README.md

Important

This repository has been integrated with a new combined pre-award-stores.

Funding service design application store

This service provides an API for accessing the Access Funding Application Store.

Developer setup guide

This service depends on the following:

• Fund store
• Notification
• Account Store
• Simple Queue Service
• Postgres database

Testing

Testing in Python repos Further information on the test data used for transactional tests is contained here

IDE Setup

Python IDE Setup

Simple Queue Service

As part of the application submission workflow, we use a FIFO AWS SQS to automate our application export to assessment.

We export the application as a 'fat' payload. This includes all application data (including metadata/attributes), this ensure assessment does not need to call application_store for additional information.

We can simulate an SQS locally when using our docker runner instance. Our docker runner uses localstack to simulate these AWS services, see here.

If messages are not consumed and deleted they will be move to the Dead-Letter_Queue, here we can inspect the message for faults and retry.

The SQS queues have a number of confiuration options, we are using the AWS SDK for Python (Boto3), see docs here.

There is an API endpoint on this service to send a submitted application to assessment:

```
/queue/{queue_name}/{application_id}
```

Database

General instructions for local db development are available here: Local database development

Useful Queries

Show count of applications by status for each round select fund_id, round_id, status, count(status) from applications group by fund_id, status, round_id; Total to be sent, by fund/round select fund_id, round_id, count(id) from applications where status not in ('SUBMITTED') group by fund_id, round_id;

Seeding Test Data

You can seed test data to use in the running application (separate to unit test data seeding). The seeding process needs a running fund-store to retrieve fund/round form section config, so it runs within the docker container for application-store within the docker runner. To run the seeding script:

1. Make sure your local docker-runner is running
2. Find the container ID of application-store by using docker ps
3. Use docker exec to get into that container: docker exec -it <container_id> bash
4. Execute the script: python scripts/seed_db_test_data.py. You will be prompted for inputs: fund, round, account_id (the UUID not email address), the status of the seeded applications and how many to create.

Testing the seeding process

Unit tests exist in test_seed_db. They are marked as skipped as they require a running fund-store to retrieve form config (no point in duplicating this for tests) so they won't run in the pipeline but are fine locally. If your local fund store runs on a non-standard port etc, edit the local_fund_store fixture in that tests file. If you want to run the tests, just comment out the skip marker.

Adding a new fund/round to the seeding process

To seed applicaitons, we need the completed form json. If you have that, skip to the end of part 1 and put that form json into the required file.

Part 1 - get the form json

1. Get a submitted application into your local DB. You can either do this manually or by running the automated tests against your local docker runner.
2. Find the application_id of that submitted application._
3. Edit the tests file to un-skip test_retrieve_test_data and then set target_app to be the application_id you just submitted.
4. Update your unit test config to point at the same DB as the docker runner. Update pytest.ini so that D:DATABASE_URL points at the docker runner application store db: D:DATABASE_URL=postgresql://postgres:[email protected]:5433/application_store # pragma: allowlist secret
5. Run the single test test_retrieve_test_data - this should output the json of all the completed forms for that application into funding-service-design-store/forms.json.
6. Copy this file into seed_data and name it <fund_short_code>_<round_short_code>_all_forms.json.
7. IMPORTANT Change the config in pytest.ini back to what it was so you don't accidentally wipe your docker runner DB next time you run tests!

Part 2 - update seeding config

1. In seed_db there is a constant called FUND_CONFIG - update this following the existing structure for your new fund/round (if it's a new round on an existing fund, just add it as another key to rounds item in that fund). You will need to know the name of the form that contains the field used to the name the application/project.
2. In the same file, update the click.option choice values for fund/round as required, to allow your new options.
3. Test it - update the unit tests to use this new config and check it works.

Builds and Deploys

Details on how our pipelines work and the release process is available here

Paketo

Paketo is used to build the docker image which gets deployed to our test and production environments. Details available here

When running the docker image generated with paketo, envs needs to contain a value for each of the following:

• ACCOUNT_STORE_API_HOST
• FUND_STORE_API_HOST
• SENTRY_DSN
• GITHUB_SHA
• DATABASE_URL

Copilot

Copilot is used for infrastructure deployment. Instructions are available here, with the following values for the application store:

• service-name: fsd-application-store
• image-name: funding-service-design-application-store