dualmapref 0.2.8__py3-none-any.whl

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (186) hide show
  1. args_parse.py +593 -0
  2. candidate_corrector.py +161 -0
  3. candidate_scorer.py +359 -0
  4. conformend/__init__.py +36 -0
  5. conformend/refiner.py +25 -0
  6. dataset/__init__.py +1 -0
  7. dataset/drugdataset.py +644 -0
  8. dataset/manifold.py +237 -0
  9. dataset/sharded_drugdataset.py +124 -0
  10. dataset/smiles2graph.py +217 -0
  11. dualmapref/__init__.py +24 -0
  12. dualmapref/refiner.py +16 -0
  13. dualmapref-0.2.8.dist-info/LICENSE +21 -0
  14. dualmapref-0.2.8.dist-info/METADATA +167 -0
  15. dualmapref-0.2.8.dist-info/RECORD +186 -0
  16. dualmapref-0.2.8.dist-info/WHEEL +5 -0
  17. dualmapref-0.2.8.dist-info/entry_points.txt +10 -0
  18. dualmapref-0.2.8.dist-info/top_level.txt +29 -0
  19. exputils.py +289 -0
  20. infer.py +23 -0
  21. infer_base_conformers.py +5589 -0
  22. infer_mean_dispersion_torsion.py +1117 -0
  23. infer_torsion.py +752 -0
  24. infer_torsional_denoiser.py +1021 -0
  25. manimol/__init__.py +30 -0
  26. manimol/__main__.py +6 -0
  27. manimol/cli.py +36 -0
  28. manimol/full_cli.py +41 -0
  29. manimol/legacy_imports.py +43 -0
  30. manimol/manifest.py +56 -0
  31. manimol/postprocess.py +148 -0
  32. manimol/priors.py +60 -0
  33. manimol/protocols.py +119 -0
  34. manimol/refiner.py +645 -0
  35. manimol/runtime.py +194 -0
  36. manimol/selection.py +85 -0
  37. manimol/torsion.py +22 -0
  38. manimol/train_cli.py +14 -0
  39. mixture_flow/src/__init__.py +1 -0
  40. mixture_flow/src/_path.py +7 -0
  41. mixture_flow/src/infer28_mixture_torsion_flow.py +1093 -0
  42. mixture_flow/src/mixture_torsion_flow.py +291 -0
  43. mixture_flow/src/summarize_mixture_debug.py +80 -0
  44. mixture_flow/src/train28_mixture_torsion_flow.py +583 -0
  45. mixture_torsion_flow.py +7 -0
  46. models/ manifold_learning.py +86 -0
  47. models/__init__.py +8 -0
  48. models/dist.py +273 -0
  49. models/dist2coords.py +23 -0
  50. models/egnn.py +107 -0
  51. models/experimental_backbones.py +364 -0
  52. models/gnnconv.py +512 -0
  53. models/kernel_inversion.py +141 -0
  54. models/kernels.py +427 -0
  55. models/losses.py +154 -0
  56. models/mean_dispersion_bridge.py +119 -0
  57. models/mean_dispersion_torsion_context.py +155 -0
  58. models/model.py +73 -0
  59. models/model0.py +2375 -0
  60. models/vis.py +128 -0
  61. pgraph_controls.py +107 -0
  62. scripts/__init__.py +2 -0
  63. scripts/analyze_basin_alignment.py +63 -0
  64. scripts/analyze_hif2a_conformer_aggregation.py +267 -0
  65. scripts/analyze_hif2a_contact_pharmacophore.py +1111 -0
  66. scripts/analyze_hif2a_fixed_budget.py +202 -0
  67. scripts/analyze_hif2a_pharmacophore.py +405 -0
  68. scripts/analyze_hif2a_pharmacophore_match_metric.py +386 -0
  69. scripts/analyze_p_control_candidate_flow.py +154 -0
  70. scripts/analyze_scaffold_basin.py +83 -0
  71. scripts/audit_dispersion_proposals.py +208 -0
  72. scripts/audit_etflow_protocol.py +258 -0
  73. scripts/audit_external_overlap.py +150 -0
  74. scripts/audit_oldbest_fulltrainval_split.py +82 -0
  75. scripts/audit_pdbbind_contact_pharmacophores.py +146 -0
  76. scripts/audit_qm9_geomol_test_split.py +94 -0
  77. scripts/audit_stage1_stage2_cross.py +131 -0
  78. scripts/build_geom_drugs_lowref_diagnostic.py +84 -0
  79. scripts/build_qm9_backfill_manifest.py +99 -0
  80. scripts/check_etflow_runtime.py +56 -0
  81. scripts/compare_prior_sources.py +122 -0
  82. scripts/compose_binding_mode_figure.py +87 -0
  83. scripts/compose_hif2a_contact_pharmacophore_suppfig.py +255 -0
  84. scripts/compose_hif2a_functional_geometry_pipeline.py +266 -0
  85. scripts/compose_manimol_internal_architecture.py +291 -0
  86. scripts/diagnose_flow_basin_intervention.py +54 -0
  87. scripts/diagnose_hif2a_pharmacophore_budget.py +185 -0
  88. scripts/diagnose_route2_p_ranking.py +123 -0
  89. scripts/discover_binding_mode_inputs.py +30 -0
  90. scripts/evaluate_functional_prior_stage1.py +230 -0
  91. scripts/evaluate_geomanimol.py +74 -0
  92. scripts/evaluate_geomanimol_flow.py +23 -0
  93. scripts/evaluate_geomanimol_multiscaffold.py +46 -0
  94. scripts/evaluate_geomanimol_residual.py +45 -0
  95. scripts/evaluate_geomanimol_torsion.py +40 -0
  96. scripts/evaluate_hif2a_bioactive_enrichment.py +91 -0
  97. scripts/evaluate_hif2a_pms_suite.py +440 -0
  98. scripts/evaluate_pdbbind_contact_pms.py +155 -0
  99. scripts/evaluate_platinum_external.py +114 -0
  100. scripts/evaluate_platinum_pool.py +82 -0
  101. scripts/evaluate_route2_calibrated_p.py +104 -0
  102. scripts/evaluate_route2_library_components.py +596 -0
  103. scripts/evaluate_route2_p_basin_budget.py +288 -0
  104. scripts/evaluate_stage1_checkpoint_only.py +30 -0
  105. scripts/export_dispersion_bond_audit.py +136 -0
  106. scripts/export_top20_shards.py +207 -0
  107. scripts/external_validation_metrics.py +152 -0
  108. scripts/fit_topology_prior.py +117 -0
  109. scripts/frozen_dispersion_audit.py +281 -0
  110. scripts/geomanimol_debug.py +50 -0
  111. scripts/geomanimol_smoke.py +37 -0
  112. scripts/infer_etflow_matched.py +524 -0
  113. scripts/inspect_checkpoint.py +12 -0
  114. scripts/inspect_conformer_capacity.py +71 -0
  115. scripts/inspect_pool.py +12 -0
  116. scripts/make_figure3_selection_loss.py +474 -0
  117. scripts/make_figure3_selection_summary.py +275 -0
  118. scripts/make_geom_debug_subset.py +62 -0
  119. scripts/make_low_reference_test_subset.py +70 -0
  120. scripts/merge_stage1_into_stage2.py +82 -0
  121. scripts/place_hif2a_candidates_for_evaluation.py +66 -0
  122. scripts/plot_hif2a_low_budget_efficiency.py +100 -0
  123. scripts/plot_prior_validation.py +48 -0
  124. scripts/precompute_stage2_assignments.py +90 -0
  125. scripts/prepare_bace_activity.py +98 -0
  126. scripts/prepare_bace_etkdg_benchmark.py +90 -0
  127. scripts/prepare_figure5_4x4_cases.py +470 -0
  128. scripts/prepare_geom_drugs_geomol_top20.py +472 -0
  129. scripts/prepare_pdbbind_core_raw.py +90 -0
  130. scripts/prepare_platinum_raw.py +74 -0
  131. scripts/prepare_pubchem_etkdg.py +54 -0
  132. scripts/rebuild_hif2a_final_sdf_and_metrics.py +55 -0
  133. scripts/rebuild_hif2a_fixed_pool_sweep.py +175 -0
  134. scripts/rebuild_hif2a_strict.py +220 -0
  135. scripts/recompute_hif2a_bootstrap.py +63 -0
  136. scripts/reference_free_geometry.py +112 -0
  137. scripts/render_binding_modes.py +413 -0
  138. scripts/render_figure5_4x4.py +330 -0
  139. scripts/render_figure6_pharmacophore_pymol.py +162 -0
  140. scripts/render_hif2a_contact_pharmacophore_pymol.py +414 -0
  141. scripts/render_hif2a_figure11.py +397 -0
  142. scripts/run_bace_2d_baseline.py +73 -0
  143. scripts/run_drugs_validation_matrix.py +225 -0
  144. scripts/run_external_reference_free.py +53 -0
  145. scripts/run_manimol_finalk_sensitivity.py +106 -0
  146. scripts/run_manimol_oversampling_sensitivity.py +109 -0
  147. scripts/run_pdisp_matched_intervention.py +116 -0
  148. scripts/run_stage1_stage2_cross20.py +83 -0
  149. scripts/score_hif2a_candidates_with_pgraph.py +221 -0
  150. scripts/summarize_dispersion_intervention.py +88 -0
  151. scripts/test_mean_dispersion_bridge.py +50 -0
  152. scripts/train_dispersion_head_diagnostic.py +320 -0
  153. scripts/train_dispersion_head_streaming.py +519 -0
  154. scripts/train_geomanimol.py +100 -0
  155. scripts/train_geomanimol_flow.py +26 -0
  156. scripts/train_geomanimol_residual.py +70 -0
  157. scripts/train_geomanimol_torsion.py +50 -0
  158. scripts/train_quantile_representation_diagnostic.py +204 -0
  159. scripts/validate_mean_dispersion_prior.py +173 -0
  160. scripts/validate_packed_shard.py +41 -0
  161. scripts/validate_stage1_prior.py +221 -0
  162. scripts/validate_top20_shards.py +36 -0
  163. scripts/validation/__init__.py +1 -0
  164. scripts/validation/hif2a_downstream_baselines.py +105 -0
  165. scripts/validation/hif2a_pairwise_ranker.py +218 -0
  166. scripts/verify_binding_mode_figure.py +47 -0
  167. torsion_manifold.py +281 -0
  168. train.py +78 -0
  169. train_graph_encoder.py +206 -0
  170. train_local_torsion_corrector.py +511 -0
  171. train_mean_dispersion_torsion_prior.py +329 -0
  172. train_probability_basin.py +235 -0
  173. train_torsion_generator.py +1631 -0
  174. train_torsion_manifold_prior.py +258 -0
  175. train_torsional_denoiser.py +415 -0
  176. trainer.py +1038 -0
  177. utils/__init__.py +7 -0
  178. utils/checkpoint.py +40 -0
  179. utils/device.py +17 -0
  180. utils/helpers.py +98 -0
  181. utils/kabsch.py +29 -0
  182. utils/lookup_table.py +164 -0
  183. utils/metrics.py +50 -0
  184. utils/optuna.py +111 -0
  185. utils/save_mol.py +366 -0
  186. utils/util.py +36 -0
@@ -0,0 +1,470 @@
1
+ #!/usr/bin/env python3
2
+ """Prepare traceable coordinates for the Figure 5 4x4 qualitative comparison.
3
+
4
+ The script generates ETKDG and ET-Flow candidates, reads an existing MANIMOL
5
+ candidate ensemble, recomputes symmetry-aware heavy-atom RMSD with one RDKit
6
+ pipeline, and writes pair SDFs for rendering. It never edits the manuscript
7
+ figure directly.
8
+ """
9
+
10
+ from __future__ import annotations
11
+
12
+ import argparse
13
+ import csv
14
+ import json
15
+ import os
16
+ import sys
17
+ import time
18
+ from dataclasses import dataclass
19
+ from pathlib import Path
20
+ from typing import Dict, Iterable, List, Optional, Sequence, Tuple
21
+
22
+ import numpy as np
23
+ from rdkit import Chem, RDLogger
24
+ from rdkit.Chem import AllChem, rdMolAlign, rdMolDescriptors
25
+ from rdkit.Geometry import Point3D
26
+
27
+ RDLogger.DisableLog("rdApp.warning")
28
+
29
+
30
+ @dataclass
31
+ class CandidateResult:
32
+ method: str
33
+ rmsd: float
34
+ mol: Chem.Mol
35
+ num_candidates: int
36
+ note: str = ""
37
+
38
+
39
+ def parse_args() -> argparse.Namespace:
40
+ parser = argparse.ArgumentParser(
41
+ description="Generate and screen Figure 5 ETKDG/ET-Flow/MANIMOL qualitative cases."
42
+ )
43
+ parser.add_argument("--dataset_dir", default="data/dataset-1")
44
+ parser.add_argument("--csv_name", default="dataset-1_with-index.csv")
45
+ parser.add_argument("--sdf_name", default="dataset-1.sdf")
46
+ parser.add_argument("--manimol_selected_dir", default="outputs/fastsmcg_dataset1/top10/selected_sdf")
47
+ parser.add_argument("--output_dir", default="outputs/figure5_4x4")
48
+ parser.add_argument("--candidate_budget", type=int, default=10)
49
+ parser.add_argument("--seed", type=int, default=20260729)
50
+ parser.add_argument("--limit_molecules", type=int, default=0)
51
+ parser.add_argument("--start_index", type=int, default=0)
52
+ parser.add_argument("--max_cases", type=int, default=4)
53
+ parser.add_argument("--etflow_model", default="drugs-o3")
54
+ parser.add_argument("--etflow_steps", type=int, default=50)
55
+ parser.add_argument("--etflow_batch_size", type=int, default=1)
56
+ parser.add_argument("--etflow_device", default="cpu")
57
+ parser.add_argument("--etflow_cache", default="")
58
+ parser.add_argument("--etkdg_mmff", action="store_true")
59
+ parser.add_argument("--skip_etflow", action="store_true")
60
+ parser.add_argument("--strict_order", action="store_true", default=True)
61
+ parser.add_argument("--relax_order_if_needed", action="store_true")
62
+ parser.add_argument("--rmsd_max_matches", type=int, default=100000)
63
+ parser.add_argument("--min_etkdg_rmsd", type=float, default=0.0)
64
+ parser.add_argument("--max_manimol_rmsd", type=float, default=999.0)
65
+ return parser.parse_args()
66
+
67
+
68
+ def safe_name(text: str) -> str:
69
+ return "".join(ch if ch.isalnum() or ch in "._-" else "_" for ch in str(text))
70
+
71
+
72
+ def strip_index_suffix(title: str) -> str:
73
+ head, sep, tail = str(title).rpartition("_")
74
+ return head if sep and tail.isdigit() else str(title)
75
+
76
+
77
+ def read_dataset_rows(csv_path: Path) -> List[dict]:
78
+ with csv_path.open(newline="", encoding="utf-8") as handle:
79
+ return list(csv.DictReader(handle))
80
+
81
+
82
+ def read_reference_sdf(sdf_path: Path) -> Dict[str, Chem.Mol]:
83
+ refs: Dict[str, Chem.Mol] = {}
84
+ for mol in Chem.SDMolSupplier(str(sdf_path), removeHs=False):
85
+ if mol is None:
86
+ continue
87
+ name = mol.GetProp("_Name") if mol.HasProp("_Name") else ""
88
+ if name:
89
+ refs[name] = mol
90
+ return refs
91
+
92
+
93
+ def heavy_copy(mol: Chem.Mol) -> Chem.Mol:
94
+ out = Chem.RemoveHs(Chem.Mol(mol), sanitize=False)
95
+ Chem.SanitizeMol(out)
96
+ return out
97
+
98
+
99
+ def mol_with_single_conf(template: Chem.Mol, source: Chem.Mol, conf_id: int = 0, name: str = "") -> Chem.Mol:
100
+ out = Chem.Mol(template)
101
+ out.RemoveAllConformers()
102
+ src = source.GetConformer(conf_id)
103
+ conf = Chem.Conformer(out.GetNumAtoms())
104
+ if source.GetNumAtoms() != out.GetNumAtoms():
105
+ raise ValueError(f"atom-count mismatch: source={source.GetNumAtoms()} template={out.GetNumAtoms()}")
106
+ for atom_idx in range(out.GetNumAtoms()):
107
+ pos = src.GetAtomPosition(atom_idx)
108
+ conf.SetAtomPosition(atom_idx, Point3D(float(pos.x), float(pos.y), float(pos.z)))
109
+ out.AddConformer(conf, assignId=True)
110
+ if name:
111
+ out.SetProp("_Name", name)
112
+ return out
113
+
114
+
115
+ def mol_from_positions(template: Chem.Mol, positions: np.ndarray, name: str) -> Chem.Mol:
116
+ if positions.shape[0] != template.GetNumAtoms():
117
+ raise ValueError(f"position atom-count mismatch: positions={positions.shape[0]} template={template.GetNumAtoms()}")
118
+ out = Chem.Mol(template)
119
+ out.RemoveAllConformers()
120
+ conf = Chem.Conformer(out.GetNumAtoms())
121
+ for atom_idx, xyz in enumerate(positions):
122
+ conf.SetAtomPosition(atom_idx, Point3D(float(xyz[0]), float(xyz[1]), float(xyz[2])))
123
+ out.AddConformer(conf, assignId=True)
124
+ out.SetProp("_Name", name)
125
+ return out
126
+
127
+
128
+ def best_rmsd_and_aligned(ref: Chem.Mol, pred: Chem.Mol, max_matches: int) -> Tuple[float, Chem.Mol]:
129
+ ref_h = heavy_copy(ref)
130
+ pred_h = heavy_copy(pred)
131
+ if ref_h.GetNumAtoms() != pred_h.GetNumAtoms():
132
+ raise ValueError(f"heavy atom-count mismatch: ref={ref_h.GetNumAtoms()} pred={pred_h.GetNumAtoms()}")
133
+ aligned = Chem.Mol(pred_h)
134
+ rmsd = rdMolAlign.GetBestRMS(aligned, ref_h, maxMatches=int(max_matches))
135
+ return float(rmsd), aligned
136
+
137
+
138
+ def choose_best(ref: Chem.Mol, candidates: Sequence[Chem.Mol], max_matches: int, method: str) -> CandidateResult:
139
+ best: Optional[CandidateResult] = None
140
+ for idx, cand in enumerate(candidates):
141
+ try:
142
+ rmsd, aligned = best_rmsd_and_aligned(ref, cand, max_matches)
143
+ except Exception:
144
+ continue
145
+ aligned.SetProp("_Name", f"{method}_best")
146
+ aligned.SetProp("method", method)
147
+ aligned.SetProp("candidate_index", str(idx))
148
+ aligned.SetProp("rmsd_to_reference", f"{rmsd:.8f}")
149
+ if best is None or rmsd < best.rmsd:
150
+ best = CandidateResult(method=method, rmsd=rmsd, mol=aligned, num_candidates=len(candidates))
151
+ if best is None:
152
+ raise ValueError(f"no valid {method} candidates after atom/RMSD checks")
153
+ return best
154
+
155
+
156
+ def generate_etkdg_candidates(ref: Chem.Mol, budget: int, seed: int, use_mmff: bool) -> List[Chem.Mol]:
157
+ heavy = heavy_copy(ref)
158
+ work = Chem.AddHs(heavy, addCoords=False)
159
+ params = AllChem.ETKDGv3()
160
+ params.randomSeed = int(seed)
161
+ params.useRandomCoords = True
162
+ params.pruneRmsThresh = -1.0
163
+ conf_ids = list(AllChem.EmbedMultipleConfs(work, numConfs=int(budget), params=params))
164
+ out: List[Chem.Mol] = []
165
+ for conf_id in conf_ids:
166
+ if use_mmff:
167
+ try:
168
+ AllChem.MMFFOptimizeMolecule(work, confId=int(conf_id), maxIters=250)
169
+ except Exception:
170
+ pass
171
+ no_h = Chem.RemoveHs(work, sanitize=False)
172
+ Chem.SanitizeMol(no_h)
173
+ if no_h.GetNumConformers() > 1:
174
+ # RemoveHs keeps all conformers; select the corresponding one by order.
175
+ tmp = Chem.Mol(no_h)
176
+ conf = no_h.GetConformer(int(conf_id) if conf_id < no_h.GetNumConformers() else 0)
177
+ single = Chem.Mol(heavy)
178
+ single.RemoveAllConformers()
179
+ new_conf = Chem.Conformer(single.GetNumAtoms())
180
+ for atom_idx in range(single.GetNumAtoms()):
181
+ pos = conf.GetAtomPosition(atom_idx)
182
+ new_conf.SetAtomPosition(atom_idx, pos)
183
+ single.AddConformer(new_conf, assignId=True)
184
+ out.append(single)
185
+ else:
186
+ out.append(no_h)
187
+ return out[: int(budget)]
188
+
189
+
190
+ def load_manimol_candidates(selected_dir: Path, mol_title: str, budget: int) -> List[Chem.Mol]:
191
+ selected_path = selected_dir / f"{safe_name(mol_title)}_selected.sdf"
192
+ if selected_path.exists():
193
+ mols = [m for m in Chem.SDMolSupplier(str(selected_path), removeHs=False) if m is not None]
194
+ if not mols:
195
+ raise ValueError(f"empty MANIMOL selected SDF: {selected_path}")
196
+ return mols[: int(budget)]
197
+
198
+ best_pair_path = selected_dir / f"{safe_name(mol_title)}_best_pair.sdf"
199
+ if best_pair_path.exists():
200
+ mols = [m for m in Chem.SDMolSupplier(str(best_pair_path), removeHs=False) if m is not None]
201
+ if len(mols) < 2:
202
+ raise ValueError(f"MANIMOL best-pair SDF should contain reference and generated conformers: {best_pair_path}")
203
+ return [mols[1]]
204
+
205
+ raise FileNotFoundError(f"{selected_path} or {best_pair_path}")
206
+
207
+
208
+ def load_etflow_model(args: argparse.Namespace):
209
+ if args.skip_etflow:
210
+ return None
211
+ from etflow import BaseFlow
212
+
213
+ cache = args.etflow_cache or None
214
+ return BaseFlow.from_default(model=args.etflow_model, device=args.etflow_device, cache=cache)
215
+
216
+
217
+ def generate_etflow_candidates(model, smiles: str, template_ref: Chem.Mol, args: argparse.Namespace) -> List[Chem.Mol]:
218
+ if model is None:
219
+ raise RuntimeError("ET-Flow generation was disabled")
220
+ result = model.predict(
221
+ [smiles],
222
+ max_batch_size=int(args.etflow_batch_size),
223
+ num_samples=int(args.candidate_budget),
224
+ n_timesteps=int(args.etflow_steps),
225
+ seed=int(args.seed),
226
+ device=str(args.etflow_device),
227
+ as_mol=True,
228
+ )
229
+ item = result[smiles] if isinstance(result, dict) else result
230
+ candidates: List[Chem.Mol] = []
231
+ if isinstance(item, np.ndarray):
232
+ heavy_ref = heavy_copy(template_ref)
233
+ arr = item
234
+ if arr.ndim != 3:
235
+ raise ValueError(f"unexpected ET-Flow coordinate array shape: {arr.shape}")
236
+ for idx, pos in enumerate(arr[: int(args.candidate_budget)]):
237
+ candidates.append(mol_from_positions(heavy_ref, np.asarray(pos), f"ET-Flow_{idx:03d}"))
238
+ else:
239
+ if isinstance(item, Chem.Mol):
240
+ base = heavy_copy(item)
241
+ mols = [mol_with_single_conf(base, base, cid, f"ET-Flow_{cid:03d}") for cid in range(base.GetNumConformers())]
242
+ else:
243
+ mols = [heavy_copy(m) for m in list(item)]
244
+ candidates = mols[: int(args.candidate_budget)]
245
+ return candidates
246
+
247
+
248
+ def write_pair_sdf(path: Path, ref: Chem.Mol, result: CandidateResult) -> None:
249
+ path.parent.mkdir(parents=True, exist_ok=True)
250
+ ref_h = heavy_copy(ref)
251
+ ref_h.SetProp("_Name", "reference")
252
+ ref_h.SetProp("role", "reference")
253
+ gen = Chem.Mol(result.mol)
254
+ gen.SetProp("_Name", f"{result.method}_best")
255
+ gen.SetProp("role", "generated")
256
+ gen.SetProp("method", result.method)
257
+ gen.SetProp("rmsd_to_reference", f"{result.rmsd:.8f}")
258
+ writer = Chem.SDWriter(str(path))
259
+ writer.write(ref_h)
260
+ writer.write(gen)
261
+ writer.close()
262
+
263
+
264
+ def write_csv(path: Path, rows: List[dict]) -> None:
265
+ path.parent.mkdir(parents=True, exist_ok=True)
266
+ if not rows:
267
+ path.write_text("", encoding="utf-8")
268
+ return
269
+ keys: List[str] = []
270
+ for row in rows:
271
+ for key in row:
272
+ if key not in keys:
273
+ keys.append(key)
274
+ with path.open("w", newline="", encoding="utf-8") as handle:
275
+ writer = csv.DictWriter(handle, fieldnames=keys)
276
+ writer.writeheader()
277
+ writer.writerows(rows)
278
+
279
+
280
+ def diversity_bin(row: dict) -> Tuple[int, int, int]:
281
+ return (
282
+ int(float(row.get("num_rotatable_bonds", 0))) // 3,
283
+ int(float(row.get("num_heavy_atoms", 0))) // 10,
284
+ len(str(row.get("SMILES", ""))) // 30,
285
+ )
286
+
287
+
288
+ def select_cases(screening_rows: List[dict], max_cases: int, min_etkdg_rmsd: float = 0.0, max_manimol_rmsd: float = 999.0) -> List[dict]:
289
+ valid = [r for r in screening_rows if str(r.get("ordering_valid", "")).lower() == "true"]
290
+ preferred = [
291
+ r
292
+ for r in valid
293
+ if float(r["ETKDG_RMSD"]) >= float(min_etkdg_rmsd)
294
+ and float(r["MANIMOL_RMSD"]) <= float(max_manimol_rmsd)
295
+ ]
296
+ if len(preferred) >= max_cases:
297
+ valid = preferred
298
+ valid.sort(key=lambda r: (-(float(r["ETKDG_RMSD"]) - float(r["MANIMOL_RMSD"])), float(r["MANIMOL_RMSD"])))
299
+ chosen: List[dict] = []
300
+ used_bins = set()
301
+ for row in valid:
302
+ b = diversity_bin(row)
303
+ if b in used_bins and len(valid) - len(chosen) > max_cases - len(chosen):
304
+ continue
305
+ chosen.append(row)
306
+ used_bins.add(b)
307
+ if len(chosen) >= max_cases:
308
+ return chosen
309
+ for row in valid:
310
+ if row not in chosen:
311
+ chosen.append(row)
312
+ if len(chosen) >= max_cases:
313
+ break
314
+ return chosen
315
+
316
+
317
+ def main() -> int:
318
+ args = parse_args()
319
+ out_dir = Path(args.output_dir)
320
+ pair_dir = out_dir / "pairs"
321
+ out_dir.mkdir(parents=True, exist_ok=True)
322
+ log_path = out_dir / "figure5_generation_log.txt"
323
+ log_lines: List[str] = []
324
+ started = time.time()
325
+ log_lines.append(f"started={time.strftime('%Y-%m-%d %H:%M:%S')}")
326
+ log_lines.append(f"python={sys.executable}")
327
+ log_lines.append(f"candidate_budget={args.candidate_budget}")
328
+ log_lines.append(f"seed={args.seed}")
329
+
330
+ rows = read_dataset_rows(Path(args.dataset_dir) / args.csv_name)
331
+ if args.start_index:
332
+ rows = rows[int(args.start_index) :]
333
+ if args.limit_molecules:
334
+ rows = rows[: int(args.limit_molecules)]
335
+ refs = read_reference_sdf(Path(args.dataset_dir) / args.sdf_name)
336
+ etflow_model = None
337
+ try:
338
+ etflow_model = load_etflow_model(args)
339
+ log_lines.append(f"etflow_model={args.etflow_model if etflow_model is not None else 'skipped'}")
340
+ except Exception as exc:
341
+ log_lines.append(f"ET-Flow model load failed: {type(exc).__name__}: {exc}")
342
+ (out_dir / "figure5_generation_log.txt").write_text("\n".join(log_lines) + "\n", encoding="utf-8")
343
+ raise
344
+
345
+ screening: List[dict] = []
346
+ result_cache: Dict[str, Dict[str, CandidateResult]] = {}
347
+ ref_cache: Dict[str, Chem.Mol] = {}
348
+ for idx, row in enumerate(rows):
349
+ mol_title = str(row["mol_title"])
350
+ base_title = strip_index_suffix(mol_title)
351
+ ref = refs.get(base_title)
352
+ if ref is None:
353
+ screening.append({"molecule_id": mol_title, "SMILES": row.get("smiles", ""), "skipped_reason": "missing_reference"})
354
+ continue
355
+ try:
356
+ rb = rdMolDescriptors.CalcNumRotatableBonds(heavy_copy(ref))
357
+ heavy = heavy_copy(ref).GetNumAtoms()
358
+ etkdg = choose_best(
359
+ ref,
360
+ generate_etkdg_candidates(ref, args.candidate_budget, args.seed + idx, args.etkdg_mmff),
361
+ args.rmsd_max_matches,
362
+ "ETKDG",
363
+ )
364
+ etflow = None
365
+ if etflow_model is not None:
366
+ etflow = choose_best(
367
+ ref,
368
+ generate_etflow_candidates(etflow_model, str(row["smiles"]), ref, args),
369
+ args.rmsd_max_matches,
370
+ "ET-Flow",
371
+ )
372
+ manimol = choose_best(
373
+ ref,
374
+ load_manimol_candidates(Path(args.manimol_selected_dir), mol_title, args.candidate_budget),
375
+ args.rmsd_max_matches,
376
+ "MANIMOL",
377
+ )
378
+ if etflow is None:
379
+ ordering_valid = bool(etkdg.rmsd > manimol.rmsd)
380
+ else:
381
+ ordering_valid = bool(etkdg.rmsd > etflow.rmsd > manimol.rmsd)
382
+ rec = {
383
+ "molecule_id": mol_title,
384
+ "SMILES": row.get("smiles", ""),
385
+ "num_heavy_atoms": int(heavy),
386
+ "num_rotatable_bonds": int(rb),
387
+ "candidate_budget": int(args.candidate_budget),
388
+ "ETKDG_RMSD": etkdg.rmsd,
389
+ "ETFlow_RMSD": "" if etflow is None else etflow.rmsd,
390
+ "MANIMOL_RMSD": manimol.rmsd,
391
+ "ordering_valid": str(ordering_valid),
392
+ "selected_for_figure": "False",
393
+ "skipped_reason": "",
394
+ }
395
+ screening.append(rec)
396
+ result_cache[mol_title] = {"ETKDG": etkdg, "MANIMOL": manimol}
397
+ if etflow is not None:
398
+ result_cache[mol_title]["ET-Flow"] = etflow
399
+ ref_cache[mol_title] = ref
400
+ middle = "" if etflow is None else f" ET-Flow={etflow.rmsd:.3f}"
401
+ print(
402
+ f"[{idx+1}/{len(rows)}] {mol_title}: "
403
+ f"ETKDG={etkdg.rmsd:.3f}{middle} MANIMOL={manimol.rmsd:.3f} valid={ordering_valid}"
404
+ )
405
+ except Exception as exc:
406
+ screening.append(
407
+ {
408
+ "molecule_id": mol_title,
409
+ "SMILES": row.get("smiles", ""),
410
+ "selected_for_figure": "False",
411
+ "ordering_valid": "False",
412
+ "skipped_reason": f"{type(exc).__name__}: {exc}",
413
+ }
414
+ )
415
+ print(f"[{idx+1}/{len(rows)}] skip {mol_title}: {type(exc).__name__}: {exc}")
416
+
417
+ chosen = select_cases(screening, args.max_cases, args.min_etkdg_rmsd, args.max_manimol_rmsd)
418
+ if len(chosen) < args.max_cases:
419
+ msg = f"Only {len(chosen)} molecules satisfied ETKDG > ET-Flow > MANIMOL; requested {args.max_cases}."
420
+ log_lines.append(msg)
421
+ print(msg)
422
+ if args.strict_order and not args.relax_order_if_needed:
423
+ write_csv(out_dir / "figure5_case_screening.csv", screening)
424
+ write_csv(out_dir / "figure5_selected_cases.csv", chosen)
425
+ log_lines.append(f"elapsed_sec={time.time() - started:.2f}")
426
+ log_path.write_text("\n".join(log_lines) + "\n", encoding="utf-8")
427
+ return 2
428
+
429
+ selected_ids = {row["molecule_id"] for row in chosen}
430
+ for rec in screening:
431
+ if rec.get("molecule_id") in selected_ids:
432
+ rec["selected_for_figure"] = "True"
433
+ for row_idx, rec in enumerate(chosen, start=1):
434
+ mol_title = rec["molecule_id"]
435
+ ref = ref_cache[mol_title]
436
+ for method, result in result_cache[mol_title].items():
437
+ write_pair_sdf(pair_dir / safe_name(mol_title) / f"{safe_name(method)}.sdf", ref, result)
438
+ ref_heavy = heavy_copy(ref)
439
+ ref_path = pair_dir / safe_name(mol_title) / "Reference.sdf"
440
+ writer = Chem.SDWriter(str(ref_path))
441
+ ref_heavy.SetProp("_Name", "reference")
442
+ writer.write(ref_heavy)
443
+ writer.close()
444
+ rec["row_label"] = f"({chr(96 + row_idx)}) {mol_title}"
445
+
446
+ write_csv(out_dir / "figure5_case_screening.csv", screening)
447
+ write_csv(out_dir / "figure5_selected_cases.csv", chosen)
448
+ render_config = {
449
+ "candidate_budget": int(args.candidate_budget),
450
+ "seed": int(args.seed),
451
+ "rmsd_protocol": "RDKit GetBestRMS on heavy-atom molecules after RemoveHs; maxMatches=%d" % int(args.rmsd_max_matches),
452
+ "methods": ["Reference", "ETKDG", "MANIMOL"] if args.skip_etflow else ["Reference", "ETKDG", "ET-Flow", "MANIMOL"],
453
+ "colors": {
454
+ "reference": "neutral medium gray",
455
+ "ETKDG": "muted orange",
456
+ "ET-Flow": "muted teal",
457
+ "MANIMOL": "deep academic blue",
458
+ },
459
+ "pair_sdf_dir": str(pair_dir),
460
+ "selected_case_count": len(chosen),
461
+ }
462
+ (out_dir / "figure5_render_config.json").write_text(json.dumps(render_config, indent=2), encoding="utf-8")
463
+ log_lines.append(f"selected_cases={len(chosen)}")
464
+ log_lines.append(f"elapsed_sec={time.time() - started:.2f}")
465
+ log_path.write_text("\n".join(log_lines) + "\n", encoding="utf-8")
466
+ return 0
467
+
468
+
469
+ if __name__ == "__main__":
470
+ raise SystemExit(main())