Add magnitude-limited catalog completeness utilities

[nikosarcevic]

This PR adds utilities for modelling magnitude-limited catalog completeness from luminosity functions.

It introduces functionality to:

• convert apparent magnitude limits into absolute magnitude limits
• integrate luminosity functions over observed and missing magnitude ranges
• compute catalog-completeness and out-of-catalog fractions
• support both scalar and array redshift inputs
• validate inputs with clear errors

This is intended to provide the LFKit-side machinery needed for magnitude-limited catalog selection modelling while keeping the implementation science-use-case agnostic.

Changes

• Adds catalog_completeness photometry utilities.
• Adds shared typing utilities.
• Adds validation tests for scalar and array inputs.
• Adds tests checking completeness and out-of-catalog fractions are bounded between 0 and 1.
• Adds a fake magnitude-limited demo catalog and generator script.
• Adds documentation and runnable examples for luminosity-function and catalog-completeness workflows.
• Updates API docs to expose the new completeness functionality.

Closes #6

• Scalar and array redshifts work.
• Apparent magnitude limits can be converted to absolute magnitude limits.
• LF integrals over observed and missing magnitude ranges are supported.
• Catalog-completeness and out-of-catalog fractions are returned.
• Fractions are bounded between 0 and 1.
• Invalid inputs raise clear errors.
• Basic tests are included.

Notes

Tabulated/user-supplied luminosity-function arrays are not included in this PR. The current implementation supports analytic/callable LF models, and tabulated LF support can be added later as a separate extension without changing the catalog-completeness interface