Closed muhanadz closed 2 years ago
I wonder if, for each of your examples you could test both the auto-detect and the # header
detection.
You'd need to create a row or two of data (instead of just headers) and create a csv file buffer...
from omero_metadata.cli import MetadataControl
column_types = MetadataControl.detect_headers(csv_buffer)
My understanding is that both either the pandas based detection or the header row parsing will first a column_types
list. This variable is then passed as an argument the HeaderResolver
and resolved into OMERO.table columns. The goal of the class introduced in this PR is to add coverage for the last part of the logic.
@will-moore should https://github.com/ome/omero-metadata/pull/75#issuecomment-1129071045 not be a completely separate unit test, possibly re-using the logic of the tests introduced in https://github.com/ome/omero-metadata/pull/72 to test #header
-> column_types
?
@sbesson Yes, I guess so. I just seemed that we've already got a good set of column names and corresponding column types in these tests that could be re-used for testing auto-header. But happy for this to happen however it's easiest.
While testing the README changes in https://github.com/ome/omero-metadata/pull/71, written some tests to assert the resolved column types from
omero metadata populate
.These tests provide insight and could help us align the expected behavior versus the executed behavior.