BUG - SWE - failure to write CDF when using latest SWAPI input (Phase D)

[jtniehof]

Description of the issue

When using the latest swapi_l3a_proton-sw inputs that include swp_flags, writing the swe_l3_sci CDF fails because there are no metadata for swp_flags.

I don't recall if it was intended to carry these through (in which case we should put the variable in the SWE metadata) or not (in which case I'm guessing we need to delete swp_flags from the data set).

Steps to reproduce the issue

imap_swe_l3_sci-e979d33c_20260120_v001.json has the dependencies for one particular failed run.

Expected vs Actual behavior

No response

Code Snippet (If applicable)

Traceback (most recent call last):
  File "//imap_l3_data_processor.py", line 180, in <module>
    raise e
  File "//imap_l3_data_processor.py", line 177, in <module>
    imap_l3_processor()
  File "//imap_l3_data_processor.py", line 127, in imap_l3_processor
    paths = processor.process()
            ^^^^^^^^^^^^^^^^^^^
  File "/imap_l3_processing/swe/swe_processor.py", line 38, in process
    product = save_data(output_data)
              ^^^^^^^^^^^^^^^^^^^^^^
  File "/imap_l3_processing/utils.py", line 100, in save_data
    write_cdf(file_path_str, data, attribute_manager)
  File "/imap_l3_processing/cdf/cdf_utils.py", line 25, in write_cdf
    variable_attributes = attribute_manager.get_variable_attributes(var_name)
                          ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/.venv/lib/python3.12/site-packages/sammi/cdf_attribute_manager.py", line 475, in get_variable_attributes
    if attr_name in self._variable_attributes[variable_name]:
                    ~~~~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^
KeyError: 'swp_flags'

Additional notes, affected areas, and suggested fixes

No response

[jtniehof]

8c27006 explicitly propagates through the swp_flags, so clearly the intention is to have them on the SWE output. We either need to copy the SWAPI metadata into the SWE metadata, or reference it somehow. I'm disturbed this wasn't caught by a test; that indicates there may be something more subtle going on. Are we potentially not testing save_data from the SWE processor with a recent SWAPI input?

[pleasant-menlo]

Fixed in 4707765. SWE processing is now completing successfully in prod.