boltz-vsynthes 1.0.27__py3-none-any.whl → 1.0.29__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.
- boltz/data/parse/pdb.py +6 -7
- boltz/data/parse/pdb_download.py +6 -8
- boltz/data/parse/schema.py +3 -5
- boltz/data/parse/sdf.py +5 -7
- {boltz_vsynthes-1.0.27.dist-info → boltz_vsynthes-1.0.29.dist-info}/METADATA +1 -1
- {boltz_vsynthes-1.0.27.dist-info → boltz_vsynthes-1.0.29.dist-info}/RECORD +10 -10
- {boltz_vsynthes-1.0.27.dist-info → boltz_vsynthes-1.0.29.dist-info}/WHEEL +0 -0
- {boltz_vsynthes-1.0.27.dist-info → boltz_vsynthes-1.0.29.dist-info}/entry_points.txt +0 -0
- {boltz_vsynthes-1.0.27.dist-info → boltz_vsynthes-1.0.29.dist-info}/licenses/LICENSE +0 -0
- {boltz_vsynthes-1.0.27.dist-info → boltz_vsynthes-1.0.29.dist-info}/top_level.txt +0 -0
boltz/data/parse/pdb.py
CHANGED
@@ -5,6 +5,7 @@ from typing import Optional
|
|
5
5
|
from Bio import PDB
|
6
6
|
from Bio.PDB.PDBParser import PDBParser
|
7
7
|
from Bio.PDB.Polypeptide import PPBuilder
|
8
|
+
from Bio.Data.IUPACData import protein_letters_3to1
|
8
9
|
from rdkit import Chem
|
9
10
|
from rdkit.Chem.rdchem import Mol
|
10
11
|
|
@@ -17,7 +18,7 @@ def parse_pdb(
|
|
17
18
|
ccd: dict[str, Mol],
|
18
19
|
mol_dir: Path,
|
19
20
|
boltz2: bool = False,
|
20
|
-
) ->
|
21
|
+
) -> dict:
|
21
22
|
"""Parse a PDB file.
|
22
23
|
|
23
24
|
Parameters
|
@@ -33,8 +34,8 @@ def parse_pdb(
|
|
33
34
|
|
34
35
|
Returns
|
35
36
|
-------
|
36
|
-
|
37
|
-
|
37
|
+
dict
|
38
|
+
Dictionary containing sequences and bonds.
|
38
39
|
"""
|
39
40
|
# Read PDB file
|
40
41
|
parser = PDBParser(QUIET=True)
|
@@ -56,10 +57,8 @@ def parse_pdb(
|
|
56
57
|
}
|
57
58
|
})
|
58
59
|
|
59
|
-
|
60
|
+
return {
|
60
61
|
"sequences": sequences,
|
61
62
|
"bonds": [],
|
62
63
|
"version": 1,
|
63
|
-
}
|
64
|
-
|
65
|
-
return parse_boltz_schema(pdb_path.stem, data, ccd, mol_dir, boltz2)
|
64
|
+
}
|
boltz/data/parse/pdb_download.py
CHANGED
@@ -55,7 +55,7 @@ def parse_pdb_id(
|
|
55
55
|
mol_dir: Path,
|
56
56
|
cache_dir: Path,
|
57
57
|
boltz2: bool = False,
|
58
|
-
) ->
|
58
|
+
) -> dict:
|
59
59
|
"""Parse a PDB file by ID.
|
60
60
|
|
61
61
|
Parameters
|
@@ -73,12 +73,12 @@ def parse_pdb_id(
|
|
73
73
|
|
74
74
|
Returns
|
75
75
|
-------
|
76
|
-
|
77
|
-
|
76
|
+
dict
|
77
|
+
Dictionary containing sequences and bonds.
|
78
78
|
"""
|
79
79
|
# Download PDB file
|
80
80
|
pdb_path = download_pdb(pdb_id, cache_dir)
|
81
|
-
|
81
|
+
|
82
82
|
# Read PDB file
|
83
83
|
parser = PDBParser(QUIET=True)
|
84
84
|
structure = parser.get_structure("protein", str(pdb_path))
|
@@ -99,10 +99,8 @@ def parse_pdb_id(
|
|
99
99
|
}
|
100
100
|
})
|
101
101
|
|
102
|
-
|
102
|
+
return {
|
103
103
|
"sequences": sequences,
|
104
104
|
"bonds": [],
|
105
105
|
"version": 1,
|
106
|
-
}
|
107
|
-
|
108
|
-
return parse_boltz_schema(pdb_id, data, ccd, mol_dir, boltz2)
|
106
|
+
}
|
boltz/data/parse/schema.py
CHANGED
@@ -1024,12 +1024,12 @@ def parse_boltz_schema( # noqa: C901, PLR0915, PLR0912
|
|
1024
1024
|
# This is a PDB ID
|
1025
1025
|
from boltz.data.parse.pdb_download import parse_pdb_id
|
1026
1026
|
target = parse_pdb_id(pdb_path.stem, ccd, mol_dir, pdb_path.parent)
|
1027
|
-
seq = target
|
1027
|
+
seq = target["sequences"][0]["protein"]["sequence"] # Access sequence from dict
|
1028
1028
|
else:
|
1029
1029
|
# This is a PDB file
|
1030
1030
|
from boltz.data.parse.pdb import parse_pdb
|
1031
1031
|
target = parse_pdb(pdb_path, ccd, mol_dir)
|
1032
|
-
seq = target
|
1032
|
+
seq = target["sequences"][0]["protein"]["sequence"] # Access sequence from dict
|
1033
1033
|
else:
|
1034
1034
|
msg = f"Protein must have either 'sequence' or 'pdb' field: {item}"
|
1035
1035
|
raise ValueError(msg)
|
@@ -1041,9 +1041,7 @@ def parse_boltz_schema( # noqa: C901, PLR0915, PLR0912
|
|
1041
1041
|
sdf_path = Path(item[entity_type]["sdf"])
|
1042
1042
|
from boltz.data.parse.sdf import parse_sdf
|
1043
1043
|
target = parse_sdf(sdf_path, ccd, mol_dir)
|
1044
|
-
|
1045
|
-
seq = target.sequences[0]
|
1046
|
-
print(seq)
|
1044
|
+
seq = target["sequences"][0]["ligand"]["smiles"] # Access sequence from dict
|
1047
1045
|
elif "ccd" in item[entity_type]:
|
1048
1046
|
seq = str(item[entity_type]["ccd"])
|
1049
1047
|
else:
|
boltz/data/parse/sdf.py
CHANGED
@@ -48,7 +48,7 @@ def parse_sdf(
|
|
48
48
|
ccd: dict[str, Mol],
|
49
49
|
mol_dir: Path,
|
50
50
|
boltz2: bool = False,
|
51
|
-
) ->
|
51
|
+
) -> dict:
|
52
52
|
"""Parse an SDF file.
|
53
53
|
|
54
54
|
Parameters
|
@@ -64,8 +64,8 @@ def parse_sdf(
|
|
64
64
|
|
65
65
|
Returns
|
66
66
|
-------
|
67
|
-
|
68
|
-
|
67
|
+
dict
|
68
|
+
Dictionary containing sequences and bonds.
|
69
69
|
"""
|
70
70
|
# Process SDF file
|
71
71
|
mol_dict = _process_sdf(str(sdf_path))
|
@@ -81,10 +81,8 @@ def parse_sdf(
|
|
81
81
|
}
|
82
82
|
})
|
83
83
|
|
84
|
-
|
84
|
+
return {
|
85
85
|
"sequences": sequences,
|
86
86
|
"bonds": [],
|
87
87
|
"version": 1,
|
88
|
-
}
|
89
|
-
|
90
|
-
return parse_boltz_schema(sdf_path.stem, data, ccd, mol_dir, boltz2)
|
88
|
+
}
|
@@ -38,10 +38,10 @@ boltz/data/parse/csv.py,sha256=Hcq8rJW2njczahEr8jfd_o-zxLaNSgJ3YIoC9srIqpw,2518
|
|
38
38
|
boltz/data/parse/fasta.py,sha256=taI4s_CqPtyF0XaLJAsVAJHCL0GXm2g1g8Qeccdxikk,3906
|
39
39
|
boltz/data/parse/mmcif.py,sha256=25kEXCkx-OuaawAs7cdz0fxdRu5_CCO0AV00u84PrjQ,36822
|
40
40
|
boltz/data/parse/mmcif_with_constraints.py,sha256=WHYZckSqUwu-Nb9vmVmxHmC7uxwVrF7AVUeVKsc5wGQ,51473
|
41
|
-
boltz/data/parse/pdb.py,sha256=
|
42
|
-
boltz/data/parse/pdb_download.py,sha256=
|
43
|
-
boltz/data/parse/schema.py,sha256=
|
44
|
-
boltz/data/parse/sdf.py,sha256=
|
41
|
+
boltz/data/parse/pdb.py,sha256=iybk4p2UgUy_ABGprDq_xxyPSdm1HAZsGTM0lhxVEwM,1654
|
42
|
+
boltz/data/parse/pdb_download.py,sha256=wge-scX-lOatX0q83W1wOsaql99rYp-6uGWSHEc995M,2718
|
43
|
+
boltz/data/parse/schema.py,sha256=CcK-9F7h3fG7VI5zMNzhi2Yjqfq30z-ndBstO_2mx8o,60908
|
44
|
+
boltz/data/parse/sdf.py,sha256=fs3MQVClDcCzxJaeVYiDuoh-fUrYc8Tcd5Bz8ws3FKI,2052
|
45
45
|
boltz/data/parse/yaml.py,sha256=GRFRMtDD4PQ4PIpA_S1jj0vRaEu2LlZd_g4rN1zUrNo,1505
|
46
46
|
boltz/data/sample/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
|
47
47
|
boltz/data/sample/cluster.py,sha256=9Sx8qP7zGZOAyEspwYFtCTbGTBZnuN-zfCKFbbA_6oI,8175
|
@@ -107,9 +107,9 @@ boltz/model/optim/scheduler.py,sha256=nB4jz0CZ4pR4n08LQngExL_pNycIdYI8AXVoHPnZWQ
|
|
107
107
|
boltz/model/potentials/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
|
108
108
|
boltz/model/potentials/potentials.py,sha256=vev8Vjfs-ML1hyrdv_R8DynG4wSFahJ6nzPWp7CYQqw,17507
|
109
109
|
boltz/model/potentials/schedules.py,sha256=m7XJjfuF9uTX3bR9VisXv1rvzJjxiD8PobXRpcBBu1c,968
|
110
|
-
boltz_vsynthes-1.0.
|
111
|
-
boltz_vsynthes-1.0.
|
112
|
-
boltz_vsynthes-1.0.
|
113
|
-
boltz_vsynthes-1.0.
|
114
|
-
boltz_vsynthes-1.0.
|
115
|
-
boltz_vsynthes-1.0.
|
110
|
+
boltz_vsynthes-1.0.29.dist-info/licenses/LICENSE,sha256=8GZ_1eZsUeG6jdqgJJxtciWzADfgLEV4LY8sKUOsJhc,1102
|
111
|
+
boltz_vsynthes-1.0.29.dist-info/METADATA,sha256=2NGfvz95nz21zJeUKD-q9ymclZarT3dPCq4dxFXQztw,7171
|
112
|
+
boltz_vsynthes-1.0.29.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
|
113
|
+
boltz_vsynthes-1.0.29.dist-info/entry_points.txt,sha256=n5a5I35ntu9lmyr16oZgHPFY0b0YxjiixY7m7nbMTLc,41
|
114
|
+
boltz_vsynthes-1.0.29.dist-info/top_level.txt,sha256=MgU3Jfb-ctWm07YGMts68PMjSh9v26D0gfG3dFRmVFA,6
|
115
|
+
boltz_vsynthes-1.0.29.dist-info/RECORD,,
|
File without changes
|
File without changes
|
File without changes
|
File without changes
|