Convert entire backend to use BIGINT primary keys

[wpfl-dbt]

Context

The new ingest process moves hashing to the server for inserting results. This means we can work with integers locally, and can use them in the database too, which should speed everything up, as there's a lot less data sloshing about.

Changes proposed in this pull request

erDiagram
    Sources {
        bigint resolution_id PK,FK
        string alias
        string schema
        string table
        string id
        jsonb indices
    }
    Clusters {
        bigint cluster_id PK,FK
        bytes cluster_hash
        bigint dataset FK
        array[string] source_pk
    }
    Contains {
        bigint parent PK,FK
        bigint child PK,FK
    }
    Probabilities {
        bigint model PK,FK
        bigint cluster PK,FK
        float probability
    }
    Resolutions {
        bigint resolution_id PK,FK
        bytes resolution_hash
        enum type
        string name
        string description
        float truth
    }
    ResolutionFrom {
        bigint parent PK,FK
        bigint child PK,FK
        int level
        float truth_cache
    }

    Sources |o--|| Resolutions : ""
    Sources ||--o{ Clusters : ""
    Clusters ||--o{ Contains : "parent, child"
    Clusters ||--o{ Probabilities : ""
    Resolutions ||--o{ Probabilities : ""
    Resolutions ||--o{ ResolutionFrom : "child, parent"

Loading

• Move entire ORM to use integer primary keys
• Temporarily adapt client to pass tests with this change

Guidance to review

Start with the ORM and new README diagram, then the server side changes, then the client.

I've chosen to rename hash to id on coming out of Matchbox. I'm worried this is too generic. What do you think?

Checklist:

• My code follows the style guidelines of this project
• New and existing unit tests pass locally with my changes

[leo-mazzone]

We have reviewed on a call, all looks good enough