Drop Pydantic v1#93
Conversation
Codecov Report❌ Patch coverage is
Additional details and impacted files@@ Coverage Diff @@
## main #93 +/- ##
==========================================
+ Coverage 97.66% 99.51% +1.85%
==========================================
Files 11 11
Lines 1071 1038 -33
==========================================
- Hits 1046 1033 -13
+ Misses 25 5 -20
Flags with carried forward coverage won't be shown. Click here to find out more. ☔ View full report in Codecov by Harness. 🚀 New features to boost your workflow:
|
Documentation build overview
37 files changed ·
|
Keys have historically matched on (id, mult, associated_handler, bond_order) only. After dropping the pydantic v1 shim, matching uses PotentialKey's full-field equality, which also compares virtual_site_type and cosmetic_attributes -- fields a user cannot reconstruct from a SMIRKS. Such keys are then silently ignored in exclude (and would raise KeyError in include). Covers both vulnerable fields via the exclude path: virtual_site_type (a v-site key) and cosmetic_attributes (any handler; here a cosmetic attribute on a Bonds parameter). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
j-wags
left a comment
j-wags
left a comment
There was a problem hiding this comment.
I didn't find any problems myself, but I let claude do a once-over and it spotted a sneaky regression wrt vsites and cosmetic attributes. I worked with claude to generate minimal tests expressing this behavior.
I lean slightly toward favoring fixing this regression (since it could cause user surprise) but I'd also be fine with A) just listing it in the release notes (and removing the new tests I committed) or B) erroring in cases where it could make a difference (though this could be hard to detect)
2 participants
Description
Fixes #91
Status