Wrap NUTS for non-linear fitting features

[Copilot]

• Add NUTS algorithm support for non-linear fitting features (BazinFit, LinexpFit, VillarFit)
• Three new algorithms: "nuts", "nuts-ceres", "nuts-lmsder"
• NUTS enabled unconditionally (no cargo feature flag)
• Parameters: nuts_ntune and nuts_niter
• Tests: precision, algorithm membership, custom params, benchmark classes
• CHANGELOG updated
• All lints pass (cargo fmt, clippy, ruff)
• Fix CI: Bump MSRV from 1.85 to 1.88 (required by nuts-derive which uses let chains stabilized in 1.88)
• Fix clippy collapsible_if lint triggered by MSRV bump enabling let chains
• Merge main (f4ec272) into branch — proper git merge commit with two parents

Original prompt

---

This section details on the original issue you should resolve

<issue_title>Wrap NUTS for non-linear fitting features</issue_title>
<issue_description></issue_description>

<agent_instructions>light-curve-feature now supports NUTS for non-linear fit features. We should add its support here, in the same way we support mcmc, e.g. add three new "algorithms" for non-linear features. Please do not forget to add tests, run them, update changelog, and run formatter and linter after you've done.</agent_instructions>

Comments on the Issue (you are @copilot in this section)

• Fixes Wrap NUTS for non-linear fitting features #631

---

✨ Let Copilot coding agent set things up for you — coding agent works faster and does higher quality work when set up for your repo.

[hombit]

@copilot Good job.

1. Let's enable nuts unconditionally, no need in a Cargo feature
2. I think we should homogenize argument names. Let's keep nuts_ntune, but use nuts_niter instead of nuts_ndraws.

[Copilot]

@copilot Good job.

1. Let's enable nuts unconditionally, no need in a Cargo feature
2. I think we should homogenize argument names. Let's keep nuts_ntune, but use nuts_niter instead of `nuts_ndr...

Done in fe1c7c1:

1. nuts is now enabled unconditionally — removed the cargo feature flag, added nuts directly to light-curve-feature dependency features, removed all #[cfg(feature = "nuts")] guards, and removed _built_with_nuts from lib.rs.
2. Renamed nuts_ndraws → nuts_niter everywhere (Rust signature, Python tests, changelog).

[codecov]

Codecov Report

❌ Patch coverage is 83.33333% with 3 lines in your changes missing coverage. Please review.
✅ Project coverage is 82.78%. Comparing base (f4ec272) to head (388bd6c).
⚠️ Report is 1 commits behind head on main.

Files with missing lines	Patch %	Lines
light-curve/src/features.rs	83.33%	3 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #635      +/-   ##
==========================================
+ Coverage   82.70%   82.78%   +0.07%     
==========================================
  Files          10       10              
  Lines        2608     2620      +12     
==========================================
+ Hits         2157     2169      +12     
  Misses        451      451              

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[codspeed-hq]

Merging this PR will not alter performance

⚠️ Unknown Walltime execution environment detected

Using the Walltime instrument on standard Hosted Runners will lead to inconsistent data.

For the most accurate results, we recommend using CodSpeed Macro Runners: bare-metal machines fine-tuned for performance measurement consistency.

✅ 106 untouched benchmarks

---

_{Comparing copilot/add-nuts-support-nonlinear-fitting (045f9dc) with main (098edef)}

[hombit]

@copilot, please fix CI failures

[Copilot]

@copilot, please fix CI failures

Fixed in 045f9dc. The msrv-build job failed because nuts-derive 0.1.0 uses let chains (if let ... && ...) which were stabilized in Rust 1.88, but the MSRV was 1.85. Bumped MSRV from 1.85 to 1.88 and relaxed the comfy-table version constraint (which was pinned to <7.2 due to the old MSRV).

The publish workflow failure (Build cp310-macosx_x86_64) was a transient homebrew download issue (failed to fetch metis and gsl packages), unrelated to this PR.

Warning

Firewall rules blocked me from connecting to one or more addresses (expand for details)

I tried to connect to the following addresses, but was blocked by firewall rules:

• https://api.github.com/repos/light-curve/light-curve-python/actions/jobs/64721940074/logs
  • Triggering command: /usr/bin/curl curl -s REDACTED (http block)

If you need me to access, download, or install something from one of these locations, you can either:

• Configure Actions setup steps to set up my environment, which run before the firewall is enabled
• Add the appropriate URLs or hosts to the custom allowlist in this repository's Copilot coding agent settings (admins only)

[hombit]

@copilot cargo clippy lint fails

[Copilot]

@copilot cargo clippy lint fails

Fixed in 85e0650. The collapsible_if lint was triggered because the MSRV bump to 1.88 enabled let chains, so clippy now suggests collapsing if let Some(sigma) = sigma { if t.len() != sigma.len() { ... } } into a single if let with &&.

[hombit]

@copilot solve merge conflicts with main

[Copilot]

@copilot solve merge conflicts with main

Resolved in bfd52fd. The conflict was in light-curve/Cargo.toml — main bumped pyo3 to 0.28.2, pyo3-arrow to 0.17, and arrow-* to 58. I applied those version bumps while keeping our comfy-table = "7" (no upper bound needed since MSRV is 1.88). Also brought in the non-conflicting changes from main (upload-artifact v7, download-artifact v8, ruff v0.15.4).

Note: this applies the changes from main on top of our branch rather than a true git merge/rebase, so GitHub may still show conflicts in the PR. If so, you may need to merge main from the GitHub UI or locally.

[hombit]

@copilot no, you merge/rebase on main

[Copilot]

@copilot no, you merge/rebase on main

Done in e7354cc — a proper merge commit with two parents (bfd52fd from our branch and f4ec272 from main). The only conflict was comfy-table in Cargo.toml (kept our "7" without the <7.2 upper bound, since MSRV is 1.88) and a Cargo.lock ordering conflict (kept our version including pulp v0.21.5 required by nuts-rs).