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 +1,155 @@
1
+ """Interpretable bridge from pairwise mean--dispersion priors to torsions."""
2
+
3
+ from __future__ import annotations
4
+
5
+ import torch
6
+
7
+ from models.model0 import _torsion_rows
8
+
9
+
10
+ def controlled_dispersion_for_width(dispersion_prior: torch.Tensor, args) -> torch.Tensor:
11
+ """Apply a matched intervention to the dispersion map used for widths.
12
+
13
+ ``constant`` preserves the molecule-level off-diagonal mean while removing
14
+ pair-specific structure. ``shuffle`` preserves the upper-triangular
15
+ marginal exactly while breaking pair correspondence and reconstructing a
16
+ symmetric map. Both controls leave the diagonal at zero.
17
+ """
18
+ if dispersion_prior is None:
19
+ return None
20
+ mode = str(getattr(args, "dispersion_control", "normal") or "normal").lower()
21
+ if mode == "normal":
22
+ return dispersion_prior
23
+ if dispersion_prior.dim() != 2 or dispersion_prior.size(0) != dispersion_prior.size(1):
24
+ raise ValueError("dispersion_prior must be square")
25
+ n = dispersion_prior.size(0)
26
+ eye = torch.eye(n, dtype=torch.bool, device=dispersion_prior.device)
27
+ off_diag = ~eye
28
+ if mode == "constant":
29
+ values = dispersion_prior[off_diag]
30
+ constant = values.mean() if values.numel() else dispersion_prior.new_zeros(())
31
+ return torch.where(off_diag, constant.expand_as(dispersion_prior), torch.zeros_like(dispersion_prior))
32
+ if mode not in {"shuffle", "random"}:
33
+ raise ValueError(f"Unknown dispersion_control={mode!r}")
34
+ seed = int(getattr(args, "dispersion_seed", 20260913))
35
+ seed = (seed + int(dispersion_prior.numel()) * 1013 + int(dispersion_prior.shape[0]) * 9176) % (2**31 - 1)
36
+ generator = torch.Generator(device=dispersion_prior.device)
37
+ generator.manual_seed(seed)
38
+ upper = torch.triu_indices(n, n, offset=1, device=dispersion_prior.device)
39
+ values = dispersion_prior[upper[0], upper[1]]
40
+ if mode == "shuffle":
41
+ values = values[torch.randperm(values.numel(), generator=generator, device=values.device)]
42
+ else:
43
+ # Match the learned empirical marginal exactly, but independently
44
+ # randomize pair assignment; this is an auxiliary control.
45
+ values = values[torch.randperm(values.numel(), generator=generator, device=values.device)]
46
+ shuffled = torch.zeros_like(dispersion_prior)
47
+ shuffled[upper[0], upper[1]] = values
48
+ shuffled[upper[1], upper[0]] = values
49
+ return shuffled
50
+
51
+
52
+ def _symmetric(matrix: torch.Tensor, upper: float) -> torch.Tensor:
53
+ matrix = torch.as_tensor(matrix)
54
+ if matrix.dim() != 2 or matrix.size(0) != matrix.size(1):
55
+ raise ValueError("pairwise prior must be square")
56
+ if not torch.isfinite(matrix).all():
57
+ raise ValueError("pairwise prior contains non-finite values")
58
+ matrix = 0.5 * (matrix + matrix.t())
59
+ matrix = matrix.clamp(0.0, upper)
60
+ return matrix.masked_fill(torch.eye(matrix.size(0), dtype=torch.bool, device=matrix.device), 0.0)
61
+
62
+
63
+ def masked_context_pool(context: torch.Tensor, valid_mask: torch.Tensor) -> torch.Tensor:
64
+ """Pool torsion contexts while excluding padded/invalid torsion rows."""
65
+ if context.dim() == 2:
66
+ context = context.unsqueeze(0)
67
+ if valid_mask.dim() == 1:
68
+ valid_mask = valid_mask.unsqueeze(0)
69
+ if context.dim() != 3 or valid_mask.shape != context.shape[:2]:
70
+ raise ValueError("context and valid_mask shapes are incompatible")
71
+ weights = valid_mask.to(dtype=context.dtype).unsqueeze(-1)
72
+ return (context * weights).sum(dim=1) / weights.sum(dim=1).clamp_min(1.0)
73
+
74
+
75
+ def mean_dispersion_torsion_context(mean_prior, dispersion_prior, mol, torsion_index, max_torsions):
76
+ """Return [mean(mu), mean(v), q90(mu), max(mu)] for each bond cut."""
77
+ mean_prior = _symmetric(mean_prior, 1.0)
78
+ dispersion_prior = _symmetric(dispersion_prior, 0.25)
79
+ if mean_prior.shape != dispersion_prior.shape:
80
+ raise ValueError("mean and dispersion priors must have matching shapes")
81
+ context = torch.zeros(int(max_torsions), 4, dtype=mean_prior.dtype, device=mean_prior.device)
82
+ valid = torch.zeros(int(max_torsions), dtype=torch.bool, device=mean_prior.device)
83
+ rows = _torsion_rows(torsion_index)
84
+ if mol is None or rows is None or rows.numel() == 0:
85
+ return context, valid
86
+
87
+ n_atoms = min(int(mean_prior.size(0)), int(mol.GetNumAtoms()))
88
+ adjacency = [[] for _ in range(n_atoms)]
89
+ for bond in mol.GetBonds():
90
+ left, right = int(bond.GetBeginAtomIdx()), int(bond.GetEndAtomIdx())
91
+ if left < n_atoms and right < n_atoms:
92
+ adjacency[left].append(right)
93
+ adjacency[right].append(left)
94
+ for index, row in enumerate(rows[: int(max_torsions)]):
95
+ cut_left, cut_right = int(row[1]), int(row[2])
96
+ if min(cut_left, cut_right) < 0 or max(cut_left, cut_right) >= n_atoms:
97
+ continue
98
+ stack, left_atoms = [cut_left], {cut_left}
99
+ while stack:
100
+ atom = stack.pop()
101
+ for neighbor in adjacency[atom]:
102
+ if {atom, neighbor} == {cut_left, cut_right} or neighbor in left_atoms:
103
+ continue
104
+ left_atoms.add(neighbor)
105
+ stack.append(neighbor)
106
+ right_atoms = [atom for atom in range(n_atoms) if atom not in left_atoms]
107
+ if not left_atoms or not right_atoms:
108
+ continue
109
+ left = torch.tensor(sorted(left_atoms), dtype=torch.long, device=mean_prior.device)
110
+ right = torch.tensor(right_atoms, dtype=torch.long, device=mean_prior.device)
111
+ mu = mean_prior[left][:, right].reshape(-1)
112
+ disp = dispersion_prior[left][:, right].reshape(-1)
113
+ context[index] = torch.stack((mu.mean(), disp.mean(), torch.quantile(mu, 0.90), mu.max()))
114
+ valid[index] = True
115
+ return context, valid
116
+
117
+
118
+ def dispersion_torsion_widths(
119
+ dispersion_prior,
120
+ mol,
121
+ torsion_index,
122
+ max_torsions,
123
+ scale=0.35,
124
+ min_width=0.5,
125
+ max_width=2.0,
126
+ ):
127
+ """Map pairwise dispersion to bounded per-torsion proposal widths.
128
+
129
+ The score for a rotatable bond is the mean ``P_disp`` value across the
130
+ two fragments created by cutting that bond. Scores are standardized
131
+ within a molecule and converted to a bounded log-width, so the geometric
132
+ mean width is approximately one and candidate counts are unchanged.
133
+ """
134
+ if dispersion_prior is None:
135
+ return torch.ones(int(max_torsions)), torch.zeros(int(max_torsions)), torch.zeros(
136
+ int(max_torsions), dtype=torch.bool
137
+ )
138
+ zero_mean = torch.zeros_like(dispersion_prior)
139
+ context, valid = mean_dispersion_torsion_context(
140
+ zero_mean,
141
+ dispersion_prior,
142
+ mol,
143
+ torsion_index,
144
+ max_torsions,
145
+ )
146
+ scores = context[:, 1]
147
+ widths = torch.ones_like(scores)
148
+ if int(valid.sum().item()) >= 2:
149
+ valid_scores = scores[valid]
150
+ center = valid_scores.median()
151
+ spread = (valid_scores - center).abs().median().clamp_min(1e-6)
152
+ z = (scores - center) / (2.0 * spread)
153
+ widths = torch.exp(float(scale) * z).clamp(float(min_width), float(max_width))
154
+ widths = torch.where(valid, widths, torch.ones_like(widths))
155
+ return widths, scores, valid
models/model.py ADDED
@@ -0,0 +1,73 @@
1
+ import torch
2
+ from torch import nn
3
+ import torch.nn.functional as F
4
+ from torch_geometric.nn import global_mean_pool,global_add_pool,global_max_pool
5
+ from torch_scatter import scatter_add, scatter_mean
6
+ import numpy as np
7
+ from .gnnconv import GNN_node,MLP
8
+ from torch_geometric.utils import to_scipy_sparse_matrix
9
+ from scipy.sparse.csgraph import shortest_path
10
+
11
+
12
+ def compute_high_dim_adj(edge_index: torch.LongTensor,
13
+ num_nodes: int,
14
+ sigma_H: float) -> torch.Tensor:
15
+
16
+ #
17
+ A = to_scipy_sparse_matrix(edge_index, num_nodes=num_nodes) # shape [N,N]
18
+ #
19
+ dist_matrix = shortest_path(A, directed=False, unweighted=True) # numpy [N,N]
20
+ # 3.
21
+ W = np.exp(- dist_matrix / sigma_H)
22
+ # turn to torch.Tensor
23
+ return torch.from_numpy(W).float()
24
+
25
+
26
+ def compute_low_dim_adj(pos: torch.Tensor,
27
+ sigma_L: float) -> torch.Tensor:
28
+
29
+
30
+ diff = pos.unsqueeze(1) - pos.unsqueeze(0) # [N,1,D] - [1,N,D] -> [N,N,D]
31
+ dist = diff.norm(dim=-1) # [N,N]
32
+
33
+ W = torch.exp(- dist / sigma_L)
34
+ return W
35
+
36
+ class GNNEncoder(nn.Module):
37
+ def __init__(self, args, config):
38
+ super(GNNEncoder, self).__init__()
39
+ self.args = args
40
+ self.config = config
41
+ emb_dim = args.emb_dim
42
+ self.gnn = GNN_node(num_layer=args.layer, emb_dim=args.emb_dim,
43
+ drop_ratio=args.dropout, gnn_type=args.gnn_type)
44
+ self.mlp =MLP(input_dim=emb_dim, hidden_dim=args.mlp_hidden, output_dim=3,num_layers=args.mlp_layer)
45
+ if args.pooling_type == 'mean':
46
+ self.pool = global_mean_pool
47
+ elif args.pooling_type == 'add':
48
+ self.pool = global_add_pool
49
+ elif args.pooling_type =='max':
50
+ self.pool = global_max_pool
51
+ else:
52
+ raise ValueError(f"Unknown pooling method: {args.pooling_type}")
53
+
54
+
55
+ def forward(self, data):
56
+ node_feat = self.gnn(data)
57
+ graph_feat = self.pool(node_feat, data.batch)
58
+ W_H = compute_high_dim_adj(data.edge_index, data.num_nodes, self.args.sigma_H)
59
+
60
+ #print(f"node_feat shape: {node_feat.shape}")
61
+
62
+ pred_pos = self.mlp(node_feat)
63
+ W_L = compute_low_dim_adj(pred_pos, self.args.sigma_L)
64
+
65
+ #print(f"data.pos shape: {data.pos.shape}")
66
+ #print(f"pred_pos shape: {pred_pos.shape}")
67
+
68
+ pos_loss = F.mse_loss(pred_pos, data.pos) # pos_pred, pos_true: [N, 3]
69
+ device = node_feat.device # or data.x.device
70
+ W_H = W_H.to(device)
71
+ W_L = W_L.to(device)
72
+ mani_loss= torch.sum((W_H - W_L) ** 2)
73
+ return pred_pos, graph_feat, pos_loss,mani_loss