update approach to isotonic calibration

[malcolmbarrett]

• update isoreg calibration
• remove news file until submission

[Copilot]

Pull request overview

This PR updates the isotonic regression calibration approach in ps_calibrate() to use a custom weighted pool-adjacent-violators algorithm (PAVA) instead of stats::isoreg(). The new implementation follows the two-step isotonic regression method described in van der Laan et al. (2024) and matches the behavior of WeightIt::calibrate(method = "isoreg").

Changes:

• Implements custom pava_weighted() function for isotonic regression without grouping tied x-values
• Updates ps_calibrate() to use two-step calibration: separate isotonic regressions for treated and control groups
• Adds comprehensive unit tests for pava_weighted() covering edge cases and expected behavior
• Removes stats::isoreg import from NAMESPACE (no longer needed)
• Removes NEWS.md file temporarily until ready for submission

Reviewed changes

Copilot reviewed 5 out of 5 changed files in this pull request and generated 2 comments.

Show a summary per file

File	Description
R/ps_calibration.R	Adds new pava_weighted() internal function and updates ps_calibrate() to use two-step isotonic regression approach
tests/testthat/test-ps_calibration.R	Adds 8 comprehensive unit tests for pava_weighted() function covering various edge cases
NAMESPACE	Removes stats::isoreg import that is no longer used
NEWS.md	Removes file temporarily (per PR description, will be added back for submission)
.Rbuildignore	Adds .beads to build ignore list (likely a development tool configuration)

---

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