@bobfrankston/mailx-service 0.1.47

package/word-html.d.ts

@@ -0,0 +1,41 @@
+/**
+ * Word's own HTML export → a body fragment compose can take.
+ *
+ * 2026-09-13 20:10 EDT — Claude Code (Fable 5.1), at Bob's direction ("go
+ * ahead and build the export fragment route").
+ *
+ * The Edit-in-Word autosave sidecar calls Word's `Range.ExportFragment(path,
+ * wdFormatFilteredHTML)` over COM on every change, which writes Word's
+ * filtered-HTML rendering of the document WITHOUT changing the document's
+ * own file or format. That HTML keeps colours, fonts, alignment, spacing and
+ * paragraph borders as inline styles — the things mammoth's docx→HTML throws
+ * away (measured on a real reply 2026-09-13: mammoth returned only <p>, <a>
+ * and <br>). This module turns that export into a fragment: body content
+ * only, Word's class and comment cruft removed, sidecar image files inlined.
+ */
+/**
+ * Paragraph formatting that the generated .docx carries as its DOCUMENT
+ * DEFAULTS, as Word's export spells it.
+ *
+ * 2026-09-27 16:45 EDT — Claude Code (Fable 5.1), after Bob's report ("when I
+ * roundtrip through word I now see spurious style settings").
+ * html-to-docx hardcodes `<w:pPrDefault>` in the styles.xml it writes:
+ * `<w:spacing w:after="120" w:line="240" w:lineRule="atLeast"/>` — 6pt after,
+ * line height 12pt. Word's export states its own p.MsoNormal in a <style>
+ * block and puts the difference on every paragraph inline, so each paragraph
+ * came back as `style='margin-bottom:6.0pt;line-height:12.0pt'`. Nobody chose
+ * those values, and a fixed 12pt line height cramps any reader whose font is
+ * larger. They are the baseline of the document we handed to Word, not an edit.
+ *
+ * The values belong to html-to-docx (1.8.0), which has no option for them. If
+ * the library changes them the declarations reappear in compose; re-measure
+ * with tests/word-docx-probe.mjs.
+ */
+export declare const DOCX_BASELINE_PARAGRAPH_STYLE: Record<string, string>;
+/** Remove inline declarations that only restate the generated document's own
+ *  defaults. A value the user set in Word differs from the baseline and is
+ *  kept. A style attribute left with nothing in it is removed. */
+export declare function dropBaselineStyles(html: string): string;
+/** Read a Word filtered-HTML export and return body-content HTML for compose. */
+export declare function wordHtmlToFragment(htmPath: string): string;
+//# sourceMappingURL=word-html.d.ts.map

package/word-html.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"word-html.d.ts","sourceRoot":"","sources":["word-html.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;GAcG;AA8CH;;;;;;;;;;;;;;;;;GAiBG;AACH,eAAO,MAAM,6BAA6B,EAAE,MAAM,CAAC,MAAM,EAAE,MAAM,CAGhE,CAAC;AAEF;;kEAEkE;AAClE,wBAAgB,kBAAkB,CAAC,IAAI,EAAE,MAAM,GAAG,MAAM,CAcvD;AAED,iFAAiF;AACjF,wBAAgB,kBAAkB,CAAC,OAAO,EAAE,MAAM,GAAG,MAAM,CAiB1D"}

package/word-html.js

@@ -0,0 +1,122 @@
+/**
+ * Word's own HTML export → a body fragment compose can take.
+ *
+ * 2026-09-13 20:10 EDT — Claude Code (Fable 5.1), at Bob's direction ("go
+ * ahead and build the export fragment route").
+ *
+ * The Edit-in-Word autosave sidecar calls Word's `Range.ExportFragment(path,
+ * wdFormatFilteredHTML)` over COM on every change, which writes Word's
+ * filtered-HTML rendering of the document WITHOUT changing the document's
+ * own file or format. That HTML keeps colours, fonts, alignment, spacing and
+ * paragraph borders as inline styles — the things mammoth's docx→HTML throws
+ * away (measured on a real reply 2026-09-13: mammoth returned only <p>, <a>
+ * and <br>). This module turns that export into a fragment: body content
+ * only, Word's class and comment cruft removed, sidecar image files inlined.
+ */
+import fs from "node:fs";
+import path from "node:path";
+/** Decode Word's export honouring its own charset meta (utf-8 in every
+ *  export seen so far, but Word has written windows-1252 in the past). */
+function decodeWordHtml(bytes) {
+    const probe = bytes.subarray(0, 2048).toString("latin1");
+    const m = /charset=["']?([\w-]+)/i.exec(probe);
+    const charset = (m?.[1] || "utf-8").toLowerCase();
+    try {
+        return new TextDecoder(charset).decode(bytes);
+    }
+    catch (e) {
+        // An unknown label falls back to utf-8 rather than failing the
+        // mirror; the caller logs the fragment length either way.
+        console.warn(`  [word-edit] unknown charset "${charset}" in Word export — decoding as utf-8 (${e?.message || e})`);
+        return bytes.toString("utf-8");
+    }
+}
+const IMAGE_MIME = {
+    ".png": "image/png", ".jpg": "image/jpeg", ".jpeg": "image/jpeg", ".gif": "image/gif",
+    ".bmp": "image/bmp", ".webp": "image/webp", ".svg": "image/svg+xml", ".emf": "image/emf", ".wmf": "image/wmf",
+};
+/** Turn `<img src="x_files/image001.png">` into a data: URI. Word writes
+ *  images next to the export in a `<name>_files` folder; compose already
+ *  converts data: images to cid parts on send. A missing file is left as
+ *  is and logged — the text still mirrors. */
+function inlineImages(html, baseDir) {
+    return html.replace(/<img\b[^>]*\bsrc\s*=\s*["']?([^"'\s>]+)["']?[^>]*>/gi, (tag, src) => {
+        if (/^(data|https?|cid):/i.test(src))
+            return tag;
+        const file = path.resolve(baseDir, decodeURIComponent(src));
+        const mime = IMAGE_MIME[path.extname(file).toLowerCase()];
+        if (!mime)
+            return tag;
+        try {
+            const data = fs.readFileSync(file).toString("base64");
+            return tag.replace(src, `data:${mime};base64,${data}`);
+        }
+        catch (e) {
+            console.warn(`  [word-edit] image ${src} from the Word export could not be read: ${e?.message || e}`);
+            return tag;
+        }
+    });
+}
+/**
+ * Paragraph formatting that the generated .docx carries as its DOCUMENT
+ * DEFAULTS, as Word's export spells it.
+ *
+ * 2026-09-27 16:45 EDT — Claude Code (Fable 5.1), after Bob's report ("when I
+ * roundtrip through word I now see spurious style settings").
+ * html-to-docx hardcodes `<w:pPrDefault>` in the styles.xml it writes:
+ * `<w:spacing w:after="120" w:line="240" w:lineRule="atLeast"/>` — 6pt after,
+ * line height 12pt. Word's export states its own p.MsoNormal in a <style>
+ * block and puts the difference on every paragraph inline, so each paragraph
+ * came back as `style='margin-bottom:6.0pt;line-height:12.0pt'`. Nobody chose
+ * those values, and a fixed 12pt line height cramps any reader whose font is
+ * larger. They are the baseline of the document we handed to Word, not an edit.
+ *
+ * The values belong to html-to-docx (1.8.0), which has no option for them. If
+ * the library changes them the declarations reappear in compose; re-measure
+ * with tests/word-docx-probe.mjs.
+ */
+export const DOCX_BASELINE_PARAGRAPH_STYLE = {
+    "margin-bottom": "6.0pt",
+    "line-height": "12.0pt",
+};
+/** Remove inline declarations that only restate the generated document's own
+ *  defaults. A value the user set in Word differs from the baseline and is
+ *  kept. A style attribute left with nothing in it is removed. */
+export function dropBaselineStyles(html) {
+    return html.replace(/\sstyle\s*=\s*(["'])([\s\S]*?)\1/gi, (attr, quote, css) => {
+        const declarations = css.split(";").map(d => d.trim()).filter(d => d);
+        const kept = declarations.filter(d => {
+            const colon = d.indexOf(":");
+            if (colon < 0)
+                return true;
+            const property = d.slice(0, colon).trim().toLowerCase();
+            const value = d.slice(colon + 1).trim().toLowerCase();
+            return DOCX_BASELINE_PARAGRAPH_STYLE[property] !== value;
+        });
+        if (kept.length === declarations.length)
+            return attr;
+        if (kept.length === 0)
+            return "";
+        return ` style=${quote}${kept.join(";")}${quote}`;
+    });
+}
+/** Read a Word filtered-HTML export and return body-content HTML for compose. */
+export function wordHtmlToFragment(htmPath) {
+    const html = decodeWordHtml(fs.readFileSync(htmPath));
+    const bodyMatch = html.match(/<body[^>]*>([\s\S]*?)<\/body>/i);
+    let inner = bodyMatch ? bodyMatch[1] : html;
+    inner = inner
+        // Conditional comments around fake-list bullets, and any other comment.
+        .replace(/<!--[\s\S]*?-->/g, "")
+        // Word's paragraph/run classes; their styles live in the export's
+        // <style> block, which does not travel with the fragment.
+        .replace(/\s+class\s*=\s*["']?Mso[\w-]*["']?/gi, "")
+        // Office namespace tags (<o:p></o:p> paragraph marks) and their content-free shells.
+        .replace(/<\/?o:[^>]*>/gi, "")
+        // Word wraps the whole export in a section div; compose has no use for it.
+        .replace(/^\s*<div\s+class\s*=\s*["']?WordSection\d*["']?[^>]*>([\s\S]*)<\/div>\s*$/i, "$1");
+    inner = dropBaselineStyles(inner);
+    inner = inlineImages(inner, path.dirname(htmPath));
+    return inner.trim();
+}
+//# sourceMappingURL=word-html.js.map

package/word-html.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"word-html.js","sourceRoot":"","sources":["word-html.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;GAcG;AAEH,OAAO,EAAE,MAAM,SAAS,CAAC;AACzB,OAAO,IAAI,MAAM,WAAW,CAAC;AAE7B;0EAC0E;AAC1E,SAAS,cAAc,CAAC,KAAa;IACjC,MAAM,KAAK,GAAG,KAAK,CAAC,QAAQ,CAAC,CAAC,EAAE,IAAI,CAAC,CAAC,QAAQ,CAAC,QAAQ,CAAC,CAAC;IACzD,MAAM,CAAC,GAAG,wBAAwB,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;IAC/C,MAAM,OAAO,GAAG,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,IAAI,OAAO,CAAC,CAAC,WAAW,EAAE,CAAC;IAClD,IAAI,CAAC;QACD,OAAO,IAAI,WAAW,CAAC,OAAO,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC;IAClD,CAAC;IAAC,OAAO,CAAM,EAAE,CAAC;QACd,+DAA+D;QAC/D,0DAA0D;QAC1D,OAAO,CAAC,IAAI,CAAC,kCAAkC,OAAO,yCAAyC,CAAC,EAAE,OAAO,IAAI,CAAC,GAAG,CAAC,CAAC;QACnH,OAAO,KAAK,CAAC,QAAQ,CAAC,OAAO,CAAC,CAAC;IACnC,CAAC;AACL,CAAC;AAED,MAAM,UAAU,GAA2B;IACvC,MAAM,EAAE,WAAW,EAAE,MAAM,EAAE,YAAY,EAAE,OAAO,EAAE,YAAY,EAAE,MAAM,EAAE,WAAW;IACrF,MAAM,EAAE,WAAW,EAAE,OAAO,EAAE,YAAY,EAAE,MAAM,EAAE,eAAe,EAAE,MAAM,EAAE,WAAW,EAAE,MAAM,EAAE,WAAW;CAChH,CAAC;AAEF;;;8CAG8C;AAC9C,SAAS,YAAY,CAAC,IAAY,EAAE,OAAe;IAC/C,OAAO,IAAI,CAAC,OAAO,CAAC,sDAAsD,EAAE,CAAC,GAAG,EAAE,GAAW,EAAE,EAAE;QAC7F,IAAI,sBAAsB,CAAC,IAAI,CAAC,GAAG,CAAC;YAAE,OAAO,GAAG,CAAC;QACjD,MAAM,IAAI,GAAG,IAAI,CAAC,OAAO,CAAC,OAAO,EAAE,kBAAkB,CAAC,GAAG,CAAC,CAAC,CAAC;QAC5D,MAAM,IAAI,GAAG,UAAU,CAAC,IAAI,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC,WAAW,EAAE,CAAC,CAAC;QAC1D,IAAI,CAAC,IAAI;YAAE,OAAO,GAAG,CAAC;QACtB,IAAI,CAAC;YACD,MAAM,IAAI,GAAG,EAAE,CAAC,YAAY,CAAC,IAAI,CAAC,CAAC,QAAQ,CAAC,QAAQ,CAAC,CAAC;YACtD,OAAO,GAAG,CAAC,OAAO,CAAC,GAAG,EAAE,QAAQ,IAAI,WAAW,IAAI,EAAE,CAAC,CAAC;QAC3D,CAAC;QAAC,OAAO,CAAM,EAAE,CAAC;YACd,OAAO,CAAC,IAAI,CAAC,uBAAuB,GAAG,4CAA4C,CAAC,EAAE,OAAO,IAAI,CAAC,EAAE,CAAC,CAAC;YACtG,OAAO,GAAG,CAAC;QACf,CAAC;IACL,CAAC,CAAC,CAAC;AACP,CAAC;AAED;;;;;;;;;;;;;;;;;GAiBG;AACH,MAAM,CAAC,MAAM,6BAA6B,GAA2B;IACjE,eAAe,EAAE,OAAO;IACxB,aAAa,EAAE,QAAQ;CAC1B,CAAC;AAEF;;kEAEkE;AAClE,MAAM,UAAU,kBAAkB,CAAC,IAAY;IAC3C,OAAO,IAAI,CAAC,OAAO,CAAC,oCAAoC,EAAE,CAAC,IAAY,EAAE,KAAa,EAAE,GAAW,EAAE,EAAE;QACnG,MAAM,YAAY,GAAG,GAAG,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC;QACtE,MAAM,IAAI,GAAG,YAAY,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE;YACjC,MAAM,KAAK,GAAG,CAAC,CAAC,OAAO,CAAC,GAAG,CAAC,CAAC;YAC7B,IAAI,KAAK,GAAG,CAAC;gBAAE,OAAO,IAAI,CAAC;YAC3B,MAAM,QAAQ,GAAG,CAAC,CAAC,KAAK,CAAC,CAAC,EAAE,KAAK,CAAC,CAAC,IAAI,EAAE,CAAC,WAAW,EAAE,CAAC;YACxD,MAAM,KAAK,GAAG,CAAC,CAAC,KAAK,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,WAAW,EAAE,CAAC;YACtD,OAAO,6BAA6B,CAAC,QAAQ,CAAC,KAAK,KAAK,CAAC;QAC7D,CAAC,CAAC,CAAC;QACH,IAAI,IAAI,CAAC,MAAM,KAAK,YAAY,CAAC,MAAM;YAAE,OAAO,IAAI,CAAC;QACrD,IAAI,IAAI,CAAC,MAAM,KAAK,CAAC;YAAE,OAAO,EAAE,CAAC;QACjC,OAAO,UAAU,KAAK,GAAG,IAAI,CAAC,IAAI,CAAC,GAAG,CAAC,GAAG,KAAK,EAAE,CAAC;IACtD,CAAC,CAAC,CAAC;AACP,CAAC;AAED,iFAAiF;AACjF,MAAM,UAAU,kBAAkB,CAAC,OAAe;IAC9C,MAAM,IAAI,GAAG,cAAc,CAAC,EAAE,CAAC,YAAY,CAAC,OAAO,CAAC,CAAC,CAAC;IACtD,MAAM,SAAS,GAAG,IAAI,CAAC,KAAK,CAAC,gCAAgC,CAAC,CAAC;IAC/D,IAAI,KAAK,GAAG,SAAS,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC;IAC5C,KAAK,GAAG,KAAK;QACT,wEAAwE;SACvE,OAAO,CAAC,kBAAkB,EAAE,EAAE,CAAC;QAChC,kEAAkE;QAClE,0DAA0D;SACzD,OAAO,CAAC,sCAAsC,EAAE,EAAE,CAAC;QACpD,qFAAqF;SACpF,OAAO,CAAC,gBAAgB,EAAE,EAAE,CAAC;QAC9B,2EAA2E;SAC1E,OAAO,CAAC,4EAA4E,EAAE,IAAI,CAAC,CAAC;IACjG,KAAK,GAAG,kBAAkB,CAAC,KAAK,CAAC,CAAC;IAClC,KAAK,GAAG,YAAY,CAAC,KAAK,EAAE,IAAI,CAAC,OAAO,CAAC,OAAO,CAAC,CAAC,CAAC;IACnD,OAAO,KAAK,CAAC,IAAI,EAAE,CAAC;AACxB,CAAC"}