SequenceUploadQueue add_to_upload_queue incorrectly describes column_external_ids

https://github.com/cognitedata/python-extractor-utils/blob/3e99e07943a7bd0c16ca3f4bd224a81c2052480f/cognite/extractorutils/uploader/time_series.py#L428-L452

The parameter description states:
column_external_ids: List of external id for the columns of the sequence

But the parameter takes in:
column_external_ids: Optional[List[dict]] = None

As opposed to a List[str]. It then will try to plug it into the columns parameter SequenceData(id=id, external_id=id, rows=rows, columns=column_external_ids) which will later result in error:

return [cast(str, c.get("externalId")) for c in self.columns]

SequenceData is expecting an API like dict for columns:
https://docs.cognite.com/api/v1/#tag/Sequences/operation/createSequence

Possible options:

1. Change the parameter for add_to_upload_queue from column_external_ids to columns but that will change the look and feel. The SDK feel is to insert a list of str as column_external_ids.
2. Keep it as column_external_ids and in add_to_upload_queue, generate the API like dict for columns from the list of str column_external_ids.
3. Within the SequenceUploadQueue, there is a function set_sequence_column_definition that sets self.column_definitions but it isn't used unless a create is being called. It feels super odd when you want to create on missing Sequences and call set_sequence_column_definition with column definitions, but when calling add_to_upload_queue, you must additionally add column definitions again.

Start with add_to_upload_queue in cognite/extractorutils/uploader/time_series.py at the linked lines, then inspect SequenceData and set_sequence_column_definition. Compare the documented column_external_ids contract with the columns value passed to SequenceData and the referenced Sequences API format. Done means choosing one consistent contract and updating the affected behavior and description accordingly.