Adding feature: iodaconverters for bufr files.

[cohen-seth]

Description

This PR adds the base suite and tasks for integrating the IODA converter 'bufr2ioda.x' within the swell workflow. This is only for BUFR files specifically.

This PR addresses the issues link below:

• Adding Bufr2IODA to SWELL #435
• BUFR2IODA Conversion #378 - the section starting at '1. Offline converter: once SWELL the steps above are complete:'

I used existing files for ncdiag as templates for the flow.cylc and get_bufr.py files.

Purpose of scripts:

(1) src/swell/tasks/bufr_to_ioda.py: Executes the conversion using both get_bufr.py and get_bufr_iodaconv_yaml.py

(2) src/swell/tasks/get_bufr.py : Locates the bufr files and creates links in the swell experiments run folder(s).

(3) src/swell/tasks/get_bufr_iodaconv_yaml.py : The ioda converter (bufr2ioda.x) requires a yaml file within which is the data input path, the data output path and information about variables which are specific to observation type (i.e. AMSUA vs GPSRO vs ...) - bufr2ioda.x converts ONE file at a time so get_bufr_iodaconv_yaml.py enables the process of locating the input bufr file, identifying the observation type, specifying the output path, and generating the yaml file based on the observation type. The current method uses template yaml files from the iodaconv/test/testinput/ folder; these files are generated by swell when it builds JEDI so the relative paths for them should stay consistent.

(4) src/swell/suites/convert_bufr/flow.cylc : the suite for all the above

Dependencies

1. IODA Converters (JEDI) must be included in the jedi build
2. The yaml files from the iodaconv/test/testinput/ folder must be included in the jedi build