network-planning
Plan OpenFE components, atom mappings, ligand networks, and RBFE/RHFE alchemical networks before execution.
- License 许可
- MIT
- 最近更新
- 2026-09-03
- Stars
- ★ 351
- Forks
- 27
01 前言 Front Matter
- name
- network-planning
- license
- MIT
- metadata
- { "disco-role": "operating" }
- description
- Plan OpenFE components, atom mappings, ligand networks, and RBFE/RHFE alchemical networks before execution.
- disable-model-invocation
- true
02 资源文件 5
- 参考文档references/api-reference.md10.7 KB
- 参考文档references/data-formats.md6.4 KB
- 参考文档references/troubleshooting.md7.9 KB
- 参考文档references/workflows.md10.1 KB
- 脚本scripts/validate_ligand_inputs.py13.3 KB
03 SKILL.md 正文
Network Planning
Use this sub-skill when a task is about building OpenFE inputs before simulation: components, atom mappings, ligand networks, RBFE/RHFE alchemical networks, or deciding whether to use API planning, CLI planning, or manual network construction.
Start Here
- Validate ligand/protein inputs with validate_ligand_inputs.py before planning or diagnosing mapper failures.
- Use API Reference for constructors, mapper/scorer choices, ligand network functions, and RBFE/RHFE planner data flow.
- Use Workflows for end-to-end planning recipes, difficult diagnostics, and protocol handoff decisions.
- Use Data Formats for SDF/MOL2/PDB/PDBx, GraphML ligand networks, alchemical-network JSON, and transformation JSON expectations.
- Use Troubleshooting for missing optional dependencies, invalid ligand names, unmapped edges, disconnected networks, charged transformations, malformed files, and immutable objects.
Scope
This sub-skill covers:
SmallMoleculeComponent,ProteinComponent,ProteinMembraneComponent,SolventComponent,ChemicalSystem,LigandAtomMapping, andLigandNetworkconstruction concepts.LomapAtomMapper,KartografAtomMapper,PersesAtomMapper,lomap_scorers,perses_scorers, andopenfe.setup.ligand_network_planningfunctions.RBFEAlchemicalNetworkPlanner,RHFEAlchemicalNetworkPlanner, and theRelativeAlchemicalNetworkPlannerdata flow from ligands toAlchemicalNetwork, including the difference between Python planner defaults and CLI planning defaults.- Pre-execution route decisions: API vs CLI planning, topology choice, mapper/scorer tradeoffs, and when to hand a plan to protocol or CLI execution guidance.
Route elsewhere:
- Protocol settings, OpenMM backend choices, adaptive settings details, and simulation configuration -> protocols.
- CLI flag syntax,
openfe plan-*,openfe quickrun, repeat job construction, andgather*command use -> cli-workflows. - Post-run estimates, uncertainties, failed edge summaries, and result aggregation -> results-analysis.
Safe Default Pattern
- Do input checks first; do not run OpenMM while diagnosing planning failures.
- Prefer unique, explicit ligand names before creating networks; duplicate or empty names can create ambiguous radial centers and duplicate transformation labels.
- Start with a maximal network to debug mapper coverage, then choose minimal spanning, redundant minimal spanning, radial, or explicit edges based on campaign needs.
- Treat OpenFE components, mappings, and networks as immutable; rebuild them from corrected inputs rather than mutating them in place.
- For charged transformations, identify the affected ligand pairs during planning and hand protocol-setting decisions to protocols.
04 同仓库更多 Skills
passnet-triton-opt
Make a single PassNet Triton kernel fast AND numerically faithful: performance model (when a replacement can win at all), block/grid/warp tuning, autotune policy, launch overhead, and per-op numeric recipes to pass the dtype baseline tolerances. Use when a pass matches and is correct but speedup below expected, or when correctness fails by small numeric margins.
passnet-solve
END-TO-END playbook for solving one PassNet sample: analyze the computation graphs, decide the optimization strategy (what to fuse, what to replace, what to leave alone), drive the iteration loop, and maximize the sample score. This is the ENTRY skill — invoke it first for any PassNet optimization task; it tells you when to use passnet-pattern-fusion, passnet-triton-opt and passnet-feedback.
passnet-skill
PassNet GPU kernel optimization via compiler passes. Design and implement Triton-based optimization passes, create pass files under ./pass_dir/, self-evaluate with pass_evaluator, and iterate to maximize GPU speedup.
passnet-pattern-fusion
Author PassNet pass files: write patterns that actually MATCH the FX graph, pick fusion regions (which ops to absorb into one Triton kernel), and structure multi-pass submissions with the mandatory shared-dispatch architecture. Use when creating or fixing pattern()/replacement_args()/replacement_func() files, when a pass "failed to match", or when deciding how to fuse multiple kernels/ops.
passnet-orchestrate
Round/batch planning and decision support for PassNet graph optimization. Use for multi-sample triage, worker allocation, eval-budget policy, and post-eval keep/revert decisions. For solving one concrete sample end to end, use passnet-solve as the entry skill; consult this skill only when you need broader planning or a specific decision gate.
passnet-feedback
Fast, deterministic feedback for PassNet work: pre-flight pattern/match verification WITHOUT burning a GPU evaluation, per-node bottleneck analysis of a sample's graphs, and eval-log parsing into per-variant status + estimated score + failure classification. Use BEFORE every GPU evaluation (check_pattern), at the START of a sample (analyze_graph), and AFTER every evaluation (parse_eval_log).