dsh-plugin-lookatstudy 0.11.1 → 0.12.1

package/lib/{pptx-parser-CD5pR2cj.mjs → pptx-parser-B9IDkiGM.mjs}

@@ -21,6 +21,49 @@ function shapeTexts(xml) {
 function slideNumber(name) {
 	return Number(name.match(/slide(\d+)\.xml$/i)?.[1] ?? 0);
 }
+/**
+* One `a:tbl` block → GFM markdown table, or "" when the table is an
+* all-empty placeholder. Cells: text via a:p runs; gridSpan advances the
+* column index (spanned columns padded); vMerge continuation cells carry no
+* text of their own. Pipes escaped, inner whitespace collapsed.
+*/
+function tableToMarkdown(tblXml) {
+	const rows = [];
+	const trRe = /<a:tr\b[^>]*>([\s\S]*?)<\/a:tr>/g;
+	let trm;
+	while ((trm = trRe.exec(tblXml)) !== null) {
+		const cells = [];
+		const tcRe = /<a:tc\b([^>]*)>([\s\S]*?)<\/a:tc>/g;
+		let col = 0;
+		let tcm;
+		while ((tcm = tcRe.exec(trm[1])) !== null) {
+			const attrs = tcm[1] ?? "";
+			const span = Math.max(1, Number(attrs.match(/gridSpan="(\d+)"/)?.[1] ?? 1));
+			const vMerge = /vMerge="1"/.test(attrs);
+			const raw = shapeTexts(tcm[2] ?? "").join(" ").trim().replace(/\|/g, "\\|").replace(/\s+/g, " ");
+			if (!vMerge) cells.push({
+				col,
+				text: raw
+			});
+			col += span;
+		}
+		rows.push(cells);
+	}
+	if (rows.filter((cs) => cs.length > 0).length === 0) return "";
+	if (rows.every((cs) => cs.every((c) => !c.text))) return "";
+	const sorted = rows.filter((cs) => cs.length > 0).map((cs) => cs.sort((a, b) => a.col - b.col));
+	const width = Math.max(...sorted.map((cs) => cs[cs.length - 1].col + 1));
+	const line = (cs) => {
+		const texts = Array.from({ length: width }, () => "");
+		for (const c of cs) texts[c.col] = c.text;
+		return `| ${texts.join(" | ")} |`;
+	};
+	return [
+		line(sorted[0]),
+		`| ${Array.from({ length: width }, () => "---").join(" | ")} |`,
+		...sorted.slice(1).map(line)
+	].join("\n");
+}
 /** .pptx → markdown(每张 slide 一个 ##)。失败抛错,调用方按"无内容"兜底。 */
 function parsePptx(buf) {
 	const entries = readZip(buf);
@@ -30,9 +73,14 @@ function parsePptx(buf) {
 	for (const name of slideNames) {
 		const xml = readZipText(entries, name);
 		const no = slideNumber(name);
-		const texts = shapeTexts(xml);
+		const tables = [];
+		const texts = shapeTexts(xml.replace(/<a:tbl\b[\s\S]*?<\/a:tbl>/g, (tbl) => {
+			const md = tableToMarkdown(tbl);
+			if (md) tables.push(md);
+			return "";
+		}));
 		const title = texts[0] ?? "";
-		const body = texts.slice(1);
+		const body = [...texts.slice(1), ...tables];
 		lines.push(`\n## Slide ${no}: ${title || "(无标题)"}\n`);
 		if (body.length) lines.push(body.join("\n\n"));
 		const notes = shapeTexts(readZipText(entries, `ppt/notesSlides/notesSlide${no}.xml`));

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "dsh-plugin-lookatstudy",
-  "version": "0.11.1",
+  "version": "0.12.1",
   "packageManager": "pnpm@11.7.0",
   "description": "Turn any markdown, local folder, or GitHub learning repo into a guided course inside DeepSeek Harness: gated skill-tree progression, BKT mastery tracking, SM-2 spaced repetition.",
   "type": "module",

package/lib/epub-parser-DvlKap-d.mjs

@@ -1,112 +0,0 @@
-import { n as htmlToMarkdown } from "./html-article-C6nhqJbX.mjs";
-import { n as readZipText, t as readZip } from "./zip-reader-KnRrq0av.mjs";
-//#region src/vendor/epub-parser.ts
-/** 提取某标签的全部出现(OPF 的属性顺序不定,先抓整标签再逐个提属性)。 */
-function tags(xml, tagName) {
-	const out = [];
-	const re = new RegExp(`<${tagName}\\b[^>]*>`, "g");
-	let m;
-	while ((m = re.exec(xml)) !== null) {
-		const raw = m[0];
-		const attrs = {};
-		const attrRe = /([\w:-]+)\s*=\s*"([^"]*)"/g;
-		let a;
-		while ((a = attrRe.exec(raw)) !== null) attrs[a[1]] = a[2];
-		out.push({
-			raw,
-			attrs
-		});
-	}
-	return out;
-}
-function firstTagText(xml, tagName) {
-	return xml.match(new RegExp(`<${tagName}[^>]*>([^<]*)</${tagName}>`, "i"))?.[1]?.trim() ?? "";
-}
-/** zip 内路径归一:posix 分隔 + href 相对 OPF 目录解析。 */
-function resolveZipPath(opfPath, href) {
-	const cleanHref = decodeURIComponent(href.split("#")[0] ?? href).replace(/\\/g, "/");
-	if (!opfPath.includes("/")) return cleanHref.replace(/^\.\//, "");
-	const parts = `${opfPath.slice(0, opfPath.lastIndexOf("/"))}/${cleanHref}`.split("/");
-	const resolved = [];
-	for (const p of parts) if (p === "..") resolved.pop();
-	else if (p !== "." && p !== "") resolved.push(p);
-	return resolved.join("/");
-}
-/** EPUB3 nav.xhtml 或 EPUB2 toc.ncx → href(去 fragment)→ 章节标题。 */
-function parseTocLabels(tocXml, isNcx) {
-	const labels = /* @__PURE__ */ new Map();
-	if (isNcx) {
-		const blockRe = /<navPoint\b[\s\S]*?<\/navPoint>/g;
-		let m;
-		while ((m = blockRe.exec(tocXml)) !== null) {
-			const label = firstTagText(m[0], "text");
-			const src = m[0].match(/<content[^>]+src="([^"]+)"/i)?.[1];
-			if (label && src) labels.set(decodeURIComponent(src.split("#")[0].replace(/\\/g, "/")), label);
-		}
-	} else {
-		const aRe = /<a\b[^>]*href="([^"]+)"[^>]*>([\s\S]*?)<\/a>/gi;
-		let m;
-		while ((m = aRe.exec(tocXml)) !== null) {
-			const label = (m[2] ?? "").replace(/<[^>]+>/g, "").trim();
-			if (label) labels.set(decodeURIComponent((m[1] ?? "").split("#")[0].replace(/\\/g, "/")), label);
-		}
-	}
-	return labels;
-}
-function sanitizeFileName(title) {
-	return (title || "chapter").replace(/[\\/:*?"<>|#\s]+/g, "-").replace(/^-+|-+$/g, "").slice(0, 50) || "chapter";
-}
-function parseEpub(buf) {
-	const entries = readZip(buf);
-	const read = (p) => readZipText(entries, p);
-	const opfPath = read("META-INF/container.xml").match(/full-path="([^"]+)"/i)?.[1];
-	if (!opfPath) throw new Error("epub 结构异常:找不到 container.xml 里的 OPF 路径");
-	const opf = read(opfPath);
-	if (!opf) throw new Error(`epub 结构异常:OPF 文件缺失(${opfPath})`);
-	const bookTitle = firstTagText(opf, "dc:title") || "未命名电子书";
-	const manifest = /* @__PURE__ */ new Map();
-	for (const t of tags(opf, "item")) manifest.set(t.attrs["id"] ?? "", {
-		href: t.attrs["href"] ?? "",
-		mediaType: (t.attrs["media-type"] ?? "").toLowerCase(),
-		properties: t.attrs["properties"] ?? ""
-	});
-	const spineIds = tags(opf, "itemref").map((t) => t.attrs["idref"] ?? "").filter(Boolean);
-	const spineTocId = opf.match(/<spine\b[^>]*\btoc="([^"]+)"/i)?.[1];
-	let tocLabels = /* @__PURE__ */ new Map();
-	const navItem = [...manifest.values()].find((it) => it.properties.split(/\s+/).includes("nav"));
-	if (navItem?.href) tocLabels = parseTocLabels(read(resolveZipPath(opfPath, navItem.href)), false);
-	if (tocLabels.size === 0 && spineTocId && manifest.has(spineTocId)) tocLabels = parseTocLabels(read(resolveZipPath(opfPath, manifest.get(spineTocId).href)), true);
-	const opfDirKey = (href) => decodeURIComponent(href.split("#")[0] ?? href).replace(/\\/g, "/");
-	const chapters = [];
-	let n = 0;
-	for (const id of spineIds) {
-		const item = manifest.get(id);
-		if (!item?.href) continue;
-		if (!(item.mediaType === "application/xhtml+xml" || item.mediaType === "text/html" || /\.(xhtml|html|htm)$/i.test(item.href))) continue;
-		if (item.properties.split(/\s+/).includes("nav")) continue;
-		const xhtml = read(resolveZipPath(opfPath, item.href));
-		if (!xhtml) continue;
-		const md = htmlToMarkdown(xhtml, { stripImages: true });
-		if (!md || md.replace(/[#\s>*-]/g, "").length < 8) continue;
-		n++;
-		const body = md.startsWith("# ") && md.includes("\n") ? md.slice(md.indexOf("\n") + 1).trim() : md;
-		const firstHeading = md.startsWith("# ") ? md.split("\n")[0].slice(2).trim() : "";
-		const title = tocLabels.get(opfDirKey(item.href)) || firstHeading || `第 ${n} 章`;
-		chapters.push({
-			path: `chapters/${String(n).padStart(2, "0")}-${sanitizeFileName(title)}.md`,
-			title,
-			markdown: `# ${title}\n\n${body}`
-		});
-	}
-	if (chapters.length === 0) throw new Error("epub 里没有可识别的章节文本");
-	return {
-		title: bookTitle,
-		chapters
-	};
-}
-/** 文件夹导入路径用:整本书压平成一个 markdown(全部 H1 降为 H2,给结构设计当 anchor 拆章)。 */
-function parseEpubFlat(buf) {
-	return parseEpub(buf).chapters.map((c) => c.markdown.replace(/^# /gm, "## ")).join("\n\n");
-}
-//#endregion
-export { parseEpubFlat };