@carlonicora/nestjs-neo4jsonapi 0.0.1

package/dist/foundations/chunker/services/types/pdf/interfaces/pdf-layout.interface.d.ts

@@ -0,0 +1,86 @@
+export interface PdfPage {
+    pageNumber: number;
+    width: number;
+    height: number;
+    elements: PdfLayoutElement[];
+    columns?: PdfColumn[];
+    headerRegion?: PdfRegion;
+    footerRegion?: PdfRegion;
+    contentRegion?: PdfRegion;
+}
+export interface PdfLayoutElement {
+    x: number;
+    y: number;
+    width: number;
+    height: number;
+    content: string;
+    fontSize: number;
+    fontName: string;
+    fontWeight?: string;
+    isItalic?: boolean;
+    isBold?: boolean;
+    color?: string;
+    backgroundColor?: string;
+    pageNumber: number;
+    elementId: string;
+    confidence?: number;
+}
+export interface PdfColumn {
+    x: number;
+    y: number;
+    width: number;
+    height: number;
+    elements: PdfLayoutElement[];
+    columnIndex: number;
+}
+export interface PdfRegion {
+    x: number;
+    y: number;
+    width: number;
+    height: number;
+    type: "header" | "footer" | "content" | "sidebar" | "margin";
+}
+export interface PdfBoundingBox {
+    x: number;
+    y: number;
+    width: number;
+    height: number;
+}
+export interface PdfTextLine {
+    elements: PdfLayoutElement[];
+    boundingBox: PdfBoundingBox;
+    baseline: number;
+    lineHeight: number;
+    readingOrder: number;
+}
+export interface PdfBlock {
+    type: "text" | "table" | "image" | "header" | "list" | "paragraph";
+    elements: PdfLayoutElement[];
+    boundingBox: PdfBoundingBox;
+    lines?: PdfTextLine[];
+    readingOrder: number;
+    confidence: number;
+}
+export interface PdfDocumentStructure {
+    pages: PdfPage[];
+    totalPages: number;
+    documentType?: "single_column" | "multi_column" | "mixed" | "table_heavy" | "scanned";
+    hasImages: boolean;
+    hasTables: boolean;
+    hasScannedContent: boolean;
+    confidence: number;
+}
+export interface PdfProcessingOptions {
+    enableOCR: boolean;
+    ocrConfidenceThreshold: number;
+    ocrLanguage?: string;
+    ocrImagePreprocessing?: boolean;
+    detectTables: boolean;
+    detectImages: boolean;
+    detectHeaders: boolean;
+    preserveLayout: boolean;
+    maxImageSize: number;
+    skipHeaderFooter: boolean;
+    columnDetection: boolean;
+}
+//# sourceMappingURL=pdf-layout.interface.d.ts.map

package/dist/foundations/chunker/services/types/pdf/interfaces/pdf-layout.interface.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"pdf-layout.interface.d.ts","sourceRoot":"","sources":["../../../../../../../src/foundations/chunker/services/types/pdf/interfaces/pdf-layout.interface.ts"],"names":[],"mappings":"AAAA,MAAM,WAAW,OAAO;IACtB,UAAU,EAAE,MAAM,CAAC;IACnB,KAAK,EAAE,MAAM,CAAC;IACd,MAAM,EAAE,MAAM,CAAC;IACf,QAAQ,EAAE,gBAAgB,EAAE,CAAC;IAC7B,OAAO,CAAC,EAAE,SAAS,EAAE,CAAC;IACtB,YAAY,CAAC,EAAE,SAAS,CAAC;IACzB,YAAY,CAAC,EAAE,SAAS,CAAC;IACzB,aAAa,CAAC,EAAE,SAAS,CAAC;CAC3B;AAED,MAAM,WAAW,gBAAgB;IAC/B,CAAC,EAAE,MAAM,CAAC;IACV,CAAC,EAAE,MAAM,CAAC;IACV,KAAK,EAAE,MAAM,CAAC;IACd,MAAM,EAAE,MAAM,CAAC;IACf,OAAO,EAAE,MAAM,CAAC;IAChB,QAAQ,EAAE,MAAM,CAAC;IACjB,QAAQ,EAAE,MAAM,CAAC;IACjB,UAAU,CAAC,EAAE,MAAM,CAAC;IACpB,QAAQ,CAAC,EAAE,OAAO,CAAC;IACnB,MAAM,CAAC,EAAE,OAAO,CAAC;IACjB,KAAK,CAAC,EAAE,MAAM,CAAC;IACf,eAAe,CAAC,EAAE,MAAM,CAAC;IACzB,UAAU,EAAE,MAAM,CAAC;IACnB,SAAS,EAAE,MAAM,CAAC;IAClB,UAAU,CAAC,EAAE,MAAM,CAAC;CACrB;AAED,MAAM,WAAW,SAAS;IACxB,CAAC,EAAE,MAAM,CAAC;IACV,CAAC,EAAE,MAAM,CAAC;IACV,KAAK,EAAE,MAAM,CAAC;IACd,MAAM,EAAE,MAAM,CAAC;IACf,QAAQ,EAAE,gBAAgB,EAAE,CAAC;IAC7B,WAAW,EAAE,MAAM,CAAC;CACrB;AAED,MAAM,WAAW,SAAS;IACxB,CAAC,EAAE,MAAM,CAAC;IACV,CAAC,EAAE,MAAM,CAAC;IACV,KAAK,EAAE,MAAM,CAAC;IACd,MAAM,EAAE,MAAM,CAAC;IACf,IAAI,EAAE,QAAQ,GAAG,QAAQ,GAAG,SAAS,GAAG,SAAS,GAAG,QAAQ,CAAC;CAC9D;AAED,MAAM,WAAW,cAAc;IAC7B,CAAC,EAAE,MAAM,CAAC;IACV,CAAC,EAAE,MAAM,CAAC;IACV,KAAK,EAAE,MAAM,CAAC;IACd,MAAM,EAAE,MAAM,CAAC;CAChB;AAED,MAAM,WAAW,WAAW;IAC1B,QAAQ,EAAE,gBAAgB,EAAE,CAAC;IAC7B,WAAW,EAAE,cAAc,CAAC;IAC5B,QAAQ,EAAE,MAAM,CAAC;IACjB,UAAU,EAAE,MAAM,CAAC;IACnB,YAAY,EAAE,MAAM,CAAC;CACtB;AAED,MAAM,WAAW,QAAQ;IACvB,IAAI,EAAE,MAAM,GAAG,OAAO,GAAG,OAAO,GAAG,QAAQ,GAAG,MAAM,GAAG,WAAW,CAAC;IACnE,QAAQ,EAAE,gBAAgB,EAAE,CAAC;IAC7B,WAAW,EAAE,cAAc,CAAC;IAC5B,KAAK,CAAC,EAAE,WAAW,EAAE,CAAC;IACtB,YAAY,EAAE,MAAM,CAAC;IACrB,UAAU,EAAE,MAAM,CAAC;CACpB;AAED,MAAM,WAAW,oBAAoB;IACnC,KAAK,EAAE,OAAO,EAAE,CAAC;IACjB,UAAU,EAAE,MAAM,CAAC;IACnB,YAAY,CAAC,EAAE,eAAe,GAAG,cAAc,GAAG,OAAO,GAAG,aAAa,GAAG,SAAS,CAAC;IACtF,SAAS,EAAE,OAAO,CAAC;IACnB,SAAS,EAAE,OAAO,CAAC;IACnB,iBAAiB,EAAE,OAAO,CAAC;IAC3B,UAAU,EAAE,MAAM,CAAC;CACpB;AAED,MAAM,WAAW,oBAAoB;IACnC,SAAS,EAAE,OAAO,CAAC;IACnB,sBAAsB,EAAE,MAAM,CAAC;IAC/B,WAAW,CAAC,EAAE,MAAM,CAAC;IACrB,qBAAqB,CAAC,EAAE,OAAO,CAAC;IAChC,YAAY,EAAE,OAAO,CAAC;IACtB,YAAY,EAAE,OAAO,CAAC;IACtB,aAAa,EAAE,OAAO,CAAC;IACvB,cAAc,EAAE,OAAO,CAAC;IACxB,YAAY,EAAE,MAAM,CAAC;IACrB,gBAAgB,EAAE,OAAO,CAAC;IAC1B,eAAe,EAAE,OAAO,CAAC;CAC1B"}

package/dist/foundations/chunker/services/types/pdf/interfaces/pdf-layout.interface.js

@@ -0,0 +1,3 @@
+"use strict";
+Object.defineProperty(exports, "__esModule", { value: true });
+//# sourceMappingURL=pdf-layout.interface.js.map

package/dist/foundations/chunker/services/types/pdf/interfaces/pdf-layout.interface.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"pdf-layout.interface.js","sourceRoot":"","sources":["../../../../../../../src/foundations/chunker/services/types/pdf/interfaces/pdf-layout.interface.ts"],"names":[],"mappings":""}

package/dist/foundations/chunker/services/types/pdf.service.d.ts

@@ -0,0 +1,64 @@
+import { PdfContent, PdfElement } from "./pdf/interfaces/pdf-element.interface";
+import { PdfProcessingOptions } from "./pdf/interfaces/pdf-layout.interface";
+import { Document } from "@langchain/core/documents";
+export declare class PdfService {
+    private readonly logger;
+    private readonly layoutExtractor;
+    private readonly tableExtractor;
+    private defaultProcessingOptions;
+    resetState(): void;
+    convertToMarkdown(pdfElements: PdfElement[]): string;
+    private extractElementAsMarkdown;
+    private extractTableAsMarkdown;
+    extractPdfContent(pdfPath: string, options?: Partial<PdfProcessingOptions>): Promise<PdfContent[]>;
+    private extractWithIntelligentParsing;
+    private createLayoutElementsFromText;
+    private groupElementsByPage;
+    private convertLayoutToContent;
+    private convertTableToText;
+    private analyzeAndStructureText;
+    private cleanAndNormalizeText;
+    private splitIntoSentences;
+    private reconstructLogicalParagraphs;
+    private shouldEndParagraph;
+    private isHeaderParagraph;
+    private detectHeader;
+    private detectTableLine;
+    private processTableLines;
+    private detectScannedPdf;
+    private extractWithBasicParsing;
+    private extractWithOCR;
+    private convertPdfToImages;
+    private processImageWithOCR;
+    private isGarbageOcrOutput;
+    /**
+     * Clean up OCR artifacts like stamps, margin text, and other trailing garbage
+     * Detects and removes trailing sections with consecutive problematic lines
+     * @param text Raw OCR text output
+     * @returns Cleaned text with trailing artifacts removed
+     */
+    private cleanupOcrArtifacts;
+    /**
+     * Check if a line is problematic (likely an artifact/stamp/margin text)
+     * @param line Single line of text (trimmed)
+     * @returns true if the line looks like garbage/artifact
+     */
+    private isProblematicLine;
+    /**
+     * Detect and correct image rotation using Tesseract OSD (Orientation and Script Detection)
+     * Returns the correctly oriented image buffer
+     */
+    private detectAndCorrectRotation;
+    /**
+     * Assess image quality to determine if preprocessing would help or damage the image
+     * High-quality scans with clean text should SKIP preprocessing to avoid damaging text
+     * Low-quality scans with noise/blur should USE preprocessing for enhancement
+     */
+    private shouldPreprocessImage;
+    private preprocessImageForOCR;
+    getRawElements(filePath: string, options?: Partial<PdfProcessingOptions>): Promise<any[]>;
+    load(params: {
+        filePath: string;
+    }, options?: Partial<PdfProcessingOptions>): Promise<Document[]>;
+}
+//# sourceMappingURL=pdf.service.d.ts.map

package/dist/foundations/chunker/services/types/pdf.service.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"pdf.service.d.ts","sourceRoot":"","sources":["../../../../../src/foundations/chunker/services/types/pdf.service.ts"],"names":[],"mappings":"AAIA,OAAO,EACL,UAAU,EACV,UAAU,EAKX,MAAM,wCAAwC,CAAC;AAChD,OAAO,EAAoB,oBAAoB,EAAE,MAAM,uCAAuC,CAAC;AAE/F,OAAO,EAAE,QAAQ,EAAE,MAAM,2BAA2B,CAAC;AAMrD,qBACa,UAAU;IACrB,OAAO,CAAC,QAAQ,CAAC,MAAM,CAA+B;IACtD,OAAO,CAAC,QAAQ,CAAC,eAAe,CAAyB;IACzD,OAAO,CAAC,QAAQ,CAAC,cAAc,CAAwB;IAEvD,OAAO,CAAC,wBAAwB,CAY9B;IAEK,UAAU,IAAI,IAAI;IAIlB,iBAAiB,CAAC,WAAW,EAAE,UAAU,EAAE,GAAG,MAAM;IAQ3D,OAAO,CAAC,wBAAwB;IA8BhC,OAAO,CAAC,sBAAsB;IAiCxB,iBAAiB,CAAC,OAAO,EAAE,MAAM,EAAE,OAAO,CAAC,EAAE,OAAO,CAAC,oBAAoB,CAAC,GAAG,OAAO,CAAC,UAAU,EAAE,CAAC;YA2G1F,6BAA6B;IA6D3C,OAAO,CAAC,4BAA4B;IAiDpC,OAAO,CAAC,mBAAmB;IAc3B,OAAO,CAAC,sBAAsB;IA6C9B,OAAO,CAAC,kBAAkB;IAkB1B,OAAO,CAAC,uBAAuB;IA0C/B,OAAO,CAAC,qBAAqB;IAmB7B,OAAO,CAAC,kBAAkB;IAkC1B,OAAO,CAAC,4BAA4B;IA0BpC,OAAO,CAAC,kBAAkB;IAwB1B,OAAO,CAAC,iBAAiB;IAOzB,OAAO,CAAC,YAAY;IAiCpB,OAAO,CAAC,eAAe;IAavB,OAAO,CAAC,iBAAiB;IA4CzB,OAAO,CAAC,gBAAgB;YAuFV,uBAAuB;YAgCvB,cAAc;YA0Hd,kBAAkB;YA2GlB,mBAAmB;IAoIjC,OAAO,CAAC,kBAAkB;IA4H1B;;;;;OAKG;IACH,OAAO,CAAC,mBAAmB;IAuE3B;;;;OAIG;IACH,OAAO,CAAC,iBAAiB;IAwCzB;;;OAGG;YACW,wBAAwB;IA4CtC;;;;OAIG;YACW,qBAAqB;YAyCrB,qBAAqB;IAuC7B,cAAc,CAAC,QAAQ,EAAE,MAAM,EAAE,OAAO,CAAC,EAAE,OAAO,CAAC,oBAAoB,CAAC,GAAG,OAAO,CAAC,GAAG,EAAE,CAAC;IASzF,IAAI,CAAC,MAAM,EAAE;QAAE,QAAQ,EAAE,MAAM,CAAA;KAAE,EAAE,OAAO,CAAC,EAAE,OAAO,CAAC,oBAAoB,CAAC,GAAG,OAAO,CAAC,QAAQ,EAAE,CAAC;CAevG"}