@pellux/goodvibes-sdk 0.26.8 → 0.26.10

package/dist/_internal/platform/knowledge/extractors.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"extractors.d.ts","sourceRoot":"","sources":["../../../../src/_internal/platform/knowledge/extractors.ts"],"names":[],"mappings":"AAEA,OAAO,KAAK,EAAsB,cAAc,EAAE,MAAM,uBAAuB,CAAC;AAGhF,OAAO,KAAK,EAAE,yBAAyB,EAAE,MAAM,YAAY,CAAC;AAI5D,MAAM,WAAW,yBAAyB;IACxC,QAAQ,CAAC,WAAW,EAAE,MAAM,CAAC;IAC7B,QAAQ,CAAC,MAAM,EAAE,yBAAyB,CAAC;IAC3C,QAAQ,CAAC,KAAK,CAAC,EAAE,MAAM,CAAC;IACxB,QAAQ,CAAC,OAAO,CAAC,EAAE,MAAM,CAAC;IAC1B,QAAQ,CAAC,OAAO,CAAC,EAAE,MAAM,CAAC;IAC1B,QAAQ,CAAC,QAAQ,EAAE,SAAS,MAAM,EAAE,CAAC;IACrC,QAAQ,CAAC,KAAK,EAAE,SAAS,MAAM,EAAE,CAAC;IAClC,QAAQ,CAAC,eAAe,EAAE,MAAM,CAAC;IACjC,QAAQ,CAAC,SAAS,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,CAAC;IAC5C,QAAQ,CAAC,QAAQ,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,CAAC;CAC5C;AA6hBD,wBAAsB,wBAAwB,CAC5C,QAAQ,EAAE,IAAI,CAAC,cAAc,EAAE,IAAI,GAAG,UAAU,GAAG,UAAU,CAAC,EAC9D,MAAM,EAAE,MAAM,GACb,OAAO,CAAC,yBAAyB,CAAC,CA4CpC"}
+{"version":3,"file":"extractors.d.ts","sourceRoot":"","sources":["../../../../src/_internal/platform/knowledge/extractors.ts"],"names":[],"mappings":"AAEA,OAAO,KAAK,EAAsB,cAAc,EAAE,MAAM,uBAAuB,CAAC;AAGhF,OAAO,KAAK,EAAE,yBAAyB,EAAE,MAAM,YAAY,CAAC;AAI5D,MAAM,WAAW,yBAAyB;IACxC,QAAQ,CAAC,WAAW,EAAE,MAAM,CAAC;IAC7B,QAAQ,CAAC,MAAM,EAAE,yBAAyB,CAAC;IAC3C,QAAQ,CAAC,KAAK,CAAC,EAAE,MAAM,CAAC;IACxB,QAAQ,CAAC,OAAO,CAAC,EAAE,MAAM,CAAC;IAC1B,QAAQ,CAAC,OAAO,CAAC,EAAE,MAAM,CAAC;IAC1B,QAAQ,CAAC,QAAQ,EAAE,SAAS,MAAM,EAAE,CAAC;IACrC,QAAQ,CAAC,KAAK,EAAE,SAAS,MAAM,EAAE,CAAC;IAClC,QAAQ,CAAC,eAAe,EAAE,MAAM,CAAC;IACjC,QAAQ,CAAC,SAAS,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,CAAC;IAC5C,QAAQ,CAAC,QAAQ,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,CAAC;CAC5C;AAmmBD,wBAAsB,wBAAwB,CAC5C,QAAQ,EAAE,IAAI,CAAC,cAAc,EAAE,IAAI,GAAG,UAAU,GAAG,UAAU,CAAC,EAC9D,MAAM,EAAE,MAAM,GACb,OAAO,CAAC,yBAAyB,CAAC,CA4CpC"}

package/dist/_internal/platform/knowledge/extractors.js

@@ -306,7 +306,79 @@ function extractYaml(buffer) {
         metadata: {},
     };
 }
-function extractPdf(buffer) {
+async function extractPdf(buffer) {
+    const parsed = await extractPdfWithPdfJs(buffer);
+    if (parsed)
+        return parsed;
+    return extractPdfRawStreams(buffer);
+}
+async function extractPdfWithPdfJs(buffer) {
+    try {
+        const pdfjs = await import('pdfjs-dist/legacy/build/pdf.mjs');
+        const loadingTask = pdfjs.getDocument({
+            data: new Uint8Array(buffer),
+            useSystemFonts: true,
+        });
+        const document = await loadingTask.promise;
+        const pageCount = document.numPages;
+        const pageTexts = [];
+        for (let pageNumber = 1; pageNumber <= pageCount; pageNumber += 1) {
+            const page = await document.getPage(pageNumber);
+            const content = await page.getTextContent();
+            const lines = textContentItemsToLines(content.items);
+            if (lines.length > 0)
+                pageTexts.push(lines.join('\n'));
+            page.cleanup();
+        }
+        await document.destroy();
+        const text = cleanText(pageTexts.join('\n\n'));
+        if (!text)
+            return undefined;
+        const searchText = searchTextPayload(text);
+        return {
+            extractorId: 'pdfjs',
+            format: 'pdf',
+            title: firstNonEmptyLine(text) ?? 'PDF document',
+            summary: summarizeText(text) ?? 'PDF document.',
+            excerpt: excerptText(text),
+            sections: uniqueStrings(text.split(/\n+/), 24),
+            links: uniqueStrings(Array.from(text.matchAll(/\bhttps?:\/\/[^\s)]+/g), (match) => match[0]), 50),
+            estimatedTokens: estimateTokens(text),
+            structure: {
+                pageCount,
+                extractedTextChars: text.length,
+                ...(searchText ? { searchText } : {}),
+            },
+            metadata: {
+                limitations: ['PDF text extraction does not perform OCR for scanned images.'],
+            },
+        };
+    }
+    catch {
+        return undefined;
+    }
+}
+function textContentItemsToLines(items) {
+    const lines = [];
+    let current = '';
+    for (const item of items) {
+        const record = unknownRecord(item);
+        const text = typeof record.str === 'string' ? cleanText(record.str) : '';
+        if (text)
+            current = current ? `${current} ${text}` : text;
+        if (record.hasEOL === true && current) {
+            lines.push(current);
+            current = '';
+        }
+    }
+    if (current)
+        lines.push(current);
+    return lines;
+}
+function unknownRecord(value) {
+    return value && typeof value === 'object' ? value : {};
+}
+function extractPdfRawStreams(buffer) {
     const body = buffer.toString('latin1');
     const texts = [];
     const streamRe = /stream\r?\n([\s\S]*?)\r?\nendstream/g;

package/dist/_internal/platform/knowledge/home-graph/index.d.ts

@@ -1,4 +1,4 @@
 export { HomeGraphService } from './service.js';
 export { HOME_GRAPH_NODE_KINDS, HOME_GRAPH_RELATIONS, } from './types.js';
-export type { HomeGraphAskInput, HomeGraphAskResult, HomeGraphDevicePassportResult, HomeGraphExport, HomeGraphIngestArtifactInput, HomeGraphIngestNoteInput, HomeGraphIngestResult, HomeGraphIngestUrlInput, HomeGraphKnowledgeTarget, HomeGraphLinkInput, HomeGraphLinkResult, HomeGraphNodeKind, HomeGraphObjectInput, HomeGraphObjectKind, HomeGraphProjectionInput, HomeGraphProjectionResult, HomeGraphRelation, HomeGraphReviewInput, HomeGraphSnapshotInput, HomeGraphStatus, HomeGraphSyncResult, } from './types.js';
+export type { HomeGraphAskInput, HomeGraphAskResult, HomeGraphDevicePassportResult, HomeGraphExport, HomeGraphIngestArtifactInput, HomeGraphIngestNoteInput, HomeGraphIngestResult, HomeGraphIngestUrlInput, HomeGraphKnowledgeTarget, HomeGraphLinkInput, HomeGraphLinkResult, HomeGraphNodeKind, HomeGraphObjectInput, HomeGraphObjectKind, HomeGraphProjectionInput, HomeGraphProjectionResult, HomeGraphReindexResult, HomeGraphRelation, HomeGraphReviewInput, HomeGraphSnapshotInput, HomeGraphStatus, HomeGraphSyncResult, } from './types.js';
 //# sourceMappingURL=index.d.ts.map

package/dist/_internal/platform/knowledge/home-graph/index.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../../../../../src/_internal/platform/knowledge/home-graph/index.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,gBAAgB,EAAE,MAAM,cAAc,CAAC;AAChD,OAAO,EACL,qBAAqB,EACrB,oBAAoB,GACrB,MAAM,YAAY,CAAC;AACpB,YAAY,EACV,iBAAiB,EACjB,kBAAkB,EAClB,6BAA6B,EAC7B,eAAe,EACf,4BAA4B,EAC5B,wBAAwB,EACxB,qBAAqB,EACrB,uBAAuB,EACvB,wBAAwB,EACxB,kBAAkB,EAClB,mBAAmB,EACnB,iBAAiB,EACjB,oBAAoB,EACpB,mBAAmB,EACnB,wBAAwB,EACxB,yBAAyB,EACzB,iBAAiB,EACjB,oBAAoB,EACpB,sBAAsB,EACtB,eAAe,EACf,mBAAmB,GACpB,MAAM,YAAY,CAAC"}
+{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../../../../../src/_internal/platform/knowledge/home-graph/index.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,gBAAgB,EAAE,MAAM,cAAc,CAAC;AAChD,OAAO,EACL,qBAAqB,EACrB,oBAAoB,GACrB,MAAM,YAAY,CAAC;AACpB,YAAY,EACV,iBAAiB,EACjB,kBAAkB,EAClB,6BAA6B,EAC7B,eAAe,EACf,4BAA4B,EAC5B,wBAAwB,EACxB,qBAAqB,EACrB,uBAAuB,EACvB,wBAAwB,EACxB,kBAAkB,EAClB,mBAAmB,EACnB,iBAAiB,EACjB,oBAAoB,EACpB,mBAAmB,EACnB,wBAAwB,EACxB,yBAAyB,EACzB,sBAAsB,EACtB,iBAAiB,EACjB,oBAAoB,EACpB,sBAAsB,EACtB,eAAe,EACf,mBAAmB,GACpB,MAAM,YAAY,CAAC"}

package/dist/_internal/platform/knowledge/home-graph/reindex.d.ts

@@ -0,0 +1,12 @@
+import type { ArtifactStore } from '../../artifacts/index.js';
+import type { ArtifactDescriptor } from '../../artifacts/types.js';
+import type { KnowledgeExtractionRecord, KnowledgeSourceRecord } from '../types.js';
+import type { HomeGraphReindexResult } from './types.js';
+export declare function reindexHomeGraphSources(input: {
+    readonly spaceId: string;
+    readonly sources: readonly KnowledgeSourceRecord[];
+    readonly extractionBySourceId: ReadonlyMap<string, KnowledgeExtractionRecord>;
+    readonly artifactStore: ArtifactStore;
+    readonly extract: (source: KnowledgeSourceRecord, artifact: ArtifactDescriptor) => Promise<KnowledgeExtractionRecord | undefined>;
+}): Promise<HomeGraphReindexResult>;
+//# sourceMappingURL=reindex.d.ts.map

package/dist/_internal/platform/knowledge/home-graph/reindex.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"reindex.d.ts","sourceRoot":"","sources":["../../../../../src/_internal/platform/knowledge/home-graph/reindex.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,aAAa,EAAE,MAAM,0BAA0B,CAAC;AAC9D,OAAO,KAAK,EAAE,kBAAkB,EAAE,MAAM,0BAA0B,CAAC;AACnE,OAAO,KAAK,EAAE,yBAAyB,EAAE,qBAAqB,EAAE,MAAM,aAAa,CAAC;AAEpF,OAAO,KAAK,EAAE,sBAAsB,EAAE,MAAM,YAAY,CAAC;AAEzD,wBAAsB,uBAAuB,CAAC,KAAK,EAAE;IACnD,QAAQ,CAAC,OAAO,EAAE,MAAM,CAAC;IACzB,QAAQ,CAAC,OAAO,EAAE,SAAS,qBAAqB,EAAE,CAAC;IACnD,QAAQ,CAAC,oBAAoB,EAAE,WAAW,CAAC,MAAM,EAAE,yBAAyB,CAAC,CAAC;IAC9E,QAAQ,CAAC,aAAa,EAAE,aAAa,CAAC;IACtC,QAAQ,CAAC,OAAO,EAAE,CAAC,MAAM,EAAE,qBAAqB,EAAE,QAAQ,EAAE,kBAAkB,KAAK,OAAO,CAAC,yBAAyB,GAAG,SAAS,CAAC,CAAC;CACnI,GAAG,OAAO,CAAC,sBAAsB,CAAC,CAgClC"}

package/dist/_internal/platform/knowledge/home-graph/reindex.js

@@ -0,0 +1,35 @@
+import { homeGraphExtractionNeedsRepair } from './search.js';
+export async function reindexHomeGraphSources(input) {
+    const sources = [];
+    const failures = [];
+    let scanned = 0;
+    let reparsed = 0;
+    let skipped = 0;
+    let failed = 0;
+    for (const source of input.sources) {
+        const artifactId = typeof source.artifactId === 'string' ? source.artifactId : undefined;
+        if (!artifactId)
+            continue;
+        scanned += 1;
+        const current = input.extractionBySourceId.get(source.id);
+        if (!homeGraphExtractionNeedsRepair(current)) {
+            skipped += 1;
+            continue;
+        }
+        const artifact = input.artifactStore.get(artifactId);
+        if (!artifact) {
+            failed += 1;
+            failures.push({ sourceId: source.id, error: `Unknown artifact: ${artifactId}` });
+            continue;
+        }
+        const extraction = await input.extract(source, artifact);
+        if (!extraction) {
+            failed += 1;
+            failures.push({ sourceId: source.id, error: 'Artifact extraction failed.' });
+            continue;
+        }
+        reparsed += 1;
+        sources.push(source);
+    }
+    return { ok: true, spaceId: input.spaceId, scanned, reparsed, skipped, failed, sources, failures };
+}

package/dist/_internal/platform/knowledge/home-graph/search.d.ts

@@ -9,5 +9,7 @@ export interface HomeGraphSearchState {
     readonly extractionBySourceId: ReadonlyMap<string, KnowledgeExtractionRecord>;
 }
 export declare function readHomeGraphSearchState(store: KnowledgeStore, spaceId: string): HomeGraphSearchState;
-export declare function scoreHomeGraphResults(query: string, sources: readonly KnowledgeSourceRecord[], nodes: readonly KnowledgeNodeRecord[], extractionBySourceId: (sourceId: string) => KnowledgeExtractionRecord | null | undefined, limit: number): HomeGraphSearchResult[];
+export declare function scoreHomeGraphResults(query: string, sources: readonly KnowledgeSourceRecord[], nodes: readonly KnowledgeNodeRecord[], edges: readonly KnowledgeEdgeRecord[], extractionBySourceId: (sourceId: string) => KnowledgeExtractionRecord | null | undefined, limit: number): HomeGraphSearchResult[];
+export declare function selectHomeGraphExtractionRepairCandidates(query: string, sources: readonly KnowledgeSourceRecord[], nodes: readonly KnowledgeNodeRecord[], edges: readonly KnowledgeEdgeRecord[], extractionBySourceId: (sourceId: string) => KnowledgeExtractionRecord | null | undefined, limit: number): KnowledgeSourceRecord[];
+export declare function homeGraphExtractionNeedsRepair(extraction: KnowledgeExtractionRecord | null | undefined): boolean;
 //# sourceMappingURL=search.d.ts.map

package/dist/_internal/platform/knowledge/home-graph/search.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"search.d.ts","sourceRoot":"","sources":["../../../../../src/_internal/platform/knowledge/home-graph/search.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,cAAc,EAAE,MAAM,aAAa,CAAC;AAClD,OAAO,KAAK,EACV,mBAAmB,EACnB,yBAAyB,EACzB,mBAAmB,EACnB,qBAAqB,EACtB,MAAM,aAAa,CAAC;AAErB,OAAO,KAAK,EAAE,qBAAqB,EAAE,MAAM,YAAY,CAAC;AAMxD,MAAM,WAAW,oBAAoB;IACnC,QAAQ,CAAC,OAAO,EAAE,MAAM,CAAC;IACzB,QAAQ,CAAC,OAAO,EAAE,SAAS,qBAAqB,EAAE,CAAC;IACnD,QAAQ,CAAC,KAAK,EAAE,SAAS,mBAAmB,EAAE,CAAC;IAC/C,QAAQ,CAAC,KAAK,EAAE,SAAS,mBAAmB,EAAE,CAAC;IAC/C,QAAQ,CAAC,oBAAoB,EAAE,WAAW,CAAC,MAAM,EAAE,yBAAyB,CAAC,CAAC;CAC/E;AAED,wBAAgB,wBAAwB,CAAC,KAAK,EAAE,cAAc,EAAE,OAAO,EAAE,MAAM,GAAG,oBAAoB,CAqBrG;AAED,wBAAgB,qBAAqB,CACnC,KAAK,EAAE,MAAM,EACb,OAAO,EAAE,SAAS,qBAAqB,EAAE,EACzC,KAAK,EAAE,SAAS,mBAAmB,EAAE,EACrC,oBAAoB,EAAE,CAAC,QAAQ,EAAE,MAAM,KAAK,yBAAyB,GAAG,IAAI,GAAG,SAAS,EACxF,KAAK,EAAE,MAAM,GACZ,qBAAqB,EAAE,CA+CzB"}
+{"version":3,"file":"search.d.ts","sourceRoot":"","sources":["../../../../../src/_internal/platform/knowledge/home-graph/search.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,cAAc,EAAE,MAAM,aAAa,CAAC;AAClD,OAAO,KAAK,EACV,mBAAmB,EACnB,yBAAyB,EACzB,mBAAmB,EACnB,qBAAqB,EACtB,MAAM,aAAa,CAAC;AAErB,OAAO,KAAK,EAAE,qBAAqB,EAAE,MAAM,YAAY,CAAC;AAgFxD,MAAM,WAAW,oBAAoB;IACnC,QAAQ,CAAC,OAAO,EAAE,MAAM,CAAC;IACzB,QAAQ,CAAC,OAAO,EAAE,SAAS,qBAAqB,EAAE,CAAC;IACnD,QAAQ,CAAC,KAAK,EAAE,SAAS,mBAAmB,EAAE,CAAC;IAC/C,QAAQ,CAAC,KAAK,EAAE,SAAS,mBAAmB,EAAE,CAAC;IAC/C,QAAQ,CAAC,oBAAoB,EAAE,WAAW,CAAC,MAAM,EAAE,yBAAyB,CAAC,CAAC;CAC/E;AAED,wBAAgB,wBAAwB,CAAC,KAAK,EAAE,cAAc,EAAE,OAAO,EAAE,MAAM,GAAG,oBAAoB,CAqBrG;AAED,wBAAgB,qBAAqB,CACnC,KAAK,EAAE,MAAM,EACb,OAAO,EAAE,SAAS,qBAAqB,EAAE,EACzC,KAAK,EAAE,SAAS,mBAAmB,EAAE,EACrC,KAAK,EAAE,SAAS,mBAAmB,EAAE,EACrC,oBAAoB,EAAE,CAAC,QAAQ,EAAE,MAAM,KAAK,yBAAyB,GAAG,IAAI,GAAG,SAAS,EACxF,KAAK,EAAE,MAAM,GACZ,qBAAqB,EAAE,CAwEzB;AAED,wBAAgB,yCAAyC,CACvD,KAAK,EAAE,MAAM,EACb,OAAO,EAAE,SAAS,qBAAqB,EAAE,EACzC,KAAK,EAAE,SAAS,mBAAmB,EAAE,EACrC,KAAK,EAAE,SAAS,mBAAmB,EAAE,EACrC,oBAAoB,EAAE,CAAC,QAAQ,EAAE,MAAM,KAAK,yBAAyB,GAAG,IAAI,GAAG,SAAS,EACxF,KAAK,EAAE,MAAM,GACZ,qBAAqB,EAAE,CA6BzB;AA0ID,wBAAgB,8BAA8B,CAAC,UAAU,EAAE,yBAAyB,GAAG,IAAI,GAAG,SAAS,GAAG,OAAO,CAWhH"}