Skip to content

Latest commit

 

History

History
77 lines (49 loc) · 2.97 KB

README.md

File metadata and controls

77 lines (49 loc) · 2.97 KB
Raw

Local setup

Installations:

  • Install the SAM CLI
  • Install Docker
  • Install node 18 and npm
  • Install AWS-CLI
  • Install Zip cli tool (in Ubuntu: sudo apt-get install zip)
  • Inside each directory install the dependecies by running npm i.

Configuring AWS

To connect your local environment to the AWS account run aws configure and provide the credentials.

Deploy infrastructure

Create secrets in the AWS secret mananager console:

  1. Google OAuth Secrets (Needed for login with Google):

    • Secret name: googlecloud
    • Secret key: BIO_GOOGLE_OAUTH_CLIENT_SECRET
    • Secret value: Your Google OAuth Client Secret (Documentation)

  2. OpenAI Secrets (Needed for using GPT as the LLM)

    • Secret name: openai
    • Secret key: OPENAI_API_KEY
    • Secret value: Your OpenAI API Key (Get it here)

Build service lambda code.

Inside service directory, npm run prod-build

Setup the Environment Variables

Add these in a .env file inside the root of the repo.

  • GOOGLE_CLIENT_ID: Your Google OAuth client id (should be in the same page as the client secret). Will be used by Congnito to perform Google login.
  • OAUTH_CALLBACK_URL: A comma separated list of URLs. These are the acceptable URLs to be redirected after performing OAuth, include localhost URLs.
  • OPENAI_ORG_ID: The org id from your open ai account. Can be found in your platform settings

Deploy cloud resources

Inside the infra directory run npx cdk deploy. This will deploy all AWS resources.

Running the service API locally

We can use the SAM cli to spin up a local lambda and API gateway instance. This way we don't have to waste time deploying while we are developing.

Setup the local lambda env vars:

Create a file service/env/dev-env-vars.json with the form:

{
  "Parameters": {
    "USER_POOL_CLIENT_ID": "<Your user pool client id in AWS>",
    "USER_POOL_ID": "<Your user pool id in AWS>",
    "OPENAI_ORG_ID": "<Your OpenAI org id>"
  }
}

Go inside the service directory and run npm start to spin it up.

There is not local database setup, so all data interaction will be with the deployed DDB. There are ways to setup DDB locally however.

Running the UI

Go to ui folder, npm i to install dependencies. To bring up the UI server in the web run npm run web inside the ui folder. This will run the web version of the UI with webpack.

To deploy the UI or use the UI with a deployed backend, we need to set the appropiate environment variables:

  • PROD_API_ENDPOINT (Lambda function URL)
  • UI_PROD_AUTH_REDIRECT_URI (Domain to redirect to after OAuth is completed)