Currently all metrics (e.g., KGE, NSE, etc.) are per-calculated with a predefined set of filters, thresholds, group-bys, etc. This was primarily done because the metrics calculations can take a long time to run depending on the complexity and data volume. Under this issue we will implement an asynchronous analytics API that allows users to submit requests and then be notified when results are ready.
Some initials concepts and thoughts:
- We already use Prefect and Prefect has some capability to be used for this use case. We should consider this.
- We have fairly robust Pydantic Model definitions for most (maybe all) metric requests. We should explore how these can be leveraged to build the API.
This issue needs more details.
Currently all metrics (e.g., KGE, NSE, etc.) are per-calculated with a predefined set of filters, thresholds, group-bys, etc. This was primarily done because the metrics calculations can take a long time to run depending on the complexity and data volume. Under this issue we will implement an asynchronous analytics API that allows users to submit requests and then be notified when results are ready.
Some initials concepts and thoughts:
This issue needs more details.