@dragon708/docmind-node 1.13.2 → 1.13.4

package/dist/index.d.ts

@@ -6,8 +6,8 @@ import { AnalyzeDocxIncludeFlags, DocxToHtmlOptions, ExtractStructuredDataFromDo
 export { AnalyzeDocxIncludeFlags, DocxEmbeddedImage, DocxEmbeddedImageConversionOptions, DocxEmbeddedImageConversionResult, DocxEmbeddedImageWebSlice, DocxImageExtractionMode, ExtractImagesFromDocxOptions, ExtractImagesFromDocxResult, ExtractStructuredDataFromDocxOptions, convertDocxEmbeddedImageToWeb, convertDocxImagesForWeb, docxImageIsBrowserRenderable, docxImageRequiresConversion, docxImageSuggestedTargetFormat, docxImageToDataUri, extractImagesFromDocx, extractStructuredDataFromDocx, isNodeJsRuntime } from '@dragon708/docmind-docx';
 import { PdfAnalyzeOptions, ExtractStructuredDataFromPdfOptions } from '@dragon708/docmind-pdf';
 export { ExtractStructuredDataFromPdfIncludeFlags, ExtractStructuredDataFromPdfOptions, extractStructuredDataFromPdf } from '@dragon708/docmind-pdf';
-import { RenderLlmTextOptions, RenderMarkdownOptions, ConvertDocxToMarkdownOptions, ConvertPdfToMarkdownOptions, ExtractMarkdownStrategy, RenderMarkdownSectionsOptions, MarkdownSection } from '@dragon708/docmind-markdown';
-export { MarkdownSection } from '@dragon708/docmind-markdown';
+import { RenderLlmTextOptions, RenderMarkdownOptions, ConvertDocxToMarkdownOptions, ConvertPdfToMarkdownOptions, ConvertHtmlToMarkdownOptions, ConvertCsvToMarkdownOptions, ConvertSpreadsheetToMarkdownOptions, ExtractMarkdownStrategy, ExtractMarkdownRoutingInfo, RenderMarkdownSectionsOptions, MarkdownSection } from '@dragon708/docmind-markdown';
+export { ConvertCsvToMarkdownOptions, ConvertCsvToMarkdownResult, ConvertHtmlToMarkdownOptions, ConvertHtmlToMarkdownResult, ConvertSpreadsheetToMarkdownOptions, ConvertSpreadsheetToMarkdownResult, CsvStringInputMode, DetectedBinaryFormat, ExtractMarkdownMediaHint, ExtractMarkdownRoutingInfo, ExtractMarkdownStrategy, HtmlStringInputMode, MarkdownSection, convertCsvToMarkdown, convertHtmlToMarkdown, convertSpreadsheetToMarkdown, detectBinaryFormat } from '@dragon708/docmind-markdown';
 
 /**
  * Options for Node public APIs (`analyzeFile`, intent methods).
@@ -54,8 +54,9 @@ interface NodeExtractStructuredDataOptions extends DocMindAnalyzeOptions {
 }
 /**
  * {@link extractMarkdown}: inherits {@link NodeExtractStructuredDataOptions}; `markdown` maps to structured-serializer
- * options passed through `extractMarkdown` in `@dragon708/docmind-markdown`. `markdownDocx` / `markdownPdf`
- * configure Mammoth→Turndown and `@cognipeer/to-markdown` (`convertPdfToMarkdown`) respectively — separate from `docx` / `pdf` used only by {@link extractStructuredData}.
+ * options passed through `extractMarkdown` in `@dragon708/docmind-markdown`. `markdownDocx` / `markdownPdf` /
+ * `markdownHtml` / `markdownCsv` / `markdownSpreadsheet` configure specialized binary routes — separate from `docx` / `pdf`
+ * used only by {@link extractStructuredData}.
  */
 interface NodeExtractMarkdownOptions extends NodeExtractStructuredDataOptions {
     readonly markdown?: RenderMarkdownOptions;
@@ -63,13 +64,20 @@ interface NodeExtractMarkdownOptions extends NodeExtractStructuredDataOptions {
     readonly markdownDocx?: ConvertDocxToMarkdownOptions;
     /** Forwarded to `extractMarkdown` → `convertPdfToMarkdown` when bytes are PDF (Node; `@cognipeer/to-markdown`). */
     readonly markdownPdf?: ConvertPdfToMarkdownOptions;
+    /** Forwarded to `extractMarkdown` → `convertHtmlToMarkdown` when bytes are detected as HTML (Node). */
+    readonly markdownHtml?: ConvertHtmlToMarkdownOptions;
+    /** Forwarded to `extractMarkdown` → `convertCsvToMarkdown` when bytes are detected as CSV (Node). */
+    readonly markdownCsv?: ConvertCsvToMarkdownOptions;
+    /** Forwarded to `extractMarkdown` → `convertSpreadsheetToMarkdown` for `.xlsx` / `.xls` (Node). */
+    readonly markdownSpreadsheet?: ConvertSpreadsheetToMarkdownOptions;
     /**
-     * Optional: receive `strategy` and merged `warnings` from `@dragon708/docmind-markdown` `extractMarkdown`
-     * (e.g. `pdf-cognipeer-specialized` vs `pdf-structured-fallback`, `docx-mammoth` vs `docx-structured-fallback`) without changing the `Promise<string>` return type.
+     * Optional: receive `strategy`, merged `warnings`, and optional {@link ExtractMarkdownRoutingInfo} from
+     * `@dragon708/docmind-markdown` `extractMarkdown` without changing the `Promise<string>` return type.
      */
     readonly onMarkdownExtract?: (info: {
         strategy: ExtractMarkdownStrategy;
         warnings: readonly string[];
+        routing?: ExtractMarkdownRoutingInfo;
     }) => void;
 }
 /**
@@ -142,23 +150,24 @@ declare function extractStructuredData(input: NodeAnalyzeInput, options?: NodeEx
  * End-to-end: {@link extractStructuredData} (for fallback + option parity) plus
  * `extractMarkdown` from `@dragon708/docmind-markdown` on `{ data, filename, mimeType }`.
  *
- * On Node, **DOCX** bytes use **Mammoth → Turndown**; **PDF** bytes use **`convertPdfToMarkdown`**
- * (`@cognipeer/to-markdown`, Node-first, no JVM). The structured envelope is always passed as `structuredFallback`
- * so specialized routes can fall back to structured → Markdown serialization with clear warnings when needed.
+ * On Node, **DOCX** bytes use **Mammoth → Turndown**; **PDF**, **HTML**, **CSV**, and **Excel** bytes use Cognipeer-backed
+ * converters in `@dragon708/docmind-markdown` (`convertPdfToMarkdown`, `convertHtmlToMarkdown`, `convertCsvToMarkdown`,
+ * `convertSpreadsheetToMarkdown`). The structured envelope is always passed as `structuredFallback` so specialized routes
+ * can fall back to structured → Markdown serialization with clear warnings when needed.
  *
- * {@link NodeExtractMarkdownOptions.onMarkdownExtract} receives `strategy` and merged `warnings` from the markdown
- * package (for example `pdf-cognipeer-specialized`, `pdf-structured-fallback`, `docx-structured-fallback`, and
- * `[docmind-markdown:extractMarkdown]` trace lines).
+ * {@link NodeExtractMarkdownOptions.onMarkdownExtract} receives `strategy`, merged `warnings`, and optional `routing`
+ * (including `routingSummary`) from the markdown package.
  *
  * @param input - Path, buffer, or {@link NodeAnalyzeInput} accepted by the Node facade.
  * @param options - Structured routing (`pdf` / `docx` / `ocr` / `normalize`), optional `markdown` serializer knobs,
- *   and optional `markdownDocx` / `markdownPdf` for the binary Markdown pipelines (distinct from structured-only `docx` / `pdf`).
+ *   and optional `markdownDocx` / `markdownPdf` / `markdownHtml` / `markdownCsv` / `markdownSpreadsheet` for binary Markdown
+ *   pipelines (distinct from structured-only `docx` / `pdf`).
  */
 declare function extractMarkdown(input: NodeAnalyzeInput, options?: NodeExtractMarkdownOptions): Promise<string>;
 /**
  * {@link extractStructuredData} then `renderLlmText` in `@dragon708/docmind-markdown` (tagged plain text for LLMs).
- * This path **does not** run the binary PDF/DOCX Markdown pipelines (no `@cognipeer/to-markdown` / Mammoth→Turndown here);
- * it linearizes the structured envelope only. Use {@link extractMarkdown} when you need specialized PDF or DOCX Markdown layout.
+ * This path **does not** run binary Cognipeer / Mammoth→Turndown Markdown pipelines (no `@cognipeer/to-markdown` here);
+ * it linearizes the structured envelope only. Use {@link extractMarkdown} when you need specialized PDF / DOCX / HTML / CSV / Excel Markdown layout on Node.
  *
  * @param options - Structured routing plus optional `llm` passed through to `renderLlmText`.
  */
@@ -275,7 +284,7 @@ type NodeExplainAnalysisPlanOptions = Omit<ExplainAnalysisPlanOptions, "intent">
 
 /**
  * Epic 1 — **Capabilities:** after resolving {@link NodeAnalyzeInput}, lists which of
- * `text` | `metadata` | `html` | `ocr` | `pages` | `structured-output` | `markdown` (hybrid package extract on Node) | `llm-text` | `structured-chunks` (split + Markdown sections)
+ * `text` | `metadata` | `html` | `ocr` | `pages` | `structured-output` | `markdown` (hybrid `extractMarkdown`: DOCX, PDF, HTML, CSV, Excel on Node via `@dragon708/docmind-markdown`) | `llm-text` | `structured-chunks` (split + Markdown sections)
  * (and image-specific ids such as `ocr-multipage`, `image-normalization`, `tiff`, `heic-node-only`, `bmp`, `gif-first-frame`) apply for that kind in Node (for PDF, `text` / `metadata` /
  * `pages` / `ocr` describe the v2 pdf-parse + PDF.js + OCR stack; see {@link buildNodeCapabilityReport}).
  * For **DOCX**, `docxEmbeddedImages` and `docxStructure` describe ZIP media and optional OOXML v2 extractors (`options.docx.include`).

package/dist/index.js

@@ -10,6 +10,7 @@ import { readFile } from 'fs/promises';
 import { basename } from 'path';
 import { fileURLToPath } from 'url';
 import { extractMarkdown as extractMarkdown$1, renderLlmText, renderMarkdownSections } from '@dragon708/docmind-markdown';
+export { convertCsvToMarkdown, convertHtmlToMarkdown, convertSpreadsheetToMarkdown, detectBinaryFormat } from '@dragon708/docmind-markdown';
 
 // src/analyze.ts
 
@@ -624,6 +625,9 @@ async function extractMarkdown(input, options) {
     markdown: markdownOpts,
     markdownDocx,
     markdownPdf,
+    markdownHtml,
+    markdownCsv,
+    markdownSpreadsheet,
     onMarkdownExtract,
     ...structuredOpts
   } = options ?? {};
@@ -642,10 +646,13 @@ async function extractMarkdown(input, options) {
       ...markdownOpts ?? {},
       ...markdownDocx !== void 0 ? { docx: markdownDocx } : {},
       ...markdownPdf !== void 0 ? { pdf: markdownPdf } : {},
+      ...markdownHtml !== void 0 ? { html: markdownHtml } : {},
+      ...markdownCsv !== void 0 ? { csv: markdownCsv } : {},
+      ...markdownSpreadsheet !== void 0 ? { spreadsheet: markdownSpreadsheet } : {},
       structuredFallback: structured
     }
   );
-  onMarkdownExtract?.({ strategy: r.strategy, warnings: r.warnings });
+  onMarkdownExtract?.({ strategy: r.strategy, warnings: r.warnings, routing: r.routing });
   return r.markdown;
 }
 async function extractLlmContent(input, options) {
@@ -724,7 +731,7 @@ function buildNodeCapabilityReport(kind) {
           "extractStructuredData: analyzePdf + PDF.js per-page text, outline, links, annotations, merged via normalizeToStructuredResult; respects pdf.ocr like analyzeFile."
         ]),
         slot("markdown", true, [
-          "extractMarkdown: `@dragon708/docmind-markdown` hybrid \u2014 PDF bytes use `@cognipeer/to-markdown` (Node-first, no Java); structured PDF (`extractStructuredData`, respects pdf.ocr) is always built as fallback."
+          "extractMarkdown: `@dragon708/docmind-markdown` hybrid \u2014 PDF/HTML/CSV/Excel bytes use `@cognipeer/to-markdown` (Node-first, no Java); DOCX uses Mammoth\u2192Turndown; structured envelope from `extractStructuredData` is always `structuredFallback` (PDF respects pdf.ocr)."
         ]),
         slot("llm-text", true, [
           "extractLlmContent: structured envelope \u2192 LLM-oriented plain text (`renderLlmText` in `@dragon708/docmind-markdown`)."
@@ -811,7 +818,7 @@ function buildNodeCapabilityReport(kind) {
           "extractStructuredData wraps UTF-8 decode in normalizeToStructuredResult (rollup text only unless you add blocks upstream)."
         ]),
         slot("markdown", true, [
-          "extractMarkdown: rollup/decoded text \u2192 Markdown (mostly paragraphs; no layout without upstream blocks)."
+          "extractMarkdown: rollup/decoded text \u2192 Markdown (mostly paragraphs; no layout without upstream blocks). If filename/MIME indicate `.html`, `.csv`, `.xlsx`/`.xls`, the markdown package routes to Cognipeer specialized converters on Node."
         ]),
         slot("llm-text", true, ["extractLlmContent: rollup \u2192 LLM plain text via the same envelope."]),
         slot("structured-chunks", true, [
@@ -1080,7 +1087,7 @@ function buildNodeExplainReport(kind, intent, pdfOcr, plan, docxInclude, ocrSlic
         limitations = [
           ...limitations,
           ...lim(
-            "Output: Markdown string from `@dragon708/docmind-markdown` extractMarkdown (PDF/DOCX binary routes on Node when applicable; structured serializer as fallback)."
+            "Output: Markdown string from `@dragon708/docmind-markdown` extractMarkdown (PDF, DOCX, HTML, CSV, Excel specialized routes on Node when bytes match; structured serializer as `structuredFallback`)."
           )
         ];
       } else if (intent === "extractLlmContent") {

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@dragon708/docmind-node",
-  "version": "1.13.2",
+  "version": "1.13.4",
   "description": "Official DocMind Node facade: analyzeFile, intent APIs, PDF/DOCX/OCR, and fs helpers.",
   "type": "module",
   "main": "./dist/index.js",
@@ -33,7 +33,7 @@
   "license": "MIT",
   "dependencies": {
     "@dragon708/docmind-docx": "^1.8.0",
-    "@dragon708/docmind-markdown": "^1.2.6",
+    "@dragon708/docmind-markdown": "^1.2.8",
     "@dragon708/docmind-ocr": "^1.1.4",
     "@dragon708/docmind-pdf": "^2.2.0",
     "@dragon708/docmind-shared": "^1.2.0"