Skip to content

Latest commit

 

History

History
70 lines (57 loc) · 2.31 KB

README.md

File metadata and controls

70 lines (57 loc) · 2.31 KB
Raw

Pre-Requirements

Setup Qdrant cloud

  1. Sign-in or sign-up at Qdrant Cloud
  2. Create a free 1GB cluster
  3. Note down the Qdrant API key and the Qdrant Cluster Url

Setup OpenAI

  1. Sign-in or sign-up at OpenAI
  2. Setup your account with enough credits
  3. Note down the OpenAI API key

Slack Bot App setup

  1. You need to be a Slack admin for your workspace.
  2. Go to Slack apps and click on “Create a new app”, select “From scratch” in the popup window, and choose a name (for me, it is “PlivoAskMe”), then click on “Create App”.
  3. Click on “New Slack Commands” and point it to your application API hosted on fly.io (you can do that later if the fly.io application is not online yet)
  4. From the left menu, click on “OAuth & Permissions” then in the “Bot Token Scopes” section, click on “Add an Oauth scope” and select “commands”
  5. Go back to the “Basic information” section in the left menu and click “Install to workspace”
  6. You should be able to see the “App Credentials” section.
  7. Note down the Slack "Verification Token"

Applicaiton settings

Copy the settings.py example file

cp settings.py.example settings.py

Edit and configure the following variables:

  • INGEST_GIT_REPO_URLS
  • INGEST_SITEMAP_URLS
  • INGEST_SITEMAP_URLS_FILTERS

Fly.io deployment

Setup the application

Create the application

fly apps create <app_name>

Copy the fly.toml example file

cp fly.toml.example fly.toml

Change the "app" with the <app_name>, and "OPENAI_MODEL" and "VECTOR_DATABASE" settings.

Configure secrets

fly secrets QDRANT_API_KEY=xxxx # replace 'xxxx' with your Qdrant API key
fly secrets OPENAI_API_KEY=xxxx # replace 'xxxx' with your OpenAI API key
fly secrets SLACK_TOKEN_ID=xxxx # replace 'xxxx' with your Slack Verification token

Deploy the app and machine

fly deploy --force-machines --local-only --region iad --vm-size shared-cpu-2x

Append data to the vector database

fly scale memory 4096 # scale up memory to ingest the data
fly ssh console --pty -C 'python3 /app/ingest.py' # collect and inject data into the vector database
fly scale memory 2048 # scale down memory

Update the app

fly deploy --local-only