Setting saturated pixels to NaN fails if image is integer

[mwcraig]

While testing some of the changes with the new photometry settings I was using an uncalibrated image. single_image_photometetry failed on this line with

File ~/Development/astronomy/stellarphot/stellarphot/photometry/photometry.py:222, in single_image_photometry(ccd_image, sourcelist, camera, observatory_location, aperture_radius, inner_annulus, outer_annulus, shift_tolerance, max_adu, fwhm_estimate, use_coordinates, include_dig_noise, reject_too_close, reject_background_outliers, passband_map, fwhm_by_fit, fname, logline)
    219     return None, None
    221 # Set high pixels to NaN
--> 222 ccd_image.data[ccd_image.data > max_adu] = np.nan
    224 # Extract necessary values from sourcelist structure
    225 star_ids = sourcelist['star_id'].value

ValueError: cannot convert float NaN to integer

I think the fix can be just converting the data to float.

[JuanCab]

Do you want to do this (cast image to float) or do you want me to do this?

[mwcraig]

Either works, but if you are interested in trying out pydantic then maybe try making a Camera setting instead. This branch on my fork has, in stellarphot/settings/models.py my intial attempt at adding units to committee (ignore the ApertureUnits class and look at Length instead.

Not directly relevant to Camera but moving in the right direction.

[JuanCab]

Submitted PR #158 to address this. Can you test your integer image against it?