feat: 🎸 local

all local in 60 seconds
This commit is contained in:
Stan Girard 2023-12-02 15:43:05 +01:00
parent b40631a0ac
commit 8382dde575
5 changed files with 967 additions and 79 deletions

View File

@ -6,6 +6,9 @@ CREATE_FIRST_USER=true # set to false if you want to create the first user manua
# login: admin@quivr.app
# password: admin
#LOCAL
#OLLAMA_API_BASE_URL=http://host.docker.internal:11434 # Uncomment to activate ollama. This is the local url for the ollama api
@ -39,9 +42,6 @@ GOOGLE_CLOUD_PROJECT=<ignore-me>
CELERY_BROKER_URL=redis://redis:6379/0
CELEBRY_BROKER_QUEUE_NAME=quivr-preview.fifo
#LOCAL
#OLLAMA_API_BASE_URL=http://host.docker.internal:11434 # local all in one remove comment to use local llm with Ollama
#RESEND
RESEND_API_KEY=<change-me>

View File

@ -70,16 +70,18 @@ Ensure you have the following installed:
> Want to use [Ollama.ai](https://ollama.ai) instead?
> Finish installing Quivr, put a randon open_api_key in the `.env` file and
> follow the instructions [here](https://brain.quivr.app/docs/Developers/selfHosted/run_fully_local) to update Quivr to use Ollama AFTER you have finished installing Quivr.
> Uncomment the following lines in the `.env` file:
> OLLAMA_API_BASE_URL
> Run the following command to start Ollama: `ollama run llama2`
> You can find more information about Ollama [here](https://ollama.ai/).
- **Step 4**: Launch the project
```bash
docker compose pull
docker compose up --build
```
docker compose up --build # if OPENAI
# docker compose -f docker-compose-ollama.yml up --build # Only if using Ollama. You need to run `ollama run llama2` first.
```
- **Step 5**: Login to the app

505
docker-compose-ollama.yml Normal file
View File

@ -0,0 +1,505 @@
version: "3.8"
services:
frontend:
pull_policy: never
build:
context: frontend
dockerfile: Dockerfile
args:
- NEXT_PUBLIC_ENV=local
- NEXT_PUBLIC_BACKEND_URL=${NEXT_PUBLIC_BACKEND_URL}
- NEXT_PUBLIC_SUPABASE_URL=${NEXT_PUBLIC_SUPABASE_URL}
- NEXT_PUBLIC_SUPABASE_ANON_KEY=${NEXT_PUBLIC_SUPABASE_ANON_KEY}
- NEXT_PUBLIC_CMS_URL=${NEXT_PUBLIC_CMS_URL}
- NEXT_PUBLIC_FRONTEND_URL=${NEXT_PUBLIC_FRONTEND_URL}
container_name: web
depends_on:
- backend-core
restart: always
ports:
- 3000:3000
backend-core:
image: backend-base
pull_policy: never
env_file:
- .env
build:
context: backend
dockerfile: Dockerfile
container_name: backend-core
command:
- "uvicorn"
- "main:app"
- "--host"
- "0.0.0.0"
- "--port"
- "5050"
- "--workers"
- "1"
restart: always
depends_on:
db:
condition: service_healthy
kong:
condition: service_healthy
ports:
- 5050:5050
redis:
image: redis:latest
container_name: redis
restart: always
ports:
- 6379:6379
worker:
pull_policy: never
image: backend-base
env_file:
- .env
build:
context: backend
dockerfile: Dockerfile
container_name: worker
command: celery -A celery_worker worker -l info
restart: always
depends_on:
- redis
beat:
image: backend-base
pull_policy: never
env_file:
- .env
build:
context: backend
dockerfile: Dockerfile
container_name: beat
command: celery -A celery_worker beat -l info
restart: always
depends_on:
- redis
flower:
image: backend-base
pull_policy: never
env_file:
- .env
build:
context: backend
dockerfile: Dockerfile
container_name: flower
command: celery -A celery_worker flower -l info --port=5555
restart: always
depends_on:
- redis
- worker
- beat
ports:
- 5555:5555
studio:
container_name: supabase-studio
image: supabase/studio:20231123-64a766a
restart: unless-stopped
healthcheck:
test:
[
"CMD",
"node",
"-e",
"require('http').get('http://localhost:3000/api/profile', (r) => {if (r.statusCode !== 200) throw new Error(r.statusCode)})"
]
timeout: 5s
interval: 5s
retries: 3
depends_on:
analytics:
condition: service_healthy
environment:
STUDIO_PG_META_URL: http://meta:8080
POSTGRES_PASSWORD: ${POSTGRES_PASSWORD}
DEFAULT_ORGANIZATION_NAME: ${STUDIO_DEFAULT_ORGANIZATION}
DEFAULT_PROJECT_NAME: ${STUDIO_DEFAULT_PROJECT}
SUPABASE_URL: http://kong:8000
SUPABASE_PUBLIC_URL: ${SUPABASE_PUBLIC_URL}
SUPABASE_ANON_KEY: ${ANON_KEY}
SUPABASE_SERVICE_KEY: ${SERVICE_ROLE_KEY}
LOGFLARE_API_KEY: ${LOGFLARE_API_KEY}
LOGFLARE_URL: http://analytics:4000
NEXT_PUBLIC_ENABLE_LOGS: true
# Comment to use Big Query backend for analytics
NEXT_ANALYTICS_BACKEND_PROVIDER: postgres
# Uncomment to use Big Query backend for analytics
# NEXT_ANALYTICS_BACKEND_PROVIDER: bigquery
kong:
container_name: supabase-kong
image: kong:2.8.1
restart: unless-stopped
# https://unix.stackexchange.com/a/294837
entrypoint: bash -c 'eval "echo \"$$(cat ~/temp.yml)\"" > ~/kong.yml && /docker-entrypoint.sh kong docker-start'
ports:
- ${KONG_HTTP_PORT}:8000/tcp
- ${KONG_HTTPS_PORT}:8443/tcp
depends_on:
analytics:
condition: service_healthy
environment:
KONG_DATABASE: "off"
KONG_DECLARATIVE_CONFIG: /home/kong/kong.yml
# https://github.com/supabase/cli/issues/14
KONG_DNS_ORDER: LAST,A,CNAME
KONG_PLUGINS: request-transformer,cors,key-auth,acl,basic-auth
KONG_NGINX_PROXY_PROXY_BUFFER_SIZE: 160k
KONG_NGINX_PROXY_PROXY_BUFFERS: 64 160k
SUPABASE_ANON_KEY: ${ANON_KEY}
SUPABASE_SERVICE_KEY: ${SERVICE_ROLE_KEY}
DASHBOARD_USERNAME: ${DASHBOARD_USERNAME}
DASHBOARD_PASSWORD: ${DASHBOARD_PASSWORD}
volumes:
# https://github.com/supabase/supabase/issues/12661
- ./volumes/api/kong.yml:/home/kong/temp.yml:ro
auth:
container_name: supabase-auth
image: supabase/gotrue:v2.99.0
depends_on:
db:
# Disable this if you are using an external Postgres database
condition: service_healthy
analytics:
condition: service_healthy
healthcheck:
test:
[
"CMD",
"wget",
"--no-verbose",
"--tries=1",
"--spider",
"http://localhost:9999/health"
]
timeout: 5s
interval: 5s
retries: 3
restart: unless-stopped
environment:
GOTRUE_API_HOST: 0.0.0.0
GOTRUE_API_PORT: 9999
API_EXTERNAL_URL: ${API_EXTERNAL_URL}
GOTRUE_DB_DRIVER: postgres
GOTRUE_DB_DATABASE_URL: postgres://supabase_auth_admin:${POSTGRES_PASSWORD}@${POSTGRES_HOST}:${POSTGRES_PORT}/${POSTGRES_DB}
GOTRUE_SITE_URL: ${SITE_URL}
GOTRUE_URI_ALLOW_LIST: ${ADDITIONAL_REDIRECT_URLS}
GOTRUE_DISABLE_SIGNUP: ${DISABLE_SIGNUP}
GOTRUE_JWT_ADMIN_ROLES: service_role
GOTRUE_JWT_AUD: authenticated
GOTRUE_JWT_DEFAULT_GROUP_NAME: authenticated
GOTRUE_JWT_EXP: ${JWT_EXPIRY}
GOTRUE_JWT_SECRET: ${JWT_SECRET}
GOTRUE_EXTERNAL_EMAIL_ENABLED: ${ENABLE_EMAIL_SIGNUP}
GOTRUE_MAILER_AUTOCONFIRM: ${ENABLE_EMAIL_AUTOCONFIRM}
# GOTRUE_MAILER_SECURE_EMAIL_CHANGE_ENABLED: true
# GOTRUE_SMTP_MAX_FREQUENCY: 1s
GOTRUE_SMTP_ADMIN_EMAIL: ${SMTP_ADMIN_EMAIL}
GOTRUE_SMTP_HOST: ${SMTP_HOST}
GOTRUE_SMTP_PORT: ${SMTP_PORT}
GOTRUE_SMTP_USER: ${SMTP_USER}
GOTRUE_SMTP_PASS: ${SMTP_PASS}
GOTRUE_SMTP_SENDER_NAME: ${SMTP_SENDER_NAME}
GOTRUE_MAILER_URLPATHS_INVITE: ${MAILER_URLPATHS_INVITE}
GOTRUE_MAILER_URLPATHS_CONFIRMATION: ${MAILER_URLPATHS_CONFIRMATION}
GOTRUE_MAILER_URLPATHS_RECOVERY: ${MAILER_URLPATHS_RECOVERY}
GOTRUE_MAILER_URLPATHS_EMAIL_CHANGE: ${MAILER_URLPATHS_EMAIL_CHANGE}
GOTRUE_EXTERNAL_PHONE_ENABLED: ${ENABLE_PHONE_SIGNUP}
GOTRUE_SMS_AUTOCONFIRM: ${ENABLE_PHONE_AUTOCONFIRM}
rest:
container_name: supabase-rest
image: postgrest/postgrest:v11.2.2
depends_on:
db:
# Disable this if you are using an external Postgres database
condition: service_healthy
analytics:
condition: service_healthy
restart: unless-stopped
environment:
PGRST_DB_URI: postgres://authenticator:${POSTGRES_PASSWORD}@${POSTGRES_HOST}:${POSTGRES_PORT}/${POSTGRES_DB}
PGRST_DB_SCHEMAS: ${PGRST_DB_SCHEMAS}
PGRST_DB_ANON_ROLE: anon
PGRST_JWT_SECRET: ${JWT_SECRET}
PGRST_DB_USE_LEGACY_GUCS: "false"
PGRST_APP_SETTINGS_JWT_SECRET: ${JWT_SECRET}
PGRST_APP_SETTINGS_JWT_EXP: ${JWT_EXPIRY}
command: "postgrest"
realtime:
# This container name looks inconsistent but is correct because realtime constructs tenant id by parsing the subdomain
container_name: realtime-dev.supabase-realtime
image: supabase/realtime:v2.25.35
depends_on:
db:
# Disable this if you are using an external Postgres database
condition: service_healthy
analytics:
condition: service_healthy
healthcheck:
test:
[
"CMD",
"bash",
"-c",
"printf \\0 > /dev/tcp/localhost/4000"
]
timeout: 5s
interval: 5s
retries: 3
restart: unless-stopped
environment:
PORT: 4000
DB_HOST: ${POSTGRES_HOST}
DB_PORT: ${POSTGRES_PORT}
DB_USER: supabase_admin
DB_PASSWORD: ${POSTGRES_PASSWORD}
DB_NAME: ${POSTGRES_DB}
DB_AFTER_CONNECT_QUERY: 'SET search_path TO _realtime'
DB_ENC_KEY: supabaserealtime
API_JWT_SECRET: ${JWT_SECRET}
FLY_ALLOC_ID: fly123
FLY_APP_NAME: realtime
SECRET_KEY_BASE: UpNVntn3cDxHJpq99YMc1T1AQgQpc8kfYTuRgBiYa15BLrx8etQoXz3gZv1/u2oq
ERL_AFLAGS: -proto_dist inet_tcp
ENABLE_TAILSCALE: "false"
DNS_NODES: "''"
command: >
sh -c "/app/bin/migrate && /app/bin/realtime eval 'Realtime.Release.seeds(Realtime.Repo)' && /app/bin/server"
storage:
container_name: supabase-storage
image: supabase/storage-api:v0.43.11
depends_on:
db:
# Disable this if you are using an external Postgres database
condition: service_healthy
rest:
condition: service_started
imgproxy:
condition: service_started
healthcheck:
test:
[
"CMD",
"wget",
"--no-verbose",
"--tries=1",
"--spider",
"http://localhost:5000/status"
]
timeout: 5s
interval: 5s
retries: 3
restart: unless-stopped
environment:
ANON_KEY: ${ANON_KEY}
SERVICE_KEY: ${SERVICE_ROLE_KEY}
POSTGREST_URL: http://rest:3000
PGRST_JWT_SECRET: ${JWT_SECRET}
DATABASE_URL: postgres://supabase_storage_admin:${POSTGRES_PASSWORD}@${POSTGRES_HOST}:${POSTGRES_PORT}/${POSTGRES_DB}
FILE_SIZE_LIMIT: 52428800
STORAGE_BACKEND: file
FILE_STORAGE_BACKEND_PATH: /var/lib/storage
TENANT_ID: stub
# TODO: https://github.com/supabase/storage-api/issues/55
REGION: stub
GLOBAL_S3_BUCKET: stub
ENABLE_IMAGE_TRANSFORMATION: "true"
IMGPROXY_URL: http://imgproxy:5001
volumes:
- ./volumes/storage:/var/lib/storage:z
imgproxy:
container_name: supabase-imgproxy
image: darthsim/imgproxy:v3.8.0
healthcheck:
test: [ "CMD", "imgproxy", "health" ]
timeout: 5s
interval: 5s
retries: 3
environment:
IMGPROXY_BIND: ":5001"
IMGPROXY_LOCAL_FILESYSTEM_ROOT: /
IMGPROXY_USE_ETAG: "true"
IMGPROXY_ENABLE_WEBP_DETECTION: ${IMGPROXY_ENABLE_WEBP_DETECTION}
volumes:
- ./volumes/storage:/var/lib/storage:z
meta:
container_name: supabase-meta
image: supabase/postgres-meta:v0.68.0
depends_on:
db:
# Disable this if you are using an external Postgres database
condition: service_healthy
analytics:
condition: service_healthy
restart: unless-stopped
environment:
PG_META_PORT: 8080
PG_META_DB_HOST: ${POSTGRES_HOST}
PG_META_DB_PORT: ${POSTGRES_PORT}
PG_META_DB_NAME: ${POSTGRES_DB}
PG_META_DB_USER: supabase_admin
PG_META_DB_PASSWORD: ${POSTGRES_PASSWORD}
functions:
container_name: supabase-edge-functions
image: supabase/edge-runtime:v1.22.4
restart: unless-stopped
depends_on:
analytics:
condition: service_healthy
environment:
JWT_SECRET: ${JWT_SECRET}
SUPABASE_URL: http://kong:8000
SUPABASE_ANON_KEY: ${ANON_KEY}
SUPABASE_SERVICE_ROLE_KEY: ${SERVICE_ROLE_KEY}
SUPABASE_DB_URL: postgresql://postgres:${POSTGRES_PASSWORD}@${POSTGRES_HOST}:${POSTGRES_PORT}/${POSTGRES_DB}
# TODO: Allow configuring VERIFY_JWT per function. This PR might help: https://github.com/supabase/cli/pull/786
VERIFY_JWT: "${FUNCTIONS_VERIFY_JWT}"
volumes:
- ./volumes/functions:/home/deno/functions:Z
command:
- start
- --main-service
- /home/deno/functions/main
analytics:
container_name: supabase-analytics
image: supabase/logflare:1.4.0
healthcheck:
test: [ "CMD", "curl", "http://localhost:4000/health" ]
timeout: 5s
interval: 5s
retries: 10
restart: unless-stopped
depends_on:
db:
# Disable this if you are using an external Postgres database
condition: service_healthy
# Uncomment to use Big Query backend for analytics
# volumes:
# - type: bind
# source: ${PWD}/gcloud.json
# target: /opt/app/rel/logflare/bin/gcloud.json
# read_only: true
environment:
LOGFLARE_NODE_HOST: 127.0.0.1
DB_USERNAME: supabase_admin
DB_DATABASE: ${POSTGRES_DB}
DB_HOSTNAME: ${POSTGRES_HOST}
DB_PORT: ${POSTGRES_PORT}
DB_PASSWORD: ${POSTGRES_PASSWORD}
DB_SCHEMA: _analytics
LOGFLARE_API_KEY: ${LOGFLARE_API_KEY}
LOGFLARE_SINGLE_TENANT: true
LOGFLARE_SUPABASE_MODE: true
LOGFLARE_MIN_CLUSTER_SIZE: 1
RELEASE_COOKIE: cookie
# Comment variables to use Big Query backend for analytics
POSTGRES_BACKEND_URL: postgresql://supabase_admin:${POSTGRES_PASSWORD}@${POSTGRES_HOST}:${POSTGRES_PORT}/${POSTGRES_DB}
POSTGRES_BACKEND_SCHEMA: _analytics
LOGFLARE_FEATURE_FLAG_OVERRIDE: multibackend=true
# Uncomment to use Big Query backend for analytics
# GOOGLE_PROJECT_ID: ${GOOGLE_PROJECT_ID}
# GOOGLE_PROJECT_NUMBER: ${GOOGLE_PROJECT_NUMBER}
ports:
- 4000:4000
entrypoint: |
sh -c `cat <<'EOF' > run.sh && sh run.sh
./logflare eval Logflare.Release.migrate
./logflare start --sname logflare
EOF
`
# Comment out everything below this point if you are using an external Postgres database
db:
container_name: supabase-db
image: supabase/postgres:15.1.0.117
healthcheck:
test: pg_isready -U postgres -h localhost
interval: 5s
timeout: 5s
retries: 10
depends_on:
vector:
condition: service_healthy
command:
- postgres
- -c
- config_file=/etc/postgresql/postgresql.conf
- -c
- log_min_messages=fatal # prevents Realtime polling queries from appearing in logs
restart: unless-stopped
ports:
# Pass down internal port because it's set dynamically by other services
- ${POSTGRES_PORT}:${POSTGRES_PORT}
environment:
POSTGRES_HOST: /var/run/postgresql
PGPORT: ${POSTGRES_PORT}
POSTGRES_PORT: ${POSTGRES_PORT}
PGPASSWORD: ${POSTGRES_PASSWORD}
POSTGRES_PASSWORD: ${POSTGRES_PASSWORD}
PGDATABASE: ${POSTGRES_DB}
POSTGRES_DB: ${POSTGRES_DB}
JWT_SECRET: ${JWT_SECRET}
JWT_EXP: ${JWT_EXPIRY}
volumes:
- ./volumes/db/realtime.sql:/docker-entrypoint-initdb.d/migrations/99-realtime.sql:Z
# Must be superuser to create event trigger
- ./volumes/db/webhooks.sql:/docker-entrypoint-initdb.d/init-scripts/98-webhooks.sql:Z
# Must be superuser to alter reserved role
- ./volumes/db/roles.sql:/docker-entrypoint-initdb.d/init-scripts/99-roles.sql:Z
# Initialize the database settings with JWT_SECRET and JWT_EXP
- ./volumes/db/jwt.sql:/docker-entrypoint-initdb.d/init-scripts/99-jwt.sql:Z
# PGDATA directory is persisted between restarts
- ./volumes/db/data:/var/lib/postgresql/data:Z
# Changes required for Analytics support
- ./volumes/db/logs.sql:/docker-entrypoint-initdb.d/migrations/99-logs.sql:Z
- ./scripts/tables-ollama.sql:/docker-entrypoint-initdb.d/seed.sql
vector:
container_name: supabase-vector
image: timberio/vector:0.28.1-alpine
healthcheck:
test:
[
"CMD",
"wget",
"--no-verbose",
"--tries=1",
"--spider",
"http://vector:9001/health"
]
timeout: 5s
interval: 5s
retries: 3
volumes:
- ./volumes/logs/vector.yml:/etc/vector/vector.yml:ro
- ${DOCKER_SOCKET_LOCATION}:/var/run/docker.sock:ro
command: [ "--config", "etc/vector/vector.yml" ]

View File

@ -31,73 +31,19 @@ Then run the following command to run Ollama in the background:
ollama run llama2
```
### Update Quivr to use Ollama
In order to have Quivr use Ollama we need to update the tables in Supabase to support the embedding format that Ollama uses. Ollama uses by default llama 2 that produces 4096 dimensional embeddings while OpenAI API produces 1536 dimensional embeddings.
Go to your supabase localhost instance [http://localhost:8000](http://localhost:8000) (default password admin/admin) and delete your table `vectors` and create a new table vectors with the following schema:
```sql
DROP TABLE IF EXISTS vectors;
CREATE TABLE IF NOT EXISTS vectors (
id UUID DEFAULT uuid_generate_v4() PRIMARY KEY,
content TEXT,
file_sha1 TEXT,
metadata JSONB,
embedding VECTOR(4096)
);
```
Then run the following command to update the table:
```sql
CREATE OR REPLACE FUNCTION match_vectors(query_embedding VECTOR(4096), match_count INT, p_brain_id UUID)
RETURNS TABLE(
id UUID,
brain_id UUID,
content TEXT,
metadata JSONB,
embedding VECTOR(4096),
similarity FLOAT
) LANGUAGE plpgsql AS $$
#variable_conflict use_column
BEGIN
RETURN QUERY
SELECT
vectors.id,
brains_vectors.brain_id,
vectors.content,
vectors.metadata,
vectors.embedding,
1 - (vectors.embedding <=> query_embedding) AS similarity
FROM
vectors
INNER JOIN
brains_vectors ON vectors.id = brains_vectors.vector_id
WHERE brains_vectors.brain_id = p_brain_id
ORDER BY
vectors.embedding <=> query_embedding
LIMIT match_count;
END;
$$;
```
This will update the match_vectors function to use the new embedding format.
## Add Ollama Model to Quivr
Now that you have your model running locally, you need to add it to Quivr.
In order to allow the user to choose between the OpenAI API and Ollama, we need to add a new model to the Quivr backend.
In order to allow the user to choose between the Ollama, we need to add a new model to the Quivr backend.
Go to supabase and in the table `user_settings` either add by default or to your user the following value to the `models` column:
```json
[
"gpt-3.5-turbo-1106",
"ollama/llama2"
"ollama/llama2",
"ollama/mistral",
]
```
@ -108,17 +54,3 @@ By adding this as default, it means that all new users will have this model by d
```sql
DELETE TABLE user_settings;
```
## Env Variables
In order to have Quivr use Ollama we need to update the env variables.
Go to `.env` and un comment the following env variables:
```bash
OLLAMA_API_BASE_URL=http://host.docker.internal:11434
```
Then go to the Quivr and you are good to go.

449
scripts/tables-ollama.sql Normal file
View File

@ -0,0 +1,449 @@
-- Create users table
CREATE TABLE IF NOT EXISTS user_daily_usage(
user_id UUID REFERENCES auth.users (id),
email TEXT,
date TEXT,
daily_requests_count INT,
PRIMARY KEY (user_id, date)
);
-- Create chats table
CREATE TABLE IF NOT EXISTS chats(
chat_id UUID DEFAULT uuid_generate_v4() PRIMARY KEY,
user_id UUID REFERENCES auth.users (id),
creation_time TIMESTAMP DEFAULT current_timestamp,
history JSONB,
chat_name TEXT
);
-- Create vector extension
CREATE EXTENSION IF NOT EXISTS vector;
-- Create vectors table
CREATE TABLE IF NOT EXISTS vectors (
id UUID DEFAULT uuid_generate_v4() PRIMARY KEY,
content TEXT,
file_sha1 TEXT,
metadata JSONB,
embedding VECTOR(4096)
);
-- Create function to match vectors
CREATE OR REPLACE FUNCTION match_vectors(query_embedding VECTOR(4096), match_count INT, p_brain_id UUID)
RETURNS TABLE(
id UUID,
brain_id UUID,
content TEXT,
metadata JSONB,
embedding VECTOR(4096),
similarity FLOAT
) LANGUAGE plpgsql AS $$
#variable_conflict use_column
BEGIN
RETURN QUERY
SELECT
vectors.id,
brains_vectors.brain_id,
vectors.content,
vectors.metadata,
vectors.embedding,
1 - (vectors.embedding <=> query_embedding) AS similarity
FROM
vectors
INNER JOIN
brains_vectors ON vectors.id = brains_vectors.vector_id
WHERE brains_vectors.brain_id = p_brain_id
ORDER BY
vectors.embedding <=> query_embedding
LIMIT match_count;
END;
$$;
-- Create stats table
CREATE TABLE IF NOT EXISTS stats (
time TIMESTAMP,
chat BOOLEAN,
embedding BOOLEAN,
details TEXT,
metadata JSONB,
id INTEGER PRIMARY KEY GENERATED ALWAYS AS IDENTITY
);
-- Create summaries table
CREATE TABLE IF NOT EXISTS summaries (
id BIGSERIAL PRIMARY KEY,
document_id UUID REFERENCES vectors(id),
content TEXT,
metadata JSONB,
embedding VECTOR(4096)
);
-- Create function to match summaries
CREATE OR REPLACE FUNCTION match_summaries(query_embedding VECTOR(4096), match_count INT, match_threshold FLOAT)
RETURNS TABLE(
id BIGINT,
document_id UUID,
content TEXT,
metadata JSONB,
embedding VECTOR(4096),
similarity FLOAT
) LANGUAGE plpgsql AS $$
#variable_conflict use_column
BEGIN
RETURN QUERY
SELECT
id,
document_id,
content,
metadata,
embedding,
1 - (summaries.embedding <=> query_embedding) AS similarity
FROM
summaries
WHERE 1 - (summaries.embedding <=> query_embedding) > match_threshold
ORDER BY
summaries.embedding <=> query_embedding
LIMIT match_count;
END;
$$;
-- Create api_keys table
CREATE TABLE IF NOT EXISTS api_keys(
key_id UUID DEFAULT gen_random_uuid() PRIMARY KEY,
user_id UUID REFERENCES auth.users (id),
api_key TEXT UNIQUE,
creation_time TIMESTAMP DEFAULT current_timestamp,
deleted_time TIMESTAMP,
is_active BOOLEAN DEFAULT true
);
--- Create prompts table
CREATE TABLE IF NOT EXISTS prompts (
id UUID DEFAULT uuid_generate_v4() PRIMARY KEY,
title VARCHAR(255),
content TEXT,
status VARCHAR(255) DEFAULT 'private'
);
DO $$
BEGIN
IF NOT EXISTS (SELECT 1 FROM pg_type WHERE typname = 'brain_type_enum') THEN
-- Create the ENUM type 'brain_type' if it doesn't exist
CREATE TYPE brain_type_enum AS ENUM ('doc', 'api');
END IF;
END $$;
--- Create brains table
CREATE TABLE IF NOT EXISTS brains (
brain_id UUID DEFAULT gen_random_uuid() PRIMARY KEY,
name TEXT NOT NULL,
status TEXT,
description TEXT,
model TEXT,
max_tokens INT,
temperature FLOAT,
prompt_id UUID REFERENCES prompts(id),
last_update TIMESTAMP DEFAULT CURRENT_TIMESTAMP,
brain_type brain_type_enum DEFAULT 'doc'
);
-- Create chat_history table
CREATE TABLE IF NOT EXISTS chat_history (
message_id UUID DEFAULT uuid_generate_v4(),
chat_id UUID REFERENCES chats(chat_id),
user_message TEXT,
assistant TEXT,
message_time TIMESTAMP DEFAULT current_timestamp,
PRIMARY KEY (chat_id, message_id),
prompt_id UUID REFERENCES prompts(id),
brain_id UUID REFERENCES brains(brain_id)
);
-- Create notification table
CREATE TABLE IF NOT EXISTS notifications (
id UUID DEFAULT gen_random_uuid() PRIMARY KEY,
datetime TIMESTAMP DEFAULT CURRENT_TIMESTAMP,
chat_id UUID REFERENCES chats(chat_id),
message TEXT,
action VARCHAR(255) NOT NULL,
status VARCHAR(255) NOT NULL
);
-- Create brains X users table
CREATE TABLE IF NOT EXISTS brains_users (
brain_id UUID,
user_id UUID,
rights VARCHAR(255),
default_brain BOOLEAN DEFAULT false,
PRIMARY KEY (brain_id, user_id),
FOREIGN KEY (user_id) REFERENCES auth.users (id),
FOREIGN KEY (brain_id) REFERENCES brains (brain_id)
);
-- Create brains X vectors table
CREATE TABLE IF NOT EXISTS brains_vectors (
brain_id UUID,
vector_id UUID,
file_sha1 TEXT,
PRIMARY KEY (brain_id, vector_id),
FOREIGN KEY (vector_id) REFERENCES vectors (id),
FOREIGN KEY (brain_id) REFERENCES brains (brain_id)
);
-- Create brains X vectors table
CREATE TABLE IF NOT EXISTS brain_subscription_invitations (
brain_id UUID,
email VARCHAR(255),
rights VARCHAR(255),
PRIMARY KEY (brain_id, email),
FOREIGN KEY (brain_id) REFERENCES brains (brain_id)
);
--- Create user_identity table
CREATE TABLE IF NOT EXISTS user_identity (
user_id UUID PRIMARY KEY,
openai_api_key VARCHAR(255)
);
-- Create the new table with 6 columns
CREATE TABLE IF NOT EXISTS api_brain_definition (
brain_id UUID REFERENCES brains(brain_id),
method VARCHAR(255) CHECK (method IN ('GET', 'POST', 'PUT', 'DELETE')),
url VARCHAR(255),
params JSON,
search_params JSON,
secrets JSON
);
CREATE OR REPLACE FUNCTION public.get_user_email_by_user_id(user_id uuid)
RETURNS TABLE (email text)
SECURITY definer
AS $$
BEGIN
RETURN QUERY SELECT au.email::text FROM auth.users au WHERE au.id = user_id;
END;
$$ LANGUAGE plpgsql;
CREATE OR REPLACE FUNCTION public.get_user_id_by_user_email(user_email text)
RETURNS TABLE (user_id uuid)
SECURITY DEFINER
AS $$
BEGIN
RETURN QUERY SELECT au.id::uuid FROM auth.users au WHERE au.email = user_email;
END;
$$ LANGUAGE plpgsql;
CREATE TABLE IF NOT EXISTS migrations (
name VARCHAR(255) PRIMARY KEY,
executed_at TIMESTAMPTZ DEFAULT current_timestamp
);
CREATE TABLE IF NOT EXISTS user_settings (
user_id UUID PRIMARY KEY,
models JSONB DEFAULT '["ollama/llama2"]'::jsonb,
daily_chat_credit INT DEFAULT 300,
max_brains INT DEFAULT 30,
max_brain_size INT DEFAULT 100000000
);
-- knowledge table
CREATE TABLE IF NOT EXISTS knowledge (
id UUID DEFAULT gen_random_uuid() PRIMARY KEY,
file_name TEXT,
url TEXT,
brain_id UUID NOT NULL REFERENCES brains(brain_id),
extension TEXT NOT NULL,
CHECK ((file_name IS NOT NULL AND url IS NULL) OR (file_name IS NULL AND url IS NOT NULL))
);
-- knowledge_vectors table
CREATE TABLE IF NOT EXISTS knowledge_vectors (
knowledge_id UUID NOT NULL REFERENCES knowledge(id),
vector_id UUID NOT NULL REFERENCES vectors(id),
embedding_model TEXT NOT NULL,
PRIMARY KEY (knowledge_id, vector_id, embedding_model)
);
-- Create the function to add user_id to the onboardings table
CREATE OR REPLACE FUNCTION public.create_user_onboarding() RETURNS TRIGGER AS $$
BEGIN
INSERT INTO public.onboardings (user_id)
VALUES (NEW.id);
RETURN NEW;
END;
$$ LANGUAGE plpgsql SECURITY definer;
-- Revoke all on function handle_new_user_onboarding() from PUBLIC;
REVOKE ALL ON FUNCTION create_user_onboarding() FROM PUBLIC;
-- Drop the trigger if it exists
DROP TRIGGER IF EXISTS create_user_onboarding_trigger ON auth.users;
-- Create the trigger on the insert into the auth.users table
CREATE TRIGGER create_user_onboarding_trigger
AFTER INSERT ON auth.users
FOR EACH ROW
EXECUTE FUNCTION public.create_user_onboarding();
-- Create the onboarding table
CREATE TABLE IF NOT EXISTS onboardings (
user_id UUID NOT NULL REFERENCES auth.users (id),
onboarding_a BOOLEAN NOT NULL DEFAULT true,
onboarding_b1 BOOLEAN NOT NULL DEFAULT true,
onboarding_b2 BOOLEAN NOT NULL DEFAULT true,
onboarding_b3 BOOLEAN NOT NULL DEFAULT true,
creation_time TIMESTAMP DEFAULT current_timestamp,
PRIMARY KEY (user_id)
);
-- Stripe settings --
-- Create extension 'wrappers' if it doesn't exist
CREATE EXTENSION IF NOT EXISTS wrappers;
-- Create foreign data wrapper 'stripe_wrapper' if it doesn't exist
DO $$
BEGIN
IF NOT EXISTS (
SELECT 1
FROM information_schema.foreign_data_wrappers
WHERE foreign_data_wrapper_name = 'stripe_wrapper'
) THEN
CREATE FOREIGN DATA WRAPPER stripe_wrapper
HANDLER stripe_fdw_handler;
END IF;
END $$;
-- Check if the server 'stripe_server' exists before creating it
DO $$
BEGIN
IF NOT EXISTS (SELECT 1 FROM pg_foreign_server WHERE srvname = 'stripe_server') THEN
CREATE SERVER stripe_server
FOREIGN DATA WRAPPER stripe_wrapper
OPTIONS (
api_key 'your_stripe_api_key' -- Replace with your Stripe API key
);
END IF;
END $$;
-- Create foreign table 'public.customers' if it doesn't exist
DO $$
BEGIN
IF NOT EXISTS (
SELECT 1
FROM information_schema.tables
WHERE table_name = 'customers'
) THEN
CREATE FOREIGN TABLE public.customers (
id text,
email text,
name text,
description text,
created timestamp,
attrs jsonb
)
SERVER stripe_server
OPTIONS (
OBJECT 'customers',
ROWID_COLUMN 'id'
);
END IF;
END $$;
-- Create table 'users' if it doesn't exist
CREATE TABLE IF NOT EXISTS public.users (
id uuid REFERENCES auth.users NOT NULL PRIMARY KEY,
email text
);
-- Create or replace function 'public.handle_new_user'
CREATE OR REPLACE FUNCTION public.handle_new_user()
RETURNS TRIGGER AS $$
BEGIN
INSERT INTO public.users (id, email)
VALUES (NEW.id, NEW.email);
RETURN NEW;
END;
$$ LANGUAGE plpgsql SECURITY DEFINER;
-- Check if the trigger 'on_auth_user_created' exists before creating it
DO $$
BEGIN
IF NOT EXISTS (SELECT 1 FROM pg_trigger WHERE tgname = 'on_auth_user_created') THEN
CREATE TRIGGER on_auth_user_created
AFTER INSERT ON auth.users
FOR EACH ROW EXECUTE FUNCTION public.handle_new_user();
END IF;
END $$;
insert into
storage.buckets (id, name)
values
('quivr', 'quivr');
CREATE POLICY "Access Quivr Storage 1jccrwz_0" ON storage.objects FOR INSERT TO anon WITH CHECK (bucket_id = 'quivr');
CREATE POLICY "Access Quivr Storage 1jccrwz_1" ON storage.objects FOR SELECT TO anon USING (bucket_id = 'quivr');
CREATE POLICY "Access Quivr Storage 1jccrwz_2" ON storage.objects FOR UPDATE TO anon USING (bucket_id = 'quivr');
CREATE POLICY "Access Quivr Storage 1jccrwz_3" ON storage.objects FOR DELETE TO anon USING (bucket_id = 'quivr');
-- Create functions for secrets in vault
CREATE OR REPLACE FUNCTION insert_secret(name text, secret text)
returns uuid
language plpgsql
security definer
set search_path = public
as $$
begin
return vault.create_secret(secret, name);
end;
$$;
create or replace function read_secret(secret_name text)
returns text
language plpgsql
security definer set search_path = public
as $$
declare
secret text;
begin
select decrypted_secret from vault.decrypted_secrets where name =
secret_name into secret;
return secret;
end;
$$;
create or replace function delete_secret(secret_name text)
returns text
language plpgsql
security definer set search_path = public
as $$
declare
deleted_rows int;
begin
delete from vault.decrypted_secrets where name = secret_name;
get diagnostics deleted_rows = row_count;
if deleted_rows = 0 then
return false;
else
return true;
end if;
end;
$$;
INSERT INTO migrations (name)
SELECT '20231128173900_remove_openai_api_key'
WHERE NOT EXISTS (
SELECT 1 FROM migrations WHERE name = '20231128173900_remove_openai_api_key'
);