Main function refactoring:
1) Convert mixed-language components from ligand generation to neural network steps into a unified Python codebase.
2)Merge make_cauldron and join_feature into one task
3)Modify the pipeline to utilize data frames for all intermediate data exchanges(deprecated CSV files). Ensure that each task's inputs and outputs are managed with data frames for full compatibility with our Ray pipeline architecture.
Note:
The bluephos_pipeline.py script operates in two modes:
Disable Ray Mode: Designed for local testing and ensures the individual tasks' correctness. This mode has been tested and is currently free of issues.
Ray Mode: Optimized for Ray's distributed processing features. Currently there are known errors that may stem from the configuration of Ray's pipeline resources or other aspects. any insights or suggestions that could contribute to a solution are welcomed.
Main function refactoring: 1) Convert mixed-language components from ligand generation to neural network steps into a unified Python codebase. 2)Merge make_cauldron and join_feature into one task 3)Modify the pipeline to utilize data frames for all intermediate data exchanges(deprecated CSV files). Ensure that each task's inputs and outputs are managed with data frames for full compatibility with our Ray pipeline architecture.
Note: The bluephos_pipeline.py script operates in two modes: