Update Bluesky component versions and add text encoding functionality

- Bump version of @pipedream/bluesky to 0.1.2
- Update action versions for create-post (0.1.0), like-post (0.0.2), retrieve-thread (0.0.2), new-follower-on-account (0.0.2), new-posts-by-author (0.0.2), and new-timeline-posts (0.0.2)
- Introduce textEncoding module for detecting facets in post content
- Remove deprecated methods from create-post action

Author: luancazarine

components/bluesky/actions/create-post/create-post.mjs

@@ -1,11 +1,12 @@
 import app from "../../bluesky.app.mjs";
 import constants from "../../common/constants.mjs";
+import textEncoding from "../../common/textEncoding.mjs";
 
 export default {
   key: "bluesky-create-post",
   name: "Create Post",
   description: "Creates a new post on Bluesky. [See the documentation](https://docs.bsky.app/docs/api/com-atproto-repo-create-record).",
-  version: "0.0.2",
+  version: "0.1.0",
   type: "action",
   props: {
     app,
@@ -15,40 +16,6 @@ export default {
       description: "The text content of the post.",
     },
   },
-  methods: {
-    parseUrls(text) {
-      const spans = [];
-      const urlRegex = /(?:[$|\W])(https?:\/\/(www\.)?[-a-zA-Z0-9@:%._+~#=]{1,256}\.[a-zA-Z0-9()]{1,6}\b([-a-zA-Z0-9()@:%_+.~#?&//=]*[-a-zA-Z0-9@%_+~#//=])?)/g;
-
-      let match;
-      while ((match = urlRegex.exec(text)) !== null) {
-        spans.push({
-          start: match.index + 1,
-          end: urlRegex.lastIndex,
-          url: match[1],
-        });
-      }
-      return spans;
-    },
-    parseFacets(text) {
-      const facets = [];
-      for (const link of this.parseUrls(text)) {
-        facets.push({
-          index: {
-            byteStart: link["start"],
-            byteEnd: link["end"],
-          },
-          features: [
-            {
-              ["$type"]: "app.bsky.richtext.facet#link",
-              uri: link["url"],
-            },
-          ],
-        });
-      }
-      return facets;
-    },
-  },
   async run({ $ }) {
     const {
       app,
@@ -62,7 +29,7 @@ export default {
         record: {
           ["$type"]: constants.RESOURCE_TYPE.POST,
           text,
-          facets: this.parseFacets(text),
+          facets: await textEncoding.detectFacets(text, app),
           createdAt: new Date().toISOString(),
         },
       },

components/bluesky/actions/like-post/like-post.mjs

@@ -5,7 +5,7 @@ export default {
   key: "bluesky-like-post",
   name: "Like Post",
   description: "Like a specific post on Bluesky. [See the documentation](https://docs.bsky.app/docs/api/com-atproto-repo-create-record).",
-  version: "0.0.1",
+  version: "0.0.2",
   type: "action",
   props: {
     app,

components/bluesky/actions/retrieve-thread/retrieve-thread.mjs

@@ -4,7 +4,7 @@ export default {
   key: "bluesky-retrieve-thread",
   name: "Retrieve Thread",
   description: "Retrieve a full thread of posts. [See the documentation](https://docs.bsky.app/docs/api/app-bsky-feed-get-post-thread).",
-  version: "0.0.1",
+  version: "0.0.2",
   type: "action",
   props: {
     app,

components/bluesky/common/textEncoding.mjs

@@ -0,0 +1,127 @@
+import TLDs from "./tlds.mjs";
+
+function isValidDomain(str) {
+  return !!TLDs.find((tld) => {
+    const i = str.lastIndexOf(tld);
+    if (i === -1) {
+      return false;
+    }
+    return str.charAt(i - 1) === "." && i === str.length - tld.length;
+  });
+}
+
+function utf16IndexToUtf8Index(i, utf16) {
+  const encoder = new TextEncoder();
+  return encoder.encode(utf16.slice(0, i)).byteLength;
+}
+
+async function detectFacets(text, app) {
+  let match;
+  const facets = [];
+
+  // mentions
+  const re1 = /(^|\s|\()(@)([a-zA-Z0-9.-]+)(\b)/g;
+  while ((match = re1.exec(text))) {
+    if (!isValidDomain(match[3]) && !match[3].endsWith(".test")) {
+      continue; // probably not a handle
+    }
+
+    const start = text.indexOf(match[3], match.index) - 1;
+    const handle = await app.resolveHandle({
+      params: {
+        handle: match[3],
+      },
+    });
+    facets.push({
+      $type: "app.bsky.richtext.facet",
+      index: {
+        byteStart: utf16IndexToUtf8Index(start, text),
+        byteEnd: utf16IndexToUtf8Index(start + match[3].length + 1, text),
+      },
+      features: [
+        {
+          $type: "app.bsky.richtext.facet#mention",
+          did: handle.did, // must be resolved afterwards
+        },
+      ],
+    });
+  }
+
+  // links
+  const re2 =
+      /(^|\s|\()((https?:\/\/[\S]+)|((?<domain>[a-z][a-z0-9]*(\.[a-z0-9]+)+)[\S]*))/gim;
+  while ((match = re2.exec(text))) {
+    let uri = match[2];
+    if (!uri.startsWith("http")) {
+      const domain = match.groups?.domain;
+      if (!domain || !isValidDomain(domain)) {
+        continue;
+      }
+      uri = `https://${uri}`;
+    }
+    const start = text.indexOf(match[2], match.index);
+    const index = {
+      start,
+      end: start + match[2].length,
+    };
+      // strip ending puncuation
+    if (/[.,;!?]$/.test(uri)) {
+      uri = uri.slice(0, -1);
+      index.end--;
+    }
+    if (/[)]$/.test(uri) && !uri.includes("(")) {
+      uri = uri.slice(0, -1);
+      index.end--;
+    }
+    facets.push({
+      index: {
+        byteStart: utf16IndexToUtf8Index(index.start, text),
+        byteEnd: utf16IndexToUtf8Index(index.end, text),
+      },
+      features: [
+        {
+          $type: "app.bsky.richtext.facet#link",
+          uri,
+        },
+      ],
+    });
+  }
+
+  const re3 = /(?:^|\s)(#[^\d\s]\S*)(?=\s)?/g;
+  while ((match = re3.exec(text))) {
+    let [
+      tag,
+    ] = match;
+    const hasLeadingSpace = /^\s/.test(tag);
+
+    tag = tag.trim().replace(/\p{P}+$/gu, ""); // strip ending punctuation
+
+    // inclusive of #, max of 64 chars
+    if (tag.length > 66) continue;
+
+    const index = match.index + (hasLeadingSpace
+      ? 1
+      : 0);
+
+    facets.push({
+      index: {
+        byteStart: utf16IndexToUtf8Index(index, text),
+        byteEnd: utf16IndexToUtf8Index(index + tag.length, text), // inclusive of last char
+      },
+      features: [
+        {
+          $type: "app.bsky.richtext.facet#tag",
+          tag: tag.replace(/^#/, ""),
+        },
+      ],
+    });
+  }
+
+  return facets.length > 0
+    ? facets
+    : undefined;
+}
+
+export default {
+  detectFacets,
+};