ovo.core.logic.design_logic_rfdiffusion

Module Contents

Functions

submit_rfdiffusion_preview

Run the RFdiffusion workflow with reduced number of diffuser timesteps.

process_workflow_results

process_rfdiffusion_design

Process a single RFdiffusion-designed backbone and its sequence designs, copying files from the scheduler output to our storage, and creating Design and DescriptorValue objects for the backbone and each sequence design.

prepare_rfdiffusion_workflow_params

get_rfdiffusion_run_parameters

prepare_custom_backbones

Prepare custom backbone input for RFdiffusion workflow

API

ovo.core.logic.design_logic_rfdiffusion.submit_rfdiffusion_preview(workflow: ovo.core.database.models_rfdiffusion.RFdiffusionWorkflow, timesteps: int, partial_diffusion: bool = False, pipeline_name: str = 'rfdiffusion-backbone', scheduler_key: str = None, **submission_args) str | None

Run the RFdiffusion workflow with reduced number of diffuser timesteps.

ovo.core.logic.design_logic_rfdiffusion.process_workflow_results(job: ovo.core.database.models.DesignJob, callback: Callable = None, extra_filenames: dict | None = None) list[ovo.core.database.models.Base]
ovo.core.logic.design_logic_rfdiffusion.process_rfdiffusion_design(pool_id: str, contig_idx: int, num_contigs: int, batch_name: str, source_backbone_path: str, backbone_descriptor_key: str, backbone_number: int, num_backbone_designs: int, num_sequence_designs: int, num_fastrelax_cycles: int, source_dir: str, destination_dir: str, refolding_primary_test: str, cyclic: bool) tuple[list[ovo.Design], dict[str, tuple[str, str]]]

Process a single RFdiffusion-designed backbone and its sequence designs, copying files from the scheduler output to our storage, and creating Design and DescriptorValue objects for the backbone and each sequence design.

Args: pool_id: Pool ID contig_idx: Index of the contig for this backbone design, used for naming and design spec. Starting at zero. num_contigs: Total number of contigs in this design job, used for formatting. batch_name: Name of the batch folder in the scheduler output where this backbone design is located, e.g. “contig1_batch1” source_dir: Workflow output directory where results are located (s3://bucket/path or local path) source_backbone_path: Path to the backbone PDB file relative to source_dir, e.g. “rfdiffusion_standardized_pdb/contig1_batch1_0_standardized.pdb” backbone_descriptor_key: Descriptor key to use for the backbone structure file backbone_number: Number of the backbone design, used as a prefix in Design ID. Starting at one for each contig. num_backbone_designs: Total number of backbone designs for this contig, used for formatting the Design ID. num_sequence_designs: Number of sequence designs generated per backbone design, used to find output files and for formatting the Design ID. num_fastrelax_cycles: Number of fastrelax sequence designs generated per backbone design, used to find output files and for formatting the Design ID. destination_dir: Directory relative to our storage where design files should be stored, e.g. “project/proj123/round1/pools/pool123/designs” refolding_primary_test: Name of the primary refolding test to read results from, e.g. “af2_model_1_ptm_tt_3red cyclic: Whether the design is macrocyclic

ovo.core.logic.design_logic_rfdiffusion.prepare_rfdiffusion_workflow_params(workflow: ovo.core.database.models_rfdiffusion.RFdiffusionWorkflow, workdir: str) dict
ovo.core.logic.design_logic_rfdiffusion.get_rfdiffusion_run_parameters(workflow: ovo.core.database.models_rfdiffusion.RFdiffusionWorkflow, partial_diffusion=False, timesteps: int = None) str
ovo.core.logic.design_logic_rfdiffusion.prepare_custom_backbones(custom_backbones: str, workdir: str) str

Prepare custom backbone input for RFdiffusion workflow