@rmdes/indiekit-endpoint-conversations 2.3.1 → 2.3.2

package/lib/matching/syndication-map.js

@@ -33,10 +33,10 @@ export async function findCanonicalPost(application, syndicationUrl) {
 export async function resolveCanonicalUrl(application, targetUrl, siteUrl) {
   // If the target is already on our domain, it's likely canonical
   if (targetUrl.startsWith(siteUrl)) {
-    return targetUrl;
+    return targetUrl.replace(/\/$/, "");
   }
 
   // Otherwise try to find via syndication reverse lookup
   const canonical = await findCanonicalPost(application, targetUrl);
-  return canonical || targetUrl;
+  return (canonical || targetUrl).replace(/\/$/, "");
 }

package/lib/polling/scheduler.js

@@ -95,6 +95,9 @@ export async function runPollCycle(indiekit, options) {
     await pollActivityPub(indiekit, stateCollection, state);
   }
 
+  // Normalize trailing slashes in canonical_url (one-time migration)
+  await normalizeCanonicalUrls(indiekit, stateCollection);
+
   // Backfill missing avatars from ap_notifications (one-time sweep per cycle)
   await backfillMissingAvatars(indiekit, stateCollection);
 
@@ -102,6 +105,78 @@ export async function runPollCycle(indiekit, options) {
   await backfillPlatformNames(indiekit, stateCollection);
 }
 
+/**
+ * One-time migration: normalize trailing slashes in canonical_url.
+ * Strips trailing slash from all canonical_url values, then removes
+ * duplicates that were created by the slash inconsistency.
+ */
+async function normalizeCanonicalUrls(indiekit, stateCollection) {
+  try {
+    const state = await stateCollection.findOne({ _id: "poll_cursors" });
+    if (state?.canonical_url_normalized) return;
+
+    const itemsCollection = indiekit.collections.get("conversation_items");
+    if (!itemsCollection) return;
+
+    // Find all items with trailing slash in canonical_url
+    const itemsWithSlash = await itemsCollection
+      .find({ canonical_url: /\/$/ })
+      .toArray();
+
+    if (itemsWithSlash.length === 0) {
+      await stateCollection.findOneAndUpdate(
+        { _id: "poll_cursors" },
+        { $set: { canonical_url_normalized: true } },
+        { upsert: true },
+      );
+      return;
+    }
+
+    let updated = 0;
+    let deduped = 0;
+
+    for (const item of itemsWithSlash) {
+      const normalized = item.canonical_url.replace(/\/$/, "");
+
+      // Check if a non-slash version already exists with same platform_id
+      const existing = await itemsCollection.findOne({
+        canonical_url: normalized,
+        platform_id: item.platform_id,
+      });
+
+      if (existing) {
+        // Duplicate — remove the trailing-slash version
+        await itemsCollection.deleteOne({ _id: item._id });
+        deduped++;
+      } else {
+        // No duplicate — just normalize the URL
+        await itemsCollection.updateOne(
+          { _id: item._id },
+          { $set: { canonical_url: normalized } },
+        );
+        updated++;
+      }
+    }
+
+    if (updated > 0 || deduped > 0) {
+      console.info(
+        `[Conversations] URL normalization: updated ${updated}, removed ${deduped} duplicates from ${itemsWithSlash.length} items with trailing slashes`,
+      );
+    }
+
+    await stateCollection.findOneAndUpdate(
+      { _id: "poll_cursors" },
+      { $set: { canonical_url_normalized: true } },
+      { upsert: true },
+    );
+  } catch (error) {
+    console.warn(
+      "[Conversations] URL normalization error:",
+      error.message,
+    );
+  }
+}
+
 /**
  * Backfill empty author.photo fields in conversation_items.
  * Tries four strategies in order:

package/lib/storage/conversation-items.js

@@ -22,6 +22,11 @@ function getCollection(application) {
 export async function upsertConversationItem(application, item) {
   const collection = getCollection(application);
 
+  // Normalize canonical_url — strip trailing slash for consistent deduplication
+  if (item.canonical_url) {
+    item.canonical_url = item.canonical_url.replace(/\/$/, "");
+  }
+
   const result = await collection.findOneAndUpdate(
     {
       canonical_url: item.canonical_url,

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@rmdes/indiekit-endpoint-conversations",
-  "version": "2.3.1",
+  "version": "2.3.2",
   "description": "Conversation aggregation endpoint for Indiekit. Backend enrichment service that polls Mastodon/Bluesky notifications and serves JF2-compatible data for the interactions page.",
   "keywords": [
     "indiekit",