@rmdes/indiekit-endpoint-posts 1.0.0-beta.34 → 1.0.0-beta.36

package/lib/controllers/edit.js

@@ -1,23 +1,48 @@
-import { getPostPropertiesByUrl } from "../utils.js";
+import { getPostPropertiesByUrl, importPostFromFile } from "../utils.js";
 
 export const editController = {
   async get(request, response, next) {
     try {
       const { url } = request.query;
-      const { application } = request.app.locals;
+      const { application, publication } = request.app.locals;
+
+      console.log("[edit] url=%s", url);
 
       if (!url) {
+        console.log("[edit] no url, redirecting to posts list");
         return response.redirect(request.baseUrl);
       }
 
+      // Try MongoDB lookup first (works for posts created via Micropub)
       const properties = await getPostPropertiesByUrl(url, application);
 
-      if (!properties) {
-        return response.redirect(request.baseUrl);
+      if (properties) {
+        const target = `${request.baseUrl}/${properties.uid}/update`;
+        console.log("[edit] found in DB, redirecting to %s", target);
+        return response.redirect(target);
+      }
+
+      console.log("[edit] not in DB, trying on-demand import");
+      console.log("[edit] publication.store=%o", !!publication.store);
+      console.log(
+        "[edit] store.options=%o",
+        publication.store?.options,
+      );
+
+      // On-demand import: post not in MongoDB, try reading from disk
+      const imported = await importPostFromFile(url, publication, application);
+
+      if (imported) {
+        const target = `${request.baseUrl}/${imported.uid}/update`;
+        console.log("[edit] imported, redirecting to %s", target);
+        return response.redirect(target);
       }
 
-      response.redirect(`${request.baseUrl}/${properties.uid}/update`);
+      // Neither in DB nor on disk — redirect to posts list
+      console.log("[edit] import failed, redirecting to posts list");
+      return response.redirect(request.baseUrl);
     } catch (error) {
+      console.error("[edit] error:", error);
       next(error);
     }
   },

package/lib/utils.js

@@ -1,4 +1,6 @@
 import { Buffer } from "node:buffer";
+import fs from "node:fs/promises";
+import path from "node:path";
 
 import { sanitise, ISO_6709_RE } from "@indiekit/util";
 import { mf2tojf2 } from "@paulrobertlloyd/mf2tojf2";
@@ -357,3 +359,251 @@ export const getSyndicateToItems = (publication, checkTargets = false) => {
         }),
   }));
 };
+
+/** Directory name → post-type mapping */
+const dirToPostType = {
+  articles: "article",
+  notes: "note",
+  likes: "like",
+  bookmarks: "bookmark",
+  photos: "photo",
+  replies: "reply",
+  reposts: "repost",
+  pages: "page",
+  videos: "video",
+  audio: "audio",
+  jams: "jam",
+  rsvps: "rsvp",
+  events: "event",
+};
+
+/**
+ * Derive file path from a post URL
+ * Supports both Eleventy format (/content/TYPE/YYYY-MM-DD-slug/)
+ * and clean Indiekit format (/TYPE/YYYY/MM/DD/slug)
+ * @param {string} url - Full URL or pathname
+ * @returns {string|false} Relative file path (e.g. "articles/2019-01-12-slug.md")
+ */
+export const deriveFilePathFromUrl = (url) => {
+  try {
+    let pathname = url;
+
+    // Extract pathname from full URL
+    if (url.startsWith("http://") || url.startsWith("https://")) {
+      pathname = new URL(url).pathname;
+    }
+
+    // Strip trailing slash
+    pathname = pathname.replace(/\/$/, "");
+
+    // Format 1: /content/TYPE/YYYY-MM-DD-slug (Eleventy output URL)
+    if (pathname.startsWith("/content/")) {
+      const relative = pathname.replace(/^\/content\//, "");
+      return `${relative}.md`;
+    }
+
+    // Format 2: /TYPE/YYYY/MM/DD/slug (clean Indiekit URL)
+    const match = pathname.match(
+      /^\/([a-z]+)\/(\d{4})\/(\d{2})\/(\d{2})\/(.+)$/,
+    );
+    if (match) {
+      const [, type, year, month, day, slug] = match;
+      return `${type}/${year}-${month}-${day}-${slug}.md`;
+    }
+
+    return false;
+  } catch {
+    return false;
+  }
+};
+
+/**
+ * Parse YAML frontmatter and content from a Markdown file string
+ * @param {string} fileContent - Raw file content
+ * @returns {{frontmatter: object, content: string}}
+ */
+const parseFrontmatterAndContent = (fileContent) => {
+  const fmMatch = fileContent.match(/^---\n([\s\S]*?)\n---\n?([\s\S]*)$/);
+  if (!fmMatch) {
+    return { frontmatter: {}, content: fileContent };
+  }
+
+  // Parse YAML manually (simple key-value and arrays)
+  const yamlStr = fmMatch[1];
+  const content = fmMatch[2] || "";
+  const frontmatter = {};
+
+  let currentKey = null;
+  let inArray = false;
+
+  for (const line of yamlStr.split("\n")) {
+    // Array item
+    if (inArray && line.startsWith("  - ")) {
+      if (!Array.isArray(frontmatter[currentKey])) {
+        frontmatter[currentKey] = [];
+      }
+      frontmatter[currentKey].push(line.replace(/^\s+- /, "").trim());
+      continue;
+    }
+
+    // Key-value pair
+    const kvMatch = line.match(/^([a-zA-Z_][a-zA-Z0-9_]*)\s*:\s*(.*)$/);
+    if (kvMatch) {
+      currentKey = kvMatch[1];
+      const value = kvMatch[2].trim();
+      inArray = false;
+
+      if (value === "") {
+        // Could be start of array or empty value
+        inArray = true;
+        frontmatter[currentKey] = [];
+      } else {
+        // Strip quotes
+        frontmatter[currentKey] = value.replace(/^["']|["']$/g, "");
+      }
+    } else {
+      inArray = false;
+    }
+  }
+
+  return { frontmatter, content };
+};
+
+/**
+ * Clean up migrated post content
+ * Strips common cruft from old posts (date headers, webmention footers)
+ * @param {string} content - Raw markdown content
+ * @returns {string} Cleaned content
+ */
+const cleanMigratedContent = (content) => {
+  let cleaned = content;
+
+  // Strip leading date-link headers: [ January 12, 2019 ](https://...)
+  cleaned = cleaned.replace(
+    /^\s*\[\s*[A-Z][a-z]+ \d{1,2},? \d{4}\s*\]\s*\(https?:\/\/[^)]+\)\s*/,
+    "",
+  );
+
+  // Strip leading H2 that duplicates the title with a link
+  cleaned = cleaned.replace(/^\s*## \[[^\]]+\]\(https?:\/\/[^)]+\)\s*\n?/, "");
+
+  // Strip "X min read" after title headers
+  cleaned = cleaned.replace(/^\s*\d+ min read\s*\n?/, "");
+
+  // Strip trailing webmention reaction sections
+  // Common pattern: a horizontal rule followed by likes/reposts/mentions
+  cleaned = cleaned.replace(
+    /\n---+\s*\n[\s\S]*?(liked|reposted|replied|mentioned|bookmarked)\s+this\b[\s\S]*$/i,
+    "",
+  );
+
+  return cleaned.trim();
+};
+
+/**
+ * Import a migrated post from disk into MongoDB (on-demand)
+ * Called when the edit controller can't find the post in the database.
+ * @param {string} url - The post's mpUrl (used to derive file path)
+ * @param {object} publication - Publication config (has store, me)
+ * @param {object} application - Application config (has collections)
+ * @returns {Promise<{uid: string}|false>} The new post's uid, or false on failure
+ */
+export const importPostFromFile = async (url, publication, application) => {
+  try {
+    const postsCollection = application?.collections?.get("posts");
+    if (!postsCollection) {
+      console.log("[import] FAIL: no posts collection");
+      return false;
+    }
+
+    // Check if already imported (race condition guard)
+    const existing = await postsCollection.findOne({ "properties.url": url });
+    if (existing) {
+      console.log("[import] already in DB by url, uid=%s", existing._id);
+      return { uid: existing._id.toString() };
+    }
+
+    // Derive file path from URL
+    const relativePath = deriveFilePathFromUrl(url);
+    if (!relativePath) {
+      console.log("[import] FAIL: could not derive path from url=%s", url);
+      return false;
+    }
+    console.log("[import] relativePath=%s", relativePath);
+
+    // Get content directory from the store
+    const contentDir = publication.store?.options?.directory;
+    if (!contentDir) {
+      console.log(
+        "[import] FAIL: no contentDir. store=%o, store.options=%o",
+        typeof publication.store,
+        publication.store?.options,
+      );
+      return false;
+    }
+
+    const absolutePath = path.join(contentDir, relativePath);
+    console.log("[import] absolutePath=%s", absolutePath);
+
+    // Check if also already imported by path
+    const existingByPath = await postsCollection.findOne({
+      path: relativePath,
+    });
+    if (existingByPath) {
+      console.log("[import] already in DB by path, uid=%s", existingByPath._id);
+      return { uid: existingByPath._id.toString() };
+    }
+
+    // Read the file
+    let fileContent;
+    try {
+      fileContent = await fs.readFile(absolutePath, "utf-8");
+    } catch (readError) {
+      console.log(
+        "[import] FAIL: could not read file %s: %s",
+        absolutePath,
+        readError.message,
+      );
+      return false;
+    }
+    console.log("[import] file read OK, length=%d", fileContent.length);
+
+    // Parse frontmatter and content
+    const { frontmatter, content } = parseFrontmatterAndContent(fileContent);
+    console.log("[import] frontmatter keys=%s", Object.keys(frontmatter));
+
+    // Derive post-type from directory name
+    const dirName = relativePath.split("/")[0];
+    const postType = dirToPostType[dirName] || "note";
+
+    // Clean the content
+    const cleanedContent = cleanMigratedContent(content);
+
+    // Build JF2 properties
+    const properties = {
+      url,
+      "post-type": postType,
+      ...(frontmatter.date && { published: frontmatter.date }),
+      ...(frontmatter.title && { name: frontmatter.title }),
+      ...(cleanedContent && { content: { text: cleanedContent } }),
+      ...(frontmatter.category && {
+        category: Array.isArray(frontmatter.category)
+          ? frontmatter.category
+          : [frontmatter.category],
+      }),
+      ...(frontmatter.visibility && { visibility: frontmatter.visibility }),
+    };
+
+    // Insert into MongoDB
+    const result = await postsCollection.insertOne({
+      path: relativePath,
+      properties,
+    });
+
+    console.log("[import] SUCCESS: inserted uid=%s", result.insertedId);
+    return { uid: result.insertedId.toString() };
+  } catch (error) {
+    console.error("[import] ERROR:", error.message, error.stack);
+    return false;
+  }
+};

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@rmdes/indiekit-endpoint-posts",
-  "version": "1.0.0-beta.34",
+  "version": "1.0.0-beta.36",
   "description": "Post management endpoint for Indiekit with syndicate form fix. View posts published by your Micropub endpoint and publish new posts to it.",
   "keywords": [
     "indiekit",