Added statistical test output to ocrreject_exam

[m-dallas]

Major improvements
• Renamed avg_ratio to combined_ratio since the value isn’t really an average.
• Added the new method combined_ratio_threshold() which generates the highest combined_ratio possible under the null hypothesis that CRs are randomly distributed across the detector given a significance level (alpha) and a mean number of pixels each CR hit is expected to impact (cr_size). This was developed in response to feedback that there should be an included way for users to immediately see if a dataset was likely overflagged.
• Added overflagged_stat to the results output. The value is True if the observed combined_ratio >= to the threshold set by combined_ratio_threshold(). Comparing to manually checked datasets from 2016-2025 shows this stat is agrees with the manual classification of the data ~93% of the time.

Bug fixes:
• There was a bug in how the extraction box region was defined. Previously it would always extend 1 pixel lower than it should. This updated version is the same as what was shown in STIS ISR 2019-02 and ensures the correct extraction box size.
• There was a bug where subarray readouts wouldn’t properly render in the html files. This was fixed.
• Trying to only show figures in an interactive environment was buggy (especially with qt), so this was changed to the “show_plot” argument which allows users to choose if they want the plots to render or just to be saved to a directory.

The tests and docstrings have been updated to reflect these changes. This also shouldn’t break the plot rendering in the STIS notebooks but it’s worth double checking.

[sean-lockwood]

Suggest making PEP8 changes as suggested via:

pycodestyle --ignore=E501,W503,W504 ocrreject_exam.py

I haven't tried running the tool yet.

[m-dallas]

Updated the script to comply with PEP8, I also double checked that the tool is giving the results we expect.