A small service to make it easier for packages.ecosyste.ms to read data about Conda Packages in different channels.
- Provide a REST interface for list of all names of packages (as json)
- Provide a REST interface for list of versions for each package (as json)
- Update info from Specs repo frequently
- Watch anaconda repos for updates
- Tell Libraries about removed versions/packages
The API includes several performance optimizations:
- Parallel HTTP fetches: Uses Typhoeus::Hydra to fetch all 19 architectures concurrently per channel
- In-memory caching: Caches merged package data and deduplicated versions to avoid repeated computation
- Redis HTTP caching: Caches HTTP responses with ETag/Last-Modified support for conditional requests
- Bandwidth optimization: 304 Not Modified responses reuse cached data, reducing bandwidth by 95%+
- HTTP responses cached for 1 hour in Redis
- ETag and Last-Modified headers used for conditional requests
- Gracefully degrades if Redis is unavailable
- Cache automatically disabled in test environment
Run bundle install to download all dependencies.
You can run a local server within a container with docker-compose docker-compose up or locally with bundle exec puma.
The server should now be running port 9292. This can be verified by going to http://localhost:9292 and verifying it sends back an 'Hello world' response.
Redis Configuration:
- Docker: Redis automatically configured via
docker-compose up - Local: Set
REDIS_URLenvironment variable (defaults toredis://localhost:6379/0) - Production: Set
REDIS_URLto your Redis instance URL
Run the unit tests using rspec locally or within a built docker container docker build -t ecosyste-ms/conda-api . && docker run -it -e PORT=9292 -p 9292:9292 ecosyste-ms/conda-api rspec.
Tests use webmock to stub HTTP requests for fast, reliable testing without external dependencies.
