Update sssom py SSSOM version to 0.12.0 #381
Conversation
|
@matentzn : All tests pass! |
matentzn
left a comment
matentzn
left a comment
There was a problem hiding this comment.
I approve of this, but see my comment for a clarifying question. You can merge at your own discretion.
…ommons/sssom-py into Update-sssom-to-0.12.0
This reverts commit 5cf3c00.
|
@matentzn : Changed the logic a little bit:
|
tests/test_reconcile.py
Outdated
| self.assertEqual(8, len(df1.index)) | ||
| df2 = deal_with_negation(self.msdf2.df) | ||
| self.assertEqual(12, len(df2.index)) | ||
| self.assertEqual(14, len(df2.index)) |
There was a problem hiding this comment.
This is what concerns me most, can you figure out what these two lines are - the update should not have change the number of rows in the reconciled part.
There was a problem hiding this comment.
Yeah now I remember. ..... The negative - positive splitting was because we wanted to enforce the rule of "negative trumps positive" in case of a confidence tie. Because we are doing this, grouping by predicate_modifier makes no sense. Hence KEY_FEATURESTRIPLE_IDS.
Restore the major refactor to what it was before. Crisis averted, thanks to you!
|
@matentzn : Ready again! |
matentzn
left a comment
matentzn
left a comment
There was a problem hiding this comment.
approved (cant approve as its my PR)
@hrshdhgd can you deal with this release? I have done no tests etc, but there was a major change in the RDF serialiser (semapv namespace was wrong), so best run some tests?