Mlflow plugin doesn't pass down run/success

[ssabdb]

Current behavior

The mlflow plugin doesn't log the graph outcome, which leads to mlflow UI appearing to report a run was a success.

hamilton/hamilton/plugins/h_mlflow.py

Line 332 in 45feeaf

mlflow.end_run()

Stack Traces

(If applicable)

Screenshots

Steps to replicate behavior

1. Use hamilton mlflow integration
2. Create and run dag with exception
3. Observe logged run_id displays as succeeded despite failure

Library & System Information

python3.11
mlflow==3.5.1
sf-hamilton[pandera,tqdm]>=1.88.0

Expected behavior

Mlflow should display the run has failed rather than succeeded.

[skrawcz]

Thanks @ssabdb, did you have a fix in mind by chance? :) or?

[priyankajoshi93094]

Hi, I’d like to work on this issue if it’s still available. Please let me know if I can proceed. Thanks!

[skrawcz]

Hi, I’d like to work on this issue if it’s still available. Please let me know if I can proceed. Thanks!

sure!

[AliNasirDawood]

Hi,I would like to know whether the bug has been resolved I would love to resolve it

[dedyoc]

Hi everyone,
I recently started exploring Hamilton, noticed this issue is still open and wanted to give it a try.
I've opened a PR here: #1541

@skrawcz just a headsup, while fixing the status tracking, I also caught and patched an AttributeError that was crashing the pre-execution hook when no inputs were passed. lmk if any adjustments are needed.