@heripo/document-processor 0.1.22 → 0.1.23

package/dist/index.d.cts

@@ -1,9 +1,63 @@
 import { LoggerMethods } from '@heripo/logger';
-import { TokenUsageReport, DoclingDocument, DocumentProcessResult, Caption, PageRange, DoclingTextItem, DoclingPictureItem, DoclingTableItem, DoclingGroupItem, ProcessedImage, ProcessedTable, ProcessedFootnote, Chapter } from '@heripo/model';
+import { TokenUsageReport, DoclingDocument, PageRange, DocumentProcessResult, Caption, DoclingTextItem, DoclingPictureItem, DoclingTableItem, DoclingGroupItem, ProcessedImage, ProcessedTable, ProcessedFootnote, Chapter } from '@heripo/model';
 import { LanguageModel } from 'ai';
 import { LLMTokenUsageAggregator, ExtendedTokenUsage } from '@heripo/shared';
 import { z } from 'zod';
 
+/**
+ * Table of Contents Entry
+ *
+ * Tree node representing the table of contents structure of a document.
+ */
+interface TocEntry {
+    /**
+     * Chapter title
+     */
+    title: string;
+    /**
+     * Hierarchy depth (1, 2, 3...)
+     */
+    level: number;
+    /**
+     * Starting page number
+     */
+    pageNo: number;
+    /**
+     * Child TOC entries
+     */
+    children?: TocEntry[];
+}
+/**
+ * TOC Area Search Result
+ */
+interface TocAreaResult {
+    /**
+     * Group or table item references corresponding to the table of contents
+     */
+    itemRefs: string[];
+    /**
+     * TOC start page
+     */
+    startPage: number;
+    /**
+     * TOC end page
+     */
+    endPage: number;
+}
+/**
+ * Page Size Information
+ */
+interface PageSizeGroup {
+    /**
+     * Size identifier (width x height)
+     */
+    sizeKey: string;
+    /**
+     * PDF page numbers with this size specification
+     */
+    pageNos: number[];
+}
+
 /**
  * DocumentProcessor Options
  */
@@ -59,6 +113,10 @@ interface DocumentProcessorOptions {
      * Maximum retry count (default: 3)
      */
     maxRetries?: number;
+    /**
+     * Maximum retry count for TOC validation correction feedback (default: 3)
+     */
+    maxValidationRetries?: number;
     /**
      * Enable fallback retry mechanism - automatically retries with fallback model on failure (default: false)
      * Set to true to enable automatic fallback retry with fallback model on component-specific model errors
@@ -76,6 +134,21 @@ interface DocumentProcessorOptions {
      */
     onTokenUsage?: (report: TokenUsageReport) => void;
 }
+/**
+ * Per-document processing inputs.
+ */
+interface DocumentProcessorProcessOptions {
+    /**
+     * Precomputed PDF page number to actual document page range mapping.
+     * When provided, automatic PageRangeParser execution is skipped.
+     */
+    pageRangeMap?: Record<number, PageRange>;
+    /**
+     * Precomputed table of contents entries.
+     * When provided, automatic TOC extraction pipeline execution is skipped.
+     */
+    tocEntries?: TocEntry[];
+}
 /**
  * DocumentProcessor
  *
@@ -121,12 +194,13 @@ interface DocumentProcessorOptions {
  *   captionParserBatchSize: 10,    // LLM caption parsing
  *   captionValidatorBatchSize: 10, // LLM caption validation
  *   maxRetries: 3,
+ *   maxValidationRetries: 3,
  * });
  *
  * const result = await processor.process(
  *   doclingDoc,
  *   'report-001',
- *   outputPath
+ *   artifactDir
  * );
  * ```
  */
@@ -142,6 +216,7 @@ declare class DocumentProcessor {
     private readonly captionParserBatchSize;
     private readonly captionValidatorBatchSize;
     private readonly maxRetries;
+    private readonly maxValidationRetries;
     private readonly enableFallbackRetry;
     private readonly abortSignal?;
     private readonly onTokenUsage?;
@@ -179,7 +254,7 @@ declare class DocumentProcessor {
      * Conversion process:
      * 1. Initialize processors and resolvers
      * 2. Normalize and filter texts
-     * 3. Clean texts and parse page ranges (parallel)
+     * 3. Clean texts and resolve page ranges
      * 4. Extract table of contents
      * 5. Convert images and tables (parallel)
      * 6. Convert chapters and link resources
@@ -188,14 +263,15 @@ declare class DocumentProcessor {
      *
      * @param doclingDoc - Original document extracted from Docling SDK
      * @param reportId - Report unique identifier
-     * @param outputPath - Path containing images and pages subdirectories (images/image_0.png, pages/page_0.png, etc.)
+     * @param artifactDir - Directory containing parser artifacts such as images/, pages/, and result.json
+     * @param processOptions - Per-document processing inputs such as manually verified page range mappings and TOC entries
      * @returns Document processing result with ProcessedDocument and token usage report
      *
      * @throws {TocExtractError} When TOC extraction fails
      * @throws {PageRangeParseError} When page range parsing fails
      * @throws {ConversionError} When error occurs during conversion
      */
-    process(doclingDoc: DoclingDocument, reportId: string, outputPath: string): Promise<DocumentProcessResult>;
+    process(doclingDoc: DoclingDocument, reportId: string, artifactDir: string, processOptions?: DocumentProcessorProcessOptions): Promise<DocumentProcessResult>;
     /**
      * Initialize all processors and resolvers
      *
@@ -361,8 +437,8 @@ interface ImageContent {
  * Subclasses: PageRangeParser, VisionTocExtractor
  */
 declare abstract class VisionLLMComponent extends BaseLLMComponent {
-    protected readonly outputPath: string;
-    constructor(logger: LoggerMethods, model: LanguageModel, componentName: string, outputPath: string, options?: VisionLLMComponentOptions, fallbackModel?: LanguageModel, aggregator?: LLMTokenUsageAggregator);
+    protected readonly artifactDir: string;
+    constructor(logger: LoggerMethods, model: LanguageModel, componentName: string, artifactDir: string, options?: VisionLLMComponentOptions, fallbackModel?: LanguageModel, aggregator?: LLMTokenUsageAggregator);
     /**
      * Call LLM with vision capabilities using LLMCaller.callVision()
      *
@@ -382,67 +458,13 @@ declare abstract class VisionLLMComponent extends BaseLLMComponent {
     /**
      * Build image content object for vision LLM messages
      *
-     * @param imagePath - Path to the image file (relative to outputPath or absolute)
+     * @param imagePath - Path to the image file (relative to artifactDir or absolute)
      * @param mimeType - MIME type of the image (default: 'image/png')
      * @returns ImageContent object for LLM message
      */
     protected buildImageContent(imagePath: string, mimeType?: string): ImageContent;
 }
 
-/**
- * Table of Contents Entry
- *
- * Tree node representing the table of contents structure of a document.
- */
-interface TocEntry {
-    /**
-     * Chapter title
-     */
-    title: string;
-    /**
-     * Hierarchy depth (1, 2, 3...)
-     */
-    level: number;
-    /**
-     * Starting page number
-     */
-    pageNo: number;
-    /**
-     * Child TOC entries
-     */
-    children?: TocEntry[];
-}
-/**
- * TOC Area Search Result
- */
-interface TocAreaResult {
-    /**
-     * Group or table item references corresponding to the table of contents
-     */
-    itemRefs: string[];
-    /**
-     * TOC start page
-     */
-    startPage: number;
-    /**
-     * TOC end page
-     */
-    endPage: number;
-}
-/**
- * Page Size Information
- */
-interface PageSizeGroup {
-    /**
-     * Size identifier (width x height)
-     */
-    sizeKey: string;
-    /**
-     * PDF page numbers with this size specification
-     */
-    pageNos: number[];
-}
-
 /**
  * CaptionParser options
  */
@@ -571,7 +593,7 @@ declare class PageRangeParser extends VisionLLMComponent {
     private readonly SAMPLE_SIZE;
     private readonly MAX_PATTERN_RETRIES;
     private readonly SIZE_TOLERANCE;
-    constructor(logger: LoggerMethods, model: LanguageModel, outputPath: string, maxRetries?: number, fallbackModel?: LanguageModel, aggregator?: LLMTokenUsageAggregator, abortSignal?: AbortSignal);
+    constructor(logger: LoggerMethods, model: LanguageModel, artifactDir: string, maxRetries?: number, fallbackModel?: LanguageModel, aggregator?: LLMTokenUsageAggregator, abortSignal?: AbortSignal);
     /**
      * Main parse method
      *
@@ -1012,6 +1034,10 @@ interface TocExtractorOptions extends BaseLLMComponentOptions {
      * Whether to skip validation entirely (default: false)
      */
     skipValidation?: boolean;
+    /**
+     * Maximum retry count for TOC validation correction feedback (default: 3)
+     */
+    maxValidationRetries?: number;
 }
 /**
  * TocExtractor
@@ -1020,16 +1046,17 @@ interface TocExtractorOptions extends BaseLLMComponentOptions {
  * Extends TextLLMComponent for standardized LLM call handling.
  *
  * When validation fails, automatically retries with correction feedback
- * up to MAX_VALIDATION_RETRIES times before throwing.
+ * up to maxValidationRetries times before throwing.
  */
 declare class TocExtractor extends TextLLMComponent {
     private readonly validationOptions?;
     private readonly skipValidation;
+    private readonly maxValidationRetries;
     constructor(logger: LoggerMethods, model: LanguageModel, options?: TocExtractorOptions, fallbackModel?: LanguageModel, abortSignal?: AbortSignal);
     /**
      * Extract TOC structure from Markdown
      *
-     * When validation fails, retries with correction feedback up to MAX_VALIDATION_RETRIES times.
+     * When validation fails, retries with correction feedback up to maxValidationRetries times.
      *
      * @param markdown - Markdown representation of TOC area
      * @param validationOverrides - Optional overrides for validation options (merged with constructor options)
@@ -1111,7 +1138,7 @@ interface VisionTocExtractorOptions extends VisionLLMComponentOptions {
 declare class VisionTocExtractor extends VisionLLMComponent {
     private readonly firstBatchSize;
     private readonly secondBatchSize;
-    constructor(logger: LoggerMethods, model: LanguageModel, outputPath: string, options?: VisionTocExtractorOptions, fallbackModel?: LanguageModel, aggregator?: LLMTokenUsageAggregator);
+    constructor(logger: LoggerMethods, model: LanguageModel, artifactDir: string, options?: VisionTocExtractorOptions, fallbackModel?: LanguageModel, aggregator?: LLMTokenUsageAggregator);
     /**
      * Extract TOC from page images
      *
@@ -1576,7 +1603,7 @@ declare class ResourceConverter {
      *
      * Runs image and table conversions in parallel, then footnotes synchronously.
      */
-    convertAll(doclingDoc: DoclingDocument, outputPath: string): Promise<{
+    convertAll(doclingDoc: DoclingDocument, artifactDir: string): Promise<{
         images: ProcessedImage[];
         tables: ProcessedTable[];
         footnotes: ProcessedFootnote[];
@@ -1584,7 +1611,7 @@ declare class ResourceConverter {
     /**
      * Convert images from DoclingDocument to ProcessedImage[]
      */
-    convertImages(doclingDoc: DoclingDocument, outputPath: string): Promise<ProcessedImage[]>;
+    convertImages(doclingDoc: DoclingDocument, artifactDir: string): Promise<ProcessedImage[]>;
     /**
      * Convert tables from DoclingDocument to ProcessedTable[]
      */
@@ -1595,4 +1622,4 @@ declare class ResourceConverter {
     convertFootnotes(doclingDoc: DoclingDocument): ProcessedFootnote[];
 }
 
-export { BaseLLMComponent, type BaseLLMComponentOptions, BaseValidator, type BaseValidatorOptions, CONTINUATION_MARKERS, CaptionParseError, CaptionParser, type CaptionParserOptions, CaptionProcessingPipeline, type CaptionProcessingPipelineDeps, CaptionValidationError, CaptionValidator, type CaptionValidatorOptions, ChapterConverter, DocumentProcessor, type DocumentProcessorOptions, type ImageContent, PAGE_NUMBER_PATTERN, PagePattern, PageRangeParseError, PageRangeParser, type PageSizeGroup, ResourceConverter, TOC_KEYWORDS, TextLLMComponent, type TocAreaResult, type TocContentValidationResult, TocContentValidationSchema, TocContentValidator, type TocContentValidatorOptions, type TocEntry, TocEntrySchema, TocExtractError, TocExtractionPipeline, type TocExtractionPipelineDeps, TocExtractor, type TocExtractorOptions, TocFinder, type TocFinderOptions, TocNotFoundError, TocParseError, type TocResponse, TocResponseSchema, VisionLLMComponent, type VisionLLMComponentOptions, type VisionTocExtractionResult, VisionTocExtractionSchema, VisionTocExtractor, type VisionTocExtractorOptions, extractMaxPageNumber };
+export { BaseLLMComponent, type BaseLLMComponentOptions, BaseValidator, type BaseValidatorOptions, CONTINUATION_MARKERS, CaptionParseError, CaptionParser, type CaptionParserOptions, CaptionProcessingPipeline, type CaptionProcessingPipelineDeps, CaptionValidationError, CaptionValidator, type CaptionValidatorOptions, ChapterConverter, DocumentProcessor, type DocumentProcessorOptions, type DocumentProcessorProcessOptions, type ImageContent, PAGE_NUMBER_PATTERN, PagePattern, PageRangeParseError, PageRangeParser, type PageSizeGroup, ResourceConverter, TOC_KEYWORDS, TextLLMComponent, type TocAreaResult, type TocContentValidationResult, TocContentValidationSchema, TocContentValidator, type TocContentValidatorOptions, type TocEntry, TocEntrySchema, TocExtractError, TocExtractionPipeline, type TocExtractionPipelineDeps, TocExtractor, type TocExtractorOptions, TocFinder, type TocFinderOptions, TocNotFoundError, TocParseError, type TocResponse, TocResponseSchema, VisionLLMComponent, type VisionLLMComponentOptions, type VisionTocExtractionResult, VisionTocExtractionSchema, VisionTocExtractor, type VisionTocExtractorOptions, extractMaxPageNumber };

package/dist/index.d.ts

@@ -1,9 +1,63 @@
 import { LoggerMethods } from '@heripo/logger';
-import { TokenUsageReport, DoclingDocument, DocumentProcessResult, Caption, PageRange, DoclingTextItem, DoclingPictureItem, DoclingTableItem, DoclingGroupItem, ProcessedImage, ProcessedTable, ProcessedFootnote, Chapter } from '@heripo/model';
+import { TokenUsageReport, DoclingDocument, PageRange, DocumentProcessResult, Caption, DoclingTextItem, DoclingPictureItem, DoclingTableItem, DoclingGroupItem, ProcessedImage, ProcessedTable, ProcessedFootnote, Chapter } from '@heripo/model';
 import { LanguageModel } from 'ai';
 import { LLMTokenUsageAggregator, ExtendedTokenUsage } from '@heripo/shared';
 import { z } from 'zod';
 
+/**
+ * Table of Contents Entry
+ *
+ * Tree node representing the table of contents structure of a document.
+ */
+interface TocEntry {
+    /**
+     * Chapter title
+     */
+    title: string;
+    /**
+     * Hierarchy depth (1, 2, 3...)
+     */
+    level: number;
+    /**
+     * Starting page number
+     */
+    pageNo: number;
+    /**
+     * Child TOC entries
+     */
+    children?: TocEntry[];
+}
+/**
+ * TOC Area Search Result
+ */
+interface TocAreaResult {
+    /**
+     * Group or table item references corresponding to the table of contents
+     */
+    itemRefs: string[];
+    /**
+     * TOC start page
+     */
+    startPage: number;
+    /**
+     * TOC end page
+     */
+    endPage: number;
+}
+/**
+ * Page Size Information
+ */
+interface PageSizeGroup {
+    /**
+     * Size identifier (width x height)
+     */
+    sizeKey: string;
+    /**
+     * PDF page numbers with this size specification
+     */
+    pageNos: number[];
+}
+
 /**
  * DocumentProcessor Options
  */
@@ -59,6 +113,10 @@ interface DocumentProcessorOptions {
      * Maximum retry count (default: 3)
      */
     maxRetries?: number;
+    /**
+     * Maximum retry count for TOC validation correction feedback (default: 3)
+     */
+    maxValidationRetries?: number;
     /**
      * Enable fallback retry mechanism - automatically retries with fallback model on failure (default: false)
      * Set to true to enable automatic fallback retry with fallback model on component-specific model errors
@@ -76,6 +134,21 @@ interface DocumentProcessorOptions {
      */
     onTokenUsage?: (report: TokenUsageReport) => void;
 }
+/**
+ * Per-document processing inputs.
+ */
+interface DocumentProcessorProcessOptions {
+    /**
+     * Precomputed PDF page number to actual document page range mapping.
+     * When provided, automatic PageRangeParser execution is skipped.
+     */
+    pageRangeMap?: Record<number, PageRange>;
+    /**
+     * Precomputed table of contents entries.
+     * When provided, automatic TOC extraction pipeline execution is skipped.
+     */
+    tocEntries?: TocEntry[];
+}
 /**
  * DocumentProcessor
  *
@@ -121,12 +194,13 @@ interface DocumentProcessorOptions {
  *   captionParserBatchSize: 10,    // LLM caption parsing
  *   captionValidatorBatchSize: 10, // LLM caption validation
  *   maxRetries: 3,
+ *   maxValidationRetries: 3,
  * });
  *
  * const result = await processor.process(
  *   doclingDoc,
  *   'report-001',
- *   outputPath
+ *   artifactDir
  * );
  * ```
  */
@@ -142,6 +216,7 @@ declare class DocumentProcessor {
     private readonly captionParserBatchSize;
     private readonly captionValidatorBatchSize;
     private readonly maxRetries;
+    private readonly maxValidationRetries;
     private readonly enableFallbackRetry;
     private readonly abortSignal?;
     private readonly onTokenUsage?;
@@ -179,7 +254,7 @@ declare class DocumentProcessor {
      * Conversion process:
      * 1. Initialize processors and resolvers
      * 2. Normalize and filter texts
-     * 3. Clean texts and parse page ranges (parallel)
+     * 3. Clean texts and resolve page ranges
      * 4. Extract table of contents
      * 5. Convert images and tables (parallel)
      * 6. Convert chapters and link resources
@@ -188,14 +263,15 @@ declare class DocumentProcessor {
      *
      * @param doclingDoc - Original document extracted from Docling SDK
      * @param reportId - Report unique identifier
-     * @param outputPath - Path containing images and pages subdirectories (images/image_0.png, pages/page_0.png, etc.)
+     * @param artifactDir - Directory containing parser artifacts such as images/, pages/, and result.json
+     * @param processOptions - Per-document processing inputs such as manually verified page range mappings and TOC entries
      * @returns Document processing result with ProcessedDocument and token usage report
      *
      * @throws {TocExtractError} When TOC extraction fails
      * @throws {PageRangeParseError} When page range parsing fails
      * @throws {ConversionError} When error occurs during conversion
      */
-    process(doclingDoc: DoclingDocument, reportId: string, outputPath: string): Promise<DocumentProcessResult>;
+    process(doclingDoc: DoclingDocument, reportId: string, artifactDir: string, processOptions?: DocumentProcessorProcessOptions): Promise<DocumentProcessResult>;
     /**
      * Initialize all processors and resolvers
      *
@@ -361,8 +437,8 @@ interface ImageContent {
  * Subclasses: PageRangeParser, VisionTocExtractor
  */
 declare abstract class VisionLLMComponent extends BaseLLMComponent {
-    protected readonly outputPath: string;
-    constructor(logger: LoggerMethods, model: LanguageModel, componentName: string, outputPath: string, options?: VisionLLMComponentOptions, fallbackModel?: LanguageModel, aggregator?: LLMTokenUsageAggregator);
+    protected readonly artifactDir: string;
+    constructor(logger: LoggerMethods, model: LanguageModel, componentName: string, artifactDir: string, options?: VisionLLMComponentOptions, fallbackModel?: LanguageModel, aggregator?: LLMTokenUsageAggregator);
     /**
      * Call LLM with vision capabilities using LLMCaller.callVision()
      *
@@ -382,67 +458,13 @@ declare abstract class VisionLLMComponent extends BaseLLMComponent {
     /**
      * Build image content object for vision LLM messages
      *
-     * @param imagePath - Path to the image file (relative to outputPath or absolute)
+     * @param imagePath - Path to the image file (relative to artifactDir or absolute)
      * @param mimeType - MIME type of the image (default: 'image/png')
      * @returns ImageContent object for LLM message
      */
     protected buildImageContent(imagePath: string, mimeType?: string): ImageContent;
 }
 
-/**
- * Table of Contents Entry
- *
- * Tree node representing the table of contents structure of a document.
- */
-interface TocEntry {
-    /**
-     * Chapter title
-     */
-    title: string;
-    /**
-     * Hierarchy depth (1, 2, 3...)
-     */
-    level: number;
-    /**
-     * Starting page number
-     */
-    pageNo: number;
-    /**
-     * Child TOC entries
-     */
-    children?: TocEntry[];
-}
-/**
- * TOC Area Search Result
- */
-interface TocAreaResult {
-    /**
-     * Group or table item references corresponding to the table of contents
-     */
-    itemRefs: string[];
-    /**
-     * TOC start page
-     */
-    startPage: number;
-    /**
-     * TOC end page
-     */
-    endPage: number;
-}
-/**
- * Page Size Information
- */
-interface PageSizeGroup {
-    /**
-     * Size identifier (width x height)
-     */
-    sizeKey: string;
-    /**
-     * PDF page numbers with this size specification
-     */
-    pageNos: number[];
-}
-
 /**
  * CaptionParser options
  */
@@ -571,7 +593,7 @@ declare class PageRangeParser extends VisionLLMComponent {
     private readonly SAMPLE_SIZE;
     private readonly MAX_PATTERN_RETRIES;
     private readonly SIZE_TOLERANCE;
-    constructor(logger: LoggerMethods, model: LanguageModel, outputPath: string, maxRetries?: number, fallbackModel?: LanguageModel, aggregator?: LLMTokenUsageAggregator, abortSignal?: AbortSignal);
+    constructor(logger: LoggerMethods, model: LanguageModel, artifactDir: string, maxRetries?: number, fallbackModel?: LanguageModel, aggregator?: LLMTokenUsageAggregator, abortSignal?: AbortSignal);
     /**
      * Main parse method
      *
@@ -1012,6 +1034,10 @@ interface TocExtractorOptions extends BaseLLMComponentOptions {
      * Whether to skip validation entirely (default: false)
      */
     skipValidation?: boolean;
+    /**
+     * Maximum retry count for TOC validation correction feedback (default: 3)
+     */
+    maxValidationRetries?: number;
 }
 /**
  * TocExtractor
@@ -1020,16 +1046,17 @@ interface TocExtractorOptions extends BaseLLMComponentOptions {
  * Extends TextLLMComponent for standardized LLM call handling.
  *
  * When validation fails, automatically retries with correction feedback
- * up to MAX_VALIDATION_RETRIES times before throwing.
+ * up to maxValidationRetries times before throwing.
  */
 declare class TocExtractor extends TextLLMComponent {
     private readonly validationOptions?;
     private readonly skipValidation;
+    private readonly maxValidationRetries;
     constructor(logger: LoggerMethods, model: LanguageModel, options?: TocExtractorOptions, fallbackModel?: LanguageModel, abortSignal?: AbortSignal);
     /**
      * Extract TOC structure from Markdown
      *
-     * When validation fails, retries with correction feedback up to MAX_VALIDATION_RETRIES times.
+     * When validation fails, retries with correction feedback up to maxValidationRetries times.
      *
      * @param markdown - Markdown representation of TOC area
      * @param validationOverrides - Optional overrides for validation options (merged with constructor options)
@@ -1111,7 +1138,7 @@ interface VisionTocExtractorOptions extends VisionLLMComponentOptions {
 declare class VisionTocExtractor extends VisionLLMComponent {
     private readonly firstBatchSize;
     private readonly secondBatchSize;
-    constructor(logger: LoggerMethods, model: LanguageModel, outputPath: string, options?: VisionTocExtractorOptions, fallbackModel?: LanguageModel, aggregator?: LLMTokenUsageAggregator);
+    constructor(logger: LoggerMethods, model: LanguageModel, artifactDir: string, options?: VisionTocExtractorOptions, fallbackModel?: LanguageModel, aggregator?: LLMTokenUsageAggregator);
     /**
      * Extract TOC from page images
      *
@@ -1576,7 +1603,7 @@ declare class ResourceConverter {
      *
      * Runs image and table conversions in parallel, then footnotes synchronously.
      */
-    convertAll(doclingDoc: DoclingDocument, outputPath: string): Promise<{
+    convertAll(doclingDoc: DoclingDocument, artifactDir: string): Promise<{
         images: ProcessedImage[];
         tables: ProcessedTable[];
         footnotes: ProcessedFootnote[];
@@ -1584,7 +1611,7 @@ declare class ResourceConverter {
     /**
      * Convert images from DoclingDocument to ProcessedImage[]
      */
-    convertImages(doclingDoc: DoclingDocument, outputPath: string): Promise<ProcessedImage[]>;
+    convertImages(doclingDoc: DoclingDocument, artifactDir: string): Promise<ProcessedImage[]>;
     /**
      * Convert tables from DoclingDocument to ProcessedTable[]
      */
@@ -1595,4 +1622,4 @@ declare class ResourceConverter {
     convertFootnotes(doclingDoc: DoclingDocument): ProcessedFootnote[];
 }
 
-export { BaseLLMComponent, type BaseLLMComponentOptions, BaseValidator, type BaseValidatorOptions, CONTINUATION_MARKERS, CaptionParseError, CaptionParser, type CaptionParserOptions, CaptionProcessingPipeline, type CaptionProcessingPipelineDeps, CaptionValidationError, CaptionValidator, type CaptionValidatorOptions, ChapterConverter, DocumentProcessor, type DocumentProcessorOptions, type ImageContent, PAGE_NUMBER_PATTERN, PagePattern, PageRangeParseError, PageRangeParser, type PageSizeGroup, ResourceConverter, TOC_KEYWORDS, TextLLMComponent, type TocAreaResult, type TocContentValidationResult, TocContentValidationSchema, TocContentValidator, type TocContentValidatorOptions, type TocEntry, TocEntrySchema, TocExtractError, TocExtractionPipeline, type TocExtractionPipelineDeps, TocExtractor, type TocExtractorOptions, TocFinder, type TocFinderOptions, TocNotFoundError, TocParseError, type TocResponse, TocResponseSchema, VisionLLMComponent, type VisionLLMComponentOptions, type VisionTocExtractionResult, VisionTocExtractionSchema, VisionTocExtractor, type VisionTocExtractorOptions, extractMaxPageNumber };
+export { BaseLLMComponent, type BaseLLMComponentOptions, BaseValidator, type BaseValidatorOptions, CONTINUATION_MARKERS, CaptionParseError, CaptionParser, type CaptionParserOptions, CaptionProcessingPipeline, type CaptionProcessingPipelineDeps, CaptionValidationError, CaptionValidator, type CaptionValidatorOptions, ChapterConverter, DocumentProcessor, type DocumentProcessorOptions, type DocumentProcessorProcessOptions, type ImageContent, PAGE_NUMBER_PATTERN, PagePattern, PageRangeParseError, PageRangeParser, type PageSizeGroup, ResourceConverter, TOC_KEYWORDS, TextLLMComponent, type TocAreaResult, type TocContentValidationResult, TocContentValidationSchema, TocContentValidator, type TocContentValidatorOptions, type TocEntry, TocEntrySchema, TocExtractError, TocExtractionPipeline, type TocExtractionPipelineDeps, TocExtractor, type TocExtractorOptions, TocFinder, type TocFinderOptions, TocNotFoundError, TocParseError, type TocResponse, TocResponseSchema, VisionLLMComponent, type VisionLLMComponentOptions, type VisionTocExtractionResult, VisionTocExtractionSchema, VisionTocExtractor, type VisionTocExtractorOptions, extractMaxPageNumber };