dsh-plugin-lookatstudy 0.8.1

package/lib/notebook-parser-ChbZBIKJ.mjs

@@ -0,0 +1,104 @@
+//#region src/vendor/notebook-parser.ts
+/** 从 kernelspec 推断代码块语言标识(用于 markdown ``` 语法高亮) */
+function inferLanguage(metadata) {
+	if (metadata.kernelspec?.language) return metadata.kernelspec.language.toLowerCase();
+	if (metadata.language_info?.name) return metadata.language_info.name.toLowerCase();
+	const name = metadata.kernelspec?.name ?? "";
+	if (name.includes("python")) return "python";
+	if (name.includes("ir") || name.includes("r-")) return "r";
+	if (name.includes("julia")) return "julia";
+	return "python";
+}
+/**
+* 把 source 字段(string 或 string[])统一成 string。
+* nbformat 的 source 可以是数组(每行一个元素)或整体字符串。
+*/
+function normalizeSource(source) {
+	if (typeof source === "string") return source;
+	if (Array.isArray(source)) return source.join("");
+	return "";
+}
+/**
+* 从 code cell 的 outputs 里提取图片(base64)。
+* nbformat output 结构:
+*   { output_type: "execute_result"|"display_data", data: { "image/png": "base64..." } }
+*   { output_type: "stream", text: "..." } — 纯文本流,不含图
+*   { output_type: "error", ename, evalue } — 错误,不含图
+*/
+function extractOutputImages(outputs, cellIndex) {
+	if (!Array.isArray(outputs)) return [];
+	const images = [];
+	for (const output of outputs) {
+		if (!output || typeof output !== "object") continue;
+		const data = output.data;
+		if (!data) continue;
+		const png = data["image/png"];
+		if (typeof png === "string" && png.length > 0) images.push({
+			base64: png,
+			mimeType: "image/png",
+			source: "notebook_output",
+			cellIndex,
+			altText: `Notebook cell ${cellIndex} 输出图`
+		});
+		const jpeg = data["image/jpeg"];
+		if (typeof jpeg === "string" && jpeg.length > 0) images.push({
+			base64: jpeg,
+			mimeType: "image/jpeg",
+			source: "notebook_output",
+			cellIndex,
+			altText: `Notebook cell ${cellIndex} 输出图`
+		});
+	}
+	return images;
+}
+/**
+* 解析一个 .ipynb JSON 字符串,转成 markdown + 图片。
+*
+* @param jsonText .ipynb 文件的原始 JSON 字符串
+* @returns { markdown, images, language, stats }
+*
+* 纯函数,不依赖 IO,可直接测。
+*/
+function parseNotebook(jsonText) {
+	let nb;
+	try {
+		nb = JSON.parse(jsonText);
+	} catch {
+		throw new Error("parseNotebook: 无效的 JSON(不是合法的 .ipynb 文件)");
+	}
+	const cells = Array.isArray(nb.cells) ? nb.cells : [];
+	const language = inferLanguage(nb.metadata ?? {});
+	const mdParts = [];
+	const images = [];
+	let markdownCells = 0;
+	let codeCells = 0;
+	for (let i = 0; i < cells.length; i++) {
+		const cell = cells[i];
+		if (!cell) continue;
+		const cellType = cell.cell_type;
+		const source = normalizeSource(cell.source);
+		if (cellType === "markdown") {
+			markdownCells++;
+			mdParts.push(source);
+		} else if (cellType === "code") {
+			codeCells++;
+			mdParts.push(`\`\`\`${language}\n${source}\n\`\`\``);
+			const cellImages = extractOutputImages(cell.outputs, i);
+			images.push(...cellImages);
+		}
+	}
+	return {
+		markdown: mdParts.join("\n\n"),
+		images,
+		language,
+		stats: {
+			markdownCells,
+			codeCells,
+			totalCells: cells.length
+		}
+	};
+}
+//#endregion
+export { parseNotebook };
+
+//# sourceMappingURL=notebook-parser-ChbZBIKJ.mjs.map

package/lib/notebook-parser-ChbZBIKJ.mjs.map

@@ -0,0 +1 @@
+{"version":3,"file":"notebook-parser-ChbZBIKJ.mjs","names":["o"],"sources":["../src/vendor/notebook-parser.ts"],"sourcesContent":["// Vendored from LookatStudy src/main/services/pure/notebook-parser.ts (MIT License, https://github.com/kaiji/LookatStudy).\n// Unmodified except this provenance header. PDF/PPTX branches resolve unavailable optional libs and are skipped per upstream try/catch.\n/**\r\n * Jupyter Notebook (.ipynb) 解析器 —— 把 JSON 格式的 notebook 转成 markdown + 图片。\r\n *\r\n * nbformat 4 结构:\r\n *   {\r\n *     cells: [\r\n *       { cell_type: \"markdown\", source: string | string[] },\r\n *       { cell_type: \"code\", source: string | string[], outputs: [\r\n *         { data: { \"image/png\": \"base64...\", \"text/plain\": \"...\" } }\r\n *       ] }\r\n *     ],\r\n *     metadata: { kernelspec: { display_name, name, language } }\r\n *   }\r\n *\r\n * 转换策略:\r\n *   - markdown cell → 原样输出(保持 Markdown 格式)\r\n *   - code cell → ```{language}\\n{code}\\n``` 代码块\r\n *   - code cell 的 image output → 提取为 ScannedImage(buffer 型,source=\"notebook_output\")\r\n *   - code cell 的 text output → 附在代码块后作为注释(可选)\r\n *\r\n * 纯函数设计,便于 verify 脚本测(JSON 解构 + cell 分拣,不依赖 IO)。\r\n */\r\n\r\n/** 解析出的图片(notebook output 或 markdown 引用) */\r\nexport interface NotebookImage {\r\n  /** base64 数据(不含 data: 前缀) */\r\n  base64: string;\r\n  mimeType: string;\r\n  /** 来源类型 */\r\n  source: \"notebook_output\" | \"markdown_ref\";\r\n  /** cell 索引(用于溯源) */\r\n  cellIndex: number;\r\n  /** alt/描述 */\r\n  altText: string;\r\n}\r\n\r\n/** notebook 解析结果 */\r\nexport interface ParsedNotebook {\r\n  /** 转换后的 markdown 文本(markdown cell + code block) */\r\n  markdown: string;\r\n  /** 提取的图片(output 内嵌图) */\r\n  images: NotebookImage[];\r\n  /** 语言(Python / R / Julia 等,从 kernelspec 推断) */\r\n  language: string;\r\n  /** cell 统计 */\r\n  stats: {\r\n    markdownCells: number;\r\n    codeCells: number;\r\n    totalCells: number;\r\n  };\r\n}\r\n\r\n/** 从 kernelspec 推断代码块语言标识(用于 markdown ``` 语法高亮) */\r\nexport function inferLanguage(metadata: {\r\n  kernelspec?: { name?: string; display_name?: string; language?: string };\r\n  language_info?: { name?: string };\r\n}): string {\r\n  // kernelspec.language 优先\r\n  if (metadata.kernelspec?.language) return metadata.kernelspec.language.toLowerCase();\r\n  // language_info.name\r\n  if (metadata.language_info?.name) return metadata.language_info.name.toLowerCase();\r\n  // kernelspec.name 推断\r\n  const name = metadata.kernelspec?.name ?? \"\";\r\n  if (name.includes(\"python\")) return \"python\";\r\n  if (name.includes(\"ir\") || name.includes(\"r-\")) return \"r\";\r\n  if (name.includes(\"julia\")) return \"julia\";\r\n  return \"python\"; // 默认 python(Jupyter 最常见)\r\n}\r\n\r\n/**\r\n * 把 source 字段(string 或 string[])统一成 string。\r\n * nbformat 的 source 可以是数组(每行一个元素)或整体字符串。\r\n */\r\nexport function normalizeSource(source: unknown): string {\r\n  if (typeof source === \"string\") return source;\r\n  if (Array.isArray(source)) return source.join(\"\");\r\n  return \"\";\r\n}\r\n\r\n/**\r\n * 从 code cell 的 outputs 里提取图片(base64)。\r\n * nbformat output 结构:\r\n *   { output_type: \"execute_result\"|\"display_data\", data: { \"image/png\": \"base64...\" } }\r\n *   { output_type: \"stream\", text: \"...\" } — 纯文本流,不含图\r\n *   { output_type: \"error\", ename, evalue } — 错误,不含图\r\n */\r\nexport function extractOutputImages(\r\n  outputs: unknown[] | undefined,\r\n  cellIndex: number,\r\n): NotebookImage[] {\r\n  if (!Array.isArray(outputs)) return [];\r\n  const images: NotebookImage[] = [];\r\n  for (const output of outputs) {\r\n    if (!output || typeof output !== \"object\") continue;\r\n    const o = output as Record<string, unknown>;\r\n    const data = o.data as Record<string, unknown> | undefined;\r\n    if (!data) continue;\r\n    // image/png\r\n    const png = data[\"image/png\"];\r\n    if (typeof png === \"string\" && png.length > 0) {\r\n      images.push({\r\n        base64: png,\r\n        mimeType: \"image/png\",\r\n        source: \"notebook_output\",\r\n        cellIndex,\r\n        altText: `Notebook cell ${cellIndex} 输出图`,\r\n      });\r\n    }\r\n    // image/jpeg\r\n    const jpeg = data[\"image/jpeg\"];\r\n    if (typeof jpeg === \"string\" && jpeg.length > 0) {\r\n      images.push({\r\n        base64: jpeg,\r\n        mimeType: \"image/jpeg\",\r\n        source: \"notebook_output\",\r\n        cellIndex,\r\n        altText: `Notebook cell ${cellIndex} 输出图`,\r\n      });\r\n    }\r\n  }\r\n  return images;\r\n}\r\n\r\n/**\r\n * 解析一个 .ipynb JSON 字符串,转成 markdown + 图片。\r\n *\r\n * @param jsonText .ipynb 文件的原始 JSON 字符串\r\n * @returns { markdown, images, language, stats }\r\n *\r\n * 纯函数,不依赖 IO,可直接测。\r\n */\r\nexport function parseNotebook(jsonText: string): ParsedNotebook {\r\n  let nb: Record<string, unknown>;\r\n  try {\r\n    nb = JSON.parse(jsonText);\r\n  } catch {\r\n    throw new Error(\"parseNotebook: 无效的 JSON(不是合法的 .ipynb 文件)\");\r\n  }\r\n\r\n  const cells = Array.isArray(nb.cells) ? (nb.cells as Record<string, unknown>[]) : [];\r\n  const metadata = (nb.metadata ?? {}) as ParsedNotebook[\"stats\"] extends never ? never : {\r\n    kernelspec?: { name?: string; display_name?: string; language?: string };\r\n    language_info?: { name?: string };\r\n  };\r\n  const language = inferLanguage(metadata as Parameters<typeof inferLanguage>[0]);\r\n\r\n  const mdParts: string[] = [];\r\n  const images: NotebookImage[] = [];\r\n  let markdownCells = 0;\r\n  let codeCells = 0;\r\n\r\n  for (let i = 0; i < cells.length; i++) {\r\n    const cell = cells[i];\r\n    if (!cell) continue;\r\n    const cellType = cell.cell_type as string | undefined;\r\n    const source = normalizeSource(cell.source);\r\n\r\n    if (cellType === \"markdown\") {\r\n      markdownCells++;\r\n      mdParts.push(source);\r\n    } else if (cellType === \"code\") {\r\n      codeCells++;\r\n      // 代码块\r\n      mdParts.push(`\\`\\`\\`${language}\\n${source}\\n\\`\\`\\``);\r\n      // 提取 output 图片\r\n      const cellImages = extractOutputImages(cell.outputs as unknown[] | undefined, i);\r\n      images.push(...cellImages);\r\n    }\r\n    // raw cell / 其他类型跳过\r\n  }\r\n\r\n  return {\r\n    markdown: mdParts.join(\"\\n\\n\"),\r\n    images,\r\n    language,\r\n    stats: {\r\n      markdownCells,\r\n      codeCells,\r\n      totalCells: cells.length,\r\n    },\r\n  };\r\n}\r\n"],"mappings":";;AAuDA,SAAgB,cAAc,UAGnB;CAET,IAAI,SAAS,YAAY,UAAU,OAAO,SAAS,WAAW,SAAS,YAAY;CAEnF,IAAI,SAAS,eAAe,MAAM,OAAO,SAAS,cAAc,KAAK,YAAY;CAEjF,MAAM,OAAO,SAAS,YAAY,QAAQ;CAC1C,IAAI,KAAK,SAAS,QAAQ,GAAG,OAAO;CACpC,IAAI,KAAK,SAAS,IAAI,KAAK,KAAK,SAAS,IAAI,GAAG,OAAO;CACvD,IAAI,KAAK,SAAS,OAAO,GAAG,OAAO;CACnC,OAAO;AACT;;;;;AAMA,SAAgB,gBAAgB,QAAyB;CACvD,IAAI,OAAO,WAAW,UAAU,OAAO;CACvC,IAAI,MAAM,QAAQ,MAAM,GAAG,OAAO,OAAO,KAAK,EAAE;CAChD,OAAO;AACT;;;;;;;;AASA,SAAgB,oBACd,SACA,WACiB;CACjB,IAAI,CAAC,MAAM,QAAQ,OAAO,GAAG,OAAO,CAAC;CACrC,MAAM,SAA0B,CAAC;CACjC,KAAK,MAAM,UAAU,SAAS;EAC5B,IAAI,CAAC,UAAU,OAAO,WAAW,UAAU;EAE3C,MAAM,OAAOA,OAAE;EACf,IAAI,CAAC,MAAM;EAEX,MAAM,MAAM,KAAK;EACjB,IAAI,OAAO,QAAQ,YAAY,IAAI,SAAS,GAC1C,OAAO,KAAK;GACV,QAAQ;GACR,UAAU;GACV,QAAQ;GACR;GACA,SAAS,iBAAiB,UAAU;EACtC,CAAC;EAGH,MAAM,OAAO,KAAK;EAClB,IAAI,OAAO,SAAS,YAAY,KAAK,SAAS,GAC5C,OAAO,KAAK;GACV,QAAQ;GACR,UAAU;GACV,QAAQ;GACR;GACA,SAAS,iBAAiB,UAAU;EACtC,CAAC;CAEL;CACA,OAAO;AACT;;;;;;;;;AAUA,SAAgB,cAAc,UAAkC;CAC9D,IAAI;CACJ,IAAI;EACF,KAAK,KAAK,MAAM,QAAQ;CAC1B,QAAQ;EACN,MAAM,IAAI,MAAM,0CAA0C;CAC5D;CAEA,MAAM,QAAQ,MAAM,QAAQ,GAAG,KAAK,IAAK,GAAG,QAAsC,CAAC;CAKnF,MAAM,WAAW,cAJC,GAAG,YAAY,CAAC,CAI4C;CAE9E,MAAM,UAAoB,CAAC;CAC3B,MAAM,SAA0B,CAAC;CACjC,IAAI,gBAAgB;CACpB,IAAI,YAAY;CAEhB,KAAK,IAAI,IAAI,GAAG,IAAI,MAAM,QAAQ,KAAK;EACrC,MAAM,OAAO,MAAM;EACnB,IAAI,CAAC,MAAM;EACX,MAAM,WAAW,KAAK;EACtB,MAAM,SAAS,gBAAgB,KAAK,MAAM;EAE1C,IAAI,aAAa,YAAY;GAC3B;GACA,QAAQ,KAAK,MAAM;EACrB,OAAO,IAAI,aAAa,QAAQ;GAC9B;GAEA,QAAQ,KAAK,SAAS,SAAS,IAAI,OAAO,SAAS;GAEnD,MAAM,aAAa,oBAAoB,KAAK,SAAkC,CAAC;GAC/E,OAAO,KAAK,GAAG,UAAU;EAC3B;CAEF;CAEA,OAAO;EACL,UAAU,QAAQ,KAAK,MAAM;EAC7B;EACA;EACA,OAAO;GACL;GACA;GACA,YAAY,MAAM;EACpB;CACF;AACF"}

package/package.json

@@ -0,0 +1,72 @@
+{
+  "name": "dsh-plugin-lookatstudy",
+  "version": "0.8.1",
+  "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",
+  "main": "lib/index.mjs",
+  "types": "lib/index.d.mts",
+  "files": [
+    "lib",
+    "cordis.patch.yml",
+    "README.md"
+  ],
+  "scripts": {
+    "build": "tsdown --config tsdown.config.ts",
+    "test": "node --import tsx --test tests/*.test.ts",
+    "verify": "node scripts/verify.mjs"
+  },
+  "keywords": [
+    "dsh-plugin",
+    "deepseek-harness",
+    "learning",
+    "tutor",
+    "spaced-repetition"
+  ],
+  "license": "MIT",
+  "dsh": {
+    "bundle": {
+      "patch": "./cordis.patch.yml"
+    },
+    "client": {
+      "inject": [
+        "slots",
+        "workspaces",
+        "sessions"
+      ],
+      "platform": "web"
+    }
+  },
+  "peerDependencies": {
+    "@deepseek-ai/cordis": ">=0.1.0-rc.0",
+    "@deepseek-ai/dsh-client-runtime": ">=0.1.0-rc.0",
+    "@deepseek-ai/dsh-client-ui-conversation": ">=0.1.0-rc.0",
+    "@deepseek-ai/dsh-tools": ">=0.1.0-rc.0",
+    "@deepseek-ai/schemastery": ">=0.1.0-rc.0",
+    "react": ">=18"
+  },
+  "exports": {
+    ".": {
+      "types": "./lib/index.d.mts",
+      "default": "./lib/index.mjs"
+    },
+    "./client": {
+      "default": "./lib/client.js"
+    },
+    "./package.json": "./package.json"
+  },
+  "devDependencies": {
+    "@types/react": "~18.3.31",
+    "react": "^18.3.1"
+  },
+  "repository": {
+    "type": "git",
+    "url": "git+https://github.com/Kaiji-Z/dsh-plugin-lookatstudy.git"
+  },
+  "bugs": {
+    "url": "https://github.com/Kaiji-Z/dsh-plugin-lookatstudy/issues"
+  },
+  "homepage": "https://github.com/Kaiji-Z/dsh-plugin-lookatstudy#readme",
+  "engines": {
+    "node": ">=22.19"
+  }
+}