…pper using STAT_ANALYSIS_JOB<n>, clean up indentation

…ing and not used in the plotting scripts

…ead of using out-of-date file that should no longer be used

… to prevent tests from passing when they should not

This reverts commit a44c866.

…ots do not break before adding other logic to StatAnalysis wrapper

…n-diff

…r setting groups of list items that can be looped over, ci-run-diff

… from the FCST/OBS_VAR<n> variables and around level values read from FCST/OBS_LEVEL_LIST, ci-run-diff

…dds quotation marks to format the values instead of handling that later in the processing

…ort HHHMMSS (3 digits of hour). Updated tests to handle example from removed configuration file in internal/tests/pytests/plotting

…n each, add comments to describe the intended behavior

… class variables

…lues into helper function to clean up logic and make it more readable, ci-run-diff

…ad offsets to be consistent with other wrappers, ci-run-diff

…ad offsets to be consistent with other wrappers, ci-run-diff

…gic and to be able to handle a loop list that contains a group of items, e.g. FCST_LEAD_LIST1, FCST_LEAD_LIST2, ci-run-diff

…tc. works as expected. Moved logic to set config_dict to function, changed default values to unset hour lists, set hour lists as needed in tests.

…eads, and vice versa. ci-run-diff

… hours are set instead of using date END + 23:59:59, ci-run-diff

…computes all combinations of runtime settings so that it is understandable by other developers, added documentation

…settings instead of performing a deep copy of self.c_dict since many of those settings are not needed for runtime settings, ci-run-diff

…not used

…he JOB args

…ed to ensure that the files processed are restricted to the time range specified in the config file with INIT/VALID_BEG/END. ci-run-diff to see if this changes any of the use case output

…as matching times to produce output when restricting the obs valid range to the range specified in the config file

…mments, fixed pylint complaints, replaced object variables not set in init, etc.

…propriate area

…ncating YYYYMMDD. If only YYYYMMDD is set in the _END time and no hours are specified in [FCST/OBS]_[INIT/VALID]_HOUR, then set HHMMSS to 23:59:59 to ensure full day is processed to preserve old behavior.

… since it is not used, users should specify the template they want explicitly

…P_ORDER = times logic no longer exists. keeping stub of function because some of the logic can be used if/when StatAnalysis becomes a RuntimeFreq wrapper to be able to specify different time frequencies to run. It will always process the full range of times specified for now

…DD_HHMMSS into constant variable since it is referenced frequently, ci-run-diff

…o set a filename template for [dump_row_file] and [out_stat_file] instead of generating a default template. Add error checking to ensure that at least 1 model is specified with MODEL<n>* variables and the corresponding filename template variable is set if [dump_row_file] or [out_stat_file] are found in any jobs. Added unit tests to ensure init set catches these cases to alert users of errors before attempting to run

…n attempting to add them to each item

…ove MODEL<n>_REFERENCE_NAME because it was only used by MakePlots wrapper

…nning build_docs.py script locally to easily see if there are any fixes needed to the docs

… list of required and optional config variables for StatAnalysis wrapper

…nd adding info to glossary items about looping over groups of list items

…n StatAnalysis wrapper

…y YYYMMDD is specified in VALID/INIT_END

…d up logic to set command line arguments to simplify how command is built, added documentation to functions and cleaned up code

…es not break existing use cases, ci-run-all-diff

…n hour lists and only set them if explicitly requested in the METplus config

…tring template substitution on any env var that will be set by wrapper will be performed before calling stat_analysis

…fcst/obs_init/valid_beg/end values from config. prevent crash when dict values are not set by using get

… read from METplus config and set in env vars for wrapped MET config file properly

…ort very old wrapped MET config files for StatAnalysis

…hat was removing first letter of sentence for CUSTOM_LOOP_LIST

…put under METPLUS_TEST_OUTPUT_BASE and added logic to remove that directory before running tests. This prevents false success of tests that generate files and checks if they exist if they were generated from a previous successful run. Write to directory with hard-coded name to prevent accidentally scrubbing other files if the env var is not set properly

…lus.py script returns an expected failure return code

…:dtcenter/METplus into feature_24i_test_failure

…fig file will automatically scrub the test output including logs and final conf only if the test succeeds. This makes it much easier to review test output to investigate issues. Tests that currently read in additional configuration files now call the metplus_config_files fixture to preserve the old behavior but those cases will not automatically clean up the output data. Eventually update those tests to set the config settings in the test instead of reading a separate file so that all tests that use a METplusConfig object will be scrubbed unless they fail

…le instead

…Analysis

…24i_test_failure

…d if the test passes, scrub files that must be generated using the old method, add documentation to pytest fixtures