SQLAlchemy connector

[webb-ben]

User story

As an NLDI Admin, I want to connect to the nldi-database via SQLAlchemy so that I am able to using python framework to structure and process the API request.

Acceptance criteria

• SQLAlchemy implementation of database connection
• SQLAlchemy supports all requests currently supported

Definition of done

• unit/functional tests added/updated
• documentation updated

[gzt5142]

An example of a sqlalchemy connection and ORM for this database is here:

data access layer -- a convenience object to hold db connect details:
https://github.com/gzt5142/nldi-crawler-py/blob/7e0aa4c6230c983a4931561ab6322c2d68ece527/src/nldi_crawler/db.py#L28

The ORM is used in two places in the crawler:

• Reading the crawler_source table https://github.com/gzt5142/nldi-crawler-py/blob/7e0aa4c6230c983a4931561ab6322c2d68ece527/src/nldi_crawler/source.py#L299
• Inserting features during ingestion: https://github.com/gzt5142/nldi-crawler-py/blob/7e0aa4c6230c983a4931561ab6322c2d68ece527/src/nldi_crawler/ingestor.py#L57

I've chosen a 'classic' binding mechanism to map a table to a simplified dataclass (rather than subclassing DeclarativeBase). This makes the ORM dependent on the dataclass rather than the other way around. Code which manipulates the feature dataclass is protected from knowing sql details this way.