Add semi_structured module

[Rajat-Sirohi]

This PR is a prototype of a larger project idea to have Gen3 support a data lake architecture. Specifically, I've added support for CRUD endpoints for semi-structured data. Additionally, I've modified the Metadata model to include optional baseid and created_date columns in order to support versioning of semi-structured data records. Version information is accessed via two additional /versions and /latest endpoints. More specific changes are detailed below:

• Add new semi_structured module
• Add baseid and created_date columns to Metadata model + corresponding migration script
• Add check to aliases module which ensures that new alias does not have the same name as an existing GUID
• Modify update_metadata() to create a record if it doesn't already exist
• Add unit tests for everything
• Add pytest-dependency package to poetry.lock (convenient for unit testing)
• Update documentation