[GH2][Correct] Gain constant needs to be float32

This fixes the reported issue at https://redmine.xfel.eu/issues/147905

Description

Corrections were failing because the Cython correction function expected float32 constants. As the gain constant was injected as float64, I made sure that the gain constant is converted to the right type before being used for correction.

Marco was also notified and he will make sure that notebook is injecting gain constant of the expected type.

How Has This Been Tested?

Tested locally and hotfixed in production

Relevant Documents (optional)

Types of changes

• Bug fix (non-breaking change which fixes an issue)

Checklist:

Reviewers

@kluyvert @hammerd

notebooks/Gotthard2/Correction_Gotthard2_NBC.ipynb

         mask[index, ...],
         g,
         const_data[mod]["Offset"],
-        const_data[mod]["RelativeGain"],
+        const_data[mod]["RelativeGain"].astype(np.float32, copy=False),
         const_data[mod]["Mask"],
         apply_offset=offset_correction,
         apply_gain=gain_correction,