Skip to content

Latest commit

 

History

History
311 lines (220 loc) · 14.3 KB

README.md

File metadata and controls

311 lines (220 loc) · 14.3 KB
Raw

LeapfrogAI UI

Getting Started

Requirements:

This application requires Supabase, and either Leapfrog API or OpenAI to function. Additionally, it can optionally use Keycloak for authentication. There are several different ways to run it, so please see the "Configuration Options" section below for more information.

Running the UI

  1. Change directories: cd src/leapfrogai_ui
  2. Create a .env file at the root of the UI project (src/leapfrogai_ui), reference the .env.example file for values to put in the .env file
  3. Install dependencies: npm install
  4. Run npm run dev -- --open

Configuration Options

It is recommended to run LeapfrogAI with UDS, but if you want to run the UI locally (on localhost, e.g. for local development), you can either:

  1. Connect to a UDS deployed version of the Leapfrog API and Supabase
    or
  2. Connect to OpenAI and UDS deployed Supabase or locally running Supabase.

Note - most data CRUD operations utilize Leapfrog API or OpenAI, but some functionality still depends on a direct connection with Supabase.

Running everything with UDS

This is the easiest way to use the UI. Follow the documentation for running the entire LeapfrogAI stack

Running UI Locally with LeapfrogAI

If running the UI locally and utilizing LeapfrogAPI, you must use the same Supabase instance that the Leapfrog API is utilizing.

  1. Connect the UI to a UDS deployed version of Supabase and Leapfrog API. Ensure these env variables are set appropriately in your .env file:
PUBLIC_SUPABASE_URL=https://supabase-kong.uds.dev
PUBLIC_SUPABASE_ANON_KEY=<anon_key>
...
LEAPFROGAI_API_BASE_URL=https://leapfrogai-api.uds.dev
DEFAULT_MODEL=llama-cpp-python # or vllm
  1. Run the frontend migrations

If you deploy the UI with UDS, the necessary database migrations will be applied. You can still run a local version of the UI, but the deployed version will have set up the database properly for you.

Further instructions will be coming soon in a future release.

Authentication

You can choose to use Keycloak (with UDS) or turn Keycloak off and just use Supabase.

When the UI and API are deployed with UDS, everything will be configured properly automatically, but if you want to run the UI outside of UDS on localhost (e.g. for development work), there are some manual configuration steps:

  1. Modify the "GOTRUE_URI_ALLOW_LIST" within Supabase.
    The Supabase UDS package has a ConfigMap called "supabase-auth-default".
    Add these values to the "GOTRUE_URI_ALLOW_LIST" (no spaces!). This variable may not exist and you will need to add it. Restart the supabase-auth pod after updating the config: http://localhost:5173/auth/callback,http://localhost:5173,http://localhost:4173/auth/callback,http://localhost:4173
    Note - Port 4173 is utilized by Playwright for E2E tests. You do not need this if you are not concerned about Playwright.
With Keycloak authentication
  1. If Supabase was deployed with UDS, it will automatically configure a Keycloak Client for you. We need to modify this client to allow localhost URIs.
    Within Keycloak, under the UDS Realm, edit the uds-supabase client.
    Under "Valid redirect URIs" add:
    http://localhost:5173/auth/callback
    http://localhost:4173/auth/callback (for Playwright tests)
  2. If you want to connect Keycloak to a locally running Supabase instance (non UDS deployed), see the "Running Supabase locally" section below.
Without Keycloak authentication
  1. To turn off Keycloak, set this .env variable: PUBLIC_DISABLE_KEYCLOAK=false
Running UI Locally with OpenAI

Set the following .env variables:

DEFAULT_MODEL=gpt-3.5-turbo
LEAPFROGAI_API_BASE_URL=https://api.openai.com/v1
#If specified, app will use OpenAI instead of Leapfrog
OPENAI_API_KEY=<your_openai_api_key>

You still need Supabase, so you can connect to UDS deployed Supabase, or run Supabase locally. To connect to UDS deployed Supabase, set these .env variables:

PUBLIC_SUPABASE_URL=https://supabase-kong.uds.dev
PUBLIC_SUPABASE_ANON_KEY=<anon_key>

Running Supabase locally:

  1. Install Supabase
  2. Run: supabase start The configuration files at src/leapfrogai_ui/supabase will ensure your Supabase is configured to work with Keycloak if you set these .env variables:
SUPABASE_AUTH_KEYCLOAK_CLIENT_ID=uds-supabase
SUPABASE_AUTH_KEYCLOAK_SECRET=<secret> #this is the client secret for the client in Keycloak
SUPABASE_AUTH_EXTERNAL_KEYCLOAK_URL=https://sso.uds.dev/realms/uds

After it starts, the Supabase API URL and Anon key are printed to the console. These are used in the .env file to connect to Supabase.

After starting supabase for the first time, you need to initialize the database with migrations and seed data:

supabase db reset

After this initial reset, if you start Supabase again it will already have the data and you don't need to run this command unless you want to restore it to the default.

Stop Supabase:

npm run supabase:stop

Warning - if switching the application from utilizing Leapfrog API to OpenAI or vice versa, and you encounter this error: Server responded with status code 431. See https://vitejs.dev/guide/troubleshooting.html#_431-request-header-fields-too-large. you need to clear your browser cookies

Building

To create a production version of the app:

npm run build

You can preview the production build with npm run preview.

Developer Notes

Tooling

Supabase

We use Supabase for authentication and a database. Application specific data (ex. user profile images, application settings like feature flags, etc..) should be stored directly in Supabase and would not normally utilize the Leapfrog API for CRUD operations.

Playwright End-to-End Tests

First install Playwright: npm init playwright@latest

To run the E2E tests:
npm run test:integration:ui Click the play button in the Playwright UI. Playwright will run it's own production build and server the app at http://localhost:4173. If you make server side changes, restart playwright for them to take effect.

Notes:

  1. Playwright tests are End-To-End tests and use the "real" full stack app. If you run these tests, they will use the configuration indicated by your .env file. See the "Configuration Options" section above to configure which database Playwright is using.
  2. If you run the tests in headless mode (npm run test:integration) you do not need the app running, it will build the app and run on port 4173.

Supabase and Keycloak Integration

The Supabase docs are inadequate for properly integrating with Keycloak. Additionally, they only support integration with the Supabase Cloud SAAS offering. Before reading the section below, first reference the Supabase docs.

The following steps are required to integrate Supabase with Keycloak for local development:

The supabase/config.toml file contains configuration options for Supabase when running it locally. When running locally, the Supabase UI dashboard does not offer all the same configuration options that the cloud version does, so you have to specify some options in this file instead.

The variables that had to be overridden were:

[auth]
site_url = "http://localhost:5173"
additional_redirect_urls = ["http://localhost:5173/auth/callback"]

[auth.external.keycloak]
enabled = true
client_id = "env(SUPABASE_AUTH_KEYCLOAK_CLIENT_ID)"
secret = "env(SUPABASE_AUTH_KEYCLOAK_SECRET)"
url= "env(SUPABASE_AUTH_EXTERNAL_KEYCLOAK_URL)"

The variables referenced in the [auth.external.keycloak] section above are in your .env file at the root of the project. You need to create a client in Keycloak to get these variables. Under a realm in Keycloak that is not the master realm (if using UDS, its "uds"):

  1. Create a new client (the client ID you use will be used in the env variables below)
  2. Turn on "Client Authentication"
  3. For "Valid redirect URLs", you need to put:
    1. http://localhost:5173/auth/callback (or the URL for the frontend app callback)
    2. http://127.0.0.1:54321/auth/v1/callback (or the URL for the Supabase callback, for locally running Supabase, DO NOT USE LOCALHOST, use 127.0.0.1)
    3. Put the same two URLs in for "Web Origins"
  4. Copy the Client Secret under the Clients -> Credentials tab and use in the env variables below
  5. You can create users under the "Users" tab and either have them verify their email (if you setup SMTP), or manually mark them as verified.
#.env
SUPABASE_AUTH_KEYCLOAK_CLIENT_ID=<client-id-for-app-from-keycloak>
SUPABASE_AUTH_KEYCLOAK_SECRET=<client-id-for-app-from-keycloak>
SUPABASE_AUTH_EXTERNAL_KEYCLOAK_URL=<URL for keycloak see notes below>

The SUPABASE_AUTH_EXTERNAL_KEYCLOAK_URL can be tricky if your Keycloak is not hosted on an actual domain. For example, if you are running Keycloak locally, you might have it at: http://localhost:8080/auth/realms/uds If you leave this as localhost:8080, when Supabase attempts to make a POST request to the Keycloak server during the auth process, it will not be able to find this address and you will likely see the error "unable to exchange external code". This comment speaks to this problem, although the solutions don't work if you are using the Supabase CLI and not running a pure docker version of Supabase.

In order to fix this, we have to edit the /etc/hosts file in the running Supabase Auth container (we can't add this through a docker compose file because we are using the Supabase CLI to start it up, migrate the db, and seed it).

Here is an example of how you can modify the /etc/hosts of the supabase container: "supabase start && docker exec -u 0 supabase_auth_supabase /bin/sh -c \"echo '100.115.154.78 keycloak.admin.uds.dev' >> /etc/hosts\" You must ensure the IP address used in this command is the correct IP for where you have Keycloak hosted. If you need to use a different Keycloak server for local development, you will need to modify this command.

If your Keycloak server is not at a hosted domain, you will also need to modify the /etc/hosts on your machine:

Example:
sudo nano /etc/hosts
*add this line (edit as required)*
xxx.xxx.xx.xx  keycloak.admin.uds.dev

Ensure the

PUBLIC_SUPABASE_URL=
PUBLIC_SUPABASE_ANON_KEY=

variables in your .env file are pointing to the correct Supabase instance.

Note - if connecting to a hosted Supabase instance, or in a cluster with networking, you will not need to override /etc/host files. The:

SUPABASE_AUTH_KEYCLOAK_CLIENT_ID=
SUPABASE_AUTH_KEYCLOAK_SECRET=
SUPABASE_AUTH_EXTERNAL_KEYCLOAK_URL=

variables will also not do anything because they are only used for locally running Supabase. You will instead need to modify the Auth provider settings directly in the Supabase dashboard and set the appropriate redirect URLs and client ID/secret.

If you use the component from @supabase/auth-ui-shared, you must ensure you supply provider and scope props: ex:

	<Auth
		supabaseClient={data.supabase}
		redirectTo={`${data.url}/auth/callback`}
		showLinks={false}
		providers={['keycloak']}
		providerScopes={{
			keycloak: 'openid'
		}}
		onlyThirdPartyProviders
		appearance={{ theme: ThemeSupa }}
	/>

If you do not use this component and need to login with a Supabase auth command directly, ensure you provide the "openid" scope with the options parameter.

The E2E tests use a fake Keycloak user. You must create this user in Keycloak, and add the username, password, and MFA secret to the .env file in order for the tests to pass. PUBLIC_DISABLE_KEYCLOAK must also be set to false to run the E2E tests.

To get the MFA secret, create your new Keycloak user with the normal Keycloak login flow (not manually through the Keycloak UI). When scanning the QR code, use an app that lets you see the url of the QR code. The secret is contained in that URL.

Login flow was adapted from this reference

Chat Data Flow

The logic for handling regular chat messages and assistant chat messages, along with persisting that data to the database is complex and deserves a detailed explanation.

Our chat page allows the user to send messages to /api/chat ("regular chat") and /api/chat/assistants ("chat with assistant"). The messages are streamed to the client so that text is progressively displayed on the screen. We use the Vercel AI SDK to handle streaming as well as response cancellation, regeneration, message editing, error handling, and more.

Messages streamed with regular chat, use the "useChat" function.
Assistants use the "useAssistants" function.
These functions do not provide the same features and handle data differently, resulting in several edge cases.

Here are a few of the big issues caused by these differences:

The useChat function does not save messages with the API to the database, we have to handle that on our own.
Messages sent with useAssistants, however, are saved to the database automatically.

Creation timestamps are handled differently depending on if they are streamed responses or if they have been saved to the database. Streamed messages have timestamps on the "createdAt" field, saved messages have timestamps on the "created_at" field. Sometimes the dates are Date strings, unix seconds, or unix milliseconds. Since dates can be returned in seconds, we lose some of the precision we would have for sorting the messages if they were returned in milliseconds. Due to this issue, there is logic in place to prevent the user from sending messages too quickly, ensuring timestamps are unique.

Additionally, streamed messages have temporary ids that do not match the ids messages are assigned when they are saved to the database. This makes editing and deleting messages challenging, so we have to keep track of both streamed messages and saved messages in client side state in the correct order. We use this state to look up the saved ids and make the appropriate API calls with the permanent ids.

While there are several automated tests for this logic, the edge cases and mocking scenarios are complex. Any modifications to this logic should be thoroughly manually tested.