@llamaindex/liteparse 2.8.1 → 2.10.0

package/README.md

@@ -58,14 +58,57 @@ const parser = new LiteParse({
   dpi: 150,                      // Rendering DPI
   outputFormat: 'json',          // "json" | "text" | "markdown"
   imageMode: 'placeholder',      // Markdown image handling: "placeholder" | "off" | "embed"
+  extractImages: true,           // Extract image bytes + metadata (default: false)
+  imageOutputDir: './images',    // Write images and return name/path metadata (optional)
   extractLinks: true,            // Render [text](url) links in markdown output
+  keepHeadersFooters: false,     // Keep running headers/footers in markdown output
+  extractVectorGraphics: false,  // Opt-in shapes + merged H/V lines per page
+  extractAnnotations: false,     // Include page annotations in structured output
+  extractFormFields: false,      // Include AcroForm widget fields and values
+  extractStructureTree: false,   // Include tagged-PDF logical structure
+  extractXfaPackets: false,      // Include raw XFA packets (name + XML content)
+  extractContentBounds: false,   // Include per-page contentBounds union bbox
+  detectScreenshotRects: false,  // Detect solid rects/lines in screenshots
   preserveVerySmallText: false,  // Keep tiny text
+  extractTextMetadata: false,    // Opt in to MCID, font metrics, colors, char codes, and trailingSpaceGenerated
   password: undefined,           // Password for protected documents
   quiet: false,                  // Suppress progress output
   numWorkers: 4,                 // Concurrent OCR workers
 });
 ```
 
+When `extractImages` is true, image extraction is enabled. `imageOutputDir` requires
+that explicit opt-in and writes the extracted bytes to disk. Each
+`result.images` entry includes its page bbox, intrinsic pixel dimensions, rotation,
+format, `name`, and `path`. Valid source JPEGs are preserved, exact duplicates reuse
+one file, and JSON CLI output contains metadata only (no base64 image data).
+`imageMode` controls Markdown presentation only and does not imply extraction. With
+`extractImages: false`, lightweight Markdown placement refs are still collected and
+`result.images` stays empty.
+
+The Node CLI uses the same snake_case JSON schema as the Rust CLI; camelCase is
+reserved for the programmatic Node API.
+
+When `extractAnnotations` is enabled, each parsed page has an `annotations`
+array containing the subtype, contents, author/title, PDF date strings,
+viewport-space rectangle and quadpoint rectangles, and URI for external link
+annotations. It is independent of `extractLinks`, which controls Markdown link
+rendering. The field is omitted when extraction is disabled.
+
+When `extractStructureTree` is enabled, each page has a `structureTree` containing
+all tagged-PDF roots and recursive elements with type, ID, actual/alternate text,
+title, typed attributes, MCIDs, children, and referenced link annotations. Untagged
+pages have an empty `roots` array; the field is omitted when disabled.
+
+Every result also carries the document's `/Info` `creator`/`producer` when
+present (API-level only, not in CLI JSON), and with `extractContentBounds`
+each page carries a `contentBounds` union bbox of its top-level content
+objects. With `extractXfaPackets`, `result.xfaPackets` lists each raw XFA
+packet (index, name, content length, XML content); non-XFA documents yield an
+empty array. Screenshots draw form-field appearances into the raster and
+report `isSolidFill` per page, plus detected solid `rects` when
+`detectScreenshotRects` is enabled.
+
 ## Parsing from Bytes
 
 Pass a `Buffer` or `Uint8Array` directly — useful for HTTP responses or in-memory data:
@@ -133,6 +176,8 @@ The npm package includes the `lit` CLI:
 ```bash
 lit parse document.pdf
 lit parse document.pdf --format json -o output.json
+lit parse document.pdf --format json --extract-annotations
+lit parse document.pdf --format json --extract-form-fields
 lit screenshot document.pdf -o ./screenshots
 lit batch-parse ./input ./output
 lit is-complex document.pdf

package/dist/cli-json.d.ts

@@ -0,0 +1,154 @@
+import type { ParseResult } from "./lib.js";
+/** Project the camelCase Node API result into the Rust CLI JSON schema. */
+export declare function parseResultToCliJson(result: ParseResult, options?: {
+    extractTextMetadata?: boolean;
+}): {
+    xfa_packets?: {
+        content?: string | undefined;
+        content_length: number;
+        name?: string | undefined;
+        index: number;
+    }[] | undefined;
+    form_type?: number | undefined;
+    image_error_count?: number | undefined;
+    images?: {
+        duplicate_of?: string | undefined;
+        page: number;
+        bbox: {
+            x: number;
+            y: number;
+            width: number;
+            height: number;
+        };
+        width: number;
+        height: number;
+        rotation: number;
+        format: string;
+        path?: string | undefined;
+        id: string;
+        name: string;
+    }[] | undefined;
+    pages: {
+        structure_tree?: {
+            roots: object[];
+        } | undefined;
+        form_fields?: {
+            selected_options?: string[] | undefined;
+            options?: string[] | undefined;
+            rect?: import("./lib.js").AnnotationRect | undefined;
+            checked?: boolean | undefined;
+            control_index?: number | undefined;
+            control_count?: number | undefined;
+            field_flags: number;
+            export_value?: string | undefined;
+            value?: string | undefined;
+            alternate_name?: string | undefined;
+            name?: string | undefined;
+            object_number?: number | undefined;
+            id: string;
+            type: string;
+            page: number;
+            annotation_index: number;
+            widget_index: number;
+        }[] | undefined;
+        annotations?: {
+            uri?: string | undefined;
+            quadpoint_rects?: {
+                x: number;
+                y: number;
+                width: number;
+                height: number;
+            }[] | undefined;
+            rect?: import("./lib.js").AnnotationRect | undefined;
+            title?: string | undefined;
+            modified?: string | undefined;
+            created?: string | undefined;
+            contents?: string | undefined;
+            subtype: string;
+        }[] | undefined;
+        vector_graphics?: {
+            shapes: {
+                has_curve: boolean;
+                fill_color?: string | undefined;
+                fill: boolean;
+                stroke_color?: string | undefined;
+                bbox: {
+                    x: number;
+                    y: number;
+                    width: number;
+                    height: number;
+                };
+                stroke: boolean;
+            }[];
+            lines: {
+                fill_color?: string | undefined;
+                fill: boolean;
+                stroke_color?: string | undefined;
+                stroke_width?: number | undefined;
+                x1: number;
+                y1: number;
+                x2: number;
+                y2: number;
+                stroke: boolean;
+            }[];
+        } | undefined;
+        complexity?: {
+            layout?: {
+                column_count: number;
+                ruled_table_count: number;
+                ruled_table_coverage: number;
+                text_table_run_count: number;
+                figure_count: number;
+                figure_coverage: number;
+                is_complex: boolean;
+                reasons: string[];
+            } | undefined;
+            is_garbled: boolean;
+            page_area: number;
+            needs_ocr: boolean;
+            reasons: string[];
+            uncovered_vector_area?: number | undefined;
+            page_number: number;
+            text_length: number;
+            text_coverage: number;
+            has_substantial_images: boolean;
+            image_block_count: number;
+            image_coverage: number;
+            largest_image_coverage: number;
+            full_page_image: boolean;
+        } | undefined;
+        text: string;
+        text_items: {
+            confidence?: number | undefined;
+            trailing_space_generated?: boolean | undefined;
+            char_codes?: number[] | undefined;
+            stroke_color?: string | undefined;
+            fill_color?: string | undefined;
+            mcid?: number | undefined;
+            font_is_buggy?: boolean | undefined;
+            text_width?: number | undefined;
+            font_weight?: number | undefined;
+            font_descent?: number | undefined;
+            font_ascent?: number | undefined;
+            font_height?: number | undefined;
+            font_size?: number | undefined;
+            font_name?: string | undefined;
+            rotation?: number | undefined;
+            text: string;
+            x: number;
+            y: number;
+            width: number;
+            height: number;
+        }[];
+        content_bounds?: {
+            x: number;
+            y: number;
+            width: number;
+            height: number;
+        } | undefined;
+        page: number;
+        width: number;
+        height: number;
+    }[];
+};
+//# sourceMappingURL=cli-json.d.ts.map

package/dist/cli-json.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"cli-json.d.ts","sourceRoot":"","sources":["../src/cli-json.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAKV,WAAW,EAKZ,MAAM,UAAU,CAAC;AAmMlB,2EAA2E;AAC3E,wBAAgB,oBAAoB,CAClC,MAAM,EAAE,WAAW,EACnB,OAAO,GAAE;IAAE,mBAAmB,CAAC,EAAE,OAAO,CAAA;CAAO;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;EAqDhD"}

package/dist/cli-json.js

@@ -0,0 +1,238 @@
+function textItemToCliJson(item, extractTextMetadata) {
+    return {
+        text: item.text,
+        x: item.x,
+        y: item.y,
+        width: item.width,
+        height: item.height,
+        ...(extractTextMetadata && item.rotation !== undefined
+            ? { rotation: item.rotation }
+            : {}),
+        ...(item.fontName !== undefined ? { font_name: item.fontName } : {}),
+        ...(item.fontSize !== undefined ? { font_size: item.fontSize } : {}),
+        ...(item.fontHeight !== undefined ? { font_height: item.fontHeight } : {}),
+        ...(item.fontAscent !== undefined ? { font_ascent: item.fontAscent } : {}),
+        ...(item.fontDescent !== undefined ? { font_descent: item.fontDescent } : {}),
+        ...(item.fontWeight !== undefined ? { font_weight: item.fontWeight } : {}),
+        ...(item.textWidth !== undefined ? { text_width: item.textWidth } : {}),
+        ...(item.fontIsBuggy !== undefined ? { font_is_buggy: item.fontIsBuggy } : {}),
+        ...(item.mcid !== undefined ? { mcid: item.mcid } : {}),
+        ...(item.fillColor !== undefined ? { fill_color: item.fillColor } : {}),
+        ...(item.strokeColor !== undefined ? { stroke_color: item.strokeColor } : {}),
+        ...(item.charCodes?.length ? { char_codes: item.charCodes } : {}),
+        ...(item.trailingSpaceGenerated
+            ? { trailing_space_generated: true }
+            : {}),
+        ...(item.confidence !== undefined ? { confidence: item.confidence } : {}),
+    };
+}
+function structureTreeElementToCliJson(element) {
+    return {
+        type: element.type,
+        ...(element.id !== undefined ? { id: element.id } : {}),
+        ...(element.actualText !== undefined ? { actual_text: element.actualText } : {}),
+        ...(element.altText !== undefined ? { alt_text: element.altText } : {}),
+        ...(element.title !== undefined ? { title: element.title } : {}),
+        ...(Object.keys(element.attributes).length ? { attributes: element.attributes } : {}),
+        marked_content_ids: element.markedContentIds,
+        children: element.children.map(structureTreeElementToCliJson),
+        annotations: element.annotations.map(annotationToCliJson),
+    };
+}
+function complexityToCliJson(complexity) {
+    return {
+        page_number: complexity.pageNumber,
+        text_length: complexity.textLength,
+        text_coverage: complexity.textCoverage,
+        has_substantial_images: complexity.hasSubstantialImages,
+        image_block_count: complexity.imageBlockCount,
+        image_coverage: complexity.imageCoverage,
+        largest_image_coverage: complexity.largestImageCoverage,
+        full_page_image: complexity.fullPageImage,
+        ...(complexity.uncoveredVectorArea !== undefined
+            ? { uncovered_vector_area: complexity.uncoveredVectorArea }
+            : {}),
+        is_garbled: complexity.isGarbled,
+        page_area: complexity.pageArea,
+        needs_ocr: complexity.needsOcr,
+        reasons: complexity.reasons,
+        ...(complexity.layout
+            ? {
+                layout: {
+                    column_count: complexity.layout.columnCount,
+                    ruled_table_count: complexity.layout.ruledTableCount,
+                    ruled_table_coverage: complexity.layout.ruledTableCoverage,
+                    text_table_run_count: complexity.layout.textTableRunCount,
+                    figure_count: complexity.layout.figureCount,
+                    figure_coverage: complexity.layout.figureCoverage,
+                    is_complex: complexity.layout.isComplex,
+                    reasons: complexity.layout.reasons,
+                },
+            }
+            : {}),
+    };
+}
+function vectorGraphicsToCliJson(vectorGraphics) {
+    return {
+        shapes: vectorGraphics.shapes.map((shape) => ({
+            bbox: {
+                x: shape.bbox.x,
+                y: shape.bbox.y,
+                width: shape.bbox.width,
+                height: shape.bbox.height,
+            },
+            stroke: shape.stroke,
+            ...(shape.strokeColor !== undefined
+                ? { stroke_color: shape.strokeColor }
+                : {}),
+            fill: shape.fill,
+            ...(shape.fillColor !== undefined
+                ? { fill_color: shape.fillColor }
+                : {}),
+            has_curve: shape.hasCurve,
+        })),
+        lines: vectorGraphics.lines.map((line) => ({
+            x1: line.x1,
+            y1: line.y1,
+            x2: line.x2,
+            y2: line.y2,
+            stroke: line.stroke,
+            ...(line.strokeWidth !== undefined
+                ? { stroke_width: line.strokeWidth }
+                : {}),
+            ...(line.strokeColor !== undefined
+                ? { stroke_color: line.strokeColor }
+                : {}),
+            fill: line.fill,
+            ...(line.fillColor !== undefined ? { fill_color: line.fillColor } : {}),
+        })),
+    };
+}
+function annotationToCliJson(annotation) {
+    return {
+        subtype: annotation.subtype,
+        ...(annotation.contents !== undefined
+            ? { contents: annotation.contents }
+            : {}),
+        ...(annotation.created !== undefined ? { created: annotation.created } : {}),
+        ...(annotation.modified !== undefined
+            ? { modified: annotation.modified }
+            : {}),
+        ...(annotation.title !== undefined ? { title: annotation.title } : {}),
+        ...(annotation.rect !== undefined ? { rect: annotation.rect } : {}),
+        ...(annotation.quadpointRects.length
+            ? {
+                quadpoint_rects: annotation.quadpointRects.map((rect) => ({
+                    x: rect.x,
+                    y: rect.y,
+                    width: rect.width,
+                    height: rect.height,
+                })),
+            }
+            : {}),
+        ...(annotation.uri !== undefined ? { uri: annotation.uri } : {}),
+    };
+}
+function formFieldToCliJson(field) {
+    return {
+        id: field.id,
+        type: field.type,
+        page: field.page,
+        annotation_index: field.annotationIndex,
+        widget_index: field.widgetIndex,
+        ...(field.objectNumber !== undefined ? { object_number: field.objectNumber } : {}),
+        ...(field.name !== undefined ? { name: field.name } : {}),
+        ...(field.alternateName !== undefined ? { alternate_name: field.alternateName } : {}),
+        ...(field.value !== undefined ? { value: field.value } : {}),
+        ...(field.exportValue !== undefined ? { export_value: field.exportValue } : {}),
+        field_flags: field.fieldFlags,
+        ...(field.controlCount !== undefined ? { control_count: field.controlCount } : {}),
+        ...(field.controlIndex !== undefined ? { control_index: field.controlIndex } : {}),
+        ...(field.checked !== undefined ? { checked: field.checked } : {}),
+        ...(field.rect !== undefined ? { rect: field.rect } : {}),
+        ...(field.options.length ? { options: field.options } : {}),
+        ...(field.selectedOptions.length ? { selected_options: field.selectedOptions } : {}),
+    };
+}
+function xfaPacketToCliJson(packet) {
+    return {
+        index: packet.index,
+        ...(packet.name !== undefined ? { name: packet.name } : {}),
+        content_length: packet.contentLength,
+        ...(packet.content !== undefined ? { content: packet.content } : {}),
+    };
+}
+function imageToCliJson(image) {
+    return {
+        id: image.id,
+        name: image.name,
+        ...(image.path !== undefined ? { path: image.path } : {}),
+        page: image.page,
+        bbox: {
+            x: image.bbox.x,
+            y: image.bbox.y,
+            width: image.bbox.width,
+            height: image.bbox.height,
+        },
+        width: image.width,
+        height: image.height,
+        rotation: image.rotation,
+        format: image.format,
+        ...(image.duplicateOf !== undefined
+            ? { duplicate_of: image.duplicateOf }
+            : {}),
+    };
+}
+/** Project the camelCase Node API result into the Rust CLI JSON schema. */
+export function parseResultToCliJson(result, options = {}) {
+    const extractTextMetadata = options.extractTextMetadata ?? false;
+    return {
+        pages: result.pages.map((page) => ({
+            page: page.pageNum,
+            width: page.width,
+            height: page.height,
+            ...(page.contentBounds
+                ? {
+                    content_bounds: {
+                        x: page.contentBounds.x,
+                        y: page.contentBounds.y,
+                        width: page.contentBounds.width,
+                        height: page.contentBounds.height,
+                    },
+                }
+                : {}),
+            text: page.text,
+            text_items: page.textItems.map((item) => textItemToCliJson(item, extractTextMetadata)),
+            ...(page.complexity
+                ? { complexity: complexityToCliJson(page.complexity) }
+                : {}),
+            ...(page.vectorGraphics
+                ? { vector_graphics: vectorGraphicsToCliJson(page.vectorGraphics) }
+                : {}),
+            ...(page.annotations !== undefined
+                ? { annotations: page.annotations.map(annotationToCliJson) }
+                : {}),
+            ...(page.formFields !== undefined
+                ? { form_fields: page.formFields.map(formFieldToCliJson) }
+                : {}),
+            ...(page.structureTree !== undefined
+                ? {
+                    structure_tree: {
+                        roots: page.structureTree.roots.map(structureTreeElementToCliJson),
+                    },
+                }
+                : {}),
+        })),
+        ...(result.images.length
+            ? { images: result.images.map(imageToCliJson) }
+            : {}),
+        ...(result.imageErrorCount
+            ? { image_error_count: result.imageErrorCount }
+            : {}),
+        ...(result.formType !== undefined ? { form_type: result.formType } : {}),
+        ...(result.xfaPackets !== undefined
+            ? { xfa_packets: result.xfaPackets.map(xfaPacketToCliJson) }
+            : {}),
+    };
+}
+//# sourceMappingURL=cli-json.js.map

package/dist/cli-json.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"cli-json.js","sourceRoot":"","sources":["../src/cli-json.ts"],"names":[],"mappings":"AAYA,SAAS,iBAAiB,CAAC,IAAc,EAAE,mBAA4B;IACrE,OAAO;QACL,IAAI,EAAE,IAAI,CAAC,IAAI;QACf,CAAC,EAAE,IAAI,CAAC,CAAC;QACT,CAAC,EAAE,IAAI,CAAC,CAAC;QACT,KAAK,EAAE,IAAI,CAAC,KAAK;QACjB,MAAM,EAAE,IAAI,CAAC,MAAM;QACnB,GAAG,CAAC,mBAAmB,IAAI,IAAI,CAAC,QAAQ,KAAK,SAAS;YACpD,CAAC,CAAC,EAAE,QAAQ,EAAE,IAAI,CAAC,QAAQ,EAAE;YAC7B,CAAC,CAAC,EAAE,CAAC;QACP,GAAG,CAAC,IAAI,CAAC,QAAQ,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,SAAS,EAAE,IAAI,CAAC,QAAQ,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;QACpE,GAAG,CAAC,IAAI,CAAC,QAAQ,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,SAAS,EAAE,IAAI,CAAC,QAAQ,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;QACpE,GAAG,CAAC,IAAI,CAAC,UAAU,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,WAAW,EAAE,IAAI,CAAC,UAAU,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;QAC1E,GAAG,CAAC,IAAI,CAAC,UAAU,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,WAAW,EAAE,IAAI,CAAC,UAAU,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;QAC1E,GAAG,CAAC,IAAI,CAAC,WAAW,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,YAAY,EAAE,IAAI,CAAC,WAAW,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;QAC7E,GAAG,CAAC,IAAI,CAAC,UAAU,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,WAAW,EAAE,IAAI,CAAC,UAAU,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;QAC1E,GAAG,CAAC,IAAI,CAAC,SAAS,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,UAAU,EAAE,IAAI,CAAC,SAAS,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;QACvE,GAAG,CAAC,IAAI,CAAC,WAAW,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,aAAa,EAAE,IAAI,CAAC,WAAW,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;QAC9E,GAAG,CAAC,IAAI,CAAC,IAAI,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,IAAI,EAAE,IAAI,CAAC,IAAI,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;QACvD,GAAG,CAAC,IAAI,CAAC,SAAS,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,UAAU,EAAE,IAAI,CAAC,SAAS,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;QACvE,GAAG,CAAC,IAAI,CAAC,WAAW,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,YAAY,EAAE,IAAI,CAAC,WAAW,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;QAC7E,GAAG,CAAC,IAAI,CAAC,SAAS,EAAE,MAAM,CAAC,CAAC,CAAC,EAAE,UAAU,EAAE,IAAI,CAAC,SAAS,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;QACjE,GAAG,CAAC,IAAI,CAAC,sBAAsB;YAC7B,CAAC,CAAC,EAAE,wBAAwB,EAAE,IAAI,EAAE;YACpC,CAAC,CAAC,EAAE,CAAC;QACP,GAAG,CAAC,IAAI,CAAC,UAAU,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,UAAU,EAAE,IAAI,CAAC,UAAU,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;KAC1E,CAAC;AACJ,CAAC;AAED,SAAS,6BAA6B,CAAC,OAA6B;IAClE,OAAO;QACL,IAAI,EAAE,OAAO,CAAC,IAAI;QAClB,GAAG,CAAC,OAAO,CAAC,EAAE,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,EAAE,EAAE,OAAO,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;QACvD,GAAG,CAAC,OAAO,CAAC,UAAU,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,WAAW,EAAE,OAAO,CAAC,UAAU,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;QAChF,GAAG,CAAC,OAAO,CAAC,OAAO,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,QAAQ,EAAE,OAAO,CAAC,OAAO,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;QACvE,GAAG,CAAC,OAAO,CAAC,KAAK,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,KAAK,EAAE,OAAO,CAAC,KAAK,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;QAChE,GAAG,CAAC,MAAM,CAAC,IAAI,CAAC,OAAO,CAAC,UAAU,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,UAAU,EAAE,OAAO,CAAC,UAAU,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;QACrF,kBAAkB,EAAE,OAAO,CAAC,gBAAgB;QAC5C,QAAQ,EAAE,OAAO,CAAC,QAAQ,CAAC,GAAG,CAAC,6BAA6B,CAAC;QAC7D,WAAW,EAAE,OAAO,CAAC,WAAW,CAAC,GAAG,CAAC,mBAAmB,CAAC;KAC1D,CAAC;AACJ,CAAC;AAED,SAAS,mBAAmB,CAAC,UAA+B;IAC1D,OAAO;QACL,WAAW,EAAE,UAAU,CAAC,UAAU;QAClC,WAAW,EAAE,UAAU,CAAC,UAAU;QAClC,aAAa,EAAE,UAAU,CAAC,YAAY;QACtC,sBAAsB,EAAE,UAAU,CAAC,oBAAoB;QACvD,iBAAiB,EAAE,UAAU,CAAC,eAAe;QAC7C,cAAc,EAAE,UAAU,CAAC,aAAa;QACxC,sBAAsB,EAAE,UAAU,CAAC,oBAAoB;QACvD,eAAe,EAAE,UAAU,CAAC,aAAa;QACzC,GAAG,CAAC,UAAU,CAAC,mBAAmB,KAAK,SAAS;YAC9C,CAAC,CAAC,EAAE,qBAAqB,EAAE,UAAU,CAAC,mBAAmB,EAAE;YAC3D,CAAC,CAAC,EAAE,CAAC;QACP,UAAU,EAAE,UAAU,CAAC,SAAS;QAChC,SAAS,EAAE,UAAU,CAAC,QAAQ;QAC9B,SAAS,EAAE,UAAU,CAAC,QAAQ;QAC9B,OAAO,EAAE,UAAU,CAAC,OAAO;QAC3B,GAAG,CAAC,UAAU,CAAC,MAAM;YACnB,CAAC,CAAC;gBACE,MAAM,EAAE;oBACN,YAAY,EAAE,UAAU,CAAC,MAAM,CAAC,WAAW;oBAC3C,iBAAiB,EAAE,UAAU,CAAC,MAAM,CAAC,eAAe;oBACpD,oBAAoB,EAAE,UAAU,CAAC,MAAM,CAAC,kBAAkB;oBAC1D,oBAAoB,EAAE,UAAU,CAAC,MAAM,CAAC,iBAAiB;oBACzD,YAAY,EAAE,UAAU,CAAC,MAAM,CAAC,WAAW;oBAC3C,eAAe,EAAE,UAAU,CAAC,MAAM,CAAC,cAAc;oBACjD,UAAU,EAAE,UAAU,CAAC,MAAM,CAAC,SAAS;oBACvC,OAAO,EAAE,UAAU,CAAC,MAAM,CAAC,OAAO;iBACnC;aACF;YACH,CAAC,CAAC,EAAE,CAAC;KACR,CAAC;AACJ,CAAC;AAED,SAAS,uBAAuB,CAAC,cAA8B;IAC7D,OAAO;QACL,MAAM,EAAE,cAAc,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC;YAC5C,IAAI,EAAE;gBACJ,CAAC,EAAE,KAAK,CAAC,IAAI,CAAC,CAAC;gBACf,CAAC,EAAE,KAAK,CAAC,IAAI,CAAC,CAAC;gBACf,KAAK,EAAE,KAAK,CAAC,IAAI,CAAC,KAAK;gBACvB,MAAM,EAAE,KAAK,CAAC,IAAI,CAAC,MAAM;aAC1B;YACD,MAAM,EAAE,KAAK,CAAC,MAAM;YACpB,GAAG,CAAC,KAAK,CAAC,WAAW,KAAK,SAAS;gBACjC,CAAC,CAAC,EAAE,YAAY,EAAE,KAAK,CAAC,WAAW,EAAE;gBACrC,CAAC,CAAC,EAAE,CAAC;YACP,IAAI,EAAE,KAAK,CAAC,IAAI;YAChB,GAAG,CAAC,KAAK,CAAC,SAAS,KAAK,SAAS;gBAC/B,CAAC,CAAC,EAAE,UAAU,EAAE,KAAK,CAAC,SAAS,EAAE;gBACjC,CAAC,CAAC,EAAE,CAAC;YACP,SAAS,EAAE,KAAK,CAAC,QAAQ;SAC1B,CAAC,CAAC;QACH,KAAK,EAAE,cAAc,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC;YACzC,EAAE,EAAE,IAAI,CAAC,EAAE;YACX,EAAE,EAAE,IAAI,CAAC,EAAE;YACX,EAAE,EAAE,IAAI,CAAC,EAAE;YACX,EAAE,EAAE,IAAI,CAAC,EAAE;YACX,MAAM,EAAE,IAAI,CAAC,MAAM;YACnB,GAAG,CAAC,IAAI,CAAC,WAAW,KAAK,SAAS;gBAChC,CAAC,CAAC,EAAE,YAAY,EAAE,IAAI,CAAC,WAAW,EAAE;gBACpC,CAAC,CAAC,EAAE,CAAC;YACP,GAAG,CAAC,IAAI,CAAC,WAAW,KAAK,SAAS;gBAChC,CAAC,CAAC,EAAE,YAAY,EAAE,IAAI,CAAC,WAAW,EAAE;gBACpC,CAAC,CAAC,EAAE,CAAC;YACP,IAAI,EAAE,IAAI,CAAC,IAAI;YACf,GAAG,CAAC,IAAI,CAAC,SAAS,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,UAAU,EAAE,IAAI,CAAC,SAAS,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;SACxE,CAAC,CAAC;KACJ,CAAC;AACJ,CAAC;AAED,SAAS,mBAAmB,CAAC,UAA8B;IACzD,OAAO;QACL,OAAO,EAAE,UAAU,CAAC,OAAO;QAC3B,GAAG,CAAC,UAAU,CAAC,QAAQ,KAAK,SAAS;YACnC,CAAC,CAAC,EAAE,QAAQ,EAAE,UAAU,CAAC,QAAQ,EAAE;YACnC,CAAC,CAAC,EAAE,CAAC;QACP,GAAG,CAAC,UAAU,CAAC,OAAO,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,OAAO,EAAE,UAAU,CAAC,OAAO,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;QAC5E,GAAG,CAAC,UAAU,CAAC,QAAQ,KAAK,SAAS;YACnC,CAAC,CAAC,EAAE,QAAQ,EAAE,UAAU,CAAC,QAAQ,EAAE;YACnC,CAAC,CAAC,EAAE,CAAC;QACP,GAAG,CAAC,UAAU,CAAC,KAAK,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,KAAK,EAAE,UAAU,CAAC,KAAK,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;QACtE,GAAG,CAAC,UAAU,CAAC,IAAI,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,IAAI,EAAE,UAAU,CAAC,IAAI,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;QACnE,GAAG,CAAC,UAAU,CAAC,cAAc,CAAC,MAAM;YAClC,CAAC,CAAC;gBACE,eAAe,EAAE,UAAU,CAAC,cAAc,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC;oBACxD,CAAC,EAAE,IAAI,CAAC,CAAC;oBACT,CAAC,EAAE,IAAI,CAAC,CAAC;oBACT,KAAK,EAAE,IAAI,CAAC,KAAK;oBACjB,MAAM,EAAE,IAAI,CAAC,MAAM;iBACpB,CAAC,CAAC;aACJ;YACH,CAAC,CAAC,EAAE,CAAC;QACP,GAAG,CAAC,UAAU,CAAC,GAAG,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,GAAG,EAAE,UAAU,CAAC,GAAG,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;KACjE,CAAC;AACJ,CAAC;AAED,SAAS,kBAAkB,CAAC,KAAgB;IAC1C,OAAO;QACL,EAAE,EAAE,KAAK,CAAC,EAAE;QACZ,IAAI,EAAE,KAAK,CAAC,IAAI;QAChB,IAAI,EAAE,KAAK,CAAC,IAAI;QAChB,gBAAgB,EAAE,KAAK,CAAC,eAAe;QACvC,YAAY,EAAE,KAAK,CAAC,WAAW;QAC/B,GAAG,CAAC,KAAK,CAAC,YAAY,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,aAAa,EAAE,KAAK,CAAC,YAAY,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;QAClF,GAAG,CAAC,KAAK,CAAC,IAAI,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,IAAI,EAAE,KAAK,CAAC,IAAI,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;QACzD,GAAG,CAAC,KAAK,CAAC,aAAa,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,cAAc,EAAE,KAAK,CAAC,aAAa,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;QACrF,GAAG,CAAC,KAAK,CAAC,KAAK,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,KAAK,EAAE,KAAK,CAAC,KAAK,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;QAC5D,GAAG,CAAC,KAAK,CAAC,WAAW,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,YAAY,EAAE,KAAK,CAAC,WAAW,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;QAC/E,WAAW,EAAE,KAAK,CAAC,UAAU;QAC7B,GAAG,CAAC,KAAK,CAAC,YAAY,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,aAAa,EAAE,KAAK,CAAC,YAAY,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;QAClF,GAAG,CAAC,KAAK,CAAC,YAAY,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,aAAa,EAAE,KAAK,CAAC,YAAY,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;QAClF,GAAG,CAAC,KAAK,CAAC,OAAO,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,OAAO,EAAE,KAAK,CAAC,OAAO,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;QAClE,GAAG,CAAC,KAAK,CAAC,IAAI,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,IAAI,EAAE,KAAK,CAAC,IAAI,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;QACzD,GAAG,CAAC,KAAK,CAAC,OAAO,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,OAAO,EAAE,KAAK,CAAC,OAAO,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;QAC3D,GAAG,CAAC,KAAK,CAAC,eAAe,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,gBAAgB,EAAE,KAAK,CAAC,eAAe,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;KACrF,CAAC;AACJ,CAAC;AAED,SAAS,kBAAkB,CAAC,MAAiB;IAC3C,OAAO;QACL,KAAK,EAAE,MAAM,CAAC,KAAK;QACnB,GAAG,CAAC,MAAM,CAAC,IAAI,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,IAAI,EAAE,MAAM,CAAC,IAAI,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;QAC3D,cAAc,EAAE,MAAM,CAAC,aAAa;QACpC,GAAG,CAAC,MAAM,CAAC,OAAO,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,OAAO,EAAE,MAAM,CAAC,OAAO,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;KACrE,CAAC;AACJ,CAAC;AAED,SAAS,cAAc,CAAC,KAAqB;IAC3C,OAAO;QACL,EAAE,EAAE,KAAK,CAAC,EAAE;QACZ,IAAI,EAAE,KAAK,CAAC,IAAI;QAChB,GAAG,CAAC,KAAK,CAAC,IAAI,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,IAAI,EAAE,KAAK,CAAC,IAAI,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;QACzD,IAAI,EAAE,KAAK,CAAC,IAAI;QAChB,IAAI,EAAE;YACJ,CAAC,EAAE,KAAK,CAAC,IAAI,CAAC,CAAC;YACf,CAAC,EAAE,KAAK,CAAC,IAAI,CAAC,CAAC;YACf,KAAK,EAAE,KAAK,CAAC,IAAI,CAAC,KAAK;YACvB,MAAM,EAAE,KAAK,CAAC,IAAI,CAAC,MAAM;SAC1B;QACD,KAAK,EAAE,KAAK,CAAC,KAAK;QAClB,MAAM,EAAE,KAAK,CAAC,MAAM;QACpB,QAAQ,EAAE,KAAK,CAAC,QAAQ;QACxB,MAAM,EAAE,KAAK,CAAC,MAAM;QACpB,GAAG,CAAC,KAAK,CAAC,WAAW,KAAK,SAAS;YACjC,CAAC,CAAC,EAAE,YAAY,EAAE,KAAK,CAAC,WAAW,EAAE;YACrC,CAAC,CAAC,EAAE,CAAC;KACR,CAAC;AACJ,CAAC;AAED,2EAA2E;AAC3E,MAAM,UAAU,oBAAoB,CAClC,MAAmB,EACnB,UAA6C,EAAE;IAE/C,MAAM,mBAAmB,GAAG,OAAO,CAAC,mBAAmB,IAAI,KAAK,CAAC;IACjE,OAAO;QACL,KAAK,EAAE,MAAM,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC;YACjC,IAAI,EAAE,IAAI,CAAC,OAAO;YAClB,KAAK,EAAE,IAAI,CAAC,KAAK;YACjB,MAAM,EAAE,IAAI,CAAC,MAAM;YACnB,GAAG,CAAC,IAAI,CAAC,aAAa;gBACpB,CAAC,CAAC;oBACE,cAAc,EAAE;wBACd,CAAC,EAAE,IAAI,CAAC,aAAa,CAAC,CAAC;wBACvB,CAAC,EAAE,IAAI,CAAC,aAAa,CAAC,CAAC;wBACvB,KAAK,EAAE,IAAI,CAAC,aAAa,CAAC,KAAK;wBAC/B,MAAM,EAAE,IAAI,CAAC,aAAa,CAAC,MAAM;qBAClC;iBACF;gBACH,CAAC,CAAC,EAAE,CAAC;YACP,IAAI,EAAE,IAAI,CAAC,IAAI;YACf,UAAU,EAAE,IAAI,CAAC,SAAS,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CACtC,iBAAiB,CAAC,IAAI,EAAE,mBAAmB,CAAC,CAC7C;YACD,GAAG,CAAC,IAAI,CAAC,UAAU;gBACjB,CAAC,CAAC,EAAE,UAAU,EAAE,mBAAmB,CAAC,IAAI,CAAC,UAAU,CAAC,EAAE;gBACtD,CAAC,CAAC,EAAE,CAAC;YACP,GAAG,CAAC,IAAI,CAAC,cAAc;gBACrB,CAAC,CAAC,EAAE,eAAe,EAAE,uBAAuB,CAAC,IAAI,CAAC,cAAc,CAAC,EAAE;gBACnE,CAAC,CAAC,EAAE,CAAC;YACP,GAAG,CAAC,IAAI,CAAC,WAAW,KAAK,SAAS;gBAChC,CAAC,CAAC,EAAE,WAAW,EAAE,IAAI,CAAC,WAAW,CAAC,GAAG,CAAC,mBAAmB,CAAC,EAAE;gBAC5D,CAAC,CAAC,EAAE,CAAC;YACP,GAAG,CAAC,IAAI,CAAC,UAAU,KAAK,SAAS;gBAC/B,CAAC,CAAC,EAAE,WAAW,EAAE,IAAI,CAAC,UAAU,CAAC,GAAG,CAAC,kBAAkB,CAAC,EAAE;gBAC1D,CAAC,CAAC,EAAE,CAAC;YACP,GAAG,CAAC,IAAI,CAAC,aAAa,KAAK,SAAS;gBAClC,CAAC,CAAC;oBACE,cAAc,EAAE;wBACd,KAAK,EAAE,IAAI,CAAC,aAAa,CAAC,KAAK,CAAC,GAAG,CAAC,6BAA6B,CAAC;qBACnE;iBACF;gBACH,CAAC,CAAC,EAAE,CAAC;SACR,CAAC,CAAC;QACH,GAAG,CAAC,MAAM,CAAC,MAAM,CAAC,MAAM;YACtB,CAAC,CAAC,EAAE,MAAM,EAAE,MAAM,CAAC,MAAM,CAAC,GAAG,CAAC,cAAc,CAAC,EAAE;YAC/C,CAAC,CAAC,EAAE,CAAC;QACP,GAAG,CAAC,MAAM,CAAC,eAAe;YACxB,CAAC,CAAC,EAAE,iBAAiB,EAAE,MAAM,CAAC,eAAe,EAAE;YAC/C,CAAC,CAAC,EAAE,CAAC;QACP,GAAG,CAAC,MAAM,CAAC,QAAQ,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,SAAS,EAAE,MAAM,CAAC,QAAQ,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;QACxE,GAAG,CAAC,MAAM,CAAC,UAAU,KAAK,SAAS;YACjC,CAAC,CAAC,EAAE,WAAW,EAAE,MAAM,CAAC,UAAU,CAAC,GAAG,CAAC,kBAAkB,CAAC,EAAE;YAC5D,CAAC,CAAC,EAAE,CAAC;KACR,CAAC;AACJ,CAAC"}

package/dist/cli.js

@@ -1,6 +1,7 @@
 #!/usr/bin/env node
 import { program } from "commander";
 import { LiteParse } from "./lib.js";
+import { parseResultToCliJson } from "./cli-json.js";
 import { readFileSync, writeFileSync, mkdirSync, readdirSync } from "node:fs";
 import { join, relative, parse as parsePath } from "node:path";
 program
@@ -47,7 +48,14 @@ program
     .option("--format <format>", 'Output format: json|text|markdown (default: "text")')
     .option("--image-mode <mode>", "How to surface raster images in markdown: off|placeholder|embed (default: placeholder)")
     .option("--image-output-dir <dir>", "Directory to write embedded images to when --image-mode embed is set")
+    .option("--extract-images", "Extract embedded image bytes and metadata")
     .option("--no-links", "Disable hyperlink extraction (emit plain anchor text)")
+    .option("--keep-headers-footers", "Keep running headers/footers in markdown output instead of stripping them")
+    .option("--extract-annotations", "Include all PDF annotations in page output")
+    .option("--extract-form-fields", "Include AcroForm widget fields and values")
+    .option("--extract-structure-tree", "Include the tagged-PDF logical structure tree")
+    .option("--extract-xfa-packets", "Include raw XFA packets (name + XML content) in JSON output")
+    .option("--extract-content-bounds", "Include each page's content_bounds in JSON output")
     .option("--ocr-server-url <url>", "HTTP OCR server URL")
     .option("--ocr-server-header <header>", 'Extra header for OCR server requests, "Name: Value" (repeatable)', collectHeader)
     .option("--no-ocr", "Disable OCR")
@@ -56,11 +64,13 @@ program
     .option("--target-pages <pages>", 'Pages to parse (e.g., "1-5,10,15-20")')
     .option("--dpi <dpi>", "Rendering DPI", parseFloat)
     .option("--preserve-small-text", "Keep very small text")
+    .option("--extract-text-metadata", "Include rich PDF text metadata in text items and JSON output")
     .option("--password <password>", "Password for encrypted documents")
     .option("--config <file>", "JSON config file path")
     .option("-q, --quiet", "Suppress progress output")
     .option("--num-workers <n>", "Number of concurrent OCR workers", parseInt)
     .option("--complexity", "Include per-page complexity signals in JSON output")
+    .option("--extract-vector-graphics", "Include page-scoped vector shapes and merged horizontal/vertical lines")
     .action(async (file, opts) => {
     try {
         const config = {};
@@ -74,8 +84,24 @@ program
             config.outputFormat = opts.format;
         if (opts.imageMode)
             config.imageMode = opts.imageMode;
+        if (opts.imageOutputDir)
+            config.imageOutputDir = opts.imageOutputDir;
+        if (opts.extractImages)
+            config.extractImages = true;
         if (opts.links === false)
             config.extractLinks = false;
+        if (opts.keepHeadersFooters)
+            config.keepHeadersFooters = true;
+        if (opts.extractAnnotations)
+            config.extractAnnotations = true;
+        if (opts.extractFormFields)
+            config.extractFormFields = true;
+        if (opts.extractStructureTree)
+            config.extractStructureTree = true;
+        if (opts.extractXfaPackets)
+            config.extractXfaPackets = true;
+        if (opts.extractContentBounds)
+            config.extractContentBounds = true;
         if (opts.ocrServerUrl)
             config.ocrServerUrl = opts.ocrServerUrl;
         if (opts.ocrServerHeader)
@@ -92,6 +118,8 @@ program
             config.dpi = opts.dpi;
         if (opts.preserveSmallText)
             config.preserveVerySmallText = true;
+        if (opts.extractTextMetadata)
+            config.extractTextMetadata = true;
         if (opts.password)
             config.password = opts.password;
         if (opts.quiet)
@@ -100,32 +128,18 @@ program
             config.numWorkers = opts.numWorkers;
         if (opts.complexity)
             config.includeComplexity = true;
+        if (opts.extractVectorGraphics)
+            config.extractVectorGraphics = true;
         // Default CLI output to text (library defaults to json)
         if (!config.outputFormat)
             config.outputFormat = "text";
         const parser = new LiteParse(config);
         const result = await parser.parse(await resolveInput(file));
         const output = config.outputFormat === "json"
-            ? JSON.stringify({
-                pages: result.pages.map((p) => ({
-                    page: p.pageNum,
-                    width: p.width,
-                    height: p.height,
-                    text: p.text,
-                    textItems: p.textItems,
-                })),
-            }, null, 2)
+            ? JSON.stringify(parseResultToCliJson(result, {
+                extractTextMetadata: config.extractTextMetadata,
+            }), null, 2)
             : result.text;
-        if (opts.imageOutputDir && result.images.length > 0) {
-            const dir = opts.imageOutputDir;
-            mkdirSync(dir, { recursive: true });
-            for (const img of result.images) {
-                writeFileSync(join(dir, `image_${img.id}.${img.format}`), img.bytes);
-            }
-            if (!opts.quiet) {
-                console.error(`[liteparse] wrote ${result.images.length} image(s) to ${dir}`);
-            }
-        }
         if (opts.output) {
             writeFileSync(opts.output, output, "utf-8");
         }
@@ -258,6 +272,14 @@ program
     .option("--password <password>", "Password for encrypted documents")
     .option("-q, --quiet", "Suppress progress output")
     .option("--num-workers <n>", "Number of concurrent OCR workers", parseInt)
+    .option("--extract-text-metadata", "Include rich PDF text metadata in text items and JSON output")
+    .option("--extract-images", "Extract embedded image bytes and metadata")
+    .option("--extract-annotations", "Include all PDF annotations in page output")
+    .option("--extract-form-fields", "Include AcroForm widget fields and values")
+    .option("--extract-structure-tree", "Include the tagged-PDF logical structure tree")
+    .option("--extract-xfa-packets", "Include raw XFA packets (name + XML content) in JSON output")
+    .option("--extract-content-bounds", "Include each page's content_bounds in JSON output")
+    .option("--extract-vector-graphics", "Include page-scoped vector shapes and merged horizontal/vertical lines")
     .action(async (inputDir, outputDir, opts) => {
     try {
         const config = {};
@@ -281,6 +303,28 @@ program
             config.quiet = true;
         if (opts.numWorkers)
             config.numWorkers = opts.numWorkers;
+        if (opts.extractTextMetadata)
+            config.extractTextMetadata = true;
+        if (opts.extractImages)
+            config.extractImages = true;
+        if (opts.extractAnnotations)
+            config.extractAnnotations = true;
+        if (opts.extractFormFields)
+            config.extractFormFields = true;
+        if (opts.extractStructureTree)
+            config.extractStructureTree = true;
+        if (opts.extractXfaPackets)
+            config.extractXfaPackets = true;
+        if (opts.extractContentBounds)
+            config.extractContentBounds = true;
+        if (opts.extractContentBounds)
+            config.extractContentBounds = true;
+        if (opts.extractXfaPackets)
+            config.extractXfaPackets = true;
+        if (opts.extractContentBounds)
+            config.extractContentBounds = true;
+        if (opts.extractVectorGraphics)
+            config.extractVectorGraphics = true;
         const parser = new LiteParse(config);
         const outExt = format === "json" ? ".json" : format === "markdown" ? ".md" : ".txt";
         mkdirSync(outputDir, { recursive: true });
@@ -308,15 +352,9 @@ program
             try {
                 const result = await parser.parse(filePath);
                 const output = format === "json"
-                    ? JSON.stringify({
-                        pages: result.pages.map((p) => ({
-                            page: p.pageNum,
-                            width: p.width,
-                            height: p.height,
-                            text: p.text,
-                            textItems: p.textItems,
-                        })),
-                    }, null, 2)
+                    ? JSON.stringify(parseResultToCliJson(result, {
+                        extractTextMetadata: config.extractTextMetadata,
+                    }), null, 2)
                     : result.text;
                 writeFileSync(outPath, output, "utf-8");
                 success++;