Conversation
- Create PanderaGenerator class in matrix_schema/generators/panderagen.py - Generate schemas for MatrixNode, MatrixEdge, UnionedNode, UnionedEdge - Integrate with Makefile via gen-pandera target (always runs) - Maintain PySpark compatibility with proper ArrayType nullable=False for list items - Preserve existing validation patterns (enum checks, unique constraints) - Auto-generate from LinkML schema with proper formatting 🤖 Generated with [Claude Code](https://claude.ai/code) Co-Authored-By: Claude <noreply@anthropic.com>
matentzn
left a comment
matentzn
left a comment
There was a problem hiding this comment.
AWeeeeeesommmmmmmeeee
THANKS!
There was a problem hiding this comment.
Should the main Makefile be manually edited? It seems the cookiecutter template should update it?
…rated schema so that we get a better diff on the PR
matentzn
left a comment
matentzn
left a comment
There was a problem hiding this comment.
Thank you, I love it. I assuming you wont merge before QC failures are dealt with :P
| return DataFrameSchema( | ||
| columns={ | ||
| "id": Column(T.StringType(), nullable=False), | ||
| "id": Column(T.StringType(), nullable=True), |
There was a problem hiding this comment.
Since you ordered the output now, why are there so many changes to the schema? for example nullable True seems like a big change?
There was a problem hiding this comment.
I'll figure out if the linkml is wrong or the schema generator is wrong. One way or another, I think id should clearly be nullable=False
There was a problem hiding this comment.
I think we need to set required: true on a whole bunch of slots
…red slot, ran gen-project
…id, for multivalued fields, and generate the enum checks in a more generic way in panderagen.py
2 participants
Summary
make gen-panderatargetKey Features
nullable=Falsefor list items.PHONYtarget ensures fresh generation on every runTest plan
make gen-panderagenerates all four schema functionsnullable=Falsefor list items🤖 Generated with Claude Code