Local setup for conda-store (core)
There are two main ways to set up your local environment and conda-store services (web UI, API server, database, etc.) for development:
- Using Docker and Docker compose: This is the recommended approach for working on
conda-store-serverlibrary. - Using standalone mode: Required for running tests and for working on the
conda-store(client) library.
You need a local copy of the conda-store repository for the development setup.
Docker setup (recommended) 🐳
Pre-requisites
Install the following dependencies before developing on conda-store:
Local deployment
To deploy conda-store locally, run the following command:
docker compose up --build -d
On a fast machine, this should take about 10 seconds, assuming the docker images have been partially built before.
Most of the conda-store docker images are built/tested for amd64(x86-64). Notice the architecture: amd64 within the docker-compose.yaml files.
There will be a performance impact when building and running on ARM architectures. Otherwise, this workflow has been shown to run and build on OSX.
The following resources will be available on the deployment:
| Resource | Localhost port | username | password |
|---|---|---|---|
| conda-store web server (UI) ✨ | localhost:8080 | admin | password |
| MinIO S3 | localhost:9000 | admin | password |
PostgreSQL (database: conda-store) | localhost:5432 | admin | password |
| Redis | localhost:6379 | - | password |
If you make any changes to conda-store-server,
run the following to have those changes in the deployment:
docker compose down -v
docker compose up --build
To stop the deployment, run:
docker compose stop
Optionally, to remove the containers, run:
docker compose rm -f
Standalone setup 💻
Pre-requisites
You need conda for this setup, you can install it with the instructions in the documentation.
Development environment
Create a conda environment with the development dependencies, and activate the environment:
conda env create -f conda-store-server/environment-dev.yaml
conda activate conda-store-server-dev
To install the conda-store-server package in editable (development) mode, run the following from the root of the repository:
python -m pip install -e ./conda-store-server
To install the conda-store package in editable (development) mode, run the following from the root of the repository:
python -m pip install -e ./conda-store
Start conda-store in standalone mode
Running conda-store in --standalone mode launches celery as a
subprocess of the web server.
python -m conda_store_server.server --standalone
Visit localhost:8080 from your web browser to access the conda-store web UI. ✨
Run the test suite ✅
You can run the codebase tests locally to verify your changes before submitting a pull request. You need Docker Compose as well as the conda development environment to run the complete set of tests.
conda-store (client)
Lint and format
Run the linting and formatting checks with hatch:
cd conda-store
hatch env run -e dev lint
Integration tests
These tests are stateful, so clear the state if you previously ran the conda-store-server service on Docker:
cd conda-store
docker compose down -v # ensure you've cleared state
docker compose up --build
Wait until the conda-store-server is running check by visiting localhost:8080.
Install conda-store (client) library in editable mode:
pip install -e .
Execute the scripts in the tests directory to run the tests:
./tests/unauthenticated-tests.sh
./tests/authenticated-tests.sh
./tests/authenticated-token-tests.sh
You need to explicitly set some environment variables to run the shebang tests:
export CONDA_STORE_URL=http://localhost:8080/conda-store
export CONDA_STORE_AUTH=basic
export CONDA_STORE_USERNAME=username
export CONDA_STORE_PASSWORD=password
./tests/shebang.sh
conda-store-server
Lint and format
Run the linting and formatting checks with hatch:
cd conda-store-server
hatch env run -e dev lint
Package build
Check that the package builds:
cd conda-store-server
hatch build
Unit tests
Run the unit tests with pytest:
cd conda-store-server
pytest
Integration tests
These tests are stateful, so clear the state if you previously ran the conda-store-server service on Docker:
cd conda-store-server
docker-compose down -v # ensure you've cleared state
docker-compose up --build
Wait until the conda-store-server is running check by visiting localhost:8080.
Run the tests with hatch:
hatch env run -e dev playwright-test
hatch env run -e dev integration-test