adrian-altner.de/scripts/migrate-posts.mjs

#!/usr/bin/env node
// One-shot migration of /Users/adrian/Developer/Websites/Content → src/content/posts/en/.
// Rewrites frontmatter to this project's schema and preserves year/month/day layout.

import { mkdirSync, readFileSync, writeFileSync, copyFileSync, existsSync } from 'node:fs';
import path from 'node:path';
import { globSync } from 'node:fs';

const SRC_ROOT = '/Users/adrian/Developer/Websites/Content';
const DEST_ROOT = '/Users/adrian/Developer/Websites/adrian-altner.de/src/content/posts/en';

// Skip list: starter duplicates and explicitly German .mdx posts.
const SKIP = new Set([
	'2022/07/08/first-post.mdx',
	'2022/07/15/second-post.mdx',
	'2022/07/22/third-post.mdx',
	'2024/07/01/using-mdx.mdx',
	'2024/07/19/markdown-style-guide.mdx',
	'2026/04/06/image-voice-memos.mdx', // superseded by .md sibling
	'2026/04/18/justified-layout.mdx',  // German
	'2026/04/18/maple-mono.mdx',        // German
	'2026/04/19/diigo-dsgvo-loeschung.mdx', // German
]);

// Category slug normalisation. Keys map from source `[[wiki-link]]` text.
const CATEGORY_MAP = {
	'on-premises-private-cloud': 'on-premises-private-cloud',
	development: 'development',
	Development: 'development',
	'in-eigener-sache': 'personal',
	projekte: 'projects',
};

// Walk recursively for .md and .mdx files.
function findContentFiles(dir, out = []) {
	const fs = require('node:fs');
	for (const name of fs.readdirSync(dir)) {
		if (name.startsWith('.')) continue;
		const full = path.join(dir, name);
		const stat = fs.statSync(full);
		if (stat.isDirectory()) findContentFiles(full, out);
		else if (/\.mdx?$/.test(name)) out.push(full);
	}
	return out;
}

// Very small inline YAML frontmatter parser. Handles the exact shapes we see
// in the source files: scalar lines, inline arrays [a, b], and block arrays
// introduced with a colon followed by `- item` lines.
function parseFrontmatter(raw) {
	const match = raw.match(/^---\n([\s\S]*?)\n---\n?([\s\S]*)$/);
	if (!match) return { data: {}, body: raw };
	const [, block, body] = match;
	const data = {};
	const lines = block.split('\n');
	let i = 0;
	while (i < lines.length) {
		const line = lines[i];
		// Strip trailing "# comment" (only if preceded by whitespace; avoids
		// eating # inside quoted strings). Simple heuristic that matches our data.
		const stripped = line.replace(/\s+#\s.*$/, '');
		const kv = stripped.match(/^(\w+):\s*(.*)$/);
		if (!kv) { i++; continue; }
		const key = kv[1];
		let value = kv[2].trim();
		if (value === '' && i + 1 < lines.length && /^\s+-\s/.test(lines[i + 1])) {
			const arr = [];
			i++;
			while (i < lines.length && /^\s+-\s/.test(lines[i])) {
				arr.push(lines[i].replace(/^\s+-\s+/, '').replace(/^["']|["']$/g, ''));
				i++;
			}
			data[key] = arr;
			continue;
		}
		if (value.startsWith('[') && value.endsWith(']')) {
			data[key] = value.slice(1, -1).split(',').map((s) => s.trim().replace(/^["']|["']$/g, '')).filter(Boolean);
		} else if (/^(true|false)$/.test(value)) {
			data[key] = value === 'true';
		} else if (/^-?\d+$/.test(value)) {
			data[key] = Number(value);
		} else {
			data[key] = value.replace(/^["']|["']$/g, '');
		}
		i++;
	}
	return { data, body };
}

function wikiLinkInner(value) {
	if (typeof value !== 'string') return undefined;
	const m = value.match(/^\[\[([^\]]+)\]\]$/);
	return m ? m[1] : value;
}

function quoteIfNeeded(s) {
	if (/[:#\[\]&*!|>'"%@`]/.test(s) || /^\s|\s$/.test(s)) {
		return `'${s.replace(/'/g, "''")}'`;
	}
	return s;
}

function formatFrontmatter(data) {
	const lines = ['---'];
	const order = [
		'title', 'description', 'pubDate', 'updatedDate',
		'heroImage', 'heroAlt', 'hideHero',
		'category',
		'tags',
		'seriesParent', 'seriesOrder',
		'url', 'repo', 'toc', 'draft',
		'translationKey',
	];
	for (const key of order) {
		if (!(key in data)) continue;
		const v = data[key];
		if (v === undefined || v === null) continue;
		if (Array.isArray(v)) {
			if (v.length === 0) continue;
			lines.push(`${key}:`);
			for (const item of v) lines.push(`  - ${quoteIfNeeded(String(item))}`);
		} else if (typeof v === 'boolean' || typeof v === 'number') {
			lines.push(`${key}: ${v}`);
		} else {
			lines.push(`${key}: ${quoteIfNeeded(String(v))}`);
		}
	}
	lines.push('---');
	return lines.join('\n') + '\n';
}

function transform(srcFile) {
	const rel = path.relative(SRC_ROOT, srcFile).replace(/\\/g, '/');
	if (SKIP.has(rel)) return null;
	const raw = readFileSync(srcFile, 'utf8');
	const { data, body } = parseFrontmatter(raw);

	const out = {};
	out.title = data.title;
	out.description = data.description;

	// Date: prefer pubDate, fall back to publishDate.
	const rawDate = data.pubDate ?? data.publishDate;
	if (rawDate !== undefined) {
		// Strings like `2026-03-01T18:57:00+01:00` or `2026-03-22` or `Jul 08 2022`
		// pass through unchanged; zod's z.coerce.date() parses them.
		out.pubDate = String(rawDate);
	}
	if (data.updatedDate) out.updatedDate = String(data.updatedDate);

	// Hero image: `heroImage` in existing format, `cover` in new format.
	if (data.heroImage) out.heroImage = data.heroImage;
	else if (data.cover) out.heroImage = data.cover;
	if (data.coverAlt) out.heroAlt = data.coverAlt;
	if (typeof data.hideHero === 'boolean') out.hideHero = data.hideHero;

	// Category: `[[wiki-link]]` → en/<mapped-slug>. Plain string (already used
	// in some existing posts) stays as en/<slug>.
	if (data.category) {
		const inner = wikiLinkInner(data.category);
		const mapped = CATEGORY_MAP[inner] ?? inner.toLowerCase();
		out.category = `en/${mapped}`;
	}

	// Series: `[[parent-slug]]` → bare parent-slug.
	if (data.seriesParent) out.seriesParent = wikiLinkInner(data.seriesParent);
	if (typeof data.seriesOrder === 'number') out.seriesOrder = data.seriesOrder;

	if (Array.isArray(data.tags) && data.tags.length > 0) out.tags = data.tags;

	if (data.url) out.url = data.url;
	if (data.repo) out.repo = data.repo;
	if (typeof data.toc === 'boolean') out.toc = data.toc;
	if (typeof data.draft === 'boolean') out.draft = data.draft;

	if (data.translationKey) out.translationKey = data.translationKey;

	const destRel = rel; // preserve year/month/day
	const destFile = path.join(DEST_ROOT, destRel);
	const frontmatter = formatFrontmatter(out);
	const content = frontmatter + '\n' + body.replace(/^\n+/, '');
	return { destFile, content, srcDir: path.dirname(srcFile), destDir: path.dirname(destFile) };
}

function main() {
	const fs = require('node:fs');
	const files = findContentFiles(SRC_ROOT);
	let written = 0, skipped = 0, assetsCopied = 0;
	const destDirs = new Set();
	for (const srcFile of files) {
		const rel = path.relative(SRC_ROOT, srcFile).replace(/\\/g, '/');
		if (SKIP.has(rel)) { skipped++; continue; }
		const t = transform(srcFile);
		if (!t) { skipped++; continue; }
		mkdirSync(t.destDir, { recursive: true });
		writeFileSync(t.destFile, t.content);
		written++;
		destDirs.add(t.srcDir + '|' + t.destDir);
	}
	// Copy accompanying asset files (images) from each source dir.
	for (const pair of destDirs) {
		const [srcDir, destDir] = pair.split('|');
		for (const name of fs.readdirSync(srcDir)) {
			if (name.startsWith('.')) continue;
			if (/\.(mdx?|DS_Store)$/i.test(name)) continue;
			const srcAsset = path.join(srcDir, name);
			const destAsset = path.join(destDir, name);
			if (!existsSync(destAsset)) {
				copyFileSync(srcAsset, destAsset);
				assetsCopied++;
			}
		}
	}
	console.log(`wrote ${written} post files, skipped ${skipped}, copied ${assetsCopied} assets`);
}

// Node ESM doesn't provide require by default — fall back to createRequire.
import { createRequire } from 'node:module';
const require = createRequire(import.meta.url);

main();