@jacobbubu/md-to-lark 1.4.11 → 1.5.1

package/README.md

@@ -112,6 +112,7 @@ Presets, Mermaid, and stage artifacts:
 npm run publish:md -- --input ./test-md/comp/comp.md --preset medium --dry-run
 npm run publish:md -- --input ./test-md/comp/comp.md --preset zh-format --dry-run
 npm run publish:md -- --input ./test-md/comp/comp.md --preset zh-format --preset ./my-preset.mjs --dry-run
+npm run publish:md -- --input ./paper.md --single-dollar-text-math --dry-run
 npm run publish:md -- --input ./tmp/generated/article.md --resource-base-dir ./source-assets --dry-run
 npm run publish:md -- --input ./test-md/mermaid.md --mermaid-target board --dry-run
 npm run publish:md -- --input ./test-md/comp/comp.md --pipeline-cache-dir ./out/debug-cache --dry-run
@@ -175,6 +176,7 @@ Guardrails:
 - Local attachment and image detection with real upload
 - Remote image download and standalone URL preparation
 - Mermaid `text-drawing` and `board` output paths
+- Opt-in single-dollar inline math parsing for LaTeX-heavy papers
 - Table width heuristics and numeric-column right alignment
 - Chinese Markdown formatting preset (`zh-format`)
 - Ordered preset composition from CLI and programmatic usage

package/dist/commands/publish-md/args.js

@@ -1,6 +1,6 @@
 function usage() {
     return [
-        'Usage: npm run publish:md -- --input <file.md|dir> [--title <doc_title_or_prefix>] [--date-prefix|--no-date-prefix] [--preset <preset_name_or_module_path>]... [--document-base-url <base_url>] [--resource-base-dir <dir>] [--folder <folder_token>] [--doc <document_id>] [--download-remote-images|--no-download-remote-images] [--yt-dlp-path <path>] [--yt-dlp-cookies-path <path>] [--pipeline-cache-dir <dir>] [--mermaid-target <text-drawing|board>] [--mermaid-board-syntax-type <int>] [--mermaid-board-style-type <int>] [--mermaid-board-diagram-type <int>] [--dry-run] [--help|-h]',
+        'Usage: npm run publish:md -- --input <file.md|dir> [--title <doc_title_or_prefix>] [--date-prefix|--no-date-prefix] [--preset <preset_name_or_module_path>]... [--document-base-url <base_url>] [--resource-base-dir <dir>] [--folder <folder_token>] [--doc <document_id>] [--download-remote-images|--no-download-remote-images] [--yt-dlp-path <path>] [--yt-dlp-cookies-path <path>] [--pipeline-cache-dir <dir>] [--single-dollar-text-math] [--mermaid-target <text-drawing|board>] [--mermaid-board-syntax-type <int>] [--mermaid-board-style-type <int>] [--mermaid-board-diagram-type <int>] [--dry-run] [--help|-h]',
         '',
         'Options:',
         '  --input   Markdown file path, or directory path (publish all *.md recursively).',
@@ -17,6 +17,7 @@ function usage() {
         '  --yt-dlp-path               Optional yt-dlp executable path for standalone URL extraction.',
         '  --yt-dlp-cookies-path       Optional cookie file path passed to yt-dlp --cookies.',
         '  --pipeline-cache-dir        Pipeline cache root directory. Default: ./out/pipeline-cache',
+        '  --single-dollar-text-math   Opt in to parsing $...$ as inline math. Default: disabled to avoid currency false positives.',
         '  --mermaid-target            Mermaid render target: text-drawing (default) or board.',
         '  --mermaid-board-syntax-type Optional integer syntax_type for board createPlantuml (default: 2).',
         '  --mermaid-board-style-type  Optional integer style_type for board createPlantuml.',
@@ -63,6 +64,7 @@ export function parsePublishMdArgs(argv, env = process.env) {
     let ytDlpPath = '';
     let ytDlpCookiesPath = '';
     let pipelineCacheDir = '';
+    let singleDollarTextMath;
     let mermaidTarget;
     let mermaidBoardSyntaxType;
     let mermaidBoardStyleType;
@@ -175,6 +177,10 @@ export function parsePublishMdArgs(argv, env = process.env) {
             i += 1;
             continue;
         }
+        if (arg === '--single-dollar-text-math' || arg === '--single-dollar-math') {
+            singleDollarTextMath = true;
+            continue;
+        }
         if (arg === '--mermaid-target') {
             const value = argv[i + 1];
             if (!value)
@@ -245,6 +251,7 @@ export function parsePublishMdArgs(argv, env = process.env) {
         ...(ytDlpPath.trim() ? { ytDlpPath: ytDlpPath.trim() } : {}),
         ...(ytDlpCookiesPath.trim() ? { ytDlpCookiesPath: ytDlpCookiesPath.trim() } : {}),
         ...(pipelineCacheDir.trim() ? { pipelineCacheDir: pipelineCacheDir.trim() } : {}),
+        ...(singleDollarTextMath === undefined ? {} : { singleDollarTextMath }),
         ...(mermaidTarget ? { mermaidTarget } : {}),
         ...(mermaidBoardSyntaxType === undefined ? {} : { mermaidBoardSyntaxType }),
         ...(mermaidBoardStyleType === undefined ? {} : { mermaidBoardStyleType }),

package/dist/commands/publish-md/command.js

@@ -73,9 +73,7 @@ export async function publishMdToLark(options, env = process.env) {
     const runtime = buildPublishRuntime(options, env, markdownPresets);
     logPublishRuntimeSummary(runtime, inputSet.markdownFiles.length, inputSet.mode);
     const normalizedDocumentId = options.documentId ? normalizeDocumentId(options.documentId) : undefined;
-    const resolveTargetDocumentId = options.dryRun || normalizedDocumentId
-        ? undefined
-        : createFolderDocumentResolver(runtime, options);
+    const resolveTargetDocumentId = options.dryRun || normalizedDocumentId ? undefined : createFolderDocumentResolver(runtime, options);
     const results = [];
     for (let index = 0; index < inputSet.markdownFiles.length; index += 1) {
         const markdownPath = inputSet.markdownFiles[index];

package/dist/commands/publish-md/pipeline-transform.js

@@ -1,4 +1,4 @@
 export { applyStandaloneAttachmentTransforms } from '../../publish/asset-adapter.js';
 export { patchBTTForMermaidAndAssets } from '../../publish/btt-patch.js';
 export { buildPipelineDocumentId } from '../../publish/ids.js';
-export { applyTableColumnWidthHeuristics, collectMermaidPatches, ensureLastBlockBttIds } from '../../publish/last-normalize.js';
+export { applyTableColumnWidthHeuristics, collectMermaidPatches, ensureLastBlockBttIds, } from '../../publish/last-normalize.js';

package/dist/interop/btt-to-last.js

@@ -1,4 +1,4 @@
-import { buildLASTIndexes, flattenTreeBlocks, fromRawBlockToLAST, } from './codec-btt-to-last.js';
+import { buildLASTIndexes, flattenTreeBlocks, fromRawBlockToLAST } from './codec-btt-to-last.js';
 import { normalizeDocumentIdToLAST } from './codec-last-to-btt.js';
 export function convertBTTToLAST(bttDoc) {
     const rawBlocks = Object.keys(bttDoc.flatBlocks ?? {}).length

package/dist/lark/docx/render-btt.js

@@ -1,5 +1,5 @@
-import { createDocumentChildren, } from './ops.js';
-import { DEFAULT_MERMAID_RENDER_CONFIG, } from './render-types.js';
+import { createDocumentChildren } from './ops.js';
+import { DEFAULT_MERMAID_RENDER_CONFIG } from './render-types.js';
 import { applyCreatedBoardMermaid, applyCreatedFileBlock, applyCreatedImageBlock } from './render-post-process.js';
 import { renderCreatedTableNode } from './render-table.js';
 import { buildCreatePayloadFromRawBlock, createRenderBatchEntry, getSourceBlockId, toObjectRecord, } from './render-payload.js';

package/dist/lark/docx/render-payload.js

@@ -124,7 +124,8 @@ export function canUseElementsOnlyPatch(rawBlock) {
     if (effectiveEntries.length === 0)
         return true;
     for (const [key, value] of effectiveEntries) {
-        if (key === 'align' && (value === 1 || value === 2 || value === 3 || value === 'left' || value === 'center' || value === 'right')) {
+        if (key === 'align' &&
+            (value === 1 || value === 2 || value === 3 || value === 'left' || value === 'center' || value === 'right')) {
             continue;
         }
         return false;
@@ -310,7 +311,12 @@ export function getSourceBlockId(node, rawBlockRecord) {
         : node.blockId;
 }
 function isBatchSafeBlockType(blockType) {
-    if (blockType === 1 || blockType === 23 || blockType === 27 || blockType === 31 || blockType === 32 || blockType === 43) {
+    if (blockType === 1 ||
+        blockType === 23 ||
+        blockType === 27 ||
+        blockType === 31 ||
+        blockType === 32 ||
+        blockType === 43) {
         return false;
     }
     return true;

package/dist/lark/docx/render-post-process.js

@@ -33,7 +33,9 @@ export async function applyCreatedImageBlock(client, documentId, createdBlockId,
         ...(typeof width === 'number' ? { width } : {}),
         ...(typeof proportionalHeight === 'number' ? { height: proportionalHeight } : {}),
         ...(image && typeof image.align === 'number' ? { align: image.align } : {}),
-        ...(image && toObjectRecord(image.caption) ? { caption: toObjectRecord(image.caption) } : {}),
+        ...(image && toObjectRecord(image.caption)
+            ? { caption: toObjectRecord(image.caption) }
+            : {}),
         ...(image && typeof image.scale === 'number' ? { scale: image.scale } : {}),
     };
     let imageToken = await uploadBinaryToNode(client, 'docx_image', createdBlockId, localPath, authOptions, mediaLimiter);

package/dist/pipeline/hast-to-last.js

@@ -145,7 +145,7 @@ function createDividerBlock(ctx, parentId) {
     });
     return blockId;
 }
-function createImageBlock(ctx, parentId, sourceUrl) {
+function createImageBlock(ctx, parentId, sourceUrl, alt = null) {
     const blockId = nextBlockId(ctx);
     const parentBlock = ctx.blocks[parentId];
     const blockBase = {
@@ -155,8 +155,15 @@ function createImageBlock(ctx, parentId, sourceUrl) {
         children: [],
         payload: createDefaultImagePayload(parentBlock?.type),
     };
+    const selectorAttrs = {};
     if (sourceUrl) {
-        blockBase.selector = { attrs: { sourceUrl } };
+        selectorAttrs.sourceUrl = sourceUrl;
+    }
+    if (alt) {
+        selectorAttrs.alt = alt;
+    }
+    if (Object.keys(selectorAttrs).length > 0) {
+        blockBase.selector = { attrs: selectorAttrs };
     }
     addBlock(ctx, blockBase);
     return blockId;
@@ -358,14 +365,31 @@ function isWhitespaceTextNode(node) {
 function getMeaningfulChildren(nodes) {
     return nodes.filter((child) => !isWhitespaceTextNode(child));
 }
-function findStandaloneImageSrcInParagraph(paragraph) {
+function extractImageSourceFromElement(element) {
+    if (element.tagName === 'img') {
+        return {
+            sourceUrl: getStringProp(element, 'src'),
+            alt: getStringProp(element, 'alt'),
+        };
+    }
+    if (element.tagName !== 'a')
+        return null;
+    const meaningfulChildren = getMeaningfulChildren(getChildren(element));
+    if (meaningfulChildren.length !== 1)
+        return null;
+    const only = meaningfulChildren[0];
+    if (!only || !isElement(only))
+        return null;
+    return extractImageSourceFromElement(only);
+}
+function findStandaloneImageInParagraph(paragraph) {
     const meaningfulChildren = getMeaningfulChildren(getChildren(paragraph));
     if (meaningfulChildren.length !== 1)
         return null;
     const only = meaningfulChildren[0];
-    if (!only || !isElement(only) || only.tagName !== 'img')
+    if (!only || !isElement(only))
         return null;
-    return getStringProp(only, 'src');
+    return extractImageSourceFromElement(only);
 }
 function parseHttpUrl(url) {
     try {
@@ -472,10 +496,11 @@ function findStandaloneRichItemInTableCell(cell) {
     }
     if (!only || !isElement(only))
         return null;
-    if (only.tagName === 'img') {
+    const imageSource = extractImageSourceFromElement(only);
+    if (imageSource) {
         return {
             kind: 'image',
-            sourceUrl: getStringProp(only, 'src'),
+            ...imageSource,
         };
     }
     if (only.tagName !== 'a')
@@ -667,7 +692,7 @@ function convertTable(ctx, table, parentId) {
             const cellBlock = ctx.blocks[cellId];
             const richItem = cell ? findStandaloneRichItemInTableCell(cell) : null;
             if (cellBlock?.type === 'table_cell' && richItem?.kind === 'image') {
-                const imageId = createImageBlock(ctx, cellId, richItem.sourceUrl);
+                const imageId = createImageBlock(ctx, cellId, richItem.sourceUrl, richItem.alt);
                 cellBlock.children = [imageId];
                 cells.push(cellId);
                 continue;
@@ -733,6 +758,44 @@ function convertUnknownElement(ctx, element, parentId) {
         ]),
     ];
 }
+function hasNonWhitespaceInline(inlines) {
+    return inlines.some((inline) => toSearchText(inline).text.trim().length > 0);
+}
+function convertParagraph(ctx, paragraph, parentId) {
+    const standaloneImage = findStandaloneImageInParagraph(paragraph);
+    if (standaloneImage?.sourceUrl) {
+        return [createImageBlock(ctx, parentId, standaloneImage.sourceUrl, standaloneImage.alt)];
+    }
+    const standaloneIframe = findStandaloneIframePayloadInParagraph(paragraph);
+    if (standaloneIframe) {
+        return [createIframeBlock(ctx, parentId, standaloneIframe.url, standaloneIframe.iframeType)];
+    }
+    const ids = [];
+    let pendingInlineNodes = [];
+    const flushText = () => {
+        if (pendingInlineNodes.length === 0)
+            return;
+        const inlines = parseInlineNodes(ctx, pendingInlineNodes);
+        pendingInlineNodes = [];
+        if (!hasNonWhitespaceInline(inlines))
+            return;
+        ids.push(createTextualBlock(ctx, 'text', parentId, inlines));
+    };
+    for (const child of getChildren(paragraph)) {
+        const image = isElement(child) ? extractImageSourceFromElement(child) : null;
+        if (image?.sourceUrl) {
+            flushText();
+            ids.push(createImageBlock(ctx, parentId, image.sourceUrl, image.alt));
+            continue;
+        }
+        pendingInlineNodes.push(child);
+    }
+    flushText();
+    if (ids.length > 0) {
+        return ids;
+    }
+    return [createTextualBlock(ctx, 'text', parentId, parseInlineNodes(ctx, getChildren(paragraph)))];
+}
 function convertBlock(ctx, node, parentId) {
     if (isWhitespaceTextNode(node)) {
         return [];
@@ -753,17 +816,8 @@ function convertBlock(ctx, node, parentId) {
         return [];
     }
     switch (node.tagName) {
-        case 'p': {
-            const standaloneImageSrc = findStandaloneImageSrcInParagraph(node);
-            if (standaloneImageSrc) {
-                return [createImageBlock(ctx, parentId, standaloneImageSrc)];
-            }
-            const standaloneIframe = findStandaloneIframePayloadInParagraph(node);
-            if (standaloneIframe) {
-                return [createIframeBlock(ctx, parentId, standaloneIframe.url, standaloneIframe.iframeType)];
-            }
-            return [createTextualBlock(ctx, 'text', parentId, parseInlineNodes(ctx, getChildren(node)))];
-        }
+        case 'p':
+            return convertParagraph(ctx, node, parentId);
         case 'h1':
         case 'h2':
         case 'h3':
@@ -789,7 +843,7 @@ function convertBlock(ctx, node, parentId) {
         case 'table':
             return convertTable(ctx, node, parentId);
         case 'img':
-            return [createImageBlock(ctx, parentId, getStringProp(node, 'src'))];
+            return [createImageBlock(ctx, parentId, getStringProp(node, 'src'), getStringProp(node, 'alt'))];
         case 'br':
             return [
                 createTextualBlock(ctx, 'text', parentId, [

package/dist/pipeline/markdown/md-to-hast.js

@@ -4,12 +4,12 @@ import remarkGfm from 'remark-gfm';
 import remarkMath from 'remark-math';
 import remarkRehype from 'remark-rehype';
 import { normalizeMarkdownBeforeParse } from './normalize-markdown.js';
-export async function markdownToHast(markdown) {
+export async function markdownToHast(markdown, options = {}) {
     const content = normalizeMarkdownBeforeParse(markdown);
     const processor = unified()
         .use(remarkParse)
         .use(remarkGfm)
-        .use(remarkMath, { singleDollarTextMath: false })
+        .use(remarkMath, { singleDollarTextMath: options.singleDollarTextMath ?? false })
         .use(remarkRehype, { allowDangerousHtml: false });
     const mdast = processor.parse(content);
     const hast = await processor.run(mdast);

package/dist/preview/markdown-terminal.js

@@ -345,6 +345,7 @@ export function renderHASTToTerminal(root, options = {}) {
     return `${ctx.lines.join('\n')}\n`;
 }
 export async function renderMarkdownToTerminal(markdown, options = {}) {
-    const hast = await markdownToHast(markdown);
+    const parseOptions = options.singleDollarTextMath === undefined ? {} : { singleDollarTextMath: options.singleDollarTextMath };
+    const hast = await markdownToHast(markdown, parseOptions);
     return renderHASTToTerminal(hast, options);
 }

package/dist/publish/btt-patch.js

@@ -44,7 +44,9 @@ function applyRawBlockPatch(rawBlock, blockId, mermaidByBlockId, assetByBlockId,
         file.media_kind = asset.mediaKind;
     }
     if (typeof file.view_type !== 'number') {
-        file.view_type = shouldUsePreviewView(path.extname(asset.fileName).toLowerCase(), asset.mediaKind ?? 'file') ? 2 : 1;
+        file.view_type = shouldUsePreviewView(path.extname(asset.fileName).toLowerCase(), asset.mediaKind ?? 'file')
+            ? 2
+            : 1;
     }
     rawBlock.file = file;
 }

package/dist/publish/index.js

@@ -2,6 +2,6 @@ export { applyStandaloneAttachmentTransforms } from './asset-adapter.js';
 export { patchBTTForMermaidAndAssets } from './btt-patch.js';
 export { buildPipelineDocumentId } from './ids.js';
 export { applyTableColumnWidthHeuristics, collectMermaidPatches, ensureLastBlockBttIds } from './last-normalize.js';
-export { processSingleMarkdownFile } from './process-file.js';
-export { buildPublishRuntime, logPublishRuntimeSummary } from './runtime.js';
+export { processSingleMarkdownFile, } from './process-file.js';
+export { buildPublishRuntime, logPublishRuntimeSummary, } from './runtime.js';
 export { buildPipelineStagePaths, ensureDir, writeJson, writeSourceStage, writePrepareStage, writePrepareLogFile, writeHastStage, writeLastStage, writeBttStage, writePublishStageArtifact, } from './stage-cache.js';

package/dist/publish/process-file.js

@@ -2,7 +2,7 @@ import { readFile } from 'node:fs/promises';
 import path from 'node:path';
 import { convertLASTToBTT } from '../interop/index.js';
 import { clearDocumentContent, normalizeDocumentId } from '../lark/docx/ops.js';
-import { renderBTTToDocument, } from '../lark/docx/render-btt.js';
+import { renderBTTToDocument } from '../lark/docx/render-btt.js';
 import { hastToLAST, markdownToHast, prepareMarkdownBeforePublish } from '../pipeline/index.js';
 import { applySingleH1TitleRule, buildTitleForMarkdown } from '../commands/publish-md/title-policy.js';
 import { applyStandaloneAttachmentTransforms } from './asset-adapter.js';
@@ -95,7 +95,7 @@ export async function processSingleMarkdownFile(params) {
     markdown = prepareResult.preparedContent;
     await writePrepareStage(stagePaths, markdown, prepareResult);
     console.error(`[prepare ${index + 1}/${inputSet.markdownFiles.length}] rewritten=${prepareResult.rewrittenCount} downloaded=${prepareResult.downloadedCount} failed=${prepareResult.failedCount} log=${prepareResult.logFilePath}`);
-    const hast = await markdownToHast(markdown);
+    const hast = await markdownToHast(markdown, runtime.markdownParseConfig);
     await writeHastStage(stagePaths, hast);
     const h1RuleResult = options.title ? {} : applySingleH1TitleRule(hast);
     const title = buildTitleForMarkdown(markdownPath, inputSet, options.title, h1RuleResult.derivedTitle, {

package/dist/publish/runtime.js

@@ -99,6 +99,9 @@ export function buildPublishRuntime(options, env, markdownPresets) {
             ...(mermaidBoardDiagramType === undefined ? {} : { diagramType: mermaidBoardDiagramType }),
         },
     };
+    const markdownParseConfig = {
+        singleDollarTextMath: options.singleDollarTextMath ?? false,
+    };
     return {
         env,
         markdownPresets,
@@ -118,6 +121,7 @@ export function buildPublishRuntime(options, env, markdownPresets) {
         downloadRemoteImages,
         ...(ytDlpPath ? { ytDlpPath } : {}),
         mermaidRenderConfig,
+        markdownParseConfig,
         prepareConfig: {
             enabled: downloadRemoteImages,
             timeoutMs: prepareTimeoutMs,
@@ -138,6 +142,9 @@ export function logPublishRuntimeSummary(runtime, inputCount, inputMode) {
     console.error(runtime.mermaidRenderConfig.target === 'board'
         ? `Mermaid: target=board syntax_type=${String(runtime.mermaidRenderConfig.board.syntaxType)} style_type=${String(runtime.mermaidRenderConfig.board.styleType ?? '(default)')} diagram_type=${String(runtime.mermaidRenderConfig.board.diagramType ?? '(default)')}`
         : 'Mermaid: target=text-drawing');
+    if (runtime.markdownParseConfig.singleDollarTextMath) {
+        console.error('Markdown parse: single_dollar_text_math=true');
+    }
     console.error(`Document URL base: ${runtime.documentBaseUrl}`);
     if (runtime.resourceBaseDir) {
         console.error(`Local asset base override: ${runtime.resourceBaseDir}`);

package/dist/shared/image-metadata.js

@@ -1,9 +1,7 @@
 import { readFileSync } from 'node:fs';
 import path from 'node:path';
 const PNG_SIGNATURE = Buffer.from([0x89, 0x50, 0x4e, 0x47, 0x0d, 0x0a, 0x1a, 0x0a]);
-const JPEG_SOF_MARKERS = new Set([
-    0xc0, 0xc1, 0xc2, 0xc3, 0xc5, 0xc6, 0xc7, 0xc9, 0xca, 0xcb, 0xcd, 0xce, 0xcf,
-]);
+const JPEG_SOF_MARKERS = new Set([0xc0, 0xc1, 0xc2, 0xc3, 0xc5, 0xc6, 0xc7, 0xc9, 0xca, 0xcb, 0xcd, 0xce, 0xcf]);
 function isPositiveInteger(value) {
     return Number.isInteger(value) && value > 0;
 }
@@ -68,9 +66,7 @@ function readJpegDimensions(buffer) {
     return null;
 }
 function readWebpDimensions(buffer) {
-    if (buffer.length < 30 ||
-        buffer.toString('ascii', 0, 4) !== 'RIFF' ||
-        buffer.toString('ascii', 8, 12) !== 'WEBP') {
+    if (buffer.length < 30 || buffer.toString('ascii', 0, 4) !== 'RIFF' || buffer.toString('ascii', 8, 12) !== 'WEBP') {
         return null;
     }
     const chunkType = buffer.toString('ascii', 12, 16);

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@jacobbubu/md-to-lark",
-  "version": "1.4.11",
+  "version": "1.5.1",
   "description": "Publish Markdown to Feishu docs with a stable pipeline.",
   "main": "dist/index.js",
   "types": "dist/index.d.ts",