Feature : Upload Accepted Predictions

[kshitijrajsharma]

What does this PR do?

• Allow users to upload accepted predictions
• Introduce a new Dockerfile specifically designed for API usage

ApprovedPrediction Table

The new ApprovedPrediction table will have the following columns:

Column Name	Description
training	Training ID checkpoint used during prediction approval
config	JSON field for storing configuration data (e.g., zoom level, vectorization parameters)
geom	Geometry field supporting point, line, and polygon predictions
approved_at	Automatic datetime field for approval timestamp

Note

No user information will be stored regarding who accepted the feature. Accepted features only be accepted if user decide to submit them with no user identification, Requires login in-order to push accepted features

Testing

To test the new functionality, ensure you have the following endpoints available:

Additional Steps

• Apply the necessary migration to complete the setup.

Thought Process

Thought process on designing this feature on fAIr is to enable us to store the accepted predictions made from fAIr so that later on it can be forwarded to tools / users as release . Once we have approved features on our db usecase will be widen , it can either be pushed to OSM , or be pushed to other platforms. This will also enable for other third party tools let user validate the fAIr predictions once they do they will push here . This feature supports the concept of turning mapper into validators

[omranlm]

We need a migration for this change, right? and the check failed for migration

[kshitijrajsharma]

We need a migration for this change, right? and the check failed for migration

Yes correct , I am trying to fix the CI before merging !