Skip to content

Lupl/fix non correspondence disengage - #96

Merged
lu-pl merged 4 commits into
mainfrom
lupl/fix-non-correspondence-disengage
Oct 6, 2026
Merged

lu-pl merged 4 commits into
mainfrom
lupl/fix-non-correspondence-disengage

Conversation

@lu-pl

@lu-pl lu-pl commented Oct 6, 2026

Copy link
Copy Markdown
Owner

Closes #94.
Closes #95 .

lu-pl added 4 commits October 5, 2026 14:07
The refactor introduces `dealiased_field_name` as variable name for referencing validation
alias-sensitive fields (scalar type and list of scalar type fields) in the planner.

Closes #94.
The `graphty` mapper essentially constructs or mediates a mapping between a model and a dataframe;
i.e. it iterates over the model fields and compiles a lazy Polars query plan that modifies the
dataframe in such a way that it's rows can be consumed by the Pydantic instance machinery.

Now, for fields that do not have a (potentially de-aliased) correspondence in the data, the
`graphty` planner should basically disengage (`continue`) and transfer responsibility to Pydantic.

Also, if the model fields iteration does not yield any dataframe modification expressions, the
planner should again transfer responsibility to Pydantic and forward the frame unchanged.

This commit implements the above two changes; this eliminates a series of causes for
`polars.ColumnNotFoundError` in the planner and also fixes the case described in #95.

Closes #95.
@lu-pl
lu-pl merged commit 77a84ef into main Oct 6, 2026
9 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

1 participant