[Bug]: Running featureform apply in local mode does not update SQLiteDB

[dvdr00t]

Expected Behavior

I have a definitions.py file in which I register a local provider and two sources: a dataset (local .csv file) and a transformation (a df_trasformation). If I edit the source code for the transformation and then run:

featureform apply --local definitions.py

I should expect the SQLiteDB to be updated with the latest modifications applied to my sources.

Actual Behavior

When I run:

featureform apply --local definitions.py

the modifications are not applied in the SQLiteDB. Deleting the .featureform folder and re-applying the definitions.py registrations, solves the problem.

Steps To Reproduce

Note that: I have Python 3.11 installed, though the form to open this issue only provides selection up to Python 3.10

I have this definitions.py file:

import featureform as ff

provider = ff.register_local()
transactions = provider.register_file(
    name="transactions",
    variant="quickstart",
    description="...",
    path="..."
)

@provider.df_transformation(inputs=[("transactions", "quickstart")], variant="quickstart")
def my_transformation(source):
    '''some description'''
    return source[["TransactionAmount"]]

If I run featureform apply --local definitions.py, the .featureform folder is correctly created and the SQLiteDB shows the correct sources registered.

If I want to change my transformation to (or anything else):

@provider.df_transformation(inputs=[("transactions", "quickstart")], variant="quickstart")
def my_transformation(source):
    '''some description'''
    return source[["TransactionAmount"]].sum()

And then I re-run featureform apply --local definitions.py, I still got the old sources registered.

What mode are you running Featureform in?

Local

What version of Python are you running?

3.10

Featureform Python Package Version

1.11.2

Featureform Helm Chart Version

No response

Kubernetes Version

No response

Relevant log output

No response