feat: migrate to weaviate

.dockerignore

@@ -21,6 +21,7 @@
 !/.npmrc
 !/tsconfig.json
 !/requirements.txt
+!/.nuxt
 
 # Ignore unnecessary files inside allowed directories
 # This should go after the allowed directories

Dockerfile

@@ -6,18 +6,21 @@ RUN pnpm install --frozen-lockfile
 COPY . .
 RUN pnpm build
 
-FROM python:3.11-slim
+FROM alpine:3.14 as weaviate
+ARG WEAVIATE_VERSION=1.24.8
+WORKDIR /app
+RUN wget https://github.com/weaviate/weaviate/releases/download/v${WEAVIATE_VERSION}/weaviate-v${WEAVIATE_VERSION}-linux-amd64.tar.gz && \
+  tar -xzf weaviate-v${WEAVIATE_VERSION}-linux-amd64.tar.gz
+
+FROM node:20-alpine
 ENV NUXT_DATA_PATH=/app/data
 ENV NUXT_MIGRATIONS_PATH=/app/migrations
 ENV NODE_ENV=production
 ENV NITRO_PORT=3000
 EXPOSE 3000
 WORKDIR /app
-RUN apt update -y && apt install curl git musl-dev -y && \
-  ln -s /usr/lib/x86_64-linux-musl/libc.so /lib/libc.musl-x86_64.so.1 && \
-  curl -sL https://deb.nodesource.com/setup_20.x | bash - && \
-  apt-get install -y nodejs && \
-  pip install chromadb
+RUN apk update && apk add git musl-dev
+COPY --from=weaviate /app/weaviate /bin/weaviate
 COPY docker/start.sh .
 COPY server/db/migrations /app/migrations
 COPY --from=builder /app/.output .output

docker-compose.yml

@@ -1,13 +1,27 @@
 version: '3.4'
 
 services:
-  chroma:
-    image: chromadb/chroma
+  weaviate:
+    command:
+      - --host
+      - 0.0.0.0
+      - --port
+      - '8080'
+      - --scheme
+      - http
+    image: cr.weaviate.io/semitechnologies/weaviate:1.24.8
     ports:
-      - 8000:8000
-    volumes:
-      - chroma_data:/var/lib/weaviate
+      - 8080:8080
     restart: on-failure:0
+    volumes:
+      - weaviate_data:/var/lib/weaviate
+    environment:
+      QUERY_DEFAULTS_LIMIT: 25
+      AUTHENTICATION_ANONYMOUS_ACCESS_ENABLED: 'true'
+      PERSISTENCE_DATA_PATH: '/var/lib/weaviate'
+      DEFAULT_VECTORIZER_MODULE: 'none'
+      ENABLE_MODULES: ''
+      CLUSTER_HOSTNAME: 'node1'
 
 volumes:
-  chroma_data:
+  weaviate_data:

docker/start.sh

@@ -1,6 +1,13 @@
-#!/bin/bash
+#!/bin/sh
 
 set -m # to make job control work
 node .output/server/index.mjs --port 3000 --host 0.0.0.0 &
-chroma run --path /app/data/chroma --port 8000 --host 0.0.0.0 &
+
+export PERSISTENCE_DATA_PATH=/app/vectorstore
+export QUERY_DEFAULTS_LIMIT=25
+export AUTHENTICATION_ANONYMOUS_ACCESS_ENABLED='true'
+export ENABLE_MODULES=''
+export DEFAULT_VECTORIZER_MODULE='none'
+export CLUSTER_HOSTNAME='node1'
+weaviate --port 8000 --host 0.0.0.0 --scheme http &
 fg %1 # gross!

nuxt.config.ts

@@ -4,7 +4,6 @@ export default defineNuxtConfig({
   telemetry: true,
   runtimeConfig: {
     ai: {
-      vectorDatabaseUrl: 'http://localhost:8000',
       token: '',
     },
     auth: {

package.json

@@ -40,8 +40,8 @@
     "@gitbeaker/rest": "^40.0.1",
     "@langchain/community": "^0.0.41",
     "@langchain/openai": "^0.0.23",
+    "@langchain/weaviate": "^0.0.1",
     "better-sqlite3": "^9.4.3",
-    "chromadb": "^1.8.1",
     "consola": "^3.2.3",
     "drizzle-orm": "^0.30.4",
     "glob": "^10.3.10",

server/api/forges/[forge_id]/index.delete.ts

@@ -1,4 +1,4 @@
-import { forgeSchema, userForgesSchema } from '~/server/schemas';
+import { forgeSchema, repoSchema, userForgesSchema } from '~/server/schemas';
 import { eq, and } from 'drizzle-orm';
 
 export default defineEventHandler(async (event) => {
@@ -15,11 +15,16 @@ export default defineEventHandler(async (event) => {
 
   await db.delete(userForgesSchema).where(eq(userForgesSchema.forgeId, forgeId)).run();
 
-  const forge = await db
+  await db
     .delete(forgeSchema)
     .where(and(eq(forgeSchema.owner, user.id), eq(forgeSchema.id, forgeId)))
     .returning()
     .get();
 
-  return forge;
+  const repos = await db.select().from(repoSchema).where(eq(repoSchema.forgeId, forgeId)).all();
+  for await (const repo of repos) {
+    await deleteRepo(repo.id);
+  }
+
+  return 'ok';
 });