Hey, I have pipeline from code, but have problem with caching. Actually clearml didn't cache already executed steps (tried to re-run pipeline from web-ui). Did I miss something?

pipe.add_function_step(
    name='download_data',
    function=download_data_step,
    function_kwargs=dict(
        dataset_version='${pipeline.dataset_version}',
        file_name='${pipeline.file_name}',
    ),
    function_return=['data_path'],
    cache_executed_step=True,
    repo=repo,
    repo_branch=repo_branch,
    working_dir=working_dir
)

Also, I have problem with pipeline execution on local machine, problem with imports. Looks like clearml do not know for imports of tasks (in my example task1.py imports utils.some_utils_functions.py). If I add absolute path of utils to sys.path it works, but I do not like this solution. The problem has gone if I execute pipeline on remote machine when I add git repo, branch and working dir as arguments to add_function_step function, but didn't manage to solve on local machine. Here is my dir structure:

tasks

• utils- some_utils_functions.py- task1.py
• task2.pypipeline_controller.py

pipe.start_locally(run_pipeline_steps_locally=True)