Rfwtools

This release add support for Jefferson Lab's Certified Software Initiative. This includes a new setup script for easier certified installation.

This release has two minor updates, but one includes a change to the API.

1. Fixes a bug when parsing capture files for a fault event where cavities were being sampled at different rates. Only data points captured at timestamps matching the first file processed were kept. Now all timestamps are kept which can introduce NaNs. However, validation routines can detect the different sample rates where as before they were missed.

1. Changes the get_capture_file_contents method to get_capture_file_names. The previous method was rarely used, and only to get a list of file names. Drop the time consuming file reads in favor of only presenting the file names.

This release fixes a few problems and updates code to work with newer dependency versions.

- The example validation routine used by DataSet.produce_example_set(), is not multithreaded and uses 8 workers. More than this didn't result in faster processing as data had to be read from disk and multiple calls had to be made to accweb.
- Updated code that used "inplace" parameters for calls to pandas DataFrames, and old_names is statsmodels.AutoReg models as they are now deprecated and threw many warnings
- Removed version lock on matplotlib. Version 3.2.1 produced some buggy behavior, but we are well past that version.
- Setup build tools to work on Windows
- Added additional docstrings
- Updated ExampleValidator interface to allow for the specification of a mya deployment at construction. Makes testing simpler.
- Updated CSV saving routines to always use '\n' as a line terminator. Saving on windows would produce blank lines otherwise.
- Other minor fixes

Note: Sphinx has a bug that autosummary does not handle inherited parameters well. They have a fix staged for the last month or two, but have not released it. I will need to update the documentation when this happens.

This release fixes a minor problem where loading an ExampleSet from a saved CSV file is difficult if the ExampleSet should contain a non-default Example type. Now the DataSet.load_example_set_csv method accepts information on the type of Example that should be created.

To reduce confusion over the circumstances under which an ExampleSet were created, the ExampleSet.save_csv method creates a header with information about the ExampleSet's internal parameters at the time of save file creation. This header is ignored by ExampleSet.load_csv, and exists for information purposes only.

Found and fixed errors in the Usage Examples of the rfwtools.extractor.* modules. This update includes no code changes.