@file-viewer/doc 2.2.4 → 2.2.6

package/dist/msdoc/constants.d.ts

@@ -2,6 +2,7 @@ export declare const DOC_CONTROL: {
     paragraph: string;
     cellMark: string;
     picture: string;
+    textBoxPlaceholder: string;
     fieldStart: string;
     fieldSeparator: string;
     fieldEnd: string;

package/dist/msdoc/constants.js

@@ -2,6 +2,7 @@ export const DOC_CONTROL = {
     paragraph: '\u000d',
     cellMark: '\u0007',
     picture: '\u0001',
+    textBoxPlaceholder: '\u0008',
     fieldStart: '\u0013',
     fieldSeparator: '\u0014',
     fieldEnd: '\u0015',

package/dist/msdoc/parser.js

@@ -123,13 +123,68 @@ function normalizePlainTextChar(ch) {
         case DOC_CONTROL.nonBreakingHyphen:
             return '-';
         case DOC_CONTROL.nonRequiredHyphen:
-            return '';
         case DOC_CONTROL.annotationRef:
+        case DOC_CONTROL.textBoxPlaceholder:
             return '';
         default:
             return ch;
     }
 }
+function normalizeParagraphText(text) {
+    return text.replaceAll(DOC_CONTROL.textBoxPlaceholder, '');
+}
+function getDocumentStoryRanges(fib, documentLength) {
+    const storyLengths = [
+        ['main', fib.fibRgLw.ccpText],
+        ['footnote', fib.fibRgLw.ccpFtn],
+        ['header', fib.fibRgLw.ccpHdd],
+        ['macro', fib.fibRgLw.ccpMcr],
+        ['annotation', fib.fibRgLw.ccpAtn],
+        ['endnote', fib.fibRgLw.ccpEdn],
+        ['textbox', fib.fibRgLw.ccpTxbx],
+        ['header-textbox', fib.fibRgLw.ccpHdrTxbx],
+    ];
+    const hasDeclaredStories = storyLengths.some(([, length]) => length > 0);
+    if (!hasDeclaredStories)
+        return [{ kind: 'main', cpStart: 0, cpEnd: documentLength }];
+    const ranges = [];
+    let cp = 0;
+    for (const [kind, rawLength] of storyLengths) {
+        const length = Math.max(0, rawLength || 0);
+        const cpStart = Math.min(cp, documentLength);
+        cp += length;
+        const cpEnd = Math.min(cp, documentLength);
+        ranges.push({ kind, cpStart, cpEnd });
+    }
+    return ranges;
+}
+function buildStoryParagraphRanges(story, documentText, papxRuns) {
+    if (story.cpEnd <= story.cpStart)
+        return [];
+    const runs = papxRuns.filter((run) => run.cpStart < story.cpEnd && run.cpEnd > story.cpStart);
+    if (runs.length) {
+        return runs.map((run) => {
+            const cpStart = Math.max(story.cpStart, run.cpStart);
+            const cpEnd = Math.min(story.cpEnd, run.cpEnd);
+            return {
+                cpStart,
+                cpEnd,
+                terminator: documentText[cpEnd - 1] || '',
+                styleId: run.styleId,
+                properties: run.properties,
+                storyKind: story.kind,
+            };
+        }).filter((range) => range.cpEnd > range.cpStart);
+    }
+    return splitParagraphRanges(documentText.slice(story.cpStart, story.cpEnd)).map((range) => ({
+        ...range,
+        cpStart: range.cpStart + story.cpStart,
+        cpEnd: range.cpEnd + story.cpStart,
+        styleId: 0,
+        properties: [],
+        storyKind: story.kind,
+    }));
+}
 function buildInlineNodes(segments, resolveAsset) {
     const output = [];
     const fieldStack = [];
@@ -207,7 +262,7 @@ function buildInlineNodes(segments, resolveAsset) {
     }
     return output;
 }
-function buildCharSegments(range, paragraphText, chpxRuns, styles, baseCharProps, resolveFont, cursorRef) {
+function buildCharSegments(range, paragraphText, chpxRuns, styles, baseCharProps, resolveFont, cursorRef, forceReadableColors = false) {
     const overlaps = getOverlappingRuns(chpxRuns, range.cpStart, range.cpEnd, cursorRef);
     const boundaries = new Set([range.cpStart, range.cpEnd]);
     for (const run of overlaps) {
@@ -227,6 +282,8 @@ function buildCharSegments(range, paragraphText, chpxRuns, styles, baseCharProps
         const charStyleProps = directState.charStyleId != null ? styles.resolveStyle(directState.charStyleId).charProps : [];
         const finalProps = mergePropertyArrays(baseCharProps, charStyleProps, directProps);
         const state = charPropsToState(finalProps);
+        if (forceReadableColors && state.colorIndex === 8)
+            state.colorIndex = 1;
         const font = resolveFont(state.fontFamilyId);
         if (font)
             state.fontFamily = font.name || font.altName || undefined;
@@ -245,6 +302,8 @@ function buildCharSegments(range, paragraphText, chpxRuns, styles, baseCharProps
     }
     if (!segments.length && paragraphText) {
         const state = charPropsToState(baseCharProps);
+        if (forceReadableColors && state.colorIndex === 8)
+            state.colorIndex = 1;
         const font = resolveFont(state.fontFamilyId);
         if (font)
             state.fontFamily = font.name || font.altName || undefined;
@@ -264,7 +323,7 @@ function buildParagraphModel(range, paragraphText, styles, fonts, chpxRuns, reso
     const tableState = tablePropsToState(tableProps);
     const baseCharProps = paraStyle.charProps;
     const resolveFont = (fontId) => fonts.byIndex(fontId);
-    const segments = buildCharSegments(range, paragraphText, chpxRuns, styles, baseCharProps, resolveFont, chpxCursor);
+    const segments = buildCharSegments(range, paragraphText, chpxRuns, styles, baseCharProps, resolveFont, chpxCursor, range.storyKind === 'textbox' || range.storyKind === 'header-textbox');
     const inlines = buildInlineNodes(segments, resolveAsset);
     return {
         id: uniqueId('para'),
@@ -601,35 +660,41 @@ export function parseMsDoc(input, options = {}) {
     const clx = parseClx(tableBytes, fib.fibRgFcLcb);
     const pieceTexts = buildPieceTextCache(wordBytes, clx);
     const documentText = pieceTexts.join('');
-    const mainStoryEnd = fib.fibRgLw.ccpText > 0 ? fib.fibRgLw.ccpText : documentText.length;
+    const stories = getDocumentStoryRanges(fib, documentText.length);
+    const mainStory = stories.find((story) => story.kind === 'main') || { kind: 'main', cpStart: 0, cpEnd: documentText.length };
+    const supplementalTextStories = stories.filter((story) => ((story.kind === 'textbox' || story.kind === 'header-textbox')
+        && normalizeParagraphText(documentText.slice(story.cpStart, story.cpEnd)).replace(/[\r\u0007]/g, '').trim().length > 0));
     const styles = parseStyles(tableBytes, fib.fibRgFcLcb);
     const fonts = parseFonts(tableBytes, fib.fibRgFcLcb);
-    const chpxRuns = readChpxRuns(wordBytes, tableBytes, fib, clx).filter((run) => run.cpStart < mainStoryEnd);
+    const relevantStoryEnd = Math.max(mainStory.cpEnd, ...supplementalTextStories.map((story) => story.cpEnd));
+    const chpxRuns = readChpxRuns(wordBytes, tableBytes, fib, clx)
+        .filter((run) => run.cpStart < relevantStoryEnd)
+        .map((run) => ({ ...run, cpEnd: Math.min(run.cpEnd, relevantStoryEnd) }));
     const papxRuns = readPapxRuns(wordBytes, tableBytes, fib, clx, dataBytes)
-        .filter((run) => run.cpStart < mainStoryEnd)
-        .map((run) => ({ ...run, cpEnd: Math.min(run.cpEnd, mainStoryEnd) }));
-    const ranges = papxRuns.length
-        ? papxRuns.map((run) => ({
-            cpStart: run.cpStart,
-            cpEnd: run.cpEnd,
-            terminator: documentText[run.cpEnd - 1] || '',
-            styleId: run.styleId,
-            properties: run.properties,
-        }))
-        : splitParagraphRanges(documentText.slice(0, mainStoryEnd)).map((range) => ({ ...range, styleId: 0, properties: [] }));
+        .filter((run) => run.cpStart < relevantStoryEnd)
+        .map((run) => ({ ...run, cpEnd: Math.min(run.cpEnd, relevantStoryEnd) }));
+    const ranges = [mainStory, ...supplementalTextStories]
+        .flatMap((story) => buildStoryParagraphRanges(story, documentText, papxRuns));
+    if (supplementalTextStories.length) {
+        pushWarning(warnings, 'Recovered text from a legacy Word text-box story; text-box positioning is unavailable, so the content was linearized after the main story.', { code: 'MSDOC_TEXTBOX_STORY_LINEARIZED', storyCount: supplementalTextStories.length });
+    }
     const objectPool = extractObjectPool(cfb);
     const assets = [];
     const usedAttachmentNames = new Set();
     const assetCache = new Map();
     const resolveAsset = createAssetResolver(dataBytes, objectPool, assets, usedAttachmentNames, assetCache, options);
     const chpxCursor = { index: 0 };
-    const paragraphs = ranges.map((range) => {
+    const paragraphs = ranges.flatMap((range) => {
         const rawParagraphText = getTextByCp(wordBytes, clx, pieceTexts, range.cpStart, range.cpEnd);
         const terminator = range.terminator === DOC_CONTROL.paragraph || range.terminator === DOC_CONTROL.cellMark ? range.terminator : '';
-        const paragraphText = terminator && rawParagraphText.endsWith(terminator)
+        const rawContent = terminator && rawParagraphText.endsWith(terminator)
             ? rawParagraphText.slice(0, -1)
             : rawParagraphText;
-        return buildParagraphModel({ ...range, terminator }, paragraphText, styles, fonts, chpxRuns, resolveAsset, chpxCursor);
+        const paragraphText = normalizeParagraphText(rawContent);
+        const isTextBoxStory = range.storyKind === 'textbox' || range.storyKind === 'header-textbox';
+        if ((rawContent.includes(DOC_CONTROL.textBoxPlaceholder) || isTextBoxStory) && !paragraphText.trim())
+            return [];
+        return [buildParagraphModel({ ...range, terminator }, paragraphText, styles, fonts, chpxRuns, resolveAsset, chpxCursor)];
     });
     normalizeRowEndOnlyTables(paragraphs);
     const blocks = buildBlocks(paragraphs);

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@file-viewer/doc",
-  "version": "2.2.4",
+  "version": "2.2.6",
   "private": false,
   "type": "module",
   "description": "Browser-native legacy MS-DOC (.doc) parser and HTML renderer extracted from the public msdoc-viewer 0.2.0 line.",
@@ -70,6 +70,7 @@
     "type-check": "tsc -b tsconfig.json",
     "verify:table-spec": "pnpm build && node scripts/verify-table-spec.mjs",
     "verify:github-34": "pnpm build && node scripts/verify-github-34.mjs",
-    "verify:github-87": "pnpm build && node scripts/verify-github-87.mjs"
+    "verify:github-87": "pnpm build && node scripts/verify-github-87.mjs",
+    "verify:github-171": "pnpm build && node scripts/verify-github-171.mjs"
   }
 }