[Docs site] Add util to generate descriptions and port into mdx files

Author: kodster28

bin/generate-descriptions.ts

@@ -0,0 +1,304 @@
+#!/usr/bin/env tsx
+
+/**
+ * This script generates descriptions for MDX files in the docs directory
+ * that don't have a description field in their frontmatter.
+ *
+ * It uses the rendered HTML from the dist directory to generate descriptions
+ * using the generateDescription function from src/util/props.ts.
+ *
+ * Usage:
+ * npm run generate-descriptions [-- --pcx-content-type <type>]
+ *
+ * Options:
+ * --pcx-content-type <type>  Filter MDX files by pcx_content_type (e.g., overview, tutorial, navigation)
+ */
+
+import fs from "fs/promises";
+import path from "path";
+import globby from "fast-glob";
+import { parse as parseHTML } from "node-html-parser";
+import { generateDescription } from "../src/util/props";
+import matter from "gray-matter";
+
+const DOCS_DIR = path.join(process.cwd(), "src/content/docs");
+const DIST_DIR = path.join(process.cwd(), "dist");
+
+// Maximum length for descriptions
+const MAX_DESCRIPTION_LENGTH = 160;
+
+/**
+ * Extracts the first paragraph from HTML content
+ */
+function extractFirstParagraph(html: string): string | undefined {
+	const dom = parseHTML(html);
+	const paragraph = dom.querySelector("p");
+
+	if (paragraph) {
+		return paragraph.textContent.trim();
+	}
+
+	return undefined;
+}
+
+/**
+ * Truncates a description to a reasonable length
+ */
+function truncateDescription(description: string): string {
+	if (description.length <= MAX_DESCRIPTION_LENGTH) {
+		return description;
+	}
+
+	// Truncate at the last space before MAX_DESCRIPTION_LENGTH
+	const truncated = description.substring(0, MAX_DESCRIPTION_LENGTH);
+	const lastSpace = truncated.lastIndexOf(" ");
+
+	if (lastSpace > 0) {
+		return truncated.substring(0, lastSpace) + "...";
+	}
+
+	return truncated + "...";
+}
+
+/**
+ * Gets the rendered HTML path for a docs file
+ */
+function getRenderedPath(docPath: string): string {
+	// Convert /src/content/docs/product/path/file.mdx to /dist/product/path/file/index.html
+	const relativePath = path.relative(DOCS_DIR, docPath);
+	const pathWithoutExt = relativePath.replace(/\.mdx$/, "");
+	return path.join(DIST_DIR, pathWithoutExt, "index.html");
+}
+
+/**
+ * Updates the frontmatter of an MDX file with a description
+ * Ensures that only the description field is modified and all other fields remain unchanged
+ * @returns boolean indicating whether the file was updated (true) or skipped (false)
+ */
+async function updateFrontmatter(
+	filePath: string,
+	description: string,
+): Promise<boolean> {
+	// Read the original file content to preserve exact formatting
+	const originalContent = await fs.readFile(filePath, "utf-8");
+
+	// Parse the frontmatter
+	const { data: frontmatter, content: mdxContent } = matter(originalContent);
+
+	// Check if the description already exists and is the same
+	if (frontmatter.description === description) {
+		console.log(
+			`⏭️ Skipped ${path.relative(process.cwd(), filePath)} (description unchanged)`,
+		);
+		return false;
+	}
+
+	// Instead of using matter.stringify which might change date formats,
+	// we'll manually update just the description field in the original content
+
+	// Extract the frontmatter section (between the first two --- markers)
+	const frontmatterMatch = originalContent.match(/^---\r?\n([\s\S]*?)\r?\n---/);
+	if (!frontmatterMatch) {
+		console.error(`Could not extract frontmatter from ${filePath}`);
+		return false;
+	}
+
+	const originalFrontmatter = frontmatterMatch[1];
+
+	// Check if description already exists in the frontmatter
+	const descriptionRegex = /^description:.*$(\r?\n(?: .*$)*)/m;
+	let newFrontmatter: string;
+
+	if (descriptionRegex.test(originalFrontmatter)) {
+		// Replace existing description
+		newFrontmatter = originalFrontmatter.replace(
+			descriptionRegex,
+			`description: >-\n  ${description.replace(/\n/g, "\n  ")}`,
+		);
+	} else {
+		// Add description at the end of frontmatter
+		newFrontmatter = `${originalFrontmatter.trim()}\ndescription: >-\n  ${description.replace(/\n/g, "\n  ")}`;
+	}
+
+	// Replace the frontmatter in the original content
+	const updatedContent = originalContent.replace(
+		/^---\r?\n[\s\S]*?\r?\n---/,
+		`---\n${newFrontmatter}\n---`,
+	);
+
+	// Write updated content back to file
+	await fs.writeFile(filePath, updatedContent, "utf-8");
+
+	console.log(`✅ Updated ${path.relative(process.cwd(), filePath)}`);
+
+	return true;
+}
+
+/**
+ * Parse command line arguments
+ */
+function parseArgs() {
+	const args = process.argv.slice(2);
+	let pcxContentType: string | undefined;
+	let showHelp = false;
+
+	for (let i = 0; i < args.length; i++) {
+		if (args[i] === "--pcx-content-type" && i + 1 < args.length) {
+			pcxContentType = args[i + 1];
+			i++; // Skip the next argument as it's the value
+		} else if (args[i] === "--help" || args[i] === "-h") {
+			showHelp = true;
+		}
+	}
+
+	return { pcxContentType, showHelp };
+}
+
+/**
+ * Main function
+ */
+function showUsage() {
+	console.log(`
+Usage: npx tsx bin/generate-descriptions.ts [options]
+
+Options:
+  --pcx-content-type <type>  Filter MDX files by pcx_content_type (e.g., overview, tutorial, navigation)
+  --help, -h                 Show this help message
+`);
+}
+
+async function main() {
+	// Parse command line arguments
+	const { pcxContentType, showHelp } = parseArgs();
+
+	if (showHelp) {
+		showUsage();
+		return;
+	}
+
+	if (pcxContentType) {
+		console.log(`Filtering by pcx_content_type: ${pcxContentType}`);
+	}
+	try {
+		// Find all MDX files in the docs directory
+		const mdxFiles = await globby("**/*.mdx", {
+			cwd: DOCS_DIR,
+			absolute: true,
+		});
+		console.log(`Found ${mdxFiles.length} MDX files in the docs directory`);
+
+		// Filter files by pcx_content_type if specified
+		let filteredMdxFiles = mdxFiles;
+		if (pcxContentType) {
+			filteredMdxFiles = [];
+			for (const mdxFile of mdxFiles) {
+				try {
+					const content = await fs.readFile(mdxFile, "utf-8");
+					const { data: frontmatter } = matter(content);
+					if (frontmatter.pcx_content_type === pcxContentType) {
+						filteredMdxFiles.push(mdxFile);
+					}
+				} catch (error) {
+					console.error(`Error reading ${mdxFile}:`, error);
+				}
+			}
+			console.log(
+				`Filtered to ${filteredMdxFiles.length} MDX files with pcx_content_type: ${pcxContentType}`,
+			);
+		}
+
+		let updatedCount = 0;
+		let skippedExistingCount = 0;
+		let skippedUnchangedCount = 0;
+		let errorCount = 0;
+
+		for (const mdxFile of filteredMdxFiles) {
+			try {
+				// Parse frontmatter
+				const content = await fs.readFile(mdxFile, "utf-8");
+				const { data: frontmatter } = matter(content);
+
+				// Skip if description already exists
+				if (frontmatter.description) {
+					skippedExistingCount++;
+					continue;
+				}
+
+				// Get the rendered HTML path
+				const renderedPath = getRenderedPath(mdxFile);
+
+				// Check if rendered HTML exists
+				try {
+					await fs.access(renderedPath);
+				} catch (error) {
+					console.warn(
+						`⚠️ Rendered HTML not found for ${path.relative(process.cwd(), mdxFile)}`,
+					);
+					errorCount++;
+					continue;
+				}
+
+				// Read rendered HTML
+				const html = await fs.readFile(renderedPath, "utf-8");
+
+				// Extract main content from HTML
+				const dom = parseHTML(html);
+				const mainContent = dom.querySelector("main")?.innerHTML || "";
+
+				if (!mainContent) {
+					console.warn(
+						`⚠️ No main content found in rendered HTML for ${path.relative(process.cwd(), mdxFile)}`,
+					);
+					errorCount++;
+					continue;
+				}
+
+				// Generate description
+				let description = await generateDescription({ html: mainContent });
+
+				// If no description was generated, try extracting the first paragraph
+				if (!description) {
+					description = extractFirstParagraph(mainContent);
+				}
+
+				// Skip if no description could be generated
+				if (!description) {
+					console.warn(
+						`⚠️ Could not generate description for ${path.relative(process.cwd(), mdxFile)}`,
+					);
+					errorCount++;
+					continue;
+				}
+
+				// Truncate description if needed
+				description = truncateDescription(description);
+
+				// Update frontmatter
+				const wasUpdated = await updateFrontmatter(mdxFile, description);
+				if (wasUpdated) {
+					updatedCount++;
+				} else {
+					skippedUnchangedCount++;
+				}
+			} catch (error) {
+				console.error(
+					`❌ Error processing ${path.relative(process.cwd(), mdxFile)}:`,
+					error,
+				);
+				errorCount++;
+			}
+		}
+
+		console.log("\n--- Summary ---");
+		console.log(`Total MDX files: ${mdxFiles.length}`);
+		console.log(`Updated: ${updatedCount}`);
+		console.log(`Skipped (already had description): ${skippedExistingCount}`);
+		console.log(`Skipped (description unchanged): ${skippedUnchangedCount}`);
+		console.log(`Errors: ${errorCount}`);
+	} catch (error) {
+		console.error("Error:", error);
+		process.exit(1);
+	}
+}
+
+main();