Fix det cal hwpss

[ykyohei]

This PR fixes issues found when I ran det_cal with hwpss subtraction for all the old observations.

1. Inconsistent number of channels between raw biasstep npy file and L3 oper book. -> updated load_and_reanalyze_bs to load fixed tones and reindex axismanager.
2. smurfgaps in biases leads to incorrect bias edges. -> Added correction in fill_zeros_biases
3. Check if data of tones is None or not to debug the error of loading fixed tones. Resolves data of tones are sometimes None #1508

[ykyohei]

Additional modifications of handling of fixed tone are done and this PR is again ready to review.
I tested this for backfilling det_cal_hwpss for old observations.
I modified load_book.py to debug #1508. I would appreciate if @kmharrington could take a look at this point.