Merge pull request #41 from inaridiy/fix/cli

Fix/cli
inaridiy · Oct 2, 2024 · f3f493c · f3f493c
2 parents cb9953f + 5e82341
commit f3f493c
Show file tree

Hide file tree

Showing 9 changed files with 185 additions and 122 deletions.
diff --git a/.changeset/lazy-actors-carry.md b/.changeset/lazy-actors-carry.md
@@ -0,0 +1,5 @@
+---
+"webforai": patch
+---
+
+Fix with { type : "json" } Error
diff --git a/packages/webforai/build.ts b/packages/webforai/build.ts
@@ -21,7 +21,7 @@ const args = arg({
 const isWatch = args["--watch"];
 
 const entryPoints = glob.sync("./src/**/*.ts", {
-	ignore: ["./src/**/*.test.ts"],
+	ignore: ["./src/**/*.test.ts", "./src/cli/**/*.ts"],
 });
 
 const addExtension = (extension = ".js", fileExtension = ".ts"): Plugin => ({

diff --git a/packages/webforai/example.md b/packages/webforai/example.md
diff --git a/packages/webforai/package.json b/packages/webforai/package.json
@@ -4,15 +4,7 @@
 	"description": "A library that provides a web interface for AI",
 	"author": "inaridiy",
 	"license": "MIT",
-	"keywords": [
-		"web",
-		"ai",
-		"html",
-		"html2md",
-		"markdown",
-		"mdast",
-		"hast"
-	],
+	"keywords": ["web", "ai", "html", "html2md", "markdown", "mdast", "hast"],
 	"repository": {
 		"type": "git",
 		"url": "https://github.com/inaridiy/webforai.git"
@@ -26,10 +18,7 @@
 		"prerelease": "pnpm build",
 		"release": "np"
 	},
-	"files": [
-		"dist",
-		"!dist/types/**/*.js"
-	],
+	"files": ["dist", "!dist/types/**/*.js"],
 	"main": "dist/cjs/index.js",
 	"type": "module",
 	"module": "dist/index.js",
@@ -69,21 +58,11 @@
 	},
 	"typesVersions": {
 		"*": {
-			"types": [
-				"./dist/types/index.d.ts"
-			],
-			"loaders/playwright": [
-				"./dist/types/loaders/playwright.d.ts"
-			],
-			"loaders/cf-puppeteer": [
-				"./dist/types/loaders/cf-puppeteer.d.ts"
-			],
-			"loaders/fetch": [
-				"./dist/types/loaders/fetch.d.ts"
-			],
-			"loaders/puppeteer": [
-				"./dist/types/loaders/puppeteer.d.ts"
-			]
+			"types": ["./dist/types/index.d.ts"],
+			"loaders/playwright": ["./dist/types/loaders/playwright.d.ts"],
+			"loaders/cf-puppeteer": ["./dist/types/loaders/cf-puppeteer.d.ts"],
+			"loaders/fetch": ["./dist/types/loaders/fetch.d.ts"],
+			"loaders/puppeteer": ["./dist/types/loaders/puppeteer.d.ts"]
 		}
 	},
 	"peerDependencies": {
@@ -104,6 +83,7 @@
 	},
 	"dependencies": {
 		"@clack/prompts": "^0.7.0",
+		"boxen": "^8.0.1",
 		"commander": "^12.1.0",
 		"hast-util-from-html": "^2.0.3",
 		"hast-util-select": "^6.0.2",
@@ -123,12 +103,12 @@
 	"devDependencies": {
 		"@cloudflare/puppeteer": "^0.0.6",
 		"@tsconfig/recommended": "^1.0.3",
-		"fastest-levenshtein": "^1.0.16",
 		"@types/hast": "^3.0.2",
 		"@types/mdast": "^4.0.2",
 		"@types/node": "^20.14.10",
 		"arg": "^5.0.2",
 		"esbuild": "^0.19.11",
+		"fastest-levenshtein": "^1.0.16",
 		"glob": "^10.3.10",
 		"ncp": "^2.0.0",
 		"np": "^9.2.0",

diff --git a/packages/webforai/src/cli/bin.ts b/packages/webforai/src/cli/bin.ts
@@ -1,6 +1,4 @@
-import { intro } from "@clack/prompts";
 import { program } from "commander";
-import pc from "picocolors";
 import packageInfo from "../../package.json";
 import { webforaiCommand } from "./commands/webforai";
 import { LOADERS, MODES } from "./constants";
@@ -18,6 +16,4 @@ program
 	.option("-d, --debug", "output extra debugging information")
 	.action(webforaiCommand);
 
-intro(pc.inverse(`webforai CLI version ${packageInfo.version}`));
-
 program.parse();
diff --git a/packages/webforai/src/cli/commands/webforai/index.ts b/packages/webforai/src/cli/commands/webforai/index.ts
@@ -1,6 +1,6 @@
 import fs from "node:fs/promises";
 import path from "node:path";
-import { intro, outro, spinner } from "@clack/prompts";
+import { intro, log, outro, spinner } from "@clack/prompts";
 import pc from "picocolors";
 import packageInfo from "../../../../package.json";
 import { htmlToMarkdown } from "../../../html-to-markdown";
@@ -21,16 +21,16 @@ export const webforaiCommand = async (
 	intro(pc.bold(pc.green(`webforai CLI version ${packageInfo.version}`)));
 
 	const sourcePath = initialPath ?? (await inputSourcePath());
-	options.debug && console.debug(`sourcePath: ${sourcePath}`);
+	options.debug && log.info(`sourcePath: ${sourcePath}`);
 
 	const loader = isUrl(sourcePath) ? options.loader ?? (await selectLoader()) : "local";
-	options.debug && console.debug(`loader: ${loader}`);
+	options.debug && log.info(`loader: ${loader}`);
 
 	const outputPath = options.output ?? (await inputOutputPath(sourcePath));
-	options.debug && console.debug(`outputPath: ${outputPath}`);
+	options.debug && log.info(`outputPath: ${outputPath}`);
 
 	const mode = options.mode ?? (await selectExtractMode());
-	options.debug && console.debug(`mode: ${mode}`);
+	options.debug && log.info(`mode: ${mode}`);
 
 	let html: string;
 	const s = spinner();
@@ -43,13 +43,13 @@ export const webforaiCommand = async (
 		console.error(error);
 		process.exit(1);
 	}
-	options.debug && console.debug(`html: ${html}`);
+	options.debug && log.info(`html: ${html}`);
 
 	const markdown = htmlToMarkdown(html, {
 		baseUrl: isUrl(sourcePath) ? sourcePath : undefined,
 		...(mode === "ai" ? aiModeOptions : readabilityModeOptions),
 	});
-	options.debug && console.debug(`markdown: ${markdown}`);
+	options.debug && log.info(`markdown: ${markdown}`);
 
 	const directory = path.dirname(outputPath);
 	const isDirectoryExists = await fs.stat(directory).then((stat) => stat.isDirectory());

diff --git a/packages/webforai/src/cli/commands/webforai/loadHtml.ts b/packages/webforai/src/cli/commands/webforai/loadHtml.ts
@@ -1,7 +1,8 @@
 import fs from "node:fs/promises";
+import { log } from "@clack/prompts";
+import boxen from "boxen";
 import pc from "picocolors";
 import { chromium } from "playwright-core";
-import playwrightPackageJson from "playwright-core/package.json";
 import { loadHtml as loadHtmlByFetch } from "../../../loaders/fetch";
 import { loadHtml as loadHtmlByPlaywright } from "../../../loaders/playwright";
 
@@ -11,49 +12,61 @@ const checkPlaywrightAvailable = async () => {
 	return !!isAvailable;
 };
 
+const getPlaywrightVersion = async () => {
+	const path = await import.meta.resolve("playwright-core/package.json");
+	const pwPackageJson = await fs
+		.readFile(path.replace("file://", ""), "utf-8")
+		.then((res) => JSON.parse(res.toString()))
+		.catch(() => null);
+	return pwPackageJson?.version;
+};
+
 export const loadHtml = async (sourcePath: string, loader: string, options: { debug?: boolean }) => {
 	if (loader === "local") {
-		options.debug && console.debug(`Loading HTML from local file: ${sourcePath}`);
+		options.debug && log.info(`Loading HTML from local file: ${sourcePath}`);
 		const content = await fs.readFile(sourcePath, "utf-8");
-		options.debug && console.debug(`HTML loaded: ${content.slice(0, 100)}`);
+		options.debug && log.info(`HTML loaded: ${content.slice(0, 100)}`);
 		return content;
 	}
 
 	if (loader === "fetch") {
-		options.debug && console.debug(`Loading HTML from URL: ${sourcePath}`);
+		options.debug && log.info(`Loading HTML from URL: ${sourcePath}`);
 		const content = await loadHtmlByFetch(sourcePath);
-		options.debug && console.debug(`HTML loaded: ${content.slice(0, 100)}`);
+		options.debug && log.info(`HTML loaded: ${content.slice(0, 100)}`);
 		return content;
 	}
 
 	if (loader === "playwright") {
-		options.debug && console.debug(`Loading HTML from playwright: ${sourcePath}`);
+		options.debug && log.info(`Loading HTML from playwright: ${sourcePath}`);
 		const isPlaywrightAvailable = await checkPlaywrightAvailable();
-		options.debug && console.debug(`Playwright available: ${isPlaywrightAvailable}`);
+		options.debug && log.info(`Playwright available: ${isPlaywrightAvailable}`);
+
+		const pwVersion = await getPlaywrightVersion();
+
 		if (!isPlaywrightAvailable) {
 			const message = [
-				"To use playwright loader, you need to install playwright.",
-				"You can install it by running:",
+				pc.bold("Error: Playwright is not available"),
 				"",
-				`	${pc.bold(`npx playwright@${playwrightPackageJson.version} install chromium`)}	`,
+				"To use the Playwright loader, please install Playwright by running:",
 				"",
-				"Hint1: If you get warning like this:",
-				"WARNING: It looks like you are running 'npx playwright install' without first installing your project's dependencies. ",
-				`${pc.bold("Ignore this warning.")}`,
+				`  npx playwright@${pwVersion} install chromium`,
 				"",
-				"Hint2: If you get message like this:",
-				"Host system is missing dependencies to run browsers.",
-				"You should install dependencies by running:",
+				"Hint 1: If you receive a warning like this:",
+				`  "WARNING: It looks like you are running 'npx playwright install' without first installing your project's dependencies."`,
+				"You can safely ignore this warning.",
 				"",
-				`	${pc.bold(`sudo npx playwright@${playwrightPackageJson.version} install-deps`)}	`,
+				"Hint 2: If you encounter the following message:",
+				`  "Host system is missing dependencies to run browsers."`,
+				"You should install the necessary dependencies by executing:",
 				"",
-				pc.gray("Note: Good luck with that."),
+				`  sudo npx playwright@${pwVersion} install-deps`,
 			];
-			console.info(message.join("\n"));
+
+			log.error(boxen(message.join("\n"), { padding: 1, borderStyle: "round" }));
 			throw new Error("Playwright is not available");
 		}
 		const content = await loadHtmlByPlaywright(sourcePath);
-		options.debug && console.debug(`HTML loaded: ${content.slice(0, 100)}`);
+		options.debug && log.info(`HTML loaded: ${content.slice(0, 100)}`);
 		return content;
 	}