v0.6.0

[zairving]

Finishing touches for v0.6.0.

• Updated Read the Docs Python version to 3.14
• Added missing links to a number of docstrings
• Fixed some formatting issues in some notebooks
• Added a new safeguard to prevent a Reducer from being instantiated if the specified out_directory already contains a reduction_parameters.json file that does not match the Reducer instance

[zairving]

This PR got a bit larger than expected. Summary of changes:

• Various documentation/docstring improvements.
• Added a dedicated timing module to hold timing routines (e.g., applying barycentric time corrections and bespoke routines for operating on astropy.timeseries.TimeSeries instances).
• Improved reduction parameter logging and added safeguards to prevent multiple Reducer instances using the same output directory but using different reduction parameters.
• Added calibration terms to Reducer.plot_noise() and S/N calculations.
• Image grouping now uses camera AND filter information - prevents issues with instruments using multiple cameras with the same filters.
• Forced photometry now works as intended. Previously, source coordinates were not being transformed correctly.
• Removed requirement for data files to contain a fit extension. Now all files in the specified data directory are assumed to be FITS files, regardless of extension. Files that cannot be opened using astropy.io.fits.open() (i.e., non-FITS files) are ignored.
• Analyzer.rebin() now works as intended. Specifically, all light curves are rebinned using a common reference time (to maintain simultaneity), and gaps are not padded with zeros.
• Updated pixel-scale for MEXMAN instrument.

[NCastro-Segura]

Hey, thanks for putting this together! The PR description undersells it a bit — there are some genuinely important fixes buried in here alongside the docs work, so good to see them land.

A few things I wanted to flag:

Forced photometry fix — this one's subtle but critical. The catalog coordinates weren't going through the per-image affine transform, so results were effectively pinned to the reference frame regardless of the actual image offset. The fix looks correct. One thing worth considering: users who ran forced photometry with earlier versions will have silently wrong light curves. Might be worth a note in the changelog so they know to rerun their reductions.

Image grouping by camera + filter — good catch. Grouping by filter alone was always going to be a problem for a multi-camera instrument where all three cameras can share the same filter name. The camera_and_filter_key() approach is clean and consistent with how the rest of the code handles this.

rebin() simultaneity and gap handling — both fixes make sense. Using t_ref as the common time_bin_start across all keys is the right call for multi-band work where you need the time grids to align. And removing gap bins via GTIs rather than padding with zeros is the scientifically correct behaviour. One minor suggestion: it'd be worth calling out the simultaneity guarantee explicitly in the rebin() docstring — it's a non-obvious property that users doing multi-band differential photometry will care about.

FITS file discovery — pragmatic change, I like it. One small thing: glob('*') will also match subdirectories, and fits.open() on a directory throws an exception that gets caught and emits a warning. It won't break anything, but the warning message will be confusing. A .is_file() check before attempting to open would clean that up.

Reducer parameter safeguard — solid addition. The json.dumps comparison works fine for now, though it's worth keeping in mind that if floating-point parameters get added in the future, formatting differences could theoretically cause false positives.

MEXMAN pixel scale — glad this got corrected. Was this from an updated instrument spec or measured on-sky?

ReadTheDocs Python 3.14 — just double-checking: is 3.14 available yet on the ubuntu-22.04 RTD image? It was still prerelease last I checked. Worth verifying the docs build doesn't break after merge.

Nothing here is a blocker. Happy to approve once you've had a chance to look at the .is_file() suggestion and the changelog note for the forced photometry fix.

[zairving]

Cheers, @NCastro-Segura!

I've implemented your suggestions for FITS file discovery and the rebin() docstring. I've also added a few small changes I've made since opening this PR.

The MEXMAN pixel scale came from observers at OAN-SPM.

I don't think there are any issues with Python 3.14 support any more. Indeed, the dev page of the docs appears to work perfectly: https://opticam.readthedocs.io/en/dev/