MAAM,cAAc,CAAC;AAS3C,OAAO,EAAE,aAAa,EAAE,MAAM,yBAAyB,CAAC;AAYxD;;;;GAIG;AACH,MAAM,CAAC,KAAK,UAAU,qBAAqB,CACzC,GAAqB,EACrB,SAAiB,EACjB,OAAe,EACf,aAAsB;IAEtB,kFAAkF;IAClF,MAAM,UAAU,GAAG,GAAG,CAAC,QAAQ,IAAI,OAAO,CAAC;IAC3C,MAAM,SAAS,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,IAAI,CAAC,GAAG,CAAC,SAAS,EAAE,UAAU,CAAC,CAAC,CAAC;IAC/D,MAAM,OAAO,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,IAAI,CAAC,GAAG,CAAC,OAAO,EAAE,UAAU,CAAC,CAAC,CAAC;IAE3D,MAAM,WAAW,GAAG,KAAK,CAAC,IAAI,CAAC,EAAE,MAAM,EAAE,OAAO,GAAG,SAAS,GAAG,CAAC,EAAE,EAAE,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,SAAS,GAAG,CAAC,CAAC,CAAC;IAE7F,mEAAmE;IACnE,MAAM,WAAW,GAAG,MAAM,aAAa,CACrC,WAAW,EACX,CAAC,OAAO,EAAE,EAAE,CAAC,kBAAkB,CAAC,GAAG,EAAE,OAAO,CAAC,EAC7C,WAAW,CAAC,iBAAiB,CAC9B,CAAC;IAEF,MAAM,QAAQ,GAAG,WAAW,CAAC,IAAI,EAAE,CAAC;IAEpC,wDAAwD;IACxD,IAAI,aAAa,EAAE,CAAC;QAClB,OAAO,kBAAkB,CAAC,QAAQ,EAAE,aAAa,CAAC,CAAC;IACrD,CAAC;IAED,OAAO,QAAQ,CAAC;AAClB,CAAC;AAED;;;;;GAKG;AACH,SAAS,uBAAuB,CAAC,QAA0B,EAAE,aAAqB;IAChF,MAAM,GAAG,GAAG,QAAQ,CAAC,SAAS,CAC5B,CAAC,EAAE,EAAE,EAAE,CAAC,EAAE,CAAC,IAAI,KAAK,SAAS,IAAI,EAAE,CAAC,IAAI,CAAC,UAAU,CAAC,GAAG,aAAa,GAAG,CAAC,CACzE,CAAC;IACF,IAAI,GAAG,KAAK,CAAC,CAAC;QAAE,OAAO,GAAG,CAAC;IAE3B,0FAA0F;IAC1F,IAAI,aAAa,CAAC,UAAU,CAAC,QAAQ,CAAC,EAAE,CAAC;QACvC,OAAO,QAAQ,CAAC,SAAS,CAAC,CAAC,EAAE,EAAE,EAAE,CAAC,EAAE,CAAC,IAAI,KAAK,SAAS,IAAI,EAAE,CAAC,IAAI,CAAC,UAAU,CAAC,aAAa,CAAC,CAAC,CAAC;IAChG,CAAC;IAED,OAAO,CAAC,CAAC,CAAC;AACZ,CAAC;AAED;;GAEG;AACH,SAAS,kBAAkB,CAAC,QAA0B,EAAE,aAAqB;IAC3E,MAAM,UAAU,GAAG,uBAAuB,CAAC,QAAQ,EAAE,aAAa,CAAC,CAAC;IACpE,OAAO,UAAU,GAAG,CAAC,CAAC,CAAC,CAAC,QAAQ,CAAC,KAAK,CAAC,UAAU,CAAC,CAAC,CAAC,CAAC,QAAQ,CAAC;AAChE,CAAC;AAED;;;;;;;;;;;;;;;;;;;;;;;GAuBG;AACH,MAAM,CAAC,KAAK,UAAU,oBAAoB,CACxC,GAAqB,EACrB,QAAgB,EAChB,iBAAyB;IAEzB,IAAI,QAAQ,GAAG,CAAC,IAAI,QAAQ,GAAG,CAAC,GAAG,CAAC,QAAQ,IAAI,CAAC,CAAC;QAAE,OAAO,EAAE,CAAC;IAE9D,MAAM,QAAQ,GAAG,MAAM,kBAAkB,CAAC,GAAG,EAAE,QAAQ,CAAC,CAAC;IACzD,MAAM,UAAU,GAAG,uBAAuB,CAAC,QAAQ,EAAE,iBAAiB,CAAC,CAAC;IACxE,OAAO,UAAU,GAAG,CAAC,CAAC,CAAC,CAAC,QAAQ,CAAC,KAAK,CAAC,CAAC,EAAE,UAAU,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC;AAC7D,CAAC;AAED;;GAEG;AACH,KAAK,UAAU,kBAAkB,CAC/B,GAAqB,EACrB,OAAe;IAEf,MAAM,IAAI,GAAG,MAAM,GAAG,CAAC,OAAO,CAAC,OAAO,CAAC,CAAC;IAExC,MAAM,CAAC,UAAU,EAAE,WAAW,CAAC,GAAG,MAAM,OAAO,CAAC,GAAG,CAAC;QAClD,IAAI,CAAC,aAAa,EAAE;QACpB,IAAI,CAAC,cAAc,CAAC,EAAE,oBAAoB,EAAE,IAAI,EAAE,CAAC;KACpD,CAAC,CAAC;IAEH,IAAI,CAAC,UAAU,EAAE,CAAC;QAChB,8CAA8C;QAC9C,OAAO,gBAAgB,CAAC,WAAW,CAAC,CAAC;IACvC,CAAC;IAED,0CAA0C;IAC1C,MAAM,OAAO,GAAG,YAAY,CAAC,WAAW,CAAC,KAAK,CAAC,CAAC;IAEhD,gDAAgD;IAChD,OAAO,cAAc,CAAC,UAAU,EAAE,OAAO,EAAE,WAAW,CAAC,MAAM,CAAC,CAAC;AACjE,CAAC;AAED;;GAEG;AACH,SAAS,YAAY,CAAC,KAAwB;IAC5C,MAAM,GAAG,GAAG,IAAI,GAAG,EAAkB,CAAC;IACtC,IAAI,SAAS,GAAkB,IAAI,CAAC;IACpC,IAAI,YAAY,GAAa,EAAE,CAAC;IAEhC,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QACzB,IAAI,eAAe,CAAC,IAAI,CAAC,EAAE,CAAC;YAC1B,IAAI,IAAI,CAAC,IAAI,KAAK,yBAAyB,IAAI,IAAI,CAAC,EAAE,KAAK,SAAS,EAAE,CAAC;gBACrE,gBAAgB;gBAChB,IAAI,SAAS,KAAK,IAAI,IAAI,YAAY,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;oBAClD,GAAG,CAAC,GAAG,CAAC,SAAS,EAAE,YAAY,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC,CAAC;gBAC5C,CAAC;gBACD,SAAS,GAAG,IAAI,CAAC,EAAE,CAAC;gBACpB,YAAY,GAAG,EAAE,CAAC;YACpB,CAAC;iBAAM,IAAI,IAAI,CAAC,IAAI,KAAK,kBAAkB,EAAE,CAAC;gBAC5C,IAAI,SAAS,KAAK,IAAI,IAAI,YAAY,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;oBAClD,GAAG,CAAC,GAAG,CAAC,SAAS,EAAE,YAAY,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC,CAAC;gBAC5C,CAAC;gBACD,SAAS,GAAG,IAAI,CAAC;gBACjB,YAAY,GAAG,EAAE,CAAC;YACpB,CAAC;QACH,CAAC;aAAM,IAAI,UAAU,CAAC,IAAI,CAAC,EAAE,CAAC;YAC5B,IAAI,SAAS,KAAK,IAAI,EAAE,CAAC;gBACvB,YAAY,CAAC,IAAI,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;gBAC5B,IAAI,IAAI,CAAC,MAAM,EAAE,CAAC;oBAChB,YAAY,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;gBAC1B,CAAC;YACH,CAAC;QACH,CAAC;IACH,CAAC;IAED,kBAAkB;IAClB,IAAI,SAAS,KAAK,IAAI,IAAI,YAAY,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QAClD,GAAG,CAAC,GAAG,CAAC,SAAS,EAAE,YAAY,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC,CAAC;IAC5C,CAAC;IAED,OAAO,GAAG,CAAC;AACb,CAAC;AAED;;GAEG;AACH,SAAS,cAAc,CACrB,IAAoB,EACpB,OAA4B,EAC5B,MAA6B;IAE7B,MAAM,QAAQ,GAAqB,EAAE,CAAC;IAEtC,IAAI,CAAC,IAAI,CAAC,QAAQ;QAAE,OAAO,QAAQ,CAAC;IAEpC,KAAK,MAAM,KAAK,IAAI,IAAI,CAAC,QAAQ,EAAE,CAAC;QAClC,IAAI,eAAe,CAAC,KAAK,CAAC,EAAE,CAAC;YAC3B,kDAAkD;YAClD,SAAS;QACX,CAAC;QAED,MAAM,IAAI,GAAG,KAAK,CAAC,IAAI,CAAC;QAExB,kDAAkD;QAClD,IAAI,IAAI,KAAK,UAAU;YAAE,SAAS;QAElC,8BAA8B;QAC9B,IAAI,WAAW,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC;YAC3B,MAAM,IAAI,GAAG,WAAW,CAAC,KAAK,EAAE,OAAO,CAAC,CAAC,IAAI,EAAE,CAAC;YAChD,IAAI,IAAI,EAAE,CAAC;gBACT,6EAA6E;gBAC7E,qCAAqC;gBACrC,MAAM,KAAK,GAAG,IAAI,CAAC,MAAM,KAAK,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;gBACnE,QAAQ,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,SAAS,EAAE,KAAK,EAAE,IAAI,EAAoB,CAAC,CAAC;YACpE,CAAC;YACD,SAAS;QACX,CAAC;QAED,YAAY;QACZ,IAAI,IAAI,KAAK,GAAG,EAAE,CAAC;YACjB,MAAM,IAAI,GAAG,WAAW,CAAC,KAAK,EAAE,OAAO,CAAC,CAAC,IAAI,EAAE,CAAC;YAChD,IAAI,CAAC,IAAI;gBAAE,SAAS;YAEpB,sBAAsB;YACtB,MAAM,SAAS,GAAG,IAAI,CAAC,KAAK,CAAC,iCAAiC,CAAC,CAAC;YAChE,IAAI,SAAS,EAAE,CAAC;gBACd,QAAQ,CAAC,IAAI,CAAC;oBACZ,IAAI,EAAE,MAAM;oBACZ,KAAK,EAAE,SAAS,CAAC,CAAC,CAAC;oBACnB,IAAI,EAAE,IAAI,CAAC,SAAS,CAAC,SAAS,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,IAAI,EAAE;iBAClC,CAAC,CAAC;gBAClB,SAAS;YACX,CAAC;YAED,QAAQ,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,WAAW,EAAE,IAAI,EAAsB,CAAC,CAAC;YAC/D,SAAS;QACX,CAAC;QAED,OAAO;QACP,IAAI,IAAI,KAAK,GAAG,EAAE,CAAC;YACjB,MAAM,KAAK,GAAG,gBAAgB,CAAC,KAAK,EAAE,OAAO,CAAC,CAAC;YAC/C,IAAI,KAAK,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;gBACrB,QAAQ,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,MAAM,EAAE,KAAK,EAAiB,CAAC,CAAC;YACxD,CAAC;YACD,SAAS;QACX,CAAC;QAED,QAAQ;QACR,IAAI,IAAI,KAAK,OAAO,EAAE,CAAC;YACrB,MAAM,KAAK,GAAG,YAAY,CAAC,KAAK,EAAE,OAAO,CAAC,CAAC;YAC3C,IAAI,KAAK,EAAE,CAAC;gBACV,QAAQ,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;YACvB,CAAC;YACD,SAAS;QACX,CAAC;QAED,mEAAmE;QACnE,IAAI,IAAI,KAAK,MAAM,EAAE,CAAC;YACpB,MAAM,IAAI,GAAG,WAAW,CAAC,KAAK,EAAE,OAAO,CAAC,CAAC,IAAI,EAAE,CAAC;YAChD,IAAI,IAAI,EAAE,CAAC;gBACT,QAAQ,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,MAAM,EAAE,IAAI,EAAE,CAAC,CAAC;YACxC,CAAC;YACD,SAAS;QACX,CAAC;QAED,oEAAoE;QACpE,MAAM,aAAa,GAAG,cAAc,CAAC,KAAK,EAAE,OAAO,EAAE,MAAM,CAAC,CAAC;QAC7D,QAAQ,CAAC,IAAI,CAAC,GAAG,aAAa,CAAC,CAAC;IAClC,CAAC;IAED,OAAO,QAAQ,CAAC;AAClB,CAAC;AAED;;GAEG;AACH,SAAS,WAAW,CAAC,IAAoB,EAAE,OAA4B;IACrE,MAAM,KAAK,GAAa,EAAE,CAAC;IAE3B,IAAI,IAAI,CAAC,QAAQ,EAAE,CAAC;QAClB,KAAK,MAAM,KAAK,IAAI,IAAI,CAAC,QAAQ,EAAE,CAAC;YAClC,IAAI,eAAe,CAAC,KAAK,CAAC,EAAE,CAAC;gBAC3B,MAAM,IAAI,GAAG,OAAO,CAAC,GAAG,CAAC,KAAK,CAAC,EAAE,CAAC,CAAC;gBACnC,IAAI,IAAI;oBAAE,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;YAC7B,CAAC;iBAAM,CAAC;gBACN,yBAAyB;gBACzB,IAAI,KAAK,CAAC,IAAI,KAAK,UAAU;oBAAE,SAAS;gBACxC,KAAK,CAAC,IAAI,CAAC,WAAW,CAAC,KAAK,EAAE,OAAO,CAAC,CAAC,CAAC;YAC1C,CAAC;QACH,CAAC;IACH,CAAC;IAED,OAAO,KAAK,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;AACxB,CAAC;AAED;;GAEG;AACH,SAAS,gBAAgB,CAAC,IAAoB,EAAE,OAA4B;IAC1E,MAAM,KAAK,GAAa,EAAE,CAAC;IAE3B,IAAI,IAAI,CAAC,QAAQ,EAAE,CAAC;QAClB,KAAK,MAAM,KAAK,IAAI,IAAI,CAAC,QAAQ,EAAE,CAAC;YAClC,IAAI,eAAe,CAAC,KAAK,CAAC;gBAAE,SAAS;YACrC,IAAI,KAAK,CAAC,IAAI,KAAK,IAAI,EAAE,CAAC;gBACxB,uCAAuC;gBACvC,MAAM,QAAQ,GAAG,mBAAmB,CAAC,KAAK,EAAE,OAAO,CAAC,CAAC;gBACrD,IAAI,QAAQ,CAAC,IAAI,EAAE,EAAE,CAAC;oBACpB,KAAK,CAAC,IAAI,CAAC,QAAQ,CAAC,IAAI,EAAE,CAAC,CAAC;gBAC9B,CAAC;YACH,CAAC;QACH,CAAC;IACH,CAAC;IAED,OAAO,KAAK,CAAC;AACf,CAAC;AAED,SAAS,mBAAmB,CAAC,EAAkB,EAAE,OAA4B;IAC3E,MAAM,KAAK,GAAa,EAAE,CAAC;IAC3B,IAAI,EAAE,CAAC,QAAQ,EAAE,CAAC;QAChB,KAAK,MAAM,KAAK,IAAI,EAAE,CAAC,QAAQ,EAAE,CAAC;YAChC,IAAI,eAAe,CAAC,KAAK,CAAC,EAAE,CAAC;gBAC3B,MAAM,IAAI,GAAG,OAAO,CAAC,GAAG,CAAC,KAAK,CAAC,EAAE,CAAC,CAAC;gBACnC,IAAI,IAAI;oBAAE,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;YAC7B,CAAC;iBAAM,IAAI,KAAK,CAAC,IAAI,KAAK,OAAO,EAAE,CAAC;gBAClC,KAAK,CAAC,IAAI,CAAC,WAAW,CAAC,KAAK,EAAE,OAAO,CAAC,CAAC,CAAC;YAC1C,CAAC;iBAAM,IAAI,KAAK,CAAC,IAAI,KAAK,KAAK,EAAE,CAAC;gBAChC,4BAA4B;YAC9B,CAAC;iBAAM,CAAC;gBACN,KAAK,CAAC,IAAI,CAAC,WAAW,CAAC,KAAK,EAAE,OAAO,CAAC,CAAC,CAAC;YAC1C,CAAC;QACH,CAAC;IACH,CAAC;IACD,OAAO,KAAK,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;AACxB,CAAC;AAED;;GAEG;AACH,SAAS,YAAY,CAAC,IAAoB,EAAE,OAA4B;IACtE,MAAM,OAAO,GAAa,EAAE,CAAC;IAC7B,MAAM,IAAI,GAAe,EAAE,CAAC;IAE5B,IAAI,CAAC,IAAI,CAAC,QAAQ;QAAE,OAAO,IAAI,CAAC;IAEhC,6EAA6E;IAC7E,MAAM,QAAQ,GAAqB,EAAE,CAAC;IACtC,MAAM,QAAQ,GAAqB,EAAE,CAAC;IAEtC,KAAK,MAAM,KAAK,IAAI,IAAI,CAAC,QAAQ,EAAE,CAAC;QAClC,IAAI,eAAe,CAAC,KAAK,CAAC;YAAE,SAAS;QAErC,IAAI,KAAK,CAAC,IAAI,KAAK,IAAI,EAAE,CAAC;YACxB,2CAA2C;YAC3C,QAAQ,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;QACvB,CAAC;aAAM,IAAI,KAAK,CAAC,IAAI,KAAK,OAAO,EAAE,CAAC;YAClC,qDAAqD;YACrD,IAAI,KAAK,CAAC,QAAQ,EAAE,CAAC;gBACnB,KAAK,MAAM,EAAE,IAAI,KAAK,CAAC,QAAQ,EAAE,CAAC;oBAChC,IAAI,CAAC,eAAe,CAAC,EAAE,CAAC,IAAI,EAAE,CAAC,IAAI,KAAK,IAAI,EAAE,CAAC;wBAC7C,QAAQ,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;oBACpB,CAAC;gBACH,CAAC;YACH,CAAC;QACH,CAAC;aAAM,IAAI,KAAK,CAAC,IAAI,KAAK,OAAO,IAAI,KAAK,CAAC,IAAI,KAAK,OAAO,EAAE,CAAC;YAC5D,yDAAyD;YACzD,IAAI,KAAK,CAAC,QAAQ,EAAE,CAAC;gBACnB,KAAK,MAAM,EAAE,IAAI,KAAK,CAAC,QAAQ,EAAE,CAAC;oBAChC,IAAI,CAAC,eAAe,CAAC,EAAE,CAAC,IAAI,EAAE,CAAC,IAAI,KAAK,IAAI,EAAE,CAAC;wBAC7C,QAAQ,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;oBACpB,CAAC;gBACH,CAAC;YACH,CAAC;QACH,CAAC;IACH,CAAC;IAED,uCAAuC;IACvC,KAAK,MAAM,EAAE,IAAI,QAAQ,EAAE,CAAC;QAC1B,IAAI,EAAE,CAAC,QAAQ,EAAE,CAAC;YAChB,MAAM,KAAK,GAAa,EAAE,CAAC;YAC3B,KAAK,MAAM,IAAI,IAAI,EAAE,CAAC,QAAQ,EAAE,CAAC;gBAC/B,IAAI,eAAe,CAAC,IAAI,CAAC;oBAAE,SAAS;gBACpC,KAAK,CAAC,IAAI,CAAC,WAAW,CAAC,IAAI,EAAE,OAAO,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC;YAChD,CAAC;YACD,IAAI,KAAK,CAAC,MAAM,GAAG,CAAC,IAAI,OAAO,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;gBAC7C,OAAO,CAAC,IAAI,CAAC,GAAG,KAAK,CAAC,CAAC;YACzB,CAAC;iBAAM,IAAI,KAAK,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;gBAC5B,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;YACnB,CAAC;QACH,CAAC;IACH,CAAC;IAED,8DAA8D;IAC9D,IAAI,iBAAiB,GAAG,KAAK,CAAC;IAC9B,IAAI,OAAO,CAAC,MAAM,KAAK,CAAC,IAAI,QAAQ,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QAChD,MAAM,QAAQ,GAAG,QAAQ,CAAC,CAAC,CAAC,CAAC;QAC7B,IAAI,QAAQ,CAAC,QAAQ,EAAE,CAAC;YACtB,MAAM,KAAK,GAAG,QAAQ,CAAC,QAAQ;iBAC5B,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,eAAe,CAAC,CAAC,CAAC,CAAC;iBAClC,KAAK,CAAC,CAAC,CAAC,EAAE,EAAE,CAAE,CAAoB,CAAC,IAAI,KAAK,IAAI,CAAC,CAAC;YACrD,IAAI,KAAK,EAAE,CAAC;gBACV,MAAM,KAAK,GAAa,EAAE,CAAC;gBAC3B,KAAK,MAAM,IAAI,IAAI,QAAQ,CAAC,QAAQ,EAAE,CAAC;oBACrC,IAAI,eAAe,CAAC,IAAI,CAAC;wBAAE,SAAS;oBACpC,KAAK,CAAC,IAAI,CAAC,WAAW,CAAC,IAAI,EAAE,OAAO,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC;gBAChD,CAAC;gBACD,OAAO,CAAC,IAAI,CAAC,GAAG,KAAK,CAAC,CAAC;gBACvB,iBAAiB,GAAG,IAAI,CAAC;YAC3B,CAAC;QACH,CAAC;IACH,CAAC;IAED,yBAAyB;IACzB,KAAK,IAAI,CAAC,GAAG,iBAAiB,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,GAAG,QAAQ,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QACjE,MAAM,EAAE,GAAG,QAAQ,CAAC,CAAC,CAAC,CAAC;QACvB,IAAI,EAAE,CAAC,QAAQ,EAAE,CAAC;YAChB,MAAM,KAAK,GAAa,EAAE,CAAC;YAC3B,KAAK,MAAM,IAAI,IAAI,EAAE,CAAC,QAAQ,EAAE,CAAC;gBAC/B,IAAI,eAAe,CAAC,IAAI,CAAC;oBAAE,SAAS;gBACpC,KAAK,CAAC,IAAI,CAAC,WAAW,CAAC,IAAI,EAAE,OAAO,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC;YAChD,CAAC;YACD,IAAI,KAAK,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;gBACrB,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;YACnB,CAAC;QACH,CAAC;IACH,CAAC;IAED,IAAI,OAAO,CAAC,MAAM,KAAK,CAAC,IAAI,IAAI,CAAC,MAAM,KAAK,CAAC;QAAE,OAAO,IAAI,CAAC;IAE3D,OAAO,EAAE,IAAI,EAAE,OAAO,EAAE,OAAO,EAAE,IAAI,EAAE,CAAC;AAC1C,CAAC;AAED;;GAEG;AACH,SAAS,gBAAgB,CAAC,WAAwB;IAChD,MAAM,KAAK,GAAa,EAAE,CAAC;IAC3B,KAAK,MAAM,IAAI,IAAI,WAAW,CAAC,KAAK,EAAE,CAAC;QACrC,IAAI,UAAU,CAAC,IAAI,CAAC,EAAE,CAAC;YACrB,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;YACrB,IAAI,IAAI,CAAC,MAAM;gBAAE,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;QACpC,CAAC;IACH,CAAC;IACD,MAAM,IAAI,GAAG,KAAK,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC,IAAI,EAAE,CAAC;IACnC,IAAI,CAAC,IAAI;QAAE,OAAO,EAAE,CAAC;IACrB,OAAO,CAAC,EAAE,IAAI,EAAE,WAAW,EAAE,IAAI,EAAE,CAAC,CAAC;AACvC,CAAC;AAED,cAAc;AACd,SAAS,UAAU,CAAC,IAAqB;IACvC,OAAO,KAAK,IAAI,IAAI,CAAC;AACvB,CAAC;AAED,SAAS,eAAe,CAAC,IAAqB;IAC5C,OAAO,MAAM,IAAI,IAAI,CAAC;AACxB,CAAC;AAED,SAAS,eAAe,CAAC,KAAyC;IAChE,OAAO,MAAM,IAAI,KAAK,IAAK,KAA2B,CAAC,IAAI,KAAK,SAAS,CAAC;AAC5E,CAAC"}
|
|
1
|
+
{"version":3,"file":"content-extractor.js","sourceRoot":"","sources":["../../src/services/content-extractor.ts"],"names":[],"mappings":"AAAA;;;GAGG;AAEH,OAAO,EAAE,WAAW,EAAE,MAAM,cAAc,CAAC;AAS3C,OAAO,EAAE,aAAa,EAAE,MAAM,yBAAyB,CAAC;AAYxD;;;;GAIG;AACH,MAAM,CAAC,KAAK,UAAU,qBAAqB,CACzC,GAAqB,EACrB,SAAiB,EACjB,OAAe,EACf,aAAsB;IAEtB,kFAAkF;IAClF,MAAM,UAAU,GAAG,GAAG,CAAC,QAAQ,IAAI,OAAO,CAAC;IAC3C,MAAM,SAAS,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,IAAI,CAAC,GAAG,CAAC,SAAS,EAAE,UAAU,CAAC,CAAC,CAAC;IAC/D,MAAM,OAAO,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,IAAI,CAAC,GAAG,CAAC,OAAO,EAAE,UAAU,CAAC,CAAC,CAAC;IAE3D,MAAM,WAAW,GAAG,KAAK,CAAC,IAAI,CAAC,EAAE,MAAM,EAAE,OAAO,GAAG,SAAS,GAAG,CAAC,EAAE,EAAE,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,SAAS,GAAG,CAAC,CAAC,CAAC;IAE7F,mEAAmE;IACnE,MAAM,WAAW,GAAG,MAAM,aAAa,CACrC,WAAW,EACX,CAAC,OAAO,EAAE,EAAE,CAAC,kBAAkB,CAAC,GAAG,EAAE,OAAO,CAAC,EAC7C,WAAW,CAAC,iBAAiB,CAC9B,CAAC;IAEF,MAAM,QAAQ,GAAG,WAAW,CAAC,IAAI,EAAE,CAAC;IAEpC,wDAAwD;IACxD,IAAI,aAAa,EAAE,CAAC;QAClB,OAAO,kBAAkB,CAAC,QAAQ,EAAE,aAAa,CAAC,CAAC;IACrD,CAAC;IAED,OAAO,QAAQ,CAAC;AAClB,CAAC;AAED;;;;;GAKG;AACH,SAAS,uBAAuB,CAAC,QAA0B,EAAE,aAAqB;IAChF,MAAM,GAAG,GAAG,QAAQ,CAAC,SAAS,CAC5B,CAAC,EAAE,EAAE,EAAE,CAAC,EAAE,CAAC,IAAI,KAAK,SAAS,IAAI,EAAE,CAAC,IAAI,CAAC,UAAU,CAAC,GAAG,aAAa,GAAG,CAAC,CACzE,CAAC;IACF,IAAI,GAAG,KAAK,CAAC,CAAC;QAAE,OAAO,GAAG,CAAC;IAE3B,0FAA0F;IAC1F,IAAI,aAAa,CAAC,UAAU,CAAC,QAAQ,CAAC,EAAE,CAAC;QACvC,OAAO,QAAQ,CAAC,SAAS,CAAC,CAAC,EAAE,EAAE,EAAE,CAAC,EAAE,CAAC,IAAI,KAAK,SAAS,IAAI,EAAE,CAAC,IAAI,CAAC,UAAU,CAAC,aAAa,CAAC,CAAC,CAAC;IAChG,CAAC;IAED,OAAO,CAAC,CAAC,CAAC;AACZ,CAAC;AAED;;GAEG;AACH,SAAS,kBAAkB,CAAC,QAA0B,EAAE,aAAqB;IAC3E,MAAM,UAAU,GAAG,uBAAuB,CAAC,QAAQ,EAAE,aAAa,CAAC,CAAC;IACpE,OAAO,UAAU,GAAG,CAAC,CAAC,CAAC,CAAC,QAAQ,CAAC,KAAK,CAAC,UAAU,CAAC,CAAC,CAAC,CAAC,QAAQ,CAAC;AAChE,CAAC;AAED;;;;;;;;;;;;;;;;;;;;;;;GAuBG;AACH,MAAM,CAAC,KAAK,UAAU,oBAAoB,CACxC,GAAqB,EACrB,QAAgB,EAChB,iBAAyB;IAEzB,IAAI,QAAQ,GAAG,CAAC,IAAI,QAAQ,GAAG,CAAC,GAAG,CAAC,QAAQ,IAAI,CAAC,CAAC;QAAE,OAAO,EAAE,CAAC;IAE9D,MAAM,QAAQ,GAAG,MAAM,kBAAkB,CAAC,GAAG,EAAE,QAAQ,CAAC,CAAC;IACzD,MAAM,UAAU,GAAG,uBAAuB,CAAC,QAAQ,EAAE,iBAAiB,CAAC,CAAC;IACxE,OAAO,UAAU,GAAG,CAAC,CAAC,CAAC,CAAC,QAAQ,CAAC,KAAK,CAAC,CAAC,EAAE,UAAU,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC;AAC7D,CAAC;AAeD;;;;;;;;;;;;;;;;;;;;;;;;;;GA0BG;AACH,MAAM,CAAC,KAAK,UAAU,mBAAmB,CACvC,GAAqB,EACrB,OAAe,EACf,gBAA0B;IAE1B,MAAM,QAAQ,GAAG,MAAM,kBAAkB,CAAC,GAAG,EAAE,OAAO,CAAC,CAAC;IAExD,MAAM,QAAQ,GAAkB,EAAE,CAAC;IACnC,IAAI,KAAK,GAAkB,IAAI,CAAC;IAChC,IAAI,KAAK,GAAG,CAAC,CAAC;IACd,IAAI,eAAe,GAAG,KAAK,CAAC;IAE5B,KAAK,MAAM,GAAG,IAAI,gBAAgB,EAAE,CAAC;QACnC,MAAM,GAAG,GAAG,uBAAuB,CAAC,QAAQ,CAAC,KAAK,CAAC,KAAK,CAAC,EAAE,GAAG,CAAC,CAAC;QAChE,IAAI,GAAG,KAAK,CAAC,CAAC,EAAE,CAAC;YACf,iFAAiF;YACjF,8EAA8E;YAC9E,gFAAgF;YAChF,KAAK,GAAG,GAAG,CAAC;YACZ,SAAS;QACX,CAAC;QACD,eAAe,GAAG,IAAI,CAAC;QACvB,MAAM,GAAG,GAAG,KAAK,GAAG,GAAG,CAAC;QACxB,IAAI,GAAG,GAAG,KAAK,EAAE,CAAC;YAChB,QAAQ,CAAC,IAAI,CAAC,EAAE,OAAO,EAAE,KAAK,EAAE,IAAI,EAAE,cAAc,CAAC,QAAQ,CAAC,KAAK,CAAC,KAAK,EAAE,GAAG,CAAC,CAAC,EAAE,CAAC,CAAC;QACtF,CAAC;QACD,KAAK,GAAG,GAAG,CAAC;QACZ,KAAK,GAAG,GAAG,CAAC;IACd,CAAC;IAED,IAAI,CAAC,eAAe,EAAE,CAAC;QACrB,OAAO,CAAC,EAAE,OAAO,EAAE,KAAK,EAAE,IAAI,EAAE,MAAM,kBAAkB,CAAC,GAAG,EAAE,OAAO,CAAC,EAAE,CAAC,CAAC,MAAM,CAC9E,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC,IAAI,EAAE,CAAC,MAAM,GAAG,CAAC,CAChC,CAAC;IACJ,CAAC;IAED,QAAQ,CAAC,IAAI,CAAC,EAAE,OAAO,EAAE,KAAK,EAAE,IAAI,EAAE,cAAc,CAAC,QAAQ,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC;IAE/E,OAAO,QAAQ,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC,IAAI,EAAE,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;AAC1D,CAAC;AAED;;GAEG;AACH,KAAK,UAAU,kBAAkB,CAAC,GAAqB,EAAE,OAAe;IACtE,MAAM,IAAI,GAAG,MAAM,GAAG,CAAC,OAAO,CAAC,OAAO,CAAC,CAAC;IACxC,MAAM,WAAW,GAAG,MAAM,IAAI,CAAC,cAAc,EAAE,CAAC;IAChD,OAAO,WAAW,CAAC,KAAK;SACrB,MAAM,CAAC,CAAC,IAAI,EAAoB,EAAE,CAAC,KAAK,IAAI,IAAI,CAAC;SACjD,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,GAAG,CAAC;SACvB,IAAI,CAAC,GAAG,CAAC,CAAC;AACf,CAAC;AAED;;GAEG;AACH,SAAS,cAAc,CAAC,QAA0B;IAChD,MAAM,KAAK,GAAa,EAAE,CAAC;IAC3B,KAAK,MAAM,EAAE,IAAI,QAAQ,EAAE,CAAC;QAC1B,QAAQ,EAAE,CAAC,IAAI,EAAE,CAAC;YAChB,KAAK,SAAS,CAAC;YACf,KAAK,WAAW,CAAC;YACjB,KAAK,MAAM;gBACT,KAAK,CAAC,IAAI,CAAC,EAAE,CAAC,IAAI,CAAC,CAAC;gBACpB,MAAM;YACR,KAAK,MAAM;gBACT,KAAK,CAAC,IAAI,CAAC,GAAG,EAAE,CAAC,KAAK,IAAI,EAAE,CAAC,IAAI,EAAE,CAAC,CAAC;gBACrC,MAAM;YACR,KAAK,MAAM;gBACT,KAAK,CAAC,IAAI,CAAC,EAAE,CAAC,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC;gBAChC,MAAM;YACR,KAAK,OAAO;gBACV,KAAK,CAAC,IAAI,CAAC,CAAC,EAAE,CAAC,OAAO,CAAC,IAAI,CAAC,GAAG,CAAC,EAAE,GAAG,EAAE,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC;gBAClF,MAAM;QACV,CAAC;IACH,CAAC;IACD,OAAO,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;AAC1B,CAAC;AAED;;GAEG;AACH,KAAK,UAAU,kBAAkB,CAC/B,GAAqB,EACrB,OAAe;IAEf,MAAM,IAAI,GAAG,MAAM,GAAG,CAAC,OAAO,CAAC,OAAO,CAAC,CAAC;IAExC,MAAM,CAAC,UAAU,EAAE,WAAW,CAAC,GAAG,MAAM,OAAO,CAAC,GAAG,CAAC;QAClD,IAAI,CAAC,aAAa,EAAE;QACpB,IAAI,CAAC,cAAc,CAAC,EAAE,oBAAoB,EAAE,IAAI,EAAE,CAAC;KACpD,CAAC,CAAC;IAEH,IAAI,CAAC,UAAU,EAAE,CAAC;QAChB,8CAA8C;QAC9C,OAAO,gBAAgB,CAAC,WAAW,CAAC,CAAC;IACvC,CAAC;IAED,0CAA0C;IAC1C,MAAM,OAAO,GAAG,YAAY,CAAC,WAAW,CAAC,KAAK,CAAC,CAAC;IAEhD,gDAAgD;IAChD,OAAO,cAAc,CAAC,UAAU,EAAE,OAAO,EAAE,WAAW,CAAC,MAAM,CAAC,CAAC;AACjE,CAAC;AAED;;GAEG;AACH,SAAS,YAAY,CAAC,KAAwB;IAC5C,MAAM,GAAG,GAAG,IAAI,GAAG,EAAkB,CAAC;IACtC,IAAI,SAAS,GAAkB,IAAI,CAAC;IACpC,IAAI,YAAY,GAAa,EAAE,CAAC;IAEhC,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QACzB,IAAI,eAAe,CAAC,IAAI,CAAC,EAAE,CAAC;YAC1B,IAAI,IAAI,CAAC,IAAI,KAAK,yBAAyB,IAAI,IAAI,CAAC,EAAE,KAAK,SAAS,EAAE,CAAC;gBACrE,gBAAgB;gBAChB,IAAI,SAAS,KAAK,IAAI,IAAI,YAAY,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;oBAClD,GAAG,CAAC,GAAG,CAAC,SAAS,EAAE,YAAY,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC,CAAC;gBAC5C,CAAC;gBACD,SAAS,GAAG,IAAI,CAAC,EAAE,CAAC;gBACpB,YAAY,GAAG,EAAE,CAAC;YACpB,CAAC;iBAAM,IAAI,IAAI,CAAC,IAAI,KAAK,kBAAkB,EAAE,CAAC;gBAC5C,IAAI,SAAS,KAAK,IAAI,IAAI,YAAY,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;oBAClD,GAAG,CAAC,GAAG,CAAC,SAAS,EAAE,YAAY,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC,CAAC;gBAC5C,CAAC;gBACD,SAAS,GAAG,IAAI,CAAC;gBACjB,YAAY,GAAG,EAAE,CAAC;YACpB,CAAC;QACH,CAAC;aAAM,IAAI,UAAU,CAAC,IAAI,CAAC,EAAE,CAAC;YAC5B,IAAI,SAAS,KAAK,IAAI,EAAE,CAAC;gBACvB,YAAY,CAAC,IAAI,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;gBAC5B,IAAI,IAAI,CAAC,MAAM,EAAE,CAAC;oBAChB,YAAY,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;gBAC1B,CAAC;YACH,CAAC;QACH,CAAC;IACH,CAAC;IAED,kBAAkB;IAClB,IAAI,SAAS,KAAK,IAAI,IAAI,YAAY,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QAClD,GAAG,CAAC,GAAG,CAAC,SAAS,EAAE,YAAY,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC,CAAC;IAC5C,CAAC;IAED,OAAO,GAAG,CAAC;AACb,CAAC;AAED;;GAEG;AACH,SAAS,cAAc,CACrB,IAAoB,EACpB,OAA4B,EAC5B,MAA6B;IAE7B,MAAM,QAAQ,GAAqB,EAAE,CAAC;IAEtC,IAAI,CAAC,IAAI,CAAC,QAAQ;QAAE,OAAO,QAAQ,CAAC;IAEpC,KAAK,MAAM,KAAK,IAAI,IAAI,CAAC,QAAQ,EAAE,CAAC;QAClC,IAAI,eAAe,CAAC,KAAK,CAAC,EAAE,CAAC;YAC3B,kDAAkD;YAClD,SAAS;QACX,CAAC;QAED,MAAM,IAAI,GAAG,KAAK,CAAC,IAAI,CAAC;QAExB,kDAAkD;QAClD,IAAI,IAAI,KAAK,UAAU;YAAE,SAAS;QAElC,8BAA8B;QAC9B,IAAI,WAAW,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC;YAC3B,MAAM,IAAI,GAAG,WAAW,CAAC,KAAK,EAAE,OAAO,CAAC,CAAC,IAAI,EAAE,CAAC;YAChD,IAAI,IAAI,EAAE,CAAC;gBACT,6EAA6E;gBAC7E,qCAAqC;gBACrC,MAAM,KAAK,GAAG,IAAI,CAAC,MAAM,KAAK,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;gBACnE,QAAQ,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,SAAS,EAAE,KAAK,EAAE,IAAI,EAAoB,CAAC,CAAC;YACpE,CAAC;YACD,SAAS;QACX,CAAC;QAED,YAAY;QACZ,IAAI,IAAI,KAAK,GAAG,EAAE,CAAC;YACjB,MAAM,IAAI,GAAG,WAAW,CAAC,KAAK,EAAE,OAAO,CAAC,CAAC,IAAI,EAAE,CAAC;YAChD,IAAI,CAAC,IAAI;gBAAE,SAAS;YAEpB,sBAAsB;YACtB,MAAM,SAAS,GAAG,IAAI,CAAC,KAAK,CAAC,iCAAiC,CAAC,CAAC;YAChE,IAAI,SAAS,EAAE,CAAC;gBACd,QAAQ,CAAC,IAAI,CAAC;oBACZ,IAAI,EAAE,MAAM;oBACZ,KAAK,EAAE,SAAS,CAAC,CAAC,CAAC;oBACnB,IAAI,EAAE,IAAI,CAAC,SAAS,CAAC,SAAS,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,IAAI,EAAE;iBAClC,CAAC,CAAC;gBAClB,SAAS;YACX,CAAC;YAED,QAAQ,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,WAAW,EAAE,IAAI,EAAsB,CAAC,CAAC;YAC/D,SAAS;QACX,CAAC;QAED,OAAO;QACP,IAAI,IAAI,KAAK,GAAG,EAAE,CAAC;YACjB,MAAM,KAAK,GAAG,gBAAgB,CAAC,KAAK,EAAE,OAAO,CAAC,CAAC;YAC/C,IAAI,KAAK,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;gBACrB,QAAQ,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,MAAM,EAAE,KAAK,EAAiB,CAAC,CAAC;YACxD,CAAC;YACD,SAAS;QACX,CAAC;QAED,QAAQ;QACR,IAAI,IAAI,KAAK,OAAO,EAAE,CAAC;YACrB,MAAM,KAAK,GAAG,YAAY,CAAC,KAAK,EAAE,OAAO,CAAC,CAAC;YAC3C,IAAI,KAAK,EAAE,CAAC;gBACV,QAAQ,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;YACvB,CAAC;YACD,SAAS;QACX,CAAC;QAED,mEAAmE;QACnE,IAAI,IAAI,KAAK,MAAM,EAAE,CAAC;YACpB,MAAM,IAAI,GAAG,WAAW,CAAC,KAAK,EAAE,OAAO,CAAC,CAAC,IAAI,EAAE,CAAC;YAChD,IAAI,IAAI,EAAE,CAAC;gBACT,QAAQ,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,MAAM,EAAE,IAAI,EAAE,CAAC,CAAC;YACxC,CAAC;YACD,SAAS;QACX,CAAC;QAED,oEAAoE;QACpE,MAAM,aAAa,GAAG,cAAc,CAAC,KAAK,EAAE,OAAO,EAAE,MAAM,CAAC,CAAC;QAC7D,QAAQ,CAAC,IAAI,CAAC,GAAG,aAAa,CAAC,CAAC;IAClC,CAAC;IAED,OAAO,QAAQ,CAAC;AAClB,CAAC;AAED;;GAEG;AACH,SAAS,WAAW,CAAC,IAAoB,EAAE,OAA4B;IACrE,MAAM,KAAK,GAAa,EAAE,CAAC;IAE3B,IAAI,IAAI,CAAC,QAAQ,EAAE,CAAC;QAClB,KAAK,MAAM,KAAK,IAAI,IAAI,CAAC,QAAQ,EAAE,CAAC;YAClC,IAAI,eAAe,CAAC,KAAK,CAAC,EAAE,CAAC;gBAC3B,MAAM,IAAI,GAAG,OAAO,CAAC,GAAG,CAAC,KAAK,CAAC,EAAE,CAAC,CAAC;gBACnC,IAAI,IAAI;oBAAE,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;YAC7B,CAAC;iBAAM,CAAC;gBACN,yBAAyB;gBACzB,IAAI,KAAK,CAAC,IAAI,KAAK,UAAU;oBAAE,SAAS;gBACxC,KAAK,CAAC,IAAI,CAAC,WAAW,CAAC,KAAK,EAAE,OAAO,CAAC,CAAC,CAAC;YAC1C,CAAC;QACH,CAAC;IACH,CAAC;IAED,OAAO,KAAK,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;AACxB,CAAC;AAED;;GAEG;AACH,SAAS,gBAAgB,CAAC,IAAoB,EAAE,OAA4B;IAC1E,MAAM,KAAK,GAAa,EAAE,CAAC;IAE3B,IAAI,IAAI,CAAC,QAAQ,EAAE,CAAC;QAClB,KAAK,MAAM,KAAK,IAAI,IAAI,CAAC,QAAQ,EAAE,CAAC;YAClC,IAAI,eAAe,CAAC,KAAK,CAAC;gBAAE,SAAS;YACrC,IAAI,KAAK,CAAC,IAAI,KAAK,IAAI,EAAE,CAAC;gBACxB,uCAAuC;gBACvC,MAAM,QAAQ,GAAG,mBAAmB,CAAC,KAAK,EAAE,OAAO,CAAC,CAAC;gBACrD,IAAI,QAAQ,CAAC,IAAI,EAAE,EAAE,CAAC;oBACpB,KAAK,CAAC,IAAI,CAAC,QAAQ,CAAC,IAAI,EAAE,CAAC,CAAC;gBAC9B,CAAC;YACH,CAAC;QACH,CAAC;IACH,CAAC;IAED,OAAO,KAAK,CAAC;AACf,CAAC;AAED,SAAS,mBAAmB,CAAC,EAAkB,EAAE,OAA4B;IAC3E,MAAM,KAAK,GAAa,EAAE,CAAC;IAC3B,IAAI,EAAE,CAAC,QAAQ,EAAE,CAAC;QAChB,KAAK,MAAM,KAAK,IAAI,EAAE,CAAC,QAAQ,EAAE,CAAC;YAChC,IAAI,eAAe,CAAC,KAAK,CAAC,EAAE,CAAC;gBAC3B,MAAM,IAAI,GAAG,OAAO,CAAC,GAAG,CAAC,KAAK,CAAC,EAAE,CAAC,CAAC;gBACnC,IAAI,IAAI;oBAAE,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;YAC7B,CAAC;iBAAM,IAAI,KAAK,CAAC,IAAI,KAAK,OAAO,EAAE,CAAC;gBAClC,KAAK,CAAC,IAAI,CAAC,WAAW,CAAC,KAAK,EAAE,OAAO,CAAC,CAAC,CAAC;YAC1C,CAAC;iBAAM,IAAI,KAAK,CAAC,IAAI,KAAK,KAAK,EAAE,CAAC;gBAChC,4BAA4B;YAC9B,CAAC;iBAAM,CAAC;gBACN,KAAK,CAAC,IAAI,CAAC,WAAW,CAAC,KAAK,EAAE,OAAO,CAAC,CAAC,CAAC;YAC1C,CAAC;QACH,CAAC;IACH,CAAC;IACD,OAAO,KAAK,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;AACxB,CAAC;AAED;;GAEG;AACH,SAAS,YAAY,CAAC,IAAoB,EAAE,OAA4B;IACtE,MAAM,OAAO,GAAa,EAAE,CAAC;IAC7B,MAAM,IAAI,GAAe,EAAE,CAAC;IAE5B,IAAI,CAAC,IAAI,CAAC,QAAQ;QAAE,OAAO,IAAI,CAAC;IAEhC,6EAA6E;IAC7E,MAAM,QAAQ,GAAqB,EAAE,CAAC;IACtC,MAAM,QAAQ,GAAqB,EAAE,CAAC;IAEtC,KAAK,MAAM,KAAK,IAAI,IAAI,CAAC,QAAQ,EAAE,CAAC;QAClC,IAAI,eAAe,CAAC,KAAK,CAAC;YAAE,SAAS;QAErC,IAAI,KAAK,CAAC,IAAI,KAAK,IAAI,EAAE,CAAC;YACxB,2CAA2C;YAC3C,QAAQ,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;QACvB,CAAC;aAAM,IAAI,KAAK,CAAC,IAAI,KAAK,OAAO,EAAE,CAAC;YAClC,qDAAqD;YACrD,IAAI,KAAK,CAAC,QAAQ,EAAE,CAAC;gBACnB,KAAK,MAAM,EAAE,IAAI,KAAK,CAAC,QAAQ,EAAE,CAAC;oBAChC,IAAI,CAAC,eAAe,CAAC,EAAE,CAAC,IAAI,EAAE,CAAC,IAAI,KAAK,IAAI,EAAE,CAAC;wBAC7C,QAAQ,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;oBACpB,CAAC;gBACH,CAAC;YACH,CAAC;QACH,CAAC;aAAM,IAAI,KAAK,CAAC,IAAI,KAAK,OAAO,IAAI,KAAK,CAAC,IAAI,KAAK,OAAO,EAAE,CAAC;YAC5D,yDAAyD;YACzD,IAAI,KAAK,CAAC,QAAQ,EAAE,CAAC;gBACnB,KAAK,MAAM,EAAE,IAAI,KAAK,CAAC,QAAQ,EAAE,CAAC;oBAChC,IAAI,CAAC,eAAe,CAAC,EAAE,CAAC,IAAI,EAAE,CAAC,IAAI,KAAK,IAAI,EAAE,CAAC;wBAC7C,QAAQ,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;oBACpB,CAAC;gBACH,CAAC;YACH,CAAC;QACH,CAAC;IACH,CAAC;IAED,uCAAuC;IACvC,KAAK,MAAM,EAAE,IAAI,QAAQ,EAAE,CAAC;QAC1B,IAAI,EAAE,CAAC,QAAQ,EAAE,CAAC;YAChB,MAAM,KAAK,GAAa,EAAE,CAAC;YAC3B,KAAK,MAAM,IAAI,IAAI,EAAE,CAAC,QAAQ,EAAE,CAAC;gBAC/B,IAAI,eAAe,CAAC,IAAI,CAAC;oBAAE,SAAS;gBACpC,KAAK,CAAC,IAAI,CAAC,WAAW,CAAC,IAAI,EAAE,OAAO,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC;YAChD,CAAC;YACD,IAAI,KAAK,CAAC,MAAM,GAAG,CAAC,IAAI,OAAO,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;gBAC7C,OAAO,CAAC,IAAI,CAAC,GAAG,KAAK,CAAC,CAAC;YACzB,CAAC;iBAAM,IAAI,KAAK,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;gBAC5B,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;YACnB,CAAC;QACH,CAAC;IACH,CAAC;IAED,8DAA8D;IAC9D,IAAI,iBAAiB,GAAG,KAAK,CAAC;IAC9B,IAAI,OAAO,CAAC,MAAM,KAAK,CAAC,IAAI,QAAQ,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QAChD,MAAM,QAAQ,GAAG,QAAQ,CAAC,CAAC,CAAC,CAAC;QAC7B,IAAI,QAAQ,CAAC,QAAQ,EAAE,CAAC;YACtB,MAAM,KAAK,GAAG,QAAQ,CAAC,QAAQ;iBAC5B,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,eAAe,CAAC,CAAC,CAAC,CAAC;iBAClC,KAAK,CAAC,CAAC,CAAC,EAAE,EAAE,CAAE,CAAoB,CAAC,IAAI,KAAK,IAAI,CAAC,CAAC;YACrD,IAAI,KAAK,EAAE,CAAC;gBACV,MAAM,KAAK,GAAa,EAAE,CAAC;gBAC3B,KAAK,MAAM,IAAI,IAAI,QAAQ,CAAC,QAAQ,EAAE,CAAC;oBACrC,IAAI,eAAe,CAAC,IAAI,CAAC;wBAAE,SAAS;oBACpC,KAAK,CAAC,IAAI,CAAC,WAAW,CAAC,IAAI,EAAE,OAAO,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC;gBAChD,CAAC;gBACD,OAAO,CAAC,IAAI,CAAC,GAAG,KAAK,CAAC,CAAC;gBACvB,iBAAiB,GAAG,IAAI,CAAC;YAC3B,CAAC;QACH,CAAC;IACH,CAAC;IAED,yBAAyB;IACzB,KAAK,IAAI,CAAC,GAAG,iBAAiB,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,GAAG,QAAQ,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QACjE,MAAM,EAAE,GAAG,QAAQ,CAAC,CAAC,CAAC,CAAC;QACvB,IAAI,EAAE,CAAC,QAAQ,EAAE,CAAC;YAChB,MAAM,KAAK,GAAa,EAAE,CAAC;YAC3B,KAAK,MAAM,IAAI,IAAI,EAAE,CAAC,QAAQ,EAAE,CAAC;gBAC/B,IAAI,eAAe,CAAC,IAAI,CAAC;oBAAE,SAAS;gBACpC,KAAK,CAAC,IAAI,CAAC,WAAW,CAAC,IAAI,EAAE,OAAO,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC;YAChD,CAAC;YACD,IAAI,KAAK,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;gBACrB,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;YACnB,CAAC;QACH,CAAC;IACH,CAAC;IAED,IAAI,OAAO,CAAC,MAAM,KAAK,CAAC,IAAI,IAAI,CAAC,MAAM,KAAK,CAAC;QAAE,OAAO,IAAI,CAAC;IAE3D,OAAO,EAAE,IAAI,EAAE,OAAO,EAAE,OAAO,EAAE,IAAI,EAAE,CAAC;AAC1C,CAAC;AAED;;GAEG;AACH,SAAS,gBAAgB,CAAC,WAAwB;IAChD,MAAM,KAAK,GAAa,EAAE,CAAC;IAC3B,KAAK,MAAM,IAAI,IAAI,WAAW,CAAC,KAAK,EAAE,CAAC;QACrC,IAAI,UAAU,CAAC,IAAI,CAAC,EAAE,CAAC;YACrB,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;YACrB,IAAI,IAAI,CAAC,MAAM;gBAAE,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;QACpC,CAAC;IACH,CAAC;IACD,MAAM,IAAI,GAAG,KAAK,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC,IAAI,EAAE,CAAC;IACnC,IAAI,CAAC,IAAI;QAAE,OAAO,EAAE,CAAC;IACrB,OAAO,CAAC,EAAE,IAAI,EAAE,WAAW,EAAE,IAAI,EAAE,CAAC,CAAC;AACvC,CAAC;AAED,cAAc;AACd,SAAS,UAAU,CAAC,IAAqB;IACvC,OAAO,KAAK,IAAI,IAAI,CAAC;AACvB,CAAC;AAED,SAAS,eAAe,CAAC,IAAqB;IAC5C,OAAO,MAAM,IAAI,IAAI,CAAC;AACxB,CAAC;AAED,SAAS,eAAe,CAAC,KAAyC;IAChE,OAAO,MAAM,IAAI,KAAK,IAAK,KAA2B,CAAC,IAAI,KAAK,SAAS,CAAC;AAC5E,CAAC"}
|
|
@@ -7,6 +7,13 @@ import type { PDFDocumentProxy } from './pdf-loader.js';
|
|
|
7
7
|
/**
|
|
8
8
|
* Build full-text search index from all PDF pages.
|
|
9
9
|
* Uses chunked parallel processing for improved performance.
|
|
10
|
+
*
|
|
11
|
+
* A page where sections start is indexed as one entry per section segment, cut at the
|
|
12
|
+
* section headings by extractPageSegments (S-8). Handing the whole page to a single
|
|
13
|
+
* section misattributed the strip above the first heading — the tail of the previous
|
|
14
|
+
* section — to the next one, so search pointed at sections where the term cannot be
|
|
15
|
+
* found by get_section / get_tables. Pages where no section starts stay on the fast
|
|
16
|
+
* raw-text path: they belong entirely to the section flowing through them.
|
|
10
17
|
*/
|
|
11
18
|
export declare function buildSearchIndex(doc: PDFDocumentProxy, sectionIndex: SectionIndex): Promise<TextIndex>;
|
|
12
19
|
/**
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"search-index.d.ts","sourceRoot":"","sources":["../../src/services/search-index.ts"],"names":[],"mappings":"AAAA;;;GAGG;AAGH,OAAO,KAAK,EAAY,SAAS,EAAE,YAAY,EAAE,SAAS,EAAE,MAAM,mBAAmB,CAAC;
|
|
1
|
+
{"version":3,"file":"search-index.d.ts","sourceRoot":"","sources":["../../src/services/search-index.ts"],"names":[],"mappings":"AAAA;;;GAGG;AAGH,OAAO,KAAK,EAAY,SAAS,EAAE,YAAY,EAAE,SAAS,EAAE,MAAM,mBAAmB,CAAC;AAItF,OAAO,KAAK,EAAE,gBAAgB,EAAY,MAAM,iBAAiB,CAAC;AAElE;;;;;;;;;;GAUG;AACH,wBAAsB,gBAAgB,CACpC,GAAG,EAAE,gBAAgB,EACrB,YAAY,EAAE,YAAY,GACzB,OAAO,CAAC,SAAS,CAAC,CA2DpB;AAYD;;;GAGG;AACH,wBAAgB,eAAe,CAC7B,KAAK,EAAE,SAAS,EAChB,KAAK,EAAE,MAAM,EACb,UAAU,EAAE,MAAM,EAClB,YAAY,EAAE,YAAY,GACzB,SAAS,EAAE,CA2Eb"}
|
|
@@ -5,29 +5,59 @@
|
|
|
5
5
|
import { CONCURRENCY } from '../config.js';
|
|
6
6
|
import { mapConcurrent } from '../utils/concurrency.js';
|
|
7
7
|
import { logger } from '../utils/logger.js';
|
|
8
|
+
import { extractPageSegments } from './content-extractor.js';
|
|
8
9
|
/**
|
|
9
10
|
* Build full-text search index from all PDF pages.
|
|
10
11
|
* Uses chunked parallel processing for improved performance.
|
|
12
|
+
*
|
|
13
|
+
* A page where sections start is indexed as one entry per section segment, cut at the
|
|
14
|
+
* section headings by extractPageSegments (S-8). Handing the whole page to a single
|
|
15
|
+
* section misattributed the strip above the first heading — the tail of the previous
|
|
16
|
+
* section — to the next one, so search pointed at sections where the term cannot be
|
|
17
|
+
* found by get_section / get_tables. Pages where no section starts stay on the fast
|
|
18
|
+
* raw-text path: they belong entirely to the section flowing through them.
|
|
11
19
|
*/
|
|
12
20
|
export async function buildSearchIndex(doc, sectionIndex) {
|
|
13
21
|
const start = Date.now();
|
|
14
22
|
const totalPages = doc.numPages;
|
|
15
23
|
logger.info('SearchIndex', `Building search index for ${totalPages} pages (concurrency: ${CONCURRENCY.searchIndex})...`);
|
|
24
|
+
// Sections starting on each page, in document order (flatOrder is document-ordered).
|
|
25
|
+
const startsByPage = new Map();
|
|
26
|
+
for (const info of sectionIndex.flatOrder) {
|
|
27
|
+
const arr = startsByPage.get(info.page);
|
|
28
|
+
if (arr) {
|
|
29
|
+
arr.push(info.sectionNumber);
|
|
30
|
+
}
|
|
31
|
+
else {
|
|
32
|
+
startsByPage.set(info.page, [info.sectionNumber]);
|
|
33
|
+
}
|
|
34
|
+
}
|
|
16
35
|
const pageNumbers = Array.from({ length: totalPages }, (_, i) => i + 1);
|
|
17
|
-
const
|
|
18
|
-
const
|
|
19
|
-
|
|
20
|
-
|
|
21
|
-
|
|
22
|
-
|
|
23
|
-
.
|
|
24
|
-
|
|
25
|
-
|
|
36
|
+
const pagesNested = await mapConcurrent(pageNumbers, async (pageNum) => {
|
|
37
|
+
const startingSections = startsByPage.get(pageNum);
|
|
38
|
+
if (!startingSections) {
|
|
39
|
+
// No section starts here — the page belongs to the section flowing through it.
|
|
40
|
+
const page = await doc.getPage(pageNum);
|
|
41
|
+
const textContent = await page.getTextContent();
|
|
42
|
+
const text = textContent.items
|
|
43
|
+
.filter((item) => 'str' in item)
|
|
44
|
+
.map((item) => item.str)
|
|
45
|
+
.join(' ');
|
|
46
|
+
const section = findSectionForPage(sectionIndex, pageNum);
|
|
47
|
+
return [{ page: pageNum, section: section?.sectionNumber || '', text }];
|
|
48
|
+
}
|
|
49
|
+
// Sections start on this page — cut it at their headings. A null owner is the
|
|
50
|
+
// strip above the first heading: it belongs to the section flowing in from the
|
|
51
|
+
// previous page.
|
|
52
|
+
const flowingIn = findSectionForPage(sectionIndex, pageNum - 1);
|
|
53
|
+
const segments = await extractPageSegments(doc, pageNum, startingSections);
|
|
54
|
+
return segments.map((seg) => ({
|
|
26
55
|
page: pageNum,
|
|
27
|
-
section: section?.sectionNumber
|
|
28
|
-
text,
|
|
29
|
-
};
|
|
56
|
+
section: seg.section ?? flowingIn?.sectionNumber ?? '',
|
|
57
|
+
text: seg.text,
|
|
58
|
+
}));
|
|
30
59
|
}, CONCURRENCY.searchIndex);
|
|
60
|
+
const pages = pagesNested.flat();
|
|
31
61
|
const buildTime = Date.now() - start;
|
|
32
62
|
logger.info('SearchIndex', `Search index built in ${buildTime}ms (${totalPages} pages)`);
|
|
33
63
|
return { pages, buildTime };
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"search-index.js","sourceRoot":"","sources":["../../src/services/search-index.ts"],"names":[],"mappings":"AAAA;;;GAGG;AAEH,OAAO,EAAE,WAAW,EAAE,MAAM,cAAc,CAAC;AAE3C,OAAO,EAAE,aAAa,EAAE,MAAM,yBAAyB,CAAC;AACxD,OAAO,EAAE,MAAM,EAAE,MAAM,oBAAoB,CAAC;
|
|
1
|
+
{"version":3,"file":"search-index.js","sourceRoot":"","sources":["../../src/services/search-index.ts"],"names":[],"mappings":"AAAA;;;GAGG;AAEH,OAAO,EAAE,WAAW,EAAE,MAAM,cAAc,CAAC;AAE3C,OAAO,EAAE,aAAa,EAAE,MAAM,yBAAyB,CAAC;AACxD,OAAO,EAAE,MAAM,EAAE,MAAM,oBAAoB,CAAC;AAC5C,OAAO,EAAE,mBAAmB,EAAE,MAAM,wBAAwB,CAAC;AAG7D;;;;;;;;;;GAUG;AACH,MAAM,CAAC,KAAK,UAAU,gBAAgB,CACpC,GAAqB,EACrB,YAA0B;IAE1B,MAAM,KAAK,GAAG,IAAI,CAAC,GAAG,EAAE,CAAC;IACzB,MAAM,UAAU,GAAG,GAAG,CAAC,QAAQ,CAAC;IAEhC,MAAM,CAAC,IAAI,CACT,aAAa,EACb,6BAA6B,UAAU,wBAAwB,WAAW,CAAC,WAAW,MAAM,CAC7F,CAAC;IAEF,qFAAqF;IACrF,MAAM,YAAY,GAAG,IAAI,GAAG,EAAoB,CAAC;IACjD,KAAK,MAAM,IAAI,IAAI,YAAY,CAAC,SAAS,EAAE,CAAC;QAC1C,MAAM,GAAG,GAAG,YAAY,CAAC,GAAG,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;QACxC,IAAI,GAAG,EAAE,CAAC;YACR,GAAG,CAAC,IAAI,CAAC,IAAI,CAAC,aAAa,CAAC,CAAC;QAC/B,CAAC;aAAM,CAAC;YACN,YAAY,CAAC,GAAG,CAAC,IAAI,CAAC,IAAI,EAAE,CAAC,IAAI,CAAC,aAAa,CAAC,CAAC,CAAC;QACpD,CAAC;IACH,CAAC;IAED,MAAM,WAAW,GAAG,KAAK,CAAC,IAAI,CAAC,EAAE,MAAM,EAAE,UAAU,EAAE,EAAE,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC;IAExE,MAAM,WAAW,GAAiB,MAAM,aAAa,CACnD,WAAW,EACX,KAAK,EAAE,OAAO,EAAE,EAAE;QAChB,MAAM,gBAAgB,GAAG,YAAY,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC;QAEnD,IAAI,CAAC,gBAAgB,EAAE,CAAC;YACtB,+EAA+E;YAC/E,MAAM,IAAI,GAAG,MAAM,GAAG,CAAC,OAAO,CAAC,OAAO,CAAC,CAAC;YACxC,MAAM,WAAW,GAAG,MAAM,IAAI,CAAC,cAAc,EAAE,CAAC;YAChD,MAAM,IAAI,GAAG,WAAW,CAAC,KAAK;iBAC3B,MAAM,CAAC,CAAC,IAAI,EAAoB,EAAE,CAAC,KAAK,IAAI,IAAI,CAAC;iBACjD,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,GAAG,CAAC;iBACvB,IAAI,CAAC,GAAG,CAAC,CAAC;YAEb,MAAM,OAAO,GAAG,kBAAkB,CAAC,YAAY,EAAE,OAAO,CAAC,CAAC;YAC1D,OAAO,CAAC,EAAE,IAAI,EAAE,OAAO,EAAE,OAAO,EAAE,OAAO,EAAE,aAAa,IAAI,EAAE,EAAE,IAAI,EAAE,CAAC,CAAC;QAC1E,CAAC;QAED,8EAA8E;QAC9E,+EAA+E;QAC/E,iBAAiB;QACjB,MAAM,SAAS,GAAG,kBAAkB,CAAC,YAAY,EAAE,OAAO,GAAG,CAAC,CAAC,CAAC;QAChE,MAAM,QAAQ,GAAG,MAAM,mBAAmB,CAAC,GAAG,EAAE,OAAO,EAAE,gBAAgB,CAAC,CAAC;QAC3E,OAAO,QAAQ,CAAC,GAAG,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,CAAC;YAC5B,IAAI,EAAE,OAAO;YACb,OAAO,EAAE,GAAG,CAAC,OAAO,IAAI,SAAS,EAAE,aAAa,IAAI,EAAE;YACtD,IAAI,EAAE,GAAG,CAAC,IAAI;SACf,CAAC,CAAC,CAAC;IACN,CAAC,EACD,WAAW,CAAC,WAAW,CACxB,CAAC;IACF,MAAM,KAAK,GAAG,WAAW,CAAC,IAAI,EAAE,CAAC;IAEjC,MAAM,SAAS,GAAG,IAAI,CAAC,GAAG,EAAE,GAAG,KAAK,CAAC;IACrC,MAAM,CAAC,IAAI,CAAC,aAAa,EAAE,yBAAyB,SAAS,OAAO,UAAU,SAAS,CAAC,CAAC;IAEzF,OAAO,EAAE,KAAK,EAAE,SAAS,EAAE,CAAC;AAC9B,CAAC;AAED;;;GAGG;AACH,SAAS,kBAAkB,CAAC,IAAY;IACtC,OAAO,IAAI;SACR,OAAO,CAAC,UAAU,EAAE,GAAG,CAAC,CAAC,iCAAiC;SAC1D,OAAO,CAAC,MAAM,EAAE,GAAG,CAAC,CAAC,CAAC,uBAAuB;AAClD,CAAC;AAED;;;GAGG;AACH,MAAM,UAAU,eAAe,CAC7B,KAAgB,EAChB,KAAa,EACb,UAAkB,EAClB,YAA0B;IAE1B,MAAM,eAAe,GAAG,kBAAkB,CAAC,KAAK,CAAC,WAAW,EAAE,CAAC,CAAC;IAChE,MAAM,UAAU,GAAG,eAAe,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;IAC5E,MAAM,IAAI,GAAgB,EAAE,CAAC;IAE7B,KAAK,MAAM,QAAQ,IAAI,KAAK,CAAC,KAAK,EAAE,CAAC;QACnC,MAAM,cAAc,GAAG,kBAAkB,CAAC,QAAQ,CAAC,IAAI,CAAC,WAAW,EAAE,CAAC,CAAC;QAEvE,+BAA+B;QAC/B,IAAI,GAAG,GAAG,cAAc,CAAC,OAAO,CAAC,eAAe,CAAC,CAAC;QAClD,IAAI,KAAK,GAAG,CAAC,CAAC;QAEd,IAAI,GAAG,KAAK,CAAC,CAAC,EAAE,CAAC;YACf,iCAAiC;YACjC,IAAI,SAAS,GAAG,cAAc,CAAC,OAAO,CAAC,eAAe,CAAC,CAAC;YACxD,OAAO,SAAS,KAAK,CAAC,CAAC,EAAE,CAAC;gBACxB,KAAK,IAAI,CAAC,CAAC,CAAC,4CAA4C;gBACxD,SAAS,GAAG,cAAc,CAAC,OAAO,CAAC,eAAe,EAAE,SAAS,GAAG,eAAe,CAAC,MAAM,CAAC,CAAC;YAC1F,CAAC;QACH,CAAC;aAAM,IAAI,UAAU,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;YACjC,mDAAmD;YACnD,MAAM,UAAU,GAAG,UAAU,CAAC,KAAK,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,cAAc,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC,CAAC;YAC7E,IAAI,CAAC,UAAU;gBAAE,SAAS;YAE1B,6CAA6C;YAC7C,KAAK,MAAM,IAAI,IAAI,UAAU,EAAE,CAAC;gBAC9B,IAAI,SAAS,GAAG,cAAc,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC;gBAC7C,OAAO,SAAS,KAAK,CAAC,CAAC,EAAE,CAAC;oBACxB,KAAK,EAAE,CAAC;oBACR,SAAS,GAAG,cAAc,CAAC,OAAO,CAAC,IAAI,EAAE,SAAS,GAAG,IAAI,CAAC,MAAM,CAAC,CAAC;gBACpE,CAAC;YACH,CAAC;YAED,kEAAkE;YAClE,GAAG,GAAG,IAAI,CAAC,GAAG,CAAC,GAAG,UAAU,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,cAAc,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC;QAC9F,CAAC;aAAM,CAAC;YACN,SAAS;QACX,CAAC;QAED,kEAAkE;QAClE,MAAM,UAAU,GAAG,oBAAoB,CAAC,QAAQ,CAAC,IAAI,EAAE,GAAG,EAAE,cAAc,CAAC,CAAC;QAC5E,MAAM,YAAY,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,UAAU,GAAG,EAAE,CAAC,CAAC;QAClD,MAAM,UAAU,GAAG,IAAI,CAAC,GAAG,CAAC,QAAQ,CAAC,IAAI,CAAC,MAAM,EAAE,UAAU,GAAG,KAAK,CAAC,MAAM,GAAG,EAAE,CAAC,CAAC;QAClF,MAAM,OAAO,GACX,CAAC,YAAY,GAAG,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE,CAAC;YAC/B,QAAQ,CAAC,IAAI,CAAC,SAAS,CAAC,YAAY,EAAE,UAAU,CAAC;YACjD,CAAC,UAAU,GAAG,QAAQ,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC;QAEnD,wBAAwB;QACxB,MAAM,WAAW,GAAG,QAAQ,CAAC,OAAO,CAAC,CAAC,CAAC,YAAY,CAAC,QAAQ,CAAC,GAAG,CAAC,QAAQ,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC;QAE/F,IAAI,CAAC,IAAI,CAAC;YACR,OAAO,EAAE,QAAQ,CAAC,OAAO;YACzB,KAAK,EAAE,WAAW,EAAE,KAAK,IAAI,EAAE;YAC/B,IAAI,EAAE,QAAQ,CAAC,IAAI;YACnB,OAAO,EAAE,OAAO,CAAC,IAAI,EAAE;YACvB,KAAK;SACN,CAAC,CAAC;IACL,CAAC;IAED,2BAA2B;IAC3B,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,KAAK,GAAG,CAAC,CAAC,KAAK,CAAC,CAAC;IAEvC,6CAA6C;IAC7C,MAAM,WAAW,GAAG,IAAI,GAAG,EAAU,CAAC;IACtC,MAAM,OAAO,GAAgB,EAAE,CAAC;IAChC,KAAK,MAAM,GAAG,IAAI,IAAI,EAAE,CAAC;QACvB,MAAM,GAAG,GAAG,GAAG,CAAC,OAAO,IAAI,QAAQ,GAAG,CAAC,IAAI,EAAE,CAAC;QAC9C,IAAI,WAAW,CAAC,GAAG,CAAC,GAAG,CAAC;YAAE,SAAS;QACnC,WAAW,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;QACrB,OAAO,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;QAClB,IAAI,OAAO,CAAC,MAAM,IAAI,UAAU;YAAE,MAAM;IAC1C,CAAC;IAED,OAAO,OAAO,CAAC;AACjB,CAAC;AAED;;GAEG;AACH,SAAS,oBAAoB,CAC3B,QAAgB,EAChB,aAAqB,EACrB,WAAmB;IAEnB,4DAA4D;IAC5D,IAAI,aAAa,IAAI,CAAC;QAAE,OAAO,CAAC,CAAC;IACjC,MAAM,KAAK,GAAG,aAAa,GAAG,WAAW,CAAC,MAAM,CAAC;IACjD,OAAO,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,KAAK,CAAC,KAAK,GAAG,QAAQ,CAAC,MAAM,CAAC,EAAE,QAAQ,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;AAC5E,CAAC;AAED;;GAEG;AACH,SAAS,kBAAkB,CACzB,YAA0B,EAC1B,IAAY;IAEZ,8CAA8C;IAC9C,MAAM,IAAI,GAAG,YAAY,CAAC,SAAS,CAAC;IACpC,IAAI,MAA6C,CAAC;IAElD,KAAK,MAAM,IAAI,IAAI,IAAI,EAAE,CAAC;QACxB,IAAI,IAAI,CAAC,IAAI,IAAI,IAAI,EAAE,CAAC;YACtB,MAAM,GAAG,EAAE,aAAa,EAAE,IAAI,CAAC,aAAa,EAAE,CAAC;QACjD,CAAC;aAAM,CAAC;YACN,MAAM;QACR,CAAC;IACH,CAAC;IAED,OAAO,MAAM,CAAC;AAChB,CAAC"}
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"table-collector.d.ts","sourceRoot":"","sources":["../../src/services/table-collector.ts"],"names":[],"mappings":"AAAA;;;;;;;;GAQG;AAGH,OAAO,KAAK,EAAE,cAAc,
|
|
1
|
+
{"version":3,"file":"table-collector.d.ts","sourceRoot":"","sources":["../../src/services/table-collector.ts"],"names":[],"mappings":"AAAA;;;;;;;;GAQG;AAGH,OAAO,KAAK,EAAE,cAAc,EAAgB,SAAS,EAAE,MAAM,mBAAmB,CAAC;AAEjF;;;;;;GAMG;AACH,wBAAgB,uBAAuB,CAAC,OAAO,EAAE,cAAc,EAAE,GAAG,SAAS,EAAE,CAmE9E"}
|
|
@@ -17,6 +17,11 @@ import { TABLE_CAPTION_START_RE } from '../config.js';
|
|
|
17
17
|
*/
|
|
18
18
|
export function collectStructTreeTables(content) {
|
|
19
19
|
const tables = [];
|
|
20
|
+
// Content index of the last `table` element folded into `tables`, so the headerless
|
|
21
|
+
// rule below can demand adjacency. A page break splits one table into consecutive
|
|
22
|
+
// `table` elements; an intervening paragraph means the next table is a different one
|
|
23
|
+
// (8.7.4.5.5: the Mesh 2 flag list must not be merged into Mesh 1's).
|
|
24
|
+
let lastTableElementIndex = -1;
|
|
20
25
|
for (let i = 0; i < content.length; i++) {
|
|
21
26
|
const element = content[i];
|
|
22
27
|
if (element.type !== 'table')
|
|
@@ -29,29 +34,63 @@ export function collectStructTreeTables(content) {
|
|
|
29
34
|
caption = prev.text;
|
|
30
35
|
}
|
|
31
36
|
}
|
|
32
|
-
|
|
33
|
-
|
|
34
|
-
|
|
35
|
-
|
|
36
|
-
|
|
37
|
-
|
|
38
|
-
|
|
39
|
-
|
|
37
|
+
if (!caption && tables.length > 0) {
|
|
38
|
+
const last = tables[tables.length - 1];
|
|
39
|
+
// Continuation, form 1: the header row is repeated (ISO restates it when a table
|
|
40
|
+
// resumes on a new page). Matching headers identify the fragment regardless of
|
|
41
|
+
// whatever lies between (the repeated header row itself is dropped).
|
|
42
|
+
if (element.headers.length > 0 && arraysEqual(last.headers, element.headers)) {
|
|
43
|
+
last.rows.push(...element.rows.map((row) => [...row]));
|
|
44
|
+
lastTableElementIndex = i;
|
|
45
|
+
continue;
|
|
46
|
+
}
|
|
47
|
+
// Continuation, form 2 (S-4): the fragment carries no usable header — either none
|
|
48
|
+
// at all, or one whose cells are all blank (image-only rows extract as ""). Headers
|
|
49
|
+
// cannot identify it, so instead require that it directly follows the previous
|
|
50
|
+
// fragment and has the same column count. A blank header row is dropped like a
|
|
51
|
+
// repeated one: whether it was a restated header or an image-only row, it holds no
|
|
52
|
+
// text either way.
|
|
53
|
+
if (i === lastTableElementIndex + 1 &&
|
|
54
|
+
isEffectivelyHeaderless(element) &&
|
|
55
|
+
columnCountOf(element.headers, element.rows) === columnCountOf(last.headers, last.rows)) {
|
|
56
|
+
last.rows.push(...element.rows.map((row) => [...row]));
|
|
57
|
+
lastTableElementIndex = i;
|
|
58
|
+
continue;
|
|
59
|
+
}
|
|
40
60
|
}
|
|
41
61
|
tables.push({
|
|
42
62
|
index: tables.length,
|
|
43
63
|
caption,
|
|
44
|
-
// Copy, never alias
|
|
45
|
-
// to `element.rows` and then pushing
|
|
46
|
-
// so every later call sees a table that
|
|
47
|
-
// get_requirements returned 4 table requirements,
|
|
48
|
-
// declared idempotent and read-only, and this made it
|
|
64
|
+
// Copy, never alias — and copy continuation rows too (above). `content` is the
|
|
65
|
+
// section content cache: handing out a reference to `element.rows` and then pushing
|
|
66
|
+
// into it rewrites the cached page itself, so every later call sees a table that
|
|
67
|
+
// has grown. get_tables followed by get_requirements returned 4 table requirements,
|
|
68
|
+
// then 7, then 10 — the tool is declared idempotent and read-only, and this made it
|
|
69
|
+
// neither. Inner row arrays alias the cache just the same: a caller editing a cell
|
|
70
|
+
// of the result must not reach the cached page.
|
|
49
71
|
headers: [...element.headers],
|
|
50
72
|
rows: element.rows.map((row) => [...row]),
|
|
51
73
|
});
|
|
74
|
+
lastTableElementIndex = i;
|
|
52
75
|
}
|
|
53
76
|
return tables;
|
|
54
77
|
}
|
|
78
|
+
/**
|
|
79
|
+
* A table fragment that headers cannot identify: none at all, or all cells blank.
|
|
80
|
+
* (Image-only header rows extract as empty strings — 8.4.3.4 / 10.6.3.)
|
|
81
|
+
*/
|
|
82
|
+
function isEffectivelyHeaderless(element) {
|
|
83
|
+
return element.headers.length === 0 || element.headers.every((h) => h.trim() === '');
|
|
84
|
+
}
|
|
85
|
+
/**
|
|
86
|
+
* Column count of a table or fragment: header width when there is a header row,
|
|
87
|
+
* otherwise the width of the first data row.
|
|
88
|
+
*/
|
|
89
|
+
function columnCountOf(headers, rows) {
|
|
90
|
+
if (headers.length > 0)
|
|
91
|
+
return headers.length;
|
|
92
|
+
return rows[0]?.length ?? 0;
|
|
93
|
+
}
|
|
55
94
|
/**
|
|
56
95
|
* Check if two string arrays are equal
|
|
57
96
|
*/
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"table-collector.js","sourceRoot":"","sources":["../../src/services/table-collector.ts"],"names":[],"mappings":"AAAA;;;;;;;;GAQG;AAEH,OAAO,EAAE,sBAAsB,EAAE,MAAM,cAAc,CAAC;AAGtD;;;;;;GAMG;AACH,MAAM,UAAU,uBAAuB,CAAC,OAAyB;IAC/D,MAAM,MAAM,GAAgB,EAAE,CAAC;IAE/B,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,OAAO,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QACxC,MAAM,OAAO,GAAG,OAAO,CAAC,CAAC,CAAC,CAAC;QAC3B,IAAI,OAAO,CAAC,IAAI,KAAK,OAAO;YAAE,SAAS;QAEvC,2CAA2C;QAC3C,IAAI,OAAO,GAAkB,IAAI,CAAC;QAClC,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC;YACV,MAAM,IAAI,GAAG,OAAO,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC;YAC5B,IAAI,IAAI,CAAC,IAAI,KAAK,WAAW,IAAI,sBAAsB,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC;gBACxE,OAAO,GAAG,IAAI,CAAC,IAAI,CAAC;YACtB,CAAC;QACH,CAAC;QAED,
|
|
1
|
+
{"version":3,"file":"table-collector.js","sourceRoot":"","sources":["../../src/services/table-collector.ts"],"names":[],"mappings":"AAAA;;;;;;;;GAQG;AAEH,OAAO,EAAE,sBAAsB,EAAE,MAAM,cAAc,CAAC;AAGtD;;;;;;GAMG;AACH,MAAM,UAAU,uBAAuB,CAAC,OAAyB;IAC/D,MAAM,MAAM,GAAgB,EAAE,CAAC;IAC/B,oFAAoF;IACpF,kFAAkF;IAClF,qFAAqF;IACrF,sEAAsE;IACtE,IAAI,qBAAqB,GAAG,CAAC,CAAC,CAAC;IAE/B,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,OAAO,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QACxC,MAAM,OAAO,GAAG,OAAO,CAAC,CAAC,CAAC,CAAC;QAC3B,IAAI,OAAO,CAAC,IAAI,KAAK,OAAO;YAAE,SAAS;QAEvC,2CAA2C;QAC3C,IAAI,OAAO,GAAkB,IAAI,CAAC;QAClC,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC;YACV,MAAM,IAAI,GAAG,OAAO,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC;YAC5B,IAAI,IAAI,CAAC,IAAI,KAAK,WAAW,IAAI,sBAAsB,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC;gBACxE,OAAO,GAAG,IAAI,CAAC,IAAI,CAAC;YACtB,CAAC;QACH,CAAC;QAED,IAAI,CAAC,OAAO,IAAI,MAAM,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;YAClC,MAAM,IAAI,GAAG,MAAM,CAAC,MAAM,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;YAEvC,iFAAiF;YACjF,+EAA+E;YAC/E,qEAAqE;YACrE,IAAI,OAAO,CAAC,OAAO,CAAC,MAAM,GAAG,CAAC,IAAI,WAAW,CAAC,IAAI,CAAC,OAAO,EAAE,OAAO,CAAC,OAAO,CAAC,EAAE,CAAC;gBAC7E,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,GAAG,OAAO,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,CAAC,GAAG,GAAG,CAAC,CAAC,CAAC,CAAC;gBACvD,qBAAqB,GAAG,CAAC,CAAC;gBAC1B,SAAS;YACX,CAAC;YAED,kFAAkF;YAClF,oFAAoF;YACpF,+EAA+E;YAC/E,+EAA+E;YAC/E,mFAAmF;YACnF,mBAAmB;YACnB,IACE,CAAC,KAAK,qBAAqB,GAAG,CAAC;gBAC/B,uBAAuB,CAAC,OAAO,CAAC;gBAChC,aAAa,CAAC,OAAO,CAAC,OAAO,EAAE,OAAO,CAAC,IAAI,CAAC,KAAK,aAAa,CAAC,IAAI,CAAC,OAAO,EAAE,IAAI,CAAC,IAAI,CAAC,EACvF,CAAC;gBACD,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,GAAG,OAAO,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,CAAC,GAAG,GAAG,CAAC,CAAC,CAAC,CAAC;gBACvD,qBAAqB,GAAG,CAAC,CAAC;gBAC1B,SAAS;YACX,CAAC;QACH,CAAC;QAED,MAAM,CAAC,IAAI,CAAC;YACV,KAAK,EAAE,MAAM,CAAC,MAAM;YACpB,OAAO;YACP,+EAA+E;YAC/E,oFAAoF;YACpF,iFAAiF;YACjF,oFAAoF;YACpF,oFAAoF;YACpF,mFAAmF;YACnF,gDAAgD;YAChD,OAAO,EAAE,CAAC,GAAG,OAAO,CAAC,OAAO,CAAC;YAC7B,IAAI,EAAE,OAAO,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,CAAC,GAAG,GAAG,CAAC,CAAC;SAC1C,CAAC,CAAC;QACH,qBAAqB,GAAG,CAAC,CAAC;IAC5B,CAAC;IAED,OAAO,MAAM,CAAC;AAChB,CAAC;AAED;;;GAGG;AACH,SAAS,uBAAuB,CAAC,OAAqB;IACpD,OAAO,OAAO,CAAC,OAAO,CAAC,MAAM,KAAK,CAAC,IAAI,OAAO,CAAC,OAAO,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,EAAE,KAAK,EAAE,CAAC,CAAC;AACvF,CAAC;AAED;;;GAGG;AACH,SAAS,aAAa,CAAC,OAAiB,EAAE,IAAgB;IACxD,IAAI,OAAO,CAAC,MAAM,GAAG,CAAC;QAAE,OAAO,OAAO,CAAC,MAAM,CAAC;IAC9C,OAAO,IAAI,CAAC,CAAC,CAAC,EAAE,MAAM,IAAI,CAAC,CAAC;AAC9B,CAAC;AAED;;GAEG;AACH,SAAS,WAAW,CAAC,CAAW,EAAE,CAAW;IAC3C,IAAI,CAAC,CAAC,MAAM,KAAK,CAAC,CAAC,MAAM;QAAE,OAAO,KAAK,CAAC;IACxC,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,CAAC,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QAClC,IAAI,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC;YAAE,OAAO,KAAK,CAAC;IAClC,CAAC;IACD,OAAO,IAAI,CAAC;AACd,CAAC"}
|