Polynomial pointing model

[axf295]

add the polynomial pointing model to mapcat

[Copilot]

Pull request overview

Adds a new polynomial-based pointing residual model to Mapcat and updates persistence/tests to support storing/retrieving this new model type alongside the existing constant model.

Changes:

• Introduces PolynomialPointingModel with coefficient fitting, prediction, and basic statistics helpers.
• Updates PointingResidualTable.residual_model to store either a constant or polynomial model via JSON (TypeAdapter-backed deserialization).
• Expands pointing tests to cover constant-model behavior and basic polynomial fitting/prediction.

Reviewed changes

Copilot reviewed 5 out of 5 changed files in this pull request and generated 7 comments.

Show a summary per file

File	Description
tests/test_pointing.py	Updates unit imports and adds tests for constant and polynomial pointing models.
mapcat/pointing/poly.py	New polynomial pointing model implementation (fit + predict + stats).
mapcat/pointing/const.py	Adjusts ConstantPointingModel.predict to subtract offsets.
mapcat/database/pointing_residual.py	Allows storing either constant or polynomial pointing model in residual_model.
mapcat/database/json.py	Switches JSON deserialization to Pydantic TypeAdapter (enables unions).

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[JBorrow]

Some minor comments. It seems kinda overcomplicated for a polynomial model but that's ok. Is there a reason we can't just use https://numpy.org/doc/stable/reference/generated/numpy.polyfit.html

[axf295]

@JBorrow
re: numpy poly fit, I think that is only 1D?

other changes have been included; removed separate ra,dec uncertianty + weights; just use a weights input since they're degenerate and the user can decide what weights to use.