granola-toolkit 0.10.0 → 0.12.0

package/README.md

@@ -214,6 +214,7 @@ Supported environment variables:
 - `TIMEOUT`
 - `CACHE_FILE`
 - `TRANSCRIPT_OUTPUT`
+- `GRANOLA_CLIENT_VERSION`
 
 ## Development Checks
 

package/dist/cli.js

@@ -5,12 +5,17 @@ import { mkdir, readFile, rm, stat, unlink, writeFile } from "node:fs/promises";
 import { homedir, platform } from "node:os";
 import { dirname, join } from "node:path";
 import { promisify } from "node:util";
+import { NodeHtmlMarkdown } from "node-html-markdown";
 import { createHash } from "node:crypto";
 //#region src/utils.ts
 const INVALID_FILENAME_CHARS = /[<>:"/\\|?*]/g;
 const CONTROL_CHARACTERS = /\p{Cc}/gu;
 const MULTIPLE_UNDERSCORES = /_+/g;
 const MULTIPLE_BLANK_LINES = /\n{3,}/g;
+const htmlMarkdown = new NodeHtmlMarkdown({
+	bulletMarker: "-",
+	ignore: ["script", "style"]
+});
 function normaliseNewlines(value) {
 	return value.replace(/\r\n?/g, "\n");
 }
@@ -117,25 +122,9 @@ function formatTimestampForTranscript(timestamp) {
 	if (Number.isNaN(parsed.getTime())) return timestamp;
 	return parsed.toISOString().slice(11, 19);
 }
-function decodeHtmlEntities(value) {
-	return value.replaceAll("&nbsp;", " ").replaceAll("&amp;", "&").replaceAll("&lt;", "<").replaceAll("&gt;", ">").replaceAll("&quot;", "\"").replaceAll("&#39;", "'");
-}
-function htmlToMarkdownFallback(value) {
-	let output = normaliseNewlines(decodeHtmlEntities(value));
-	output = output.replace(/<br\s*\/?>/gi, "\n");
-	output = output.replace(/<li\b[^>]*>/gi, "- ");
-	output = output.replace(/<\/li>/gi, "\n");
-	output = output.replace(/<h1\b[^>]*>/gi, "# ");
-	output = output.replace(/<h2\b[^>]*>/gi, "## ");
-	output = output.replace(/<h3\b[^>]*>/gi, "### ");
-	output = output.replace(/<h4\b[^>]*>/gi, "#### ");
-	output = output.replace(/<h5\b[^>]*>/gi, "##### ");
-	output = output.replace(/<h6\b[^>]*>/gi, "###### ");
-	output = output.replace(/<\/(p|div|section|article|ul|ol|blockquote|h[1-6])>/gi, "\n\n");
-	output = output.replace(/<[^>]+>/g, "");
-	output = output.replace(/[ \t]+\n/g, "\n");
-	output = output.replace(MULTIPLE_BLANK_LINES, "\n\n");
-	return output.trim();
+function htmlToMarkdown(value) {
+	if (!value.trim()) return "";
+	return normaliseNewlines(htmlMarkdown.translate(value)).replace(/[ \t]+\n/g, "\n").replace(MULTIPLE_BLANK_LINES, "\n\n").trim();
 }
 function latestDocumentTimestamp(document) {
 	const candidates = [
@@ -551,13 +540,23 @@ function parseDocument(value) {
 }
 //#endregion
 //#region src/client/granola.ts
-const USER_AGENT = "Granola/5.354.0";
-const CLIENT_VERSION = "5.354.0";
+const DEFAULT_CLIENT_VERSION = "5.354.0";
 const DOCUMENTS_URL = "https://api.granola.ai/v2/get-documents";
+function resolveClientVersion(value) {
+	return value?.trim() || process.env.GRANOLA_CLIENT_VERSION?.trim() || DEFAULT_CLIENT_VERSION;
+}
 var GranolaApiClient = class {
-	constructor(httpClient, documentsUrl = DOCUMENTS_URL) {
+	clientVersion;
+	documentsUrl;
+	constructor(httpClient, options = DOCUMENTS_URL) {
 		this.httpClient = httpClient;
-		this.documentsUrl = documentsUrl;
+		if (typeof options === "string") {
+			this.documentsUrl = options;
+			this.clientVersion = resolveClientVersion();
+			return;
+		}
+		this.documentsUrl = options.documentsUrl ?? DOCUMENTS_URL;
+		this.clientVersion = resolveClientVersion(options.clientVersion);
 	}
 	async listDocuments(options) {
 		const documents = [];
@@ -570,8 +569,8 @@ var GranolaApiClient = class {
 				offset
 			}, {
 				headers: {
-					"User-Agent": USER_AGENT,
-					"X-Client-Version": CLIENT_VERSION
+					"User-Agent": `Granola/${this.clientVersion}`,
+					"X-Client-Version": this.clientVersion
 				},
 				timeoutMs: options.timeoutMs
 			});
@@ -591,35 +590,79 @@ var GranolaApiClient = class {
 };
 //#endregion
 //#region src/client/http.ts
+const RETRYABLE_STATUS_CODES = new Set([
+	429,
+	500,
+	502,
+	503,
+	504
+]);
+function sleep(delayMs) {
+	return new Promise((resolve) => {
+		setTimeout(resolve, delayMs);
+	});
+}
+function parseRetryAfter(headerValue) {
+	if (!headerValue?.trim()) return;
+	if (/^\d+$/.test(headerValue.trim())) return Number(headerValue.trim()) * 1e3;
+	const retryAt = Date.parse(headerValue);
+	if (Number.isNaN(retryAt)) return;
+	return Math.max(0, retryAt - Date.now());
+}
 var AuthenticatedHttpClient = class {
 	fetchImpl;
 	constructor(options) {
 		this.fetchImpl = options.fetchImpl ?? fetch;
 		this.logger = options.logger;
+		this.maxRetries = options.maxRetries ?? 2;
+		this.retryBaseDelayMs = options.retryBaseDelayMs ?? 500;
+		this.retryMaxDelayMs = options.retryMaxDelayMs ?? 5e3;
+		this.sleepImpl = options.sleepImpl ?? sleep;
 		this.tokenProvider = options.tokenProvider;
 	}
 	logger;
+	maxRetries;
+	retryBaseDelayMs;
+	retryMaxDelayMs;
+	sleepImpl;
 	tokenProvider;
-	async request(options) {
+	async retry(options, attempt, reason, response) {
+		const retryAfterMs = parseRetryAfter(response?.headers.get("retry-after") ?? null);
+		const delayMs = Math.min(retryAfterMs ?? this.retryBaseDelayMs * 2 ** attempt, this.retryMaxDelayMs);
+		this.logger?.warn?.(`${reason}; retrying in ${delayMs}ms (${attempt + 1}/${this.maxRetries})`);
+		await this.sleepImpl(delayMs);
+		return this.request(options, attempt + 1);
+	}
+	async request(options, attempt = 0) {
 		const { retryOnUnauthorized = true, timeoutMs, url } = options;
 		const accessToken = await this.tokenProvider.getAccessToken();
-		const response = await this.fetchImpl(url, {
-			body: options.body,
-			headers: {
-				...options.headers,
-				Authorization: `Bearer ${accessToken}`
-			},
-			method: options.method ?? "GET",
-			signal: AbortSignal.timeout(timeoutMs)
-		});
+		let response;
+		try {
+			response = await this.fetchImpl(url, {
+				body: options.body,
+				headers: {
+					...options.headers,
+					Authorization: `Bearer ${accessToken}`
+				},
+				method: options.method ?? "GET",
+				signal: AbortSignal.timeout(timeoutMs)
+			});
+		} catch (error) {
+			if (attempt < this.maxRetries) {
+				const message = error instanceof Error ? error.message : String(error);
+				return this.retry(options, attempt, `request failed: ${message}`);
+			}
+			throw error;
+		}
 		if (response.status === 401 && retryOnUnauthorized) {
 			this.logger?.warn?.("request returned 401; invalidating token provider and retrying once");
 			await this.tokenProvider.invalidate();
 			return this.request({
 				...options,
 				retryOnUnauthorized: false
-			});
+			}, attempt);
 		}
+		if (RETRYABLE_STATUS_CODES.has(response.status) && attempt < this.maxRetries) return this.retry(options, attempt, `request returned ${response.status} ${response.statusText || ""}`.trim(), response);
 		return response;
 	}
 	async postJson(url, body, options = { timeoutMs: 3e4 }) {
@@ -998,7 +1041,7 @@ function selectNoteContent(document) {
 		content: lastViewedPanel,
 		source: "lastViewedPanel.content"
 	};
-	const originalContent = htmlToMarkdownFallback(document.lastViewedPanel?.originalContent ?? "").trim();
+	const originalContent = htmlToMarkdown(document.lastViewedPanel?.originalContent ?? "").trim();
 	if (originalContent) return {
 		content: originalContent,
 		source: "lastViewedPanel.originalContent"

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "granola-toolkit",
-  "version": "0.10.0",
+  "version": "0.12.0",
   "description": "CLI toolkit for exporting and working with Granola notes and transcripts",
   "keywords": [
     "cli",
@@ -49,6 +49,9 @@
     "typecheck": "vp exec tsc --noEmit",
     "prepare": "vp config"
   },
+  "dependencies": {
+    "node-html-markdown": "^2.0.0"
+  },
   "devDependencies": {
     "@types/node": "^25.5.2",
     "typescript": "^5.9.3",