hello! basic ClearML Pipeline question here.

let's say I've declared pipeline like so:

pipe = PipelineController(
        project="Stevens AutoScan",
        name="Dataset Formation Pipeline",
        repo=".git",
        skip_global_imports=False,
    )

then I add a step

 pipe.add_function_step(
        name="Annotation Extraction",
        function=run_annotation_extraction,
    )

after this i try to start it locally ( pipe.start_locally(run_pipeline_steps_locally=True) ). and immediately got an import error.

Additional context:
inside the module where run_annotation_extraction is defined I do import some functions from the other module.

The structure:
├── requirements.txt
├── src
│ ├── utils
│ │ ├── annotator
| | └── main.py
|. └── common
|. └── rotator.py

so i do run python -m src.utils.main
where I have imported run_annotation_extraction function, which requires quads_rotator which is imported like so: from src.common.rotator import quads_rotator in same module where run_annotation_extraction is defined.

The resulted standalone task script is turned out to be in the /tmp/ folder with imports as if it was inside the project's root. Any suggestions? Seems like I'm overlooking something pretty obvious.