@absolutejs/absolute 0.19.0-beta.506 → 0.19.0-beta.507
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/ai/index.js +42 -6
- package/dist/ai/index.js.map +5 -5
- package/package.json +1 -1
package/dist/ai/index.js
CHANGED
|
@@ -2221,10 +2221,20 @@ var extractWeightedLexicalFields = (result) => {
|
|
|
2221
2221
|
const source = result.source ?? "";
|
|
2222
2222
|
const archivePath = typeof metadata.archivePath === "string" ? metadata.archivePath : source.includes("#") ? source.split("#")[1] ?? "" : "";
|
|
2223
2223
|
const mediaSegments = Array.isArray(metadata.mediaSegments) ? metadata.mediaSegments.map((segment) => segment && typeof segment === "object" ? toFieldText(segment) : "").filter(Boolean).join(" ") : "";
|
|
2224
|
+
const spreadsheetFocus = metadata.sourceNativeKind === "spreadsheet_sheet" ? [
|
|
2225
|
+
"spreadsheet",
|
|
2226
|
+
"workbook",
|
|
2227
|
+
"worksheet",
|
|
2228
|
+
"sheet",
|
|
2229
|
+
typeof metadata.sheetName === "string" ? `sheet named ${metadata.sheetName}` : "",
|
|
2230
|
+
typeof metadata.sheetIndex === "number" ? `worksheet ${metadata.sheetIndex + 1}` : ""
|
|
2231
|
+
].filter(Boolean).join(" ") : "";
|
|
2224
2232
|
const metadataFocus = [
|
|
2225
2233
|
metadata.sourceNativeKind,
|
|
2234
|
+
spreadsheetFocus,
|
|
2226
2235
|
metadata.sheetName,
|
|
2227
2236
|
metadata.sheetNames,
|
|
2237
|
+
metadata.sheetIndex,
|
|
2228
2238
|
metadata.slideNumber,
|
|
2229
2239
|
metadata.slideTitle,
|
|
2230
2240
|
metadata.slideTitles,
|
|
@@ -2248,10 +2258,10 @@ var FIELD_WEIGHTS = {
|
|
|
2248
2258
|
archivePath: 4.2,
|
|
2249
2259
|
chunkText: 1,
|
|
2250
2260
|
mediaSegments: 3.8,
|
|
2251
|
-
metadataFocus:
|
|
2261
|
+
metadataFocus: 4.1,
|
|
2252
2262
|
metadataText: 1.4,
|
|
2253
2263
|
source: 3.4,
|
|
2254
|
-
title: 2
|
|
2264
|
+
title: 2.8
|
|
2255
2265
|
};
|
|
2256
2266
|
var getWeightedFieldTokens = (result) => {
|
|
2257
2267
|
const fields = extractWeightedLexicalFields({
|
|
@@ -2626,6 +2636,7 @@ var DOMAIN_EXPANSIONS = {
|
|
|
2626
2636
|
video: ["video", "media", "recording", "transcript", "timestamp"]
|
|
2627
2637
|
};
|
|
2628
2638
|
var TERM_EXPANSIONS = {
|
|
2639
|
+
audio: ["timestamp", "transcript", "mp3", "speaker"],
|
|
2629
2640
|
frontend: [
|
|
2630
2641
|
"frontend",
|
|
2631
2642
|
"framework",
|
|
@@ -2649,7 +2660,9 @@ var TERM_EXPANSIONS = {
|
|
|
2649
2660
|
procedure: ["recovery", "runbook"],
|
|
2650
2661
|
procedur: ["recovery", "runbook"],
|
|
2651
2662
|
receipt: ["invoice", "ocr", "pdf"],
|
|
2663
|
+
named: ["sheet", "worksheet", "title"],
|
|
2652
2664
|
sheet: ["worksheet", "workbook", "xlsx"],
|
|
2665
|
+
timestamp: ["audio", "media", "transcript", "segment"],
|
|
2653
2666
|
transcript: ["audio", "video", "media"],
|
|
2654
2667
|
workbook: ["sheet", "spreadsheet", "xlsx"]
|
|
2655
2668
|
};
|
|
@@ -2699,6 +2712,22 @@ var createHeuristicRAGQueryTransform = (options = {}) => createRAGQueryTransform
|
|
|
2699
2712
|
}
|
|
2700
2713
|
const expandedTerms = domains.flatMap((domain) => DOMAIN_EXPANSIONS[domain] ?? []);
|
|
2701
2714
|
const tokenExpansions = tokens.flatMap((token) => TERM_EXPANSIONS[token] ?? []);
|
|
2715
|
+
const spreadsheetNamedVariant = domains.includes("spreadsheet") ? uniqueQueryStrings([
|
|
2716
|
+
...tokens,
|
|
2717
|
+
"spreadsheet",
|
|
2718
|
+
"workbook",
|
|
2719
|
+
"worksheet",
|
|
2720
|
+
"sheet",
|
|
2721
|
+
"named"
|
|
2722
|
+
]).join(" ") : "";
|
|
2723
|
+
const mediaTimestampVariant = domains.includes("audio") || domains.includes("video") ? uniqueQueryStrings([
|
|
2724
|
+
...tokens,
|
|
2725
|
+
"audio",
|
|
2726
|
+
"media",
|
|
2727
|
+
"timestamp",
|
|
2728
|
+
"transcript",
|
|
2729
|
+
"segment"
|
|
2730
|
+
]).join(" ") : "";
|
|
2702
2731
|
const rewrittenQuery = uniqueQueryStrings([
|
|
2703
2732
|
...tokens,
|
|
2704
2733
|
...expandedTerms,
|
|
@@ -2709,9 +2738,15 @@ var createHeuristicRAGQueryTransform = (options = {}) => createRAGQueryTransform
|
|
|
2709
2738
|
...DOMAIN_EXPANSIONS[domain] ?? [],
|
|
2710
2739
|
...tokenExpansions
|
|
2711
2740
|
]).join(" "));
|
|
2741
|
+
if (spreadsheetNamedVariant.length > 0) {
|
|
2742
|
+
variants.push(spreadsheetNamedVariant);
|
|
2743
|
+
}
|
|
2744
|
+
if (mediaTimestampVariant.length > 0) {
|
|
2745
|
+
variants.push(mediaTimestampVariant);
|
|
2746
|
+
}
|
|
2712
2747
|
return {
|
|
2713
2748
|
query: rewrittenQuery,
|
|
2714
|
-
variants
|
|
2749
|
+
variants: uniqueQueryStrings(variants)
|
|
2715
2750
|
};
|
|
2716
2751
|
}
|
|
2717
2752
|
});
|
|
@@ -3552,6 +3587,7 @@ var createOfficeDocumentExtractor = () => ({
|
|
|
3552
3587
|
} else if (extension === ".xlsx" || extension === ".ods") {
|
|
3553
3588
|
text = spreadsheetText(entries);
|
|
3554
3589
|
const sheets = spreadsheetSheetTexts(entries);
|
|
3590
|
+
const workbookLabel = input.title ?? input.name ?? input.path ?? input.source ?? DEFAULT_BINARY_NAME;
|
|
3555
3591
|
officeMetadata = {
|
|
3556
3592
|
sheetNames: spreadsheetSheetNames(entries)
|
|
3557
3593
|
};
|
|
@@ -3568,9 +3604,9 @@ var createOfficeDocumentExtractor = () => ({
|
|
|
3568
3604
|
sheetName: sheet.name
|
|
3569
3605
|
},
|
|
3570
3606
|
source: input.source ?? input.path ?? input.name ?? `${slugify(input.title ?? DEFAULT_BINARY_NAME)}${extension || ".office"}`,
|
|
3571
|
-
text: normalizeWhitespace(`Spreadsheet sheet ${sheet.name}
|
|
3607
|
+
text: normalizeWhitespace(`Spreadsheet workbook ${workbookLabel}. ` + `Worksheet ${index + 1}. ` + `Workbook sheet named ${sheet.name}. ` + `Sheet ${sheet.name} from spreadsheet workbook ${workbookLabel}.` + `
|
|
3572
3608
|
${sheet.text}`),
|
|
3573
|
-
title: input.title ? `${input.title} \xB7 ${sheet.name}` : sheet.name
|
|
3609
|
+
title: input.title ? `${input.title} \xB7 Sheet ${sheet.name}` : `Sheet ${sheet.name}`
|
|
3574
3610
|
}));
|
|
3575
3611
|
} else if (extension === ".pptx" || extension === ".odp") {
|
|
3576
3612
|
text = presentationText(entries);
|
|
@@ -10202,5 +10238,5 @@ export {
|
|
|
10202
10238
|
aiChat
|
|
10203
10239
|
};
|
|
10204
10240
|
|
|
10205
|
-
//# debugId=
|
|
10241
|
+
//# debugId=80FE4F97C8F27FEB64756E2164756E21
|
|
10206
10242
|
//# sourceMappingURL=index.js.map
|