HTTP API backend service for Edna.
You can use Stack or Cabal to build Edna server.
Run stack build or cabal build.
We also provide a Makefile to facilitate building and testing during development.
By default it uses stack, but you can set EDNA_USE_CABAL=1 to use cabal.
edna-server calls Python code under the hood which has some dependencies.
So first of all you need to enter poetry shell with all dependencies available.
poetry commands should run from the folder with Python code, so start with cd ../analysis.
There, do poetry install to install dependencies and then poetry shell to make them available.
After that go back to this folder (cd ../backend).
Depending on the tool you used to build edna-server you need to run one of these commands:
stack exec -- edna-server -c dev-config.yamlcabal run -- edna-server -c dev-config.yaml
Alternatively, you can use make run command which does what's written above.
Set RUN_ARGUMENTS to command line options you want to pass.
Example: RUN_ARGUMENTS='-c dev-config.yaml' make run.
You can update dev-config.yaml as you wish before running.
If you use the config from this repository:
- The server endpoints will be available at
http://localhost:9000/api/*. - Swagger docs will be available at
http://localhost:9000/docs/. - Data for prometheus will be available at
http://localhost:9000/metrics/.
Note that you need to also run PostgreSQL server to make backend work.
One way to run it is to use docker-compose.yml provided in the deployment folder.
Set EDNA_DB_DEBUG=1 to enable logging of all DB actions made.
It works in tests as well.
You can pass other command line arguments, to see full list run stack exec -- edna-server -h or cabal run -- edna-server -h:
You can use environment variables to configure your server (available variables listed in .env).
NOTE: priority of configurations is following: command line arguments, environment variables, config file.
To run tests you need to have pg_tmp to run DB for tests.
Then run make test to execute tests.
This command automatically executes poetry to obtain Python dependencies.
Apart from edna-server executable we provide edna-generator.
This tool can generate sample data for Edna and put it into DB.
You can build and run it the same way as edna-server (via stack or cabal).
You can also run it using Makefile.
It has the same options as edna-server plus some additional options that specify how much data to generate.
Pass --help to get the full list of available options.
You are adviced to set logging to prod or nothing (otherwise there will be too much output in logs)
Also maybe you also want to run it on empty DB.
Example with prod logging and empty DB: RUN_ARGUMENTS='-c dev-config.yaml -l prod --init-mode enable-with-drop' make run-generator.
If DB is not empty and there are conflicts (such as already existing project name), by default they will be skipped.
Note that edna-generator is an experimental tool for developers and may be unstable/unreliable sometimes.