molclaw-diffdock-auto

# DiffDock Automated Protein-Ligand Docking Note: - Local files are not directly accessible by the server. Please upload them to the server using `drugsda-file-transfer` before execution. - For PDB file inputs, it is recommended to preprocess them using `drugsda-fix_pdb` before execution. ## Usage ### 2. DiffDock Auto Docking The description of tool *diffdock_auto*. ```tex Automate DiffDock protein–ligand docking for single or batch inputs, run inference (unless `dry_run`), and return per-complex confidence summaries and produced files for prioritization. Args: protein_path (str|None): Protein PDB file path for single docking. ligand (str|None): SMILES string or ligand file path (.sdf/.mol/.mol2/.pdb) for single docking. protein_sequence (str|None): Protein sequence to trigger ESMFold when no PDB is available. protein_ligand_csv (str|None): CSV path with `protein_path` and `ligand` columns for batch runs. complex_name (str|None): Optional complex identifier for a single docking task. inference_steps (int): Number of diffusion steps. Default: 20. samples_per_complex (int): Number of generated samples per complex. Default: 40. batch_size (int): Batch size for DiffDock inference. Default: 10. device (str): Compute device, 'cuda' or 'cpu'. Default: 'cuda'. dry_run (bool): If True, only validate inputs and skip DiffDock execution. Return: status (str): 'success' or 'error'. msg (str): Human-readable summary or error message. output_dir (str|None): Path to the run-specific output directory under tool_result/diffdoc_result. summary_metrics (dict|None): Global metrics such as num_complexes, best_confidence, average_confidence, median_confidence. quality_distribution (dict|None): Counts grouped by quality labels (Excellent/Good/Medium/Low). complex_results (dict|None): Sanitized per-complex analysis dictionaries. summary_csv (str|None): Path to docking_summary.csv when exported. files (List[str]|None): Sorted file list produced in output_dir. ``` How to use tool *diffdock_auto* : ```python response = await client.session.call_tool( "diffdock_auto", arguments={ "protein_path": "/abs/path/outputok.pdb", "ligand": "CC(C)Cc1ccc(C)cc1", "inference_steps": 20, "samples_per_complex": 40, "batch_size": 10, "device": "cuda", "dry_run": False } ) result = client.parse_result(response) summary_metrics = result["summary_metrics"] ``` #### Example parameter sets ```python # 1) Main mode: single protein-ligand docking (from readme usage/test) { "protein_path": "/abs/path/outputok.pdb", "ligand": "CC(C)Cc1ccc(C)cc1", "dry_run": False } # 2) Variant mode: batch docking with CSV input (from diffdock_auto_main usage) { "protein_ligand_csv": "/abs/path/batch_input.csv", "inference_steps": 20, "samples_per_complex": 60, "batch_size": 15, "dry_run": False } # 3) Variant mode: input validation only (dry run) {`` "protein_path": "/abs/path/outputok.pdb", "ligand": "CC(C)Cc1ccc(C)cc1", "device": "cpu", "dry_run": True } ``` --- ## ⚠ Mandatory Output File Download (L3 Principle 14) **After calling this tool, you MUST download all output structure files** from the MCP server to the local workspace using `server_file_to_base64`. A tool call is NOT considered complete until its output files have been downloaded and verified locally (`ls -la <file>` — size must be > 0). ```python import base64, os response = await client.session.call_tool( "server_file_to_base64", arguments={"file_path": result["output_file"]} # or relevant output field ) dl = client.parse_result(response) local_path = "stepNN_descriptive_name.ext" with open(local_path, "wb") as f: f.write(base64.b64decode(dl["base64_string"])) assert os.path.getsize(local_path) > 0, f"Download failed: {local_path}" ``` **Download policy:** All structure output files are **Category A (user-critical)** — essential for user verification, downstream analysis, and reproducibility. When in doubt, download. Over-collection is always preferred over under-collection. ## ⚠ CRITICAL: Cross-Molecule Ranking Prohibition **DiffDock's confidence score is ONLY valid for comparing poses of the SAME molecule.** NEVER use DiffDock confidence to rank DIFFERENT molecules against each other. For cross-molecule ranking, use QuickVina docking scores, EquiScore, or Boltz-2 binding probability instead.