Skip to content
This repository has been archived by the owner on Oct 23, 2024. It is now read-only.
/ lantern Public archive

Lantern manages a waitlist for Khoj. It used to be a lot more, but now it's simple!

Notifications You must be signed in to change notification settings

khoj-ai/lantern

Repository files navigation

Lantern

Lantern is a Django application for managing application-level services for Khoj.

Development

You should have Docker and Docker Compose installed on your system for quickest setup.

Run locally

python3 -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt
export DJANGO_DEBUG=True

Database Setup

  1. Install Postgres
brew install postgresql@15
  1. Start Postgres
brew services start postgresql@15
  1. Install utilities
brew install libpq

Add to PATH

# ~/.zshrc
export PATH="/Applications/homebrew/opt/libpq/bin:$PATH"

Create role

createuser -s postgres

Create database

createdb lantern -U postgres

Start the service

gunicorn -c config/gunicorn/dev.py

Kill the service

If you started the process in daemon mode, you can get the pid in the .pid file under config/gunicorn. Then run:

kill -9 <pid>

Build Docker Image

docker build -t lantern .

Docker Container

docker-compose up -d

Enter Docker Container

docker exec -it lantern-web-1 bash

Run migrations

python manage.py migrate

Debugging Latency

  1. Install Profiling packages

Make sure that you have the following Python packages installed:

django-extensions==3.2.3
snakeviz==2.2.0
  1. Add debugging configuration in VSCode

In your launch.json, add the following configuration:

        {
            "name": "Lantern - Profile",
            "type": "python",
            "request": "launch",
            "program": "${workspaceFolder}/manage.py",
            "args": [
                "runprofileserver", "0.0.0.0:5000", "--use-cprofile", "--prof-path", "${workspaceFolder}/profile"
            ],
            "django": true,
            "justMyCode": true
        }
  1. Add a /profile folder to your projects root directory if it doesn't already exist
  2. Make queries as normal

The binary output to represent the profile of individual queries should be outputted to the /profile directory you've created

  1. Open the profiler visualizer

From your terminal, run snakeviz profile in the root directory of the project.

About

Lantern manages a waitlist for Khoj. It used to be a lot more, but now it's simple!

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Sponsor this project

 

Packages

No packages published