Handle calibration files with extra cameras

[ErwanBeurienne]

Hello !

In this pull request, I add a new helper function, adapt_P_and_calib_params(), to handle cases where the number of cameras defined in the calibration files exceeds the number of pose folders available. This function issues clear warnings and automatically adapts the calibration parameters so that only the relevant cameras are used.

The main changes are:

1. Add adapt_P_and_calib_params() to centralize the logic for aligning calibration parameters with available pose folders.
2. Update personAssociation.py and triangulation.py to call this function when loading calibration data.
3. Improve logging messages to make it explicit when some cameras from the calibration files are ignored because they have no corresponding pose folder.

This should make the pipeline more robust to inconsistent calibration/pose configurations while preserving existing behavior when the inputs are consistent.

[davidpagnon]

Thanks, Erwan!
I'll probably have a look at it next week :)

[ErwanBeurienne]

I pushed a follow-up fix: the previous approach could mismatch camera parameters when the calib file contains non-contiguous camera IDs (e.g. cam1, cam2, cam4, cam6).

The filtering is now done by camera name (from calib file) and json pose folder directly in retrieve_calib_params() and computeP(). adapt_P_and_calib_params() was removed.

[davidpagnon]

Hi Erwan,

Let me know if I misunderstood anything:

• In common.py, you added the optional parameter json_pose_dirs_names in retrieve_calib_params and computeP, and used it in personAssociation.py and triangulation.py.
• If the parameter is not used, the behavior is unchanged. Otherwise, it associates cameras in the calibration file with those in the _jons pose files based on their names

This looks good to me, however could you address the couple of questions I asked in the review?

Thanks, and I wish you a great new year's eve!