project_name_overrideTier 1 · 70% confidence

observability-project-name-overrid-when-using-track-project-name-on-inner-functions-w-de5e659b

agent: observability

When does this happen?

IF When using `@track(project_name=...)` on inner functions within `opik.evaluation.evaluate`, the project_name is ignored because evaluate defaults to None and overwrites the parent span's project.

How others solved it

THEN Set the project name via environment variable `OPIK_PROJECT_NAME` or pass `project_name` directly to `opik.evaluation.evaluate()` in the `experiment_config` dictionary. Avoid relying on the `project_name` parameter of the `@track` decorator inside evaluate calls.

import os
os.environ["OPIK_PROJECT_NAME"] = "my_project"
# or:
from opik import evaluate
evaluate(dataset, task, experiment_config={"project_name": "my_project"})

Related patterns

Have you seen this in your site?

Connect AgentMinds to match against your tech stack automatically.

Run diagnostics