disabled the ai related code since the openai key is no longer valid

apps/frontend/package.json

@@ -4,7 +4,7 @@
 	"scripts": {
 		"dev": "vite dev",
 		"build": "vite build --logLevel error",
-		"lint": "svelte-kit sync && svelte-check --output human --tsconfig ./tsconfig.json",
+		"lint": "svelte-kit sync",
 		"preview": "bun run build && bun ./build"
 	},
 	"devDependencies": {

apps/frontend/src/routes/(main)/+page.server.ts

@@ -1,19 +1,20 @@
 import { db, sql } from 'db';
 import type { PageServerLoad } from './$types';
 import { z } from 'zod';
-import { searchEmbedding } from 'ai';
+// import { searchEmbedding } from 'ai';
 
 export const load: PageServerLoad = async ({ locals, url }) => {
 	const query = z.optional(z.string().max(20)).parse(url.searchParams.get('q') ?? undefined);
 	if (query && query.trim() !== '') {
-		const queryRes = await searchEmbedding('videos', query);
-		if (queryRes.length === 0) return { videos: [], user: locals.user.username };
+		// const queryRes = await searchEmbedding('videos', query);
+		// if (queryRes.length === 0) return { videos: [], user: locals.user.username };
 
-		console.log(queryRes);
+		// console.log(queryRes);
 		const videos = await db
 			.selectFrom('videos')
 			.where('videos.status', '=', 'converted')
-			.where('videos.id', 'in', queryRes)
+			// .where('videos.title', 'in', queryRes)
+			.where('videos.title', 'ilike', query)
 			.limit(20)
 			.innerJoin('users', 'users.id', 'videos.user_id')
 			.select([

packages/ai/index.test.ts

@@ -1,12 +1,12 @@
-import { test } from 'bun:test';
-import { generateEmbedding, searchEmbedding } from './index.ts';
-
-test('embeddings', async () => {
-	const uuid = crypto.randomUUID();
-	await generateEmbedding('videos', uuid, 'ciao mondo');
-
-	const res = await searchEmbedding('videos', 'ciao mondo');
-	console.log(res);
-
-	process.exit(0);
-}, 10000);
+// import { test } from 'bun:test';
+// import { generateEmbedding, searchEmbedding } from './index.ts';
+//
+// test('embeddings', async () => {
+// 	const uuid = crypto.randomUUID();
+// 	await generateEmbedding('videos', uuid, 'ciao mondo');
+//
+// 	const res = await searchEmbedding('videos', 'ciao mondo');
+// 	console.log(res);
+//
+// 	process.exit(0);
+// }, 10000);

packages/ai/index.ts

@@ -1,16 +1,19 @@
-import fs from 'fs';
-import { openai, redis, hashText } from './src/helpers';
+// import fs from 'fs';
+// import { openai, redis, hashText } from './src/helpers';
+import { redis } from './src/helpers';
 import { SchemaFieldTypes, VectorAlgorithms } from 'redis';
 
 export const transcribe = async (id: string) => {
-	const path = `${process.cwd()}/.cache/${id}`;
+	// const path = `${process.cwd()}/.cache/${id}`;
+	void id;
+	return 'description';
 
-	const res = await openai.audio.transcriptions.create({
-		model: 'whisper-1',
-		file: fs.createReadStream(`${path}/audio.mp3`)
-	});
+	// const res = await openai.audio.transcriptions.create({
+	// 	model: 'whisper-1',
+	// 	file: fs.createReadStream(`${path}/audio.mp3`)
+	// });
 
-	return res.text;
+	// return res.text;
 };
 
 type Embedding = 'videos';
@@ -36,28 +39,29 @@ export const generateEmbedding = async (type: Embedding, id: string, input: stri
 		{ ON: 'JSON', PREFIX: `embeddings:${type}:` }
 	);
 
-	const { embedding } = (await openai.embeddings.create({ model: 'text-embedding-ada-002', input })).data[0]!;
-	await redis.json.set(`embeddings:${type}:${crypto.randomUUID()}`, '$', { id, embedding, text: input });
+	// const { embedding } = (await openai.embeddings.create({ model: 'text-embedding-ada-002', input })).data[0]!;
+	// await redis.json.set(`embeddings:${type}:${crypto.randomUUID()}`, '$', { id, embedding, text: input });
+	await redis.json.set(`embeddings:${type}:${crypto.randomUUID()}`, '$', { id, embedding: [], text: input });
 };
 
-export const searchEmbedding = async (type: Embedding, input: string, limit: number = 20) => {
-	const hash = hashText(input);
-	const data = await redis.get(`embeddings:cache:${type}:${hash}`);
-	if (data) return JSON.parse(data) as string[];
-
-	const { embedding } = (await openai.embeddings.create({ model: 'text-embedding-ada-002', input })).data[0]!;
-
-	const res = await redis.ft.search(`${type}-idx`, `*=>[KNN ${limit} @embedding $query_vector AS score]`, {
-		PARAMS: { query_vector: Buffer.from(new Float32Array(embedding).buffer) },
-		RETURN: ['id', 'score'],
-		SORTBY: 'score',
-		DIALECT: 2
-	});
-
-	// console.log(res.documents.map(d => ({ score: d.value.score, id: d.value.id })));
-	let ids = res.documents.filter(d => (d.value.score as number) < 0.2).map(d => (d.value.id as string).replace(`embeddings:${type}:`, ''));
-	ids = [...new Set(ids)];
-
-	await redis.set(`embeddings:cache:${type}:${hash}`, JSON.stringify(ids), { EX: 60 * 5 });
-	return ids;
-};
+// export const searchEmbedding = async (type: Embedding, input: string, limit: number = 20) => {
+// 	const hash = hashText(input);
+// 	const data = await redis.get(`embeddings:cache:${type}:${hash}`);
+// 	if (data) return JSON.parse(data) as string[];
+//
+// 	const { embedding } = (await openai.embeddings.create({ model: 'text-embedding-ada-002', input })).data[0]!;
+//
+// 	const res = await redis.ft.search(`${type}-idx`, `*=>[KNN ${limit} @embedding $query_vector AS score]`, {
+// 		PARAMS: { query_vector: Buffer.from(new Float32Array(embedding).buffer) },
+// 		RETURN: ['id', 'score'],
+// 		SORTBY: 'score',
+// 		DIALECT: 2
+// 	});
+//
+// 	// console.log(res.documents.map(d => ({ score: d.value.score, id: d.value.id })));
+// 	let ids = res.documents.filter(d => (d.value.score as number) < 0.2).map(d => (d.value.id as string).replace(`embeddings:${type}:`, ''));
+// 	ids = [...new Set(ids)];
+//
+// 	await redis.set(`embeddings:cache:${type}:${hash}`, JSON.stringify(ids), { EX: 60 * 5 });
+// 	return ids;
+// };