mirror: allow fetching all usernames (#1293)

This is a minor change to the Discourse mirror so that it supports a
query to get all users from the server. It will be convenient for a
followon change which makes `update` search for every user's likes.

I also modified createGraph so that it uses the new method, which
results in code that is cleaner and slightly more efficient.

Test plan: Unit tests updated.

src/plugins/discourse/createGraph.js

@@ -130,17 +130,19 @@ export function createGraph(serverUrl: string, data: DiscourseData): Graph {
   const g = new Graph();
   const topicIdToTitle: Map<TopicId, string> = new Map();
 
+  for (const username of data.users()) {
+    g.addNode(userNode(serverUrl, username));
+  }
+
   for (const topic of data.topics()) {
     topicIdToTitle.set(topic.id, topic.title);
     g.addNode(topicNode(serverUrl, topic));
-    g.addNode(userNode(serverUrl, topic.authorUsername));
     g.addEdge(authorsTopicEdge(serverUrl, topic));
   }
 
   for (const post of data.posts()) {
     const topicTitle = topicIdToTitle.get(post.topicId) || "[unknown topic]";
     g.addNode(postNode(serverUrl, post, topicTitle));
-    g.addNode(userNode(serverUrl, post.authorUsername));
     g.addEdge(authorsPostEdge(serverUrl, post));
     g.addEdge(topicContainsPostEdge(serverUrl, post));
     let replyToPostIndex = post.replyToPostIndex;

src/plugins/discourse/createGraph.test.js

@@ -40,6 +40,16 @@ describe("plugins/discourse/createGraph", () => {
     posts(): $ReadOnlyArray<Post> {
       return this._posts;
     }
+    users(): $ReadOnlyArray<string> {
+      const users = new Set();
+      for (const {authorUsername} of this.posts()) {
+        users.add(authorUsername);
+      }
+      for (const {authorUsername} of this.topics()) {
+        users.add(authorUsername);
+      }
+      return Array.from(users);
+    }
     findPostInTopic(topicId: TopicId, indexWithinTopic: number): ?PostId {
       const post = this._posts.filter(
         (p) => p.topicId === topicId && p.indexWithinTopic === indexWithinTopic

src/plugins/discourse/mirror.js

@@ -13,7 +13,7 @@ import {
 
 // The version should be bumped any time the database schema is changed,
 // so that the cache will be properly invalidated.
-const VERSION = "discourse_mirror_v1";
+const VERSION = "discourse_mirror_v2";
 
 /**
  * An interface for retrieving all of the Discourse data at once.
@@ -45,6 +45,13 @@ export interface DiscourseData {
    * Returns undefined if no such post is available.
    */
   findPostInTopic(topicId: TopicId, indexWithinTopic: number): ?PostId;
+
+  /**
+   * Get usernames for all users.
+   *
+   * The order is unspecified.
+   */
+  users(): $ReadOnlyArray<string>;
 }
 
 /**
@@ -131,12 +138,14 @@ export class Mirror implements DiscourseData {
     db.prepare("INSERT INTO meta (zero, config) VALUES (0, ?)").run(config);
 
     const tables = [
+      "CREATE TABLE users (username TEXT PRIMARY KEY)",
       dedent`\
         CREATE TABLE topics (
             id INTEGER PRIMARY KEY,
             title TEXT NOT NULL,
             timestamp_ms INTEGER NOT NULL,
-            author_username TEXT NOT NULL
+            author_username TEXT NOT NULL,
+            FOREIGN KEY(author_username) REFERENCES users(username)
         )
       `,
       dedent`\
@@ -147,7 +156,8 @@ export class Mirror implements DiscourseData {
             topic_id INTEGER NOT NULL,
             index_within_topic INTEGER NOT NULL,
             reply_to_post_index INTEGER,
-            FOREIGN KEY(topic_id) REFERENCES topics(id)
+            FOREIGN KEY(topic_id) REFERENCES topics(id),
+            FOREIGN KEY(author_username) REFERENCES users(username)
         )
       `,
     ];
@@ -200,6 +210,13 @@ export class Mirror implements DiscourseData {
       }));
   }
 
+  users(): $ReadOnlyArray<string> {
+    return this._db
+      .prepare("SELECT username FROM users")
+      .pluck()
+      .all();
+  }
+
   findPostInTopic(topicId: TopicId, indexWithinTopic: number): ?PostId {
     return this._db
       .prepare(
@@ -237,6 +254,9 @@ export class Mirror implements DiscourseData {
         topicId,
         authorUsername,
       } = post;
+      db.prepare("INSERT OR IGNORE INTO users (username) VALUES (?)").run(
+        authorUsername
+      );
       db.prepare(
         dedent`\
           REPLACE INTO posts (
@@ -275,6 +295,9 @@ export class Mirror implements DiscourseData {
       if (topicWithPosts != null) {
         const {topic, posts} = topicWithPosts;
         const {id, title, timestampMs, authorUsername} = topic;
+        db.prepare("INSERT OR IGNORE INTO users (username) VALUES (?)").run(
+          authorUsername
+        );
         this._db
           .prepare(
             dedent`\

src/plugins/discourse/mirror.test.js

@@ -194,6 +194,21 @@ describe("plugins/discourse/mirror", () => {
     expect(mirror.posts()).toEqual(posts);
   });
 
+  it("provides usernames for all active users", async () => {
+    const {mirror, fetcher} = example();
+    fetcher.addPost(2, null, "alpha");
+    fetcher.addPost(3, null, "beta");
+    fetcher.addPost(3, 1, "alpha");
+    await mirror.update();
+    // credbot appears because it is the nominal author of all topics
+    expect(
+      mirror
+        .users()
+        .slice()
+        .sort()
+    ).toEqual(["alpha", "beta", "credbot"]);
+  });
+
   describe("update semantics", () => {
     it("only fetches new topics on `update`", async () => {
       const {mirror, fetcher} = example();