Improve ApAutoBadcols class and ap_auto_badcol script

[DaveStrickland]

The ApAutoBadcols class and associated ap_auto_badcol.py script should:

1. Catch cases where the local standard deviation is zero and not emit RuntimeWarning: divide by zero encountered in divide messages.
2. Have the ability to generate optional files (CSV, TSV?) showing computed statistics of all rows and/or columns, to allow validation of the algorithm and similar diagnostic investigation. (Prompted by strong spatial patterns in the T24 RBI-flushed master darks).
3. Ability to generate the user bad column (row, rectangle) YaML instead of requring the user to copy and edit the template, then cut/paste the output of ap_auto_badcol. into that file.
   • YaML generation should be integrated into ApAutoBadcols
   • Should by default not over-write an existing YaML file of the specified name

[DaveStrickland]

Changes:

• A "user bad pixel" format YaML file can be written, instead of the user having to cut/paste into a template they had to copy from elsewhere.
  • By default an existing user bad pixel YaML file will not be overwritten, but options have been added to allow deliberate over writes.
  • Axis (along column, along row) statistics can be optionally written as CSV files
  • A plot of the column and row statistics can also be generated
  • A bug in calculating bad rows was found and corrected