What this PR does / why we need it:

dill~=0.3.0 keeps Python 3.14 users on dill 0.3.9, which doesn't support 3.14's changes to pickle. Serializing a UDF with dill.dumps(udf, recurse=True), as Transformation.to_proto() and StreamFeatureView.to_proto() do, fails as soon as the UDF refers to a module-level global, such as pd in a pandas UDF:

TypeError: _Pickler._batch_setitems() missing 1 required positional argument: 'obj'

So an on-demand feature view written like the docs example can't be applied on 3.14:

@on_demand_feature_view(
    sources=[driver_stats, request],
    schema=[Field(name="conv_rate_plus_val", dtype=Float64)],
    mode="pandas",
)
def transformed_conv_rate(inputs: pd.DataFrame) -> pd.DataFrame:
    df = pd.DataFrame()
    df["conv_rate_plus_val"] = inputs["conv_rate"] + inputs["val_to_add"]
    return df

transformed_conv_rate.to_proto()  # TypeError on Python 3.14 with dill 0.3.9

dill added Python 3.14 support in 0.4.0 and made it formal in 0.4.1. This widens the requirement to dill>=0.3.0,<0.5 and records it in pixi.lock (regenerated with pixi lock; the locked dill stays at 0.3.9). The sdk/python/requirements lock files are unchanged and still resolve dill 0.3.9 for 3.10–3.12, so CI keeps testing the same versions; installs on 3.14 now get a dill that works.

A UDF serialized with dill 0.3.9 loads with 0.4.1 and the other way round (checked on 3.12), so existing registries and mixed client/server versions keep working.

Which issue(s) this PR fixes:

No existing issue.

Checks

• I've made sure the tests are passing.
• My commits are signed off (git commit -s)
• My PR title follows conventional commits format

Testing Strategy

• Unit tests
• Integration tests
• Manual tests
• Testing is not required for this change

Misc

CI runs Python 3.10–3.12, which is why this doesn't show up there, and I didn't add a test for the same reason: the existing on-demand feature view tests cover it on 3.14. Locally, on macOS:

• Python 3.14.7 with -e ".[test,k8s,aws,gcp]": the ODFV, transformation and stream feature view unit tests (103 tests in 11 files) have 15 failures with the TypeError above on dill 0.3.9, and none on dill 0.4.1. Five further tests need pyspark, which I didn't install, and fail either way.
• Python 3.12 with sdk/python/requirements/py3.12-ci-requirements.txt: the full sdk/python/tests/unit run gives 2773 passed, 44 skipped with dill 0.3.9, and the same with dill 0.4.1 swapped in.