How to read artifact file (CSV) programmatically?

[Giorgi]

Hello, 

can I programmatically access artifact file (csv), via artifact_uri and read it?

Tried the following, but didn't work, says no such file or directory:

mlflow.pyfunc.pandas.read_csv(artifact_uri+'/xgb-classifier-test-8/dataset_statistics.csv')

pandas.read_csv(artifact_uri+'/xgb-classifier-test-8/dataset_statistics.csv')

Thanks!

[Giorgi · Accepted solution]

Maybe there are better solutions, here is what I've found:

from mlflow.tracking import MlflowClient
 
client = MlflowClient()
pd.read_csv(client.download_artifacts(run_id, "xgb-classifier-test-8/dataset_statistics.csv"))