Store embed titles and search them

media-scrape/test/run.ts

@@ -29,6 +29,7 @@ import { isoFromSnowflake, snowflakeFromIso } from "../src/snowflake.js";
 import {
     dirSizeBytes,
     listScrapes,
+    normalizeTitleKey,
     readScrapeMeta,
     ScrapeStore,
     writeScrapeMeta
@@ -1347,6 +1348,61 @@ void (async () => {
         store.close();
     });
 
+    await test("title search matches word prefixes and exact phrases", () => {
+        const dir = join(root, "title-search");
+        mkdirSync(dir, { recursive: true });
+        const store = new ScrapeStore(dir);
+        const embed = (url: string, title: string) => ({
+            type: "video",
+            url,
+            title,
+            provider: "YouTube",
+            videoUrl: url,
+            imageUrl: null,
+            thumbnailUrl: null
+        });
+
+        store.insertPage([
+            message("1000", {
+                attachments: [attachment({ filename: "no-title.png", contentType: "image/png" })],
+                embeds: [
+                    embed("https://www.youtube.com/watch?v=a", "Man who jumps over car"),
+                    embed("https://www.youtube.com/watch?v=b", "Human cannonball"),
+                    embed("https://www.youtube.com/watch?v=c", "Woman's Day: 3 Things!")
+                ]
+            })
+        ]);
+
+        const ids = (options: Record<string, unknown>) =>
+            store
+                .queryMedia({ limit: 50, ...options })
+                .map(row => row.url.split("v=")[1])
+                .sort();
+
+        // Prefix tokens in any order, but only at word starts.
+        assert.deepEqual(ids({ titleQuery: "Ma Ca Ju" }), ["a"]);
+        assert.deepEqual(ids({ titleQuery: "man" }), ["a"]);
+        assert.deepEqual(ids({ titleQuery: "man car" }), ["a"]);
+        assert.deepEqual(ids({ titleQuery: "cannon" }), ["b"]);
+        assert.deepEqual(ids({ titleQuery: "hum" }), ["b"]);
+        assert.deepEqual(ids({ titleQuery: "an" }), []);
+
+        // Exact mode is one phrase.
+        assert.deepEqual(ids({ titleQuery: "man who jumps", titleExact: true }), ["a"]);
+        assert.deepEqual(ids({ titleQuery: "man car", titleExact: true }), []);
+
+        // Punctuation folds away.
+        assert.deepEqual(ids({ titleQuery: "woman's day" }), ["c"]);
+        assert.deepEqual(ids({ titleQuery: "3 things" }), ["c"]);
+
+        store.close();
+    });
+
+    await test("normalizeTitleKey collapses punctuation and case", () => {
+        assert.equal(normalizeTitleKey("Woman's Day: 3 Things!"), "woman s day 3 things");
+        assert.equal(normalizeTitleKey("  A---B  "), "a b");
+    });
+
     await test("provider filters are the curated archive list", () => {
         assert.ok(PROVIDER_FILTERS.length >= 12, `provider list has ${PROVIDER_FILTERS.length} entries`);
         assert.equal(new Set(PROVIDER_FILTERS).size, PROVIDER_FILTERS.length);