Skip to content

Issues & Troubleshooting Log

This log tracks known issues, user-reported bugs, and observations from internal testing for ALchemist. It is maintained by the development team.


How to Report an Issue

If you encounter a problem or have feedback, please open an issue on GitHub or email ccoatney@nrel.gov with the following information:

  • Brief description of the issue
  • Steps to reproduce (if applicable)
  • Your operating system and environment
  • Any error messages or screenshots
  • Date observed

Known Issues

Issue Date Reported Status Notes / Workarounds
BoTorch kernel hyperparameters not shown in "Next Point" dialog 2024-06-16 Open Hyperparameters are available in the console output. UI fix planned.

Resolved Issues

Issue Date Reported Date Resolved Notes
GUI not displaying fully on macOS; windows may be cut off 2024-06-16 2025-06-29 Resolved as of latest testing; GUI now displays correctly on Mac without external monitor.
Loading variables from CSV does not work; only JSON loads correctly 2025-06-29 2025-07-15 Fixed CSV parsing for integer min/max values and categorical value parsing.
Saving variables as CSV and reloading does not restore variables 2025-06-29 2025-07-15 Fixed Integer variable population and main UI update after variable definition.
Main UI "Load Variables" button fails with JSON error when loading CSV files 2025-07-15 2025-07-15 Fixed load_variables() function to properly detect and parse both JSON and CSV file formats.
Categorical variables losing values when editing in variables setup 2025-07-15 2025-07-15 Enhanced categorical editor data filtering and improved Sheet widget data handling.
Model Prediction Optimum tool: suggested experiment gives fractional value for integer variable (BoTorch backend) 2025-06-29 2025-07-15 Fixed by implementing integer rounding in optimization results. Note: BoTorch likely has native integer constraints - investigate optimize_acqf with integer_indices parameter for future improvement.
Model Prediction Optimum tool: optimizing to maximum or minimum gives same suggested values (BoTorch backend) 2025-06-29 2025-07-15 Fixed by correcting acquisition panel to use find_optimum() method instead of select_next() method.

This log is updated as issues are reported and resolved.