packet_file_to_datasets *NOT* in IDEX

[anamanica]

Change Summary

Overview

Originally, this PR was intended to switch over IDEX to using the packet_file_to_datasets function established in utils.py, however it was decided that this would be more work than it is worth. Due to the nature of the .xml file for IDEX, using this function created dimension issues. For now, IDEX will continue to use PacketParser, and RawDustEvents

I added a comment in the doc-string for packet_file_to_datasets to explain this.

New Dependencies

None

New Files

None

Deleted Files

None

Updated Files

• utils.py
  • Added comment to packet_file_to_dataset docstring.

Testing

None

closes 687 for IDEX only!

[bourque]

Thank you for trying to make this work! It sounds like you learned some things along the way so that is good!

[greglucas]

Looks good to me. Some very minor nits, that could be follow-on PRs if desired too. I think its great to add documentation about the limitations, so thank you for adding this!

[greglucas]

I think this should be either a Notes section in numpydoc, https://numpydoc.readthedocs.io/en/latest/format.html#notes or a .. note:: rst callout.
https://sublime-and-sphinx-guide.readthedocs.io/en/latest/notes_warnings.html

Some minor wordsmithing, where I'd avoid saying what we'll do with the specific classes in idex_l1.py in case someone refactors that later and forgets to update this docstring.

Suggested change

	Note: While this is a general function intended to work for all
	instruments, this cannot be used for IDEX due to the complex nature of
	the xml file. IDEX will continue to use PacketParser and RawDustEvents
	found in idex_l1.py.
	.. note::
	This function only handles packet definitions with the same variable structure across
	all packets with the same ApId. For example, this cannot be used for IDEX due to the
	conditional XML structure defined for their science packet.

One other thought that occurred to me is whether we could catch this and print out a nice warning message here so others don't also hit this roadblock. Something like creating a new dictionary that keeps track of apid: variables. Around l365 below I think something like this might work?

            if apid not in data_dict:
                # This is the first packet for this APID
                data_dict[apid] = collections.defaultdict(list)
                datatype_mapping[apid] = dict()
                variable_mapping[apid] = packet.data.keys()
             if packet.data.keys() != apid_variable_mapping[apid]:
                 raise ValueError(f"Cannot parse apid {apid} which has different variables defined across the same packet type")

[anamanica]

Sweet! Yeah, I agree with you. I changed the comment to a "Notes" section in the doc-string.

I didn't end up implementing the dimension check in this PR (I just don't quite understand the code yet, and I feel like I should before I add it and test it). Would you like me to write another issue ticket?

[greglucas]

Yeah, that'd be great. I think the code above is close to working, so if you copy that over to another issue and work from that it might be a good thing. Additionally, add a test trying to parse IDEX and showing that your ValueError is raised with pytest.raises(ValueError, match="Cannot prase apid").