caddforge-docs 0.3.0__py3-none-any.whl

caddforge_docs/site/examples/pipelines/flexible-gnina-to-openbpmd-to-md-to-abfe.yaml

@@ -0,0 +1,189 @@
+# Advanced template: supply all prepared inputs and scientific runtimes.
+version: 1
+name: flexible-gnina-openbpmd-md-abfe
+
+defaults:
+  backend: slurm
+  registry_path: runs/job_registry.sqlite
+  job_name: flexible-gnina-openbpmd-md-abfe
+
+stages:
+  - id: gnina_flexible
+    tool: docking/gnina
+    config: configs/gnina_flexible.yaml
+
+  - id: retained_ligands
+    type: filter.tabular
+    inputs:
+      # One row per ligand; best_score is lower-is-better affinity.
+      table: gnina_flexible.scores
+    config:
+      filters:
+        - column: best_score
+          op: "<="
+          value: -12.0
+      columns: [ligand, best_score]
+      output_name: retained_ligands.csv
+
+  - id: top4_flexible_poses
+    type: filter.tabular
+    inputs:
+      # One row per generated flexible pose, with complex_file/protein_file/ligand_pose_file.
+      table: gnina_flexible.flexible_complexes_manifest
+      # Retain all poses belonging to ligands that passed the best-score cutoff.
+      ids: retained_ligands.filtered_table
+    config:
+      id_column: ligand
+      ids_column: ligand
+      group_by: ligand
+      sort_by: CNNscore
+      ascending: false
+      top_n_per_group: 4
+      columns:
+        - ligand
+        - pose_rank
+        - minimizedAffinity
+        - CNNscore
+        - complex_file
+        - protein_file
+        - ligand_pose_file
+      output_name: top4_flexible_poses.csv
+
+  - id: retain_top4_complexes
+    type: filter.molecules
+    inputs:
+      # OpenBPMD should receive the full protein-ligand complex PDB.
+      molecules: gnina_flexible.flexible_complexes
+      ids: top4_flexible_poses.filtered_table
+    config:
+      path_column: complex_file
+
+  - id: openbpmd_top4
+    tool: md/openbpmd
+    config: configs/openbpmd_pose.yaml
+    inputs:
+      # Bind each selected complex PDB to OpenBPMDConfig.complex_pdb_path.
+      complex_pdb_path: retain_top4_complexes.molecules
+    scatter:
+      by: complex_pdb_path
+      chunk_size: 1
+    outputs:
+      # results.collect resolves the sibling result_record.json from this path.
+      result_record: results_json
+
+  - id: collect_openbpmd
+    type: results.collect
+    inputs:
+      results: openbpmd_top4.result_record
+      # Preserve ligand, pose_rank, and the exact flexible artifact paths.
+      table: top4_flexible_poses.filtered_table
+    config:
+      metrics:
+        comp_score: metrics.comp_score
+        pose_score: metrics.pose_score
+      output_name: openbpmd_top4_metrics.csv
+
+  - id: best_openbpmd_pose
+    type: filter.tabular
+    inputs:
+      table: collect_openbpmd.table
+    config:
+      group_by: ligand
+      sort_by: comp_score
+      ascending: true
+      top_n_per_group: 1
+      columns:
+        - ligand
+        - pose_rank
+        - comp_score
+        - pose_score
+        - complex_file
+        - protein_file
+        - ligand_pose_file
+        - result_file
+      output_name: best_openbpmd_pose_per_ligand.csv
+
+  - id: retain_best_complexes
+    type: filter.molecules
+    inputs:
+      # Plain MD receives the full protein-ligand complex PDB for the selected pose.
+      molecules: gnina_flexible.flexible_complexes
+      ids: best_openbpmd_pose.filtered_table
+    config:
+      path_column: complex_file
+
+  - id: retain_best_ligands
+    type: filter.molecules
+    inputs:
+      # Plain MD needs the ligand-only SDF for non-standard ligand parameterization.
+      molecules: gnina_flexible.flexible_ligand_poses
+      ids: best_openbpmd_pose.filtered_table
+    config:
+      path_column: ligand_pose_file
+
+  - id: plain_md_10ns
+    tool: md/plain-md
+    config:
+      simulation_time_ns: 10.0
+      trajectory_filename: traj.xtc
+      output_dir_name: output
+    inputs:
+      system_pdb_path: retain_best_complexes.molecules
+      molecule_sdf_paths: retain_best_ligands.molecules
+    scatter:
+      # Scatter over complex PDBs. The matching ligand SDF is bound as a one-item
+      # molecule_sdf_paths list for the same shard.
+      by: system_pdb_path
+      chunk_size: 1
+
+  - id: center_align_md
+    tool: md/center-and-align
+    config:
+      center_selection: protein or resname LIG
+      keep_selection: protein or resname LIG
+      align_selection: protein and backbone
+      output_trajectory_filename: traj-centered.xtc
+      output_structure_filename: traj-centered.pdb
+    inputs:
+      # The plain-md trajectory uses the prepared solvated system topology.
+      topology_path: plain_md_10ns.prepared_structure
+      trajectory_path: plain_md_10ns.trajectory
+    scatter:
+      by: trajectory_path
+      chunk_size: 1
+    outputs:
+      centered_structure: output_structure
+      centered_trajectory: output_trajectory
+
+  - id: cluster_representative
+    tool: md/cluster-representative
+    config:
+      ligand_resname: LIG
+      cluster_selection: all
+      extract_selection: all
+      last_ns: 10.0
+      output_complex_filename: representative_complex.pdb
+      output_protein_filename: representative_protein.pdb
+      output_ligand_sdf_filename: representative_ligand.sdf
+    inputs:
+      topology: center_align_md.centered_structure
+      trajectory: center_align_md.centered_trajectory
+      # The ligand template preserves bond orders, hydrogens, and formal charges in the exported SDF.
+      ligand_template: retain_best_ligands.molecules
+    scatter:
+      by: trajectory_path
+      chunk_size: 1
+
+  - id: openfe_abfe
+    tool: fep/openfe-abfe
+    config: configs/openfe_abfe.yaml
+    inputs:
+      # ABFE receives separate files from the representative frame, not the original docked pose.
+      ligand_sdf: cluster_representative.ligand
+      protein_pdb: cluster_representative.protein
+    scatter:
+      # Scatter over ligand_sdf; the matching representative protein_pdb is bound by row order.
+      by: ligand_sdf
+      chunk_size: 1
+    outputs:
+      results: results_json

caddforge_docs/site/examples/pipelines/gnina-flexible-complex-artifacts.yaml

@@ -0,0 +1,52 @@
+# Advanced template: supply all prepared inputs and scientific runtimes.
+version: 1
+name: gnina-flexible-artifacts
+
+defaults:
+  backend: local
+  registry_path: runs/job_registry.sqlite
+  job_name: gnina-flexible-artifacts
+
+stages:
+  - id: gnina_flexible
+    tool: docking/gnina
+    config: configs/gnina_flexible.yaml
+
+  - id: first_flexible_pose_per_ligand
+    type: filter.tabular
+    inputs:
+      # This is the flexible_complexes_manifest.csv emitted by Gnina.
+      table: gnina_flexible.flexible_complexes_manifest
+    config:
+      group_by: ligand
+      sort_by: pose_rank
+      ascending: true
+      top_n_per_group: 1
+      output_name: first_flexible_pose_per_ligand.csv
+
+  - id: retain_ligand_pose_sdfs
+    type: filter.molecules
+    inputs:
+      # Manifest-backed path collection from the ligand_pose_file column.
+      molecules: gnina_flexible.flexible_ligand_poses
+      ids: first_flexible_pose_per_ligand.filtered_table
+    config:
+      path_column: ligand_pose_file
+
+  - id: retain_protein_pdbs
+    type: filter.molecules
+    inputs:
+      # Manifest-backed path collection from the protein_file column.
+      molecules: gnina_flexible.flexible_proteins
+      ids: first_flexible_pose_per_ligand.filtered_table
+    config:
+      path_column: protein_file
+
+  - id: retain_complex_pdbs
+    type: filter.molecules
+    inputs:
+      # Manifest-backed path collection from the complex_file column.
+      molecules: gnina_flexible.flexible_complexes
+      ids: first_flexible_pose_per_ligand.filtered_table
+    config:
+      path_column: complex_file

caddforge_docs/site/examples/pipelines/minimize-a-multi-pose-sdf.yaml

@@ -0,0 +1,46 @@
+# Advanced template: supply all prepared inputs and scientific runtimes.
+# Poses must share receptor coordinates and contain explicit hydrogens.
+# Paths resolve relative to the submission working directory.
+version: 1
+name: pose-cleanup
+
+defaults:
+  backend: local
+  run_directory: runs/pose-cleanup
+  registry_path: runs/job_registry.sqlite
+  export_results: false
+
+stages:
+  - id: minimize
+    tool: md/minimize
+    config:
+      protein_pdb_path: receptor.pdb
+      ligand_path: poses.sdf
+      mode: ligand_only
+    scatter:
+      by: ligand_path
+      chunk_size: 1
+    outputs:
+      results: result_record
+      ligands: minimized_ligand
+
+  - id: collect
+    type: results.collect
+    inputs:
+      results: minimize.results
+    config:
+      metrics:
+        quality_passed: metrics.quality_passed
+        convergence: metrics.convergence_status
+        ligand_rmsd_angstrom: metrics.ligand_heavy_atom_rmsd_angstrom
+        minimized_ligand_file: paths.minimized_ligand
+
+  - id: accepted
+    type: filter.tabular
+    inputs:
+      table: collect.table
+    config:
+      filters:
+        - column: quality_passed
+          op: '=='
+          value: 1

caddforge_docs/site/examples/pipelines/pose-level-md-selection.yaml

@@ -0,0 +1,95 @@
+# Advanced template: supply all prepared inputs and scientific runtimes.
+version: 1
+name: pose-level-openbpmd-selection
+
+defaults:
+  backend: slurm
+  registry_path: runs/job_registry.sqlite
+  job_name: pose-level-openbpmd-selection
+
+stages:
+  - id: gnina_docking
+    tool: docking/gnina
+    config: configs/gnina_docking.yaml
+    outputs:
+      # Expose the score CSV and pose directory from Result.paths.
+      scores: best_scores_csv
+      poses: poses_dir
+
+  - id: extract_poses
+    type: poses.extract
+    inputs:
+      # Split every SDF/SDFGZ pose from Gnina into a normalized one-pose SDF file.
+      poses: gnina_docking.poses
+      # Optional but useful: carries score-table metadata into the pose manifest.
+      table: gnina_docking.scores
+    config:
+      output_name: pose_manifest.csv
+      ligand_column: ligand
+      path_column: pose_file
+
+  - id: repair_pose_hydrogens
+    type: molecules.hydrogenate
+    inputs:
+      # Hydrogenate the extracted pose path collection before handing poses to OpenBPMD.
+      molecules: extract_poses.poses
+      # Reuse the pose manifest so ligand, pose_id, and pose_file provenance are preserved.
+      table: extract_poses.manifest
+    config:
+      ligand_column: ligand
+      pose_id_column: pose_id
+      # Without a templates input, hydrogens are added directly to each pose.
+      # See the built-in stage docs for template-based bond-order repair.
+      on_template_failure: warn
+      on_failure: fail
+      add_coords: true
+      output_name: hydrogenated_pose_manifest.csv
+
+  - id: openbpmd_per_pose
+    tool: md/openbpmd
+    config: configs/openbpmd_pose.yaml
+    inputs:
+      # Bind each hydrogenated pose file to OpenBPMDConfig.ligand_path.
+      ligand_path: repair_pose_hydrogens.molecules
+    scatter:
+      # Path-collection scatter submits each ligand_path shard independently.
+      by: ligand_path
+      chunk_size: 1
+    outputs:
+      # Expose canonical per-shard result records for collection.
+      result_record: result_record_json
+
+  - id: collect_pose_metrics
+    type: results.collect
+    inputs:
+      # Collect metrics from every scattered OpenBPMD result record.
+      results: openbpmd_per_pose.result_record
+      # Join explicit hydrogenation metadata to the collected metrics.
+      table: repair_pose_hydrogens.manifest
+    config:
+      metrics:
+        comp_score: metrics.comp_score
+        pose_score: metrics.pose_score
+      output_name: openbpmd_pose_metrics.csv
+
+  - id: best_pose_per_ligand
+    type: filter.tabular
+    inputs:
+      # results.collect always emits collect_pose_metrics.table.
+      table: collect_pose_metrics.table
+    config:
+      group_by: ligand
+      sort_by: comp_score
+      ascending: true
+      top_n_per_group: 1
+      output_name: best_pose_per_ligand.csv
+
+  - id: retain_best_poses
+    type: filter.molecules
+    inputs:
+      # Filter the hydrogenated pose files, not the pre-repair Gnina poses.
+      molecules: repair_pose_hydrogens.molecules
+      ids: best_pose_per_ligand.filtered_table
+    config:
+      # Prefer exact file paths from the filtered table over filename matching.
+      path_column: output_file

caddforge_docs/site/examples/pipelines/result-collection-and-diagnostics.yaml

@@ -0,0 +1,55 @@
+# Advanced template: supply all prepared inputs and scientific runtimes.
+version: 1
+name: collect-and-diagnose
+
+defaults:
+  backend: local
+  registry_path: runs/job_registry.sqlite
+
+stages:
+  - id: scattered_transfer
+    tool: qm/transfer-free-energy
+    config: configs/transfer_free_energy.yaml
+    scatter:
+      # Split the ligand SDF into one-molecule transfer calculations.
+      by: ligand_sdf
+      chunk_size: 1
+    outputs:
+      # Expose per-shard canonical result records.
+      result_record: result_record_json
+
+  - id: collect_transfer
+    type: results.collect
+    inputs:
+      results: scattered_transfer.result_record
+    config:
+      metrics:
+        # Copy payload["metrics"]["delta_g_kcal_mol"] into a table column.
+        delta_g_kcal_mol: metrics.delta_g_kcal_mol
+      output_name: transfer_metrics.csv
+
+  - id: filter_transfer
+    type: filter.tabular
+    inputs:
+      # results.collect emits collect_solvation.table.
+      table: collect_transfer.table
+    config:
+      filters:
+        - column: delta_g_kcal_mol
+          op: "<="
+          value: -3.0
+      sort_by: delta_g_kcal_mol
+      ascending: true
+      output_name: favorable_transfer.csv
+
+  - id: distribution_report
+    tool: datascience/distribution-diagnostics
+    config:
+      # Built-in stage outputs can be bound into inline tool config fields.
+      csv_path: filter_transfer.filtered_table
+      column: delta_g_kcal_mol
+      bins: 50
+      include_log: false
+    outputs:
+      report: report_json
+      plot: plot_png

caddforge_docs/site/examples/quickstart/diagnostics.yaml

@@ -0,0 +1,8 @@
+backend: local
+csv_path: values.csv
+column: value
+bins: 10
+sample_size: 100
+random_seed: 0
+include_log: false
+results_filename: distribution_report.json