sitebase / packages / core / src / export.ts

import { mkdir, readFile, writeFile } from "node:fs/promises";
import { dirname, join, resolve } from "node:path";
import { fetchPublicationWithDocuments } from "./atproto.ts";
import {
	createHandlebars,
	DEFAULT_CONTENT_TEMPLATE,
	generateContent,
} from "./templates.ts";
import type {
	Document,
	ExportConfig,
	ExportOptions,
	ExportResult,
	Publication,
	TemplateData,
} from "./types.ts";

/**
 * Get the text content from a document, handling different content formats
 */
function getDocumentContent(doc: Document): string {
	// textContent is the raw markdown
	if (doc.textContent) {
		return doc.textContent;
	}
	// content might be a string or structured content
	if (typeof doc.content === "string") {
		return doc.content;
	}
	return "";
}

/**
 * Filter documents based on tag inclusion/exclusion rules
 */
function filterDocuments(
	documents: Document[],
	includeTags?: string[],
	excludeTags?: string[],
): { included: Document[]; skipped: number } {
	let filtered = documents;
	let skipped = 0;

	// If includeTags specified, only keep docs with at least one matching tag
	if (includeTags && includeTags.length > 0) {
		filtered = filtered.filter((doc) => {
			const docTags = doc.tags || [];
			const matches = includeTags.some((tag) => docTags.includes(tag));
			if (!matches) skipped++;
			return matches;
		});
	} else {
		// Default behavior: exclude drafts unless explicitly included
		filtered = filtered.filter((doc) => {
			const docTags = doc.tags || [];
			const isDraft = docTags.includes("draft");
			if (isDraft) skipped++;
			return !isDraft;
		});
	}

	// Exclude docs with any of the excludeTags
	if (excludeTags && excludeTags.length > 0) {
		const beforeExclude = filtered.length;
		filtered = filtered.filter((doc) => {
			const docTags = doc.tags || [];
			return !excludeTags.some((tag) => docTags.includes(tag));
		});
		skipped += beforeExclude - filtered.length;
	}

	return { included: filtered, skipped };
}

/**
 * Build template data from a document and publication
 */
function buildTemplateData(
	doc: Document,
	publication: Publication,
): TemplateData {
	return {
		title: doc.title,
		path: doc.path,
		description: doc.description,
		content: getDocumentContent(doc),
		tags: doc.tags || [],
		publishedAt: doc.publishedAt,
		updatedAt: doc.updatedAt,
		publication: {
			name: publication.name,
			url: publication.url,
			description: publication.description,
		},
	};
}

/**
 * Sanitize a filename by removing path traversal and invalid characters
 */
function sanitizeFilename(filename: string): string {
	return filename.replace(/\.\./g, "").replace(/[<>:"|?*]/g, "");
}

/**
 * Export a publication to markdown files
 */
export async function exportPublication(
	options: ExportOptions,
): Promise<ExportResult> {
	const {
		publicationUri,
		outputDir,
		filename: filenameFunction,
		contentTemplate,
		contentFunction,
		includeTags,
		excludeTags,
	} = options;

	const result: ExportResult = {
		filesWritten: [],
		documentsProcessed: 0,
		documentsSkipped: 0,
		warnings: [],
	};

	// Fetch publication and documents
	const { publication, documents } =
		await fetchPublicationWithDocuments(publicationUri);

	// Filter documents by tags
	const { included, skipped } = filterDocuments(
		documents,
		includeTags,
		excludeTags,
	);
	result.documentsSkipped = skipped;

	// Create output directory
	await mkdir(outputDir, { recursive: true });

	// Track filenames to detect conflicts
	const usedFilenames = new Set<string>();

	// Set up Handlebars (only needed if using content template)
	const hbs = createHandlebars();

	// Process each document
	for (const doc of included) {
		result.documentsProcessed++;

		const data = buildTemplateData(doc, publication);

		// Generate filename using function
		let filename = filenameFunction(data);
		filename = sanitizeFilename(filename);

		if (!filename) {
			result.warnings.push(
				`Skipping document "${doc.title}": could not generate filename`,
			);
			result.documentsSkipped++;
			continue;
		}

		// Check for filename conflicts
		if (usedFilenames.has(filename)) {
			result.warnings.push(
				`Skipping document "${doc.title}": filename conflict with "${filename}"`,
			);
			result.documentsSkipped++;
			continue;
		}
		usedFilenames.add(filename);

		// Generate content using function or template
		let content: string;
		if (contentFunction) {
			content = contentFunction(data);
		} else {
			content = generateContent(
				hbs,
				contentTemplate || DEFAULT_CONTENT_TEMPLATE,
				data,
			);
		}

		// Write file (creating subdirectories if needed)
		const filePath = join(outputDir, filename);
		const fileDir = dirname(filePath);
		await mkdir(fileDir, { recursive: true });
		await writeFile(filePath, content, "utf-8");
		result.filesWritten.push(filePath);
	}

	return result;
}

/**
 * Export a publication using a config with multiple export targets
 * @param config - The export configuration
 * @param configDir - Directory containing the config file (for resolving relative paths)
 */
export async function exportFromConfig(
	config: ExportConfig,
	configDir: string,
): Promise<ExportResult[]> {
	const results: ExportResult[] = [];

	for (const target of config.exports) {
		// Load content template from file if specified
		let contentTemplate: string | undefined;
		if (target.contentTemplate) {
			const templatePath = resolve(configDir, target.contentTemplate);
			contentTemplate = await readFile(templatePath, "utf-8");
		}

		const options: ExportOptions = {
			publicationUri: config.publicationUri,
			outputDir: resolve(configDir, target.outputDir),
			includeTags: target.includeTags,
			excludeTags: target.excludeTags,
			filename: target.filename,
			contentTemplate,
			contentFunction: target.content,
		};

		const result = await exportPublication(options);
		results.push(result);
	}

	return results;
}