Feature request: Blueprinted names for the columns with a Python data frame model.

[samcunliffe]

Hello devs! 👋 Hopefully, I'm allowed to simply open a feature request here. Please tell me if I'm missing something.

---

With SQL model blueprints, it's possible to use a blueprint variable in the column names.

With a Python model that uses the decorated def execute(...) -> pd.DataFrame: ... function, the docs say

Because SQLMesh creates tables before evaluating models, the schema of the output DataFrame is a required argument. The @model argument columns contains a dictionary of column names to types.

So I need to declare the columns in the @model decorator, but they can't be blueprinted.

Here's an minimum not-working example:

@model(
    "schema.@{name}",   # blueprinting works here!
    kind="full",
    columns={
        "@{column_name}": "text",             # ... but not here
        "TRANSFORMED_@{column_name}": "text",
    },
    blueprints=[
        {"name": "FirstTable", "column_name": "FIRSTCOLUMN"},
        {"name": "SecondTable", "column_name": "SECONDCOLUMN"},
    ],
)
def execute(
    context: ExecutionContext,
    start: datetime,
    end: datetime, 
    execution_time: datetime, 
    **kwargs: t.Any, 
) -> pd.DataFrame:
    """Execute the model."""
    # Get hold of blueprint variables for the column name
    name = context.blueprint_var("name")
    column_name = context.blueprint_var("column_name")

    source_table = f"{name}Source"
    select_statement = f"SELECT {column_name} FROM {source_table};" 
    output_column_name = f"TRANSFORMED_{column_name}"

    df = context.fetchdf(select_statement)
    df[output_column_name] = df[column_name].apply(some_fancy_transformation)
    return df

Is it possible/easy to support this pattern?