rxnmapper 0.4.2__tar.gz → 0.5.0__tar.gz
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.
- {rxnmapper-0.4.2/rxnmapper.egg-info → rxnmapper-0.5.0}/PKG-INFO +24 -35
- {rxnmapper-0.4.2 → rxnmapper-0.5.0}/README.md +18 -26
- {rxnmapper-0.4.2 → rxnmapper-0.5.0}/pyproject.toml +0 -4
- {rxnmapper-0.4.2 → rxnmapper-0.5.0}/rxnmapper/__init__.py +1 -1
- {rxnmapper-0.4.2 → rxnmapper-0.5.0}/rxnmapper/batched_mapper.py +2 -5
- {rxnmapper-0.4.2 → rxnmapper-0.5.0}/rxnmapper/core.py +23 -18
- {rxnmapper-0.4.2 → rxnmapper-0.5.0}/rxnmapper/smiles_utils.py +1 -1
- {rxnmapper-0.4.2 → rxnmapper-0.5.0}/rxnmapper/tokenization_smiles.py +70 -19
- {rxnmapper-0.4.2 → rxnmapper-0.5.0/rxnmapper.egg-info}/PKG-INFO +24 -35
- {rxnmapper-0.4.2 → rxnmapper-0.5.0}/rxnmapper.egg-info/requires.txt +0 -5
- {rxnmapper-0.4.2 → rxnmapper-0.5.0}/setup.cfg +5 -8
- {rxnmapper-0.4.2 → rxnmapper-0.5.0}/LICENSE +0 -0
- {rxnmapper-0.4.2 → rxnmapper-0.5.0}/MANIFEST.in +0 -0
- {rxnmapper-0.4.2 → rxnmapper-0.5.0}/rxnmapper/attention.py +0 -0
- {rxnmapper-0.4.2 → rxnmapper-0.5.0}/rxnmapper/models/transformers/albert_heads_8_uspto_all_1310k/config.json +0 -0
- {rxnmapper-0.4.2 → rxnmapper-0.5.0}/rxnmapper/models/transformers/albert_heads_8_uspto_all_1310k/pytorch_model.bin +0 -0
- {rxnmapper-0.4.2 → rxnmapper-0.5.0}/rxnmapper/models/transformers/albert_heads_8_uspto_all_1310k/special_tokens_map.json +0 -0
- {rxnmapper-0.4.2 → rxnmapper-0.5.0}/rxnmapper/models/transformers/albert_heads_8_uspto_all_1310k/tokenizer_config.json +0 -0
- {rxnmapper-0.4.2 → rxnmapper-0.5.0}/rxnmapper/models/transformers/albert_heads_8_uspto_all_1310k/training_args.bin +0 -0
- {rxnmapper-0.4.2 → rxnmapper-0.5.0}/rxnmapper/models/transformers/albert_heads_8_uspto_all_1310k/vocab.txt +0 -0
- {rxnmapper-0.4.2 → rxnmapper-0.5.0}/rxnmapper/py.typed +0 -0
- {rxnmapper-0.4.2 → rxnmapper-0.5.0}/rxnmapper.egg-info/SOURCES.txt +0 -0
- {rxnmapper-0.4.2 → rxnmapper-0.5.0}/rxnmapper.egg-info/dependency_links.txt +0 -0
- {rxnmapper-0.4.2 → rxnmapper-0.5.0}/rxnmapper.egg-info/not-zip-safe +0 -0
- {rxnmapper-0.4.2 → rxnmapper-0.5.0}/rxnmapper.egg-info/top_level.txt +0 -0
- {rxnmapper-0.4.2 → rxnmapper-0.5.0}/setup.py +0 -0
- {rxnmapper-0.4.2 → rxnmapper-0.5.0}/tests/__init__.py +0 -0
- {rxnmapper-0.4.2 → rxnmapper-0.5.0}/tests/test_batched_mapper.py +0 -0
- {rxnmapper-0.4.2 → rxnmapper-0.5.0}/tests/test_mapper.py +0 -0
- {rxnmapper-0.4.2 → rxnmapper-0.5.0}/tests/test_smiles_utils.py +0 -0
- {rxnmapper-0.4.2 → rxnmapper-0.5.0}/tests/utils.py +0 -0
|
@@ -1,6 +1,6 @@
|
|
|
1
1
|
Metadata-Version: 2.4
|
|
2
2
|
Name: rxnmapper
|
|
3
|
-
Version: 0.
|
|
3
|
+
Version: 0.5.0
|
|
4
4
|
Summary: Reaction atom-mapping from transformers
|
|
5
5
|
Home-page: https://github.com/rxn4chemistry/rxnmapper
|
|
6
6
|
Author: IBM RXN team
|
|
@@ -9,16 +9,14 @@ License: MIT
|
|
|
9
9
|
Classifier: Operating System :: OS Independent
|
|
10
10
|
Classifier: Intended Audience :: Developers
|
|
11
11
|
Classifier: Intended Audience :: Science/Research
|
|
12
|
-
Classifier: License :: OSI Approved :: MIT License
|
|
13
12
|
Classifier: Programming Language :: Python :: 3
|
|
14
|
-
Classifier: Programming Language :: Python :: 3.6
|
|
15
|
-
Classifier: Programming Language :: Python :: 3.7
|
|
16
|
-
Classifier: Programming Language :: Python :: 3.8
|
|
17
|
-
Classifier: Programming Language :: Python :: 3.9
|
|
18
13
|
Classifier: Programming Language :: Python :: 3.10
|
|
19
14
|
Classifier: Programming Language :: Python :: 3.11
|
|
15
|
+
Classifier: Programming Language :: Python :: 3.12
|
|
16
|
+
Classifier: Programming Language :: Python :: 3.13
|
|
17
|
+
Classifier: Programming Language :: Python :: 3.14
|
|
20
18
|
Classifier: Topic :: Software Development :: Libraries :: Python Modules
|
|
21
|
-
Requires-Python: >=3.
|
|
19
|
+
Requires-Python: >=3.10
|
|
22
20
|
Description-Content-Type: text/markdown
|
|
23
21
|
License-File: LICENSE
|
|
24
22
|
Requires-Dist: pandas>=1.0.3
|
|
@@ -38,8 +36,7 @@ Requires-Dist: sphinx==2.2.1; extra == "dev"
|
|
|
38
36
|
Requires-Dist: twine==3.1.1; extra == "dev"
|
|
39
37
|
Requires-Dist: types-setuptools>=62.6.0; extra == "dev"
|
|
40
38
|
Provides-Extra: rdkit
|
|
41
|
-
Requires-Dist: rdkit
|
|
42
|
-
Requires-Dist: rdkit>=2022.3.4; python_version >= "3.7" and extra == "rdkit"
|
|
39
|
+
Requires-Dist: rdkit>=2022.3.4; extra == "rdkit"
|
|
43
40
|
Dynamic: license-file
|
|
44
41
|
|
|
45
42
|
# Extraction of organic chemistry grammar from unsupervised learning of chemical reactions
|
|
@@ -51,37 +48,27 @@ Enable robust atom mapping on valid reaction SMILES. The atom-mapping informatio
|
|
|
51
48
|
|
|
52
49
|
## Installation
|
|
53
50
|
|
|
54
|
-
###
|
|
55
|
-
|
|
56
|
-
|
|
57
|
-
|
|
58
|
-
|
|
51
|
+
### Create virtual environment (optional)
|
|
52
|
+
|
|
53
|
+
```bash
|
|
54
|
+
python3 -m venv .venv
|
|
55
|
+
source .venv/bin/activate
|
|
59
56
|
```
|
|
60
57
|
|
|
61
|
-
###
|
|
62
|
-
You can install the package and setup the environment directly from github using:
|
|
58
|
+
### Install from pip
|
|
63
59
|
|
|
64
|
-
```
|
|
65
|
-
|
|
66
|
-
cd rxnmapper
|
|
67
|
-
conda create -n rxnmapper python=3.6 -y
|
|
68
|
-
conda activate rxnmapper
|
|
69
|
-
pip install -e .
|
|
60
|
+
```bash
|
|
61
|
+
pip install "rxnmapper[rdkit]"
|
|
70
62
|
```
|
|
71
63
|
|
|
72
|
-
|
|
64
|
+
You can leave out `[rdkit]` if RDKit is already available in your Python environment.
|
|
73
65
|
|
|
74
|
-
|
|
75
|
-
It can also be installed via Conda or Pypi:
|
|
66
|
+
### From source
|
|
76
67
|
|
|
77
68
|
```bash
|
|
78
|
-
|
|
79
|
-
|
|
80
|
-
|
|
81
|
-
# Install RDKit from Pypi
|
|
82
|
-
pip install rdkit
|
|
83
|
-
# for Python<3.7
|
|
84
|
-
# pip install rdkit-pypi
|
|
69
|
+
git clone https://github.com/rxn4chemistry/rxnmapper.git
|
|
70
|
+
cd rxnmapper
|
|
71
|
+
pip install -e ".[rdkit]"
|
|
85
72
|
```
|
|
86
73
|
|
|
87
74
|
## Usage
|
|
@@ -118,10 +105,12 @@ results = list(rxn_mapper.map_reactions_with_info(rxns)) # results as dictionar
|
|
|
118
105
|
|
|
119
106
|
### Testing
|
|
120
107
|
|
|
121
|
-
You can run the
|
|
108
|
+
You can run the test suite with:
|
|
122
109
|
|
|
123
|
-
|
|
124
|
-
|
|
110
|
+
```bash
|
|
111
|
+
pip install -e .[dev,rdkit]
|
|
112
|
+
pytest tests
|
|
113
|
+
```
|
|
125
114
|
|
|
126
115
|
## Examples
|
|
127
116
|
|
|
@@ -7,37 +7,27 @@ Enable robust atom mapping on valid reaction SMILES. The atom-mapping informatio
|
|
|
7
7
|
|
|
8
8
|
## Installation
|
|
9
9
|
|
|
10
|
-
###
|
|
11
|
-
|
|
12
|
-
|
|
13
|
-
|
|
14
|
-
|
|
10
|
+
### Create virtual environment (optional)
|
|
11
|
+
|
|
12
|
+
```bash
|
|
13
|
+
python3 -m venv .venv
|
|
14
|
+
source .venv/bin/activate
|
|
15
15
|
```
|
|
16
16
|
|
|
17
|
-
###
|
|
18
|
-
You can install the package and setup the environment directly from github using:
|
|
17
|
+
### Install from pip
|
|
19
18
|
|
|
20
|
-
```
|
|
21
|
-
|
|
22
|
-
cd rxnmapper
|
|
23
|
-
conda create -n rxnmapper python=3.6 -y
|
|
24
|
-
conda activate rxnmapper
|
|
25
|
-
pip install -e .
|
|
19
|
+
```bash
|
|
20
|
+
pip install "rxnmapper[rdkit]"
|
|
26
21
|
```
|
|
27
22
|
|
|
28
|
-
|
|
23
|
+
You can leave out `[rdkit]` if RDKit is already available in your Python environment.
|
|
29
24
|
|
|
30
|
-
|
|
31
|
-
It can also be installed via Conda or Pypi:
|
|
25
|
+
### From source
|
|
32
26
|
|
|
33
27
|
```bash
|
|
34
|
-
|
|
35
|
-
|
|
36
|
-
|
|
37
|
-
# Install RDKit from Pypi
|
|
38
|
-
pip install rdkit
|
|
39
|
-
# for Python<3.7
|
|
40
|
-
# pip install rdkit-pypi
|
|
28
|
+
git clone https://github.com/rxn4chemistry/rxnmapper.git
|
|
29
|
+
cd rxnmapper
|
|
30
|
+
pip install -e ".[rdkit]"
|
|
41
31
|
```
|
|
42
32
|
|
|
43
33
|
## Usage
|
|
@@ -74,10 +64,12 @@ results = list(rxn_mapper.map_reactions_with_info(rxns)) # results as dictionar
|
|
|
74
64
|
|
|
75
65
|
### Testing
|
|
76
66
|
|
|
77
|
-
You can run the
|
|
67
|
+
You can run the test suite with:
|
|
78
68
|
|
|
79
|
-
|
|
80
|
-
|
|
69
|
+
```bash
|
|
70
|
+
pip install -e .[dev,rdkit]
|
|
71
|
+
pytest tests
|
|
72
|
+
```
|
|
81
73
|
|
|
82
74
|
## Examples
|
|
83
75
|
|
|
@@ -4,14 +4,10 @@ build-backend = "setuptools.build_meta"
|
|
|
4
4
|
|
|
5
5
|
[tool.mypy]
|
|
6
6
|
check_untyped_defs = true
|
|
7
|
-
plugins = [
|
|
8
|
-
"numpy.typing.mypy_plugin",
|
|
9
|
-
]
|
|
10
7
|
|
|
11
8
|
[[tool.mypy.overrides]]
|
|
12
9
|
module = [
|
|
13
10
|
"pandas.*",
|
|
14
|
-
"pkg_resources.*",
|
|
15
11
|
"rdkit.*",
|
|
16
12
|
"scipy.*",
|
|
17
13
|
"tqdm.*",
|
|
@@ -1,11 +1,10 @@
|
|
|
1
1
|
import logging
|
|
2
2
|
from typing import Any, Dict, Iterable, Iterator, List, Optional
|
|
3
3
|
|
|
4
|
-
import pkg_resources
|
|
5
4
|
from rxn.utilities.containers import chunker
|
|
6
5
|
from rxn.utilities.files import PathLike
|
|
7
6
|
|
|
8
|
-
from .core import RXNMapper
|
|
7
|
+
from .core import DEFAULT_MODEL_PATH, RXNMapper
|
|
9
8
|
|
|
10
9
|
logger = logging.getLogger(__name__)
|
|
11
10
|
logger.addHandler(logging.NullHandler())
|
|
@@ -45,9 +44,7 @@ class BatchedMapper:
|
|
|
45
44
|
is an issue in the prediction (number of tokens, invalid SMILES, ...).
|
|
46
45
|
"""
|
|
47
46
|
if model_path is None:
|
|
48
|
-
model_path =
|
|
49
|
-
"rxnmapper", "models/transformers/albert_heads_8_uspto_all_1310k"
|
|
50
|
-
)
|
|
47
|
+
model_path = DEFAULT_MODEL_PATH
|
|
51
48
|
self.mapper = RXNMapper(
|
|
52
49
|
config=dict(
|
|
53
50
|
model_path=str(model_path),
|
|
@@ -8,7 +8,6 @@ from contextlib import contextmanager
|
|
|
8
8
|
from typing import Any, Dict, Iterator, List, Optional, Tuple
|
|
9
9
|
|
|
10
10
|
import numpy as np
|
|
11
|
-
import pkg_resources
|
|
12
11
|
import torch
|
|
13
12
|
from rxn.chemutils.reaction_equation import ReactionEquation
|
|
14
13
|
from rxn.chemutils.reaction_smiles import (
|
|
@@ -23,7 +22,17 @@ from .attention import AttentionScorer
|
|
|
23
22
|
from .smiles_utils import generate_atom_mapped_reaction_atoms, process_reaction
|
|
24
23
|
from .tokenization_smiles import SmilesTokenizer
|
|
25
24
|
|
|
26
|
-
MODEL_TYPE_DICT
|
|
25
|
+
MODEL_TYPE_DICT: Dict[str, Any] = {
|
|
26
|
+
"bert": BertModel,
|
|
27
|
+
"albert": AlbertModel,
|
|
28
|
+
"roberta": RobertaModel,
|
|
29
|
+
}
|
|
30
|
+
DEFAULT_MODEL_PATH = os.path.join(
|
|
31
|
+
os.path.dirname(__file__),
|
|
32
|
+
"models",
|
|
33
|
+
"transformers",
|
|
34
|
+
"albert_heads_8_uspto_all_1310k",
|
|
35
|
+
)
|
|
27
36
|
|
|
28
37
|
_logger = logging.getLogger(__name__)
|
|
29
38
|
_logger.addHandler(logging.NullHandler())
|
|
@@ -70,12 +79,7 @@ class RXNMapper:
|
|
|
70
79
|
config = {}
|
|
71
80
|
|
|
72
81
|
# Config takes "model_path", "model_type", "attention_multiplier", "head", "layers"
|
|
73
|
-
self.model_path = config.get(
|
|
74
|
-
"model_path",
|
|
75
|
-
pkg_resources.resource_filename(
|
|
76
|
-
"rxnmapper", "models/transformers/albert_heads_8_uspto_all_1310k"
|
|
77
|
-
),
|
|
78
|
-
)
|
|
82
|
+
self.model_path = config.get("model_path", DEFAULT_MODEL_PATH)
|
|
79
83
|
self.model_type = config.get("model_type", "albert")
|
|
80
84
|
self.attention_multiplier = config.get("attention_multiplier", 90.0)
|
|
81
85
|
self.head = config.get("head", 5)
|
|
@@ -94,12 +98,15 @@ class RXNMapper:
|
|
|
94
98
|
Tuple: containing model and tokenizer
|
|
95
99
|
"""
|
|
96
100
|
model_class = MODEL_TYPE_DICT[self.model_type]
|
|
97
|
-
|
|
98
|
-
|
|
99
|
-
|
|
100
|
-
|
|
101
|
-
|
|
102
|
-
)
|
|
101
|
+
model_kwargs: Dict[str, Any] = {
|
|
102
|
+
"output_attentions": True,
|
|
103
|
+
"output_past": False,
|
|
104
|
+
"output_hidden_states": False,
|
|
105
|
+
}
|
|
106
|
+
if getattr(model_class, "_supports_sdpa", False):
|
|
107
|
+
model_kwargs["attn_implementation"] = "eager"
|
|
108
|
+
|
|
109
|
+
model = model_class.from_pretrained(self.model_path, **model_kwargs)
|
|
103
110
|
|
|
104
111
|
vocab_path = None
|
|
105
112
|
|
|
@@ -107,7 +114,7 @@ class RXNMapper:
|
|
|
107
114
|
vocab_path = os.path.join(self.model_path, "vocab.txt")
|
|
108
115
|
|
|
109
116
|
tokenizer = SmilesTokenizer(
|
|
110
|
-
vocab_path,
|
|
117
|
+
vocab_path, model_max_length=model.config.max_position_embeddings
|
|
111
118
|
)
|
|
112
119
|
return (model, tokenizer)
|
|
113
120
|
|
|
@@ -134,7 +141,7 @@ class RXNMapper:
|
|
|
134
141
|
else:
|
|
135
142
|
use_head = force_head
|
|
136
143
|
|
|
137
|
-
encoded_ids = self.tokenizer
|
|
144
|
+
encoded_ids = self.tokenizer(
|
|
138
145
|
rxn_smiles_list,
|
|
139
146
|
padding=True,
|
|
140
147
|
return_tensors="pt",
|
|
@@ -148,8 +155,6 @@ class RXNMapper:
|
|
|
148
155
|
f"Reaction SMILES has {max_input_length} tokens, should be at most {max_supported_by_model}."
|
|
149
156
|
)
|
|
150
157
|
|
|
151
|
-
# suppress warning that suggests setting "attn_implementation"; doing
|
|
152
|
-
# so would break compatibility for old `transformers` versions.
|
|
153
158
|
with suppress_transformers_warnings():
|
|
154
159
|
with torch.no_grad():
|
|
155
160
|
output = self.model(**parsed_input)
|
|
@@ -330,7 +330,7 @@ def canonicalize_and_atom_map(smi: str, return_equivalent_atoms=False):
|
|
|
330
330
|
]
|
|
331
331
|
)
|
|
332
332
|
|
|
333
|
-
atom_maps_canonical = [mol.
|
|
333
|
+
atom_maps_canonical = [mol.GetAtomWithIdx(idx).GetProp("atom_map") for idx in order]
|
|
334
334
|
|
|
335
335
|
if not return_equivalent_atoms:
|
|
336
336
|
return (can_smi, atom_maps_canonical)
|
|
@@ -6,15 +6,16 @@ import collections
|
|
|
6
6
|
import logging
|
|
7
7
|
import os
|
|
8
8
|
import re
|
|
9
|
+
from typing import Dict, List, Optional
|
|
9
10
|
|
|
10
|
-
from transformers import
|
|
11
|
+
from transformers import PreTrainedTokenizer
|
|
11
12
|
|
|
12
13
|
from .smiles_utils import SMI_REGEX_PATTERN
|
|
13
14
|
|
|
14
15
|
logger = logging.getLogger(__name__)
|
|
15
16
|
|
|
16
17
|
|
|
17
|
-
class SmilesTokenizer(
|
|
18
|
+
class SmilesTokenizer(PreTrainedTokenizer):
|
|
18
19
|
r"""
|
|
19
20
|
Constructs a SmilesTokenizer.
|
|
20
21
|
Mostly copied from https://github.com/huggingface/transformers
|
|
@@ -23,24 +24,25 @@ class SmilesTokenizer(BertTokenizer):
|
|
|
23
24
|
vocab_file: Path to a SMILES character per line vocabulary file
|
|
24
25
|
"""
|
|
25
26
|
|
|
27
|
+
vocab_files_names = {"vocab_file": "vocab.txt"}
|
|
28
|
+
model_input_names = ["input_ids", "token_type_ids", "attention_mask"]
|
|
29
|
+
|
|
26
30
|
def __init__(
|
|
27
31
|
self,
|
|
28
|
-
vocab_file,
|
|
29
|
-
|
|
30
|
-
|
|
31
|
-
|
|
32
|
-
|
|
33
|
-
|
|
34
|
-
**kwargs
|
|
32
|
+
vocab_file: Optional[str],
|
|
33
|
+
unk_token: str = "[UNK]",
|
|
34
|
+
sep_token: str = "[SEP]",
|
|
35
|
+
pad_token: str = "[PAD]",
|
|
36
|
+
cls_token: str = "[CLS]",
|
|
37
|
+
mask_token: str = "[MASK]",
|
|
38
|
+
**kwargs,
|
|
35
39
|
):
|
|
36
|
-
"""Constructs a
|
|
40
|
+
"""Constructs a SmilesTokenizer.
|
|
37
41
|
|
|
38
42
|
Args:
|
|
39
43
|
**vocab_file**: Path to a SMILES character per line vocabulary file
|
|
40
44
|
"""
|
|
41
|
-
|
|
42
|
-
|
|
43
|
-
if not os.path.isfile(vocab_file):
|
|
45
|
+
if not vocab_file or not os.path.isfile(vocab_file):
|
|
44
46
|
raise ValueError("Can't find a vocab file at path '{}'.".format(vocab_file))
|
|
45
47
|
self.vocab = load_vocab(vocab_file)
|
|
46
48
|
self.highest_unused_index = max(
|
|
@@ -49,29 +51,78 @@ class SmilesTokenizer(BertTokenizer):
|
|
|
49
51
|
self.ids_to_tokens = collections.OrderedDict(
|
|
50
52
|
[(ids, tok) for tok, ids in self.vocab.items()]
|
|
51
53
|
)
|
|
52
|
-
self.basic_tokenizer = BasicSmilesTokenizer()
|
|
54
|
+
self.basic_tokenizer = BasicSmilesTokenizer()
|
|
55
|
+
|
|
56
|
+
if "max_len" in kwargs and "model_max_length" not in kwargs:
|
|
57
|
+
kwargs["model_max_length"] = kwargs.pop("max_len")
|
|
58
|
+
|
|
59
|
+
super().__init__(
|
|
60
|
+
unk_token=unk_token,
|
|
61
|
+
sep_token=sep_token,
|
|
62
|
+
pad_token=pad_token,
|
|
63
|
+
cls_token=cls_token,
|
|
64
|
+
mask_token=mask_token,
|
|
65
|
+
**kwargs,
|
|
66
|
+
)
|
|
53
67
|
self.init_kwargs["model_max_length"] = self.model_max_length
|
|
54
68
|
|
|
55
69
|
@property
|
|
56
|
-
def vocab_size(self):
|
|
70
|
+
def vocab_size(self) -> int:
|
|
57
71
|
return len(self.vocab)
|
|
58
72
|
|
|
59
73
|
@property
|
|
60
|
-
def vocab_list(self):
|
|
74
|
+
def vocab_list(self) -> List[str]:
|
|
61
75
|
return list(self.vocab.keys())
|
|
62
76
|
|
|
63
|
-
def
|
|
77
|
+
def get_vocab(self) -> Dict[str, int]:
|
|
78
|
+
return dict(self.vocab, **self.added_tokens_encoder)
|
|
79
|
+
|
|
80
|
+
def _tokenize(self, text: str, **kwargs) -> List[str]: # type: ignore[override]
|
|
64
81
|
split_tokens = [token for token in self.basic_tokenizer.tokenize(text)]
|
|
65
82
|
return split_tokens
|
|
66
83
|
|
|
67
|
-
def _convert_token_to_id(self, token):
|
|
84
|
+
def _convert_token_to_id(self, token: str) -> int:
|
|
68
85
|
"""Converts a token (str/unicode) in an id using the vocab."""
|
|
69
86
|
return self.vocab.get(token, self.vocab.get(self.unk_token))
|
|
70
87
|
|
|
71
|
-
def _convert_id_to_token(self, index):
|
|
88
|
+
def _convert_id_to_token(self, index: int) -> str:
|
|
72
89
|
"""Converts an index (integer) in a token (string/unicode) using the vocab."""
|
|
73
90
|
return self.ids_to_tokens.get(index, self.unk_token)
|
|
74
91
|
|
|
92
|
+
def build_inputs_with_special_tokens(
|
|
93
|
+
self, token_ids_0: List[int], token_ids_1: Optional[List[int]] = None
|
|
94
|
+
) -> List[int]:
|
|
95
|
+
if token_ids_1 is None:
|
|
96
|
+
return [self.cls_token_id] + token_ids_0 + [self.sep_token_id]
|
|
97
|
+
cls = [self.cls_token_id]
|
|
98
|
+
sep = [self.sep_token_id]
|
|
99
|
+
return cls + token_ids_0 + sep + token_ids_1 + sep
|
|
100
|
+
|
|
101
|
+
def get_special_tokens_mask(
|
|
102
|
+
self,
|
|
103
|
+
token_ids_0: List[int],
|
|
104
|
+
token_ids_1: Optional[List[int]] = None,
|
|
105
|
+
already_has_special_tokens: bool = False,
|
|
106
|
+
) -> List[int]:
|
|
107
|
+
if already_has_special_tokens:
|
|
108
|
+
return super().get_special_tokens_mask(
|
|
109
|
+
token_ids_0=token_ids_0,
|
|
110
|
+
token_ids_1=token_ids_1,
|
|
111
|
+
already_has_special_tokens=True,
|
|
112
|
+
)
|
|
113
|
+
if token_ids_1 is not None:
|
|
114
|
+
return [1] + ([0] * len(token_ids_0)) + [1] + ([0] * len(token_ids_1)) + [1]
|
|
115
|
+
return [1] + ([0] * len(token_ids_0)) + [1]
|
|
116
|
+
|
|
117
|
+
def create_token_type_ids_from_sequences(
|
|
118
|
+
self, token_ids_0: List[int], token_ids_1: Optional[List[int]] = None
|
|
119
|
+
) -> List[int]:
|
|
120
|
+
sep = [self.sep_token_id]
|
|
121
|
+
cls = [self.cls_token_id]
|
|
122
|
+
if token_ids_1 is None:
|
|
123
|
+
return len(cls + token_ids_0 + sep) * [0]
|
|
124
|
+
return len(cls + token_ids_0 + sep) * [0] + len(token_ids_1 + sep) * [1]
|
|
125
|
+
|
|
75
126
|
def convert_tokens_to_string(self, tokens):
|
|
76
127
|
"""Converts a sequence of tokens (string) in a single string."""
|
|
77
128
|
out_string = " ".join(tokens).replace(" ##", "").strip()
|
|
@@ -1,6 +1,6 @@
|
|
|
1
1
|
Metadata-Version: 2.4
|
|
2
2
|
Name: rxnmapper
|
|
3
|
-
Version: 0.
|
|
3
|
+
Version: 0.5.0
|
|
4
4
|
Summary: Reaction atom-mapping from transformers
|
|
5
5
|
Home-page: https://github.com/rxn4chemistry/rxnmapper
|
|
6
6
|
Author: IBM RXN team
|
|
@@ -9,16 +9,14 @@ License: MIT
|
|
|
9
9
|
Classifier: Operating System :: OS Independent
|
|
10
10
|
Classifier: Intended Audience :: Developers
|
|
11
11
|
Classifier: Intended Audience :: Science/Research
|
|
12
|
-
Classifier: License :: OSI Approved :: MIT License
|
|
13
12
|
Classifier: Programming Language :: Python :: 3
|
|
14
|
-
Classifier: Programming Language :: Python :: 3.6
|
|
15
|
-
Classifier: Programming Language :: Python :: 3.7
|
|
16
|
-
Classifier: Programming Language :: Python :: 3.8
|
|
17
|
-
Classifier: Programming Language :: Python :: 3.9
|
|
18
13
|
Classifier: Programming Language :: Python :: 3.10
|
|
19
14
|
Classifier: Programming Language :: Python :: 3.11
|
|
15
|
+
Classifier: Programming Language :: Python :: 3.12
|
|
16
|
+
Classifier: Programming Language :: Python :: 3.13
|
|
17
|
+
Classifier: Programming Language :: Python :: 3.14
|
|
20
18
|
Classifier: Topic :: Software Development :: Libraries :: Python Modules
|
|
21
|
-
Requires-Python: >=3.
|
|
19
|
+
Requires-Python: >=3.10
|
|
22
20
|
Description-Content-Type: text/markdown
|
|
23
21
|
License-File: LICENSE
|
|
24
22
|
Requires-Dist: pandas>=1.0.3
|
|
@@ -38,8 +36,7 @@ Requires-Dist: sphinx==2.2.1; extra == "dev"
|
|
|
38
36
|
Requires-Dist: twine==3.1.1; extra == "dev"
|
|
39
37
|
Requires-Dist: types-setuptools>=62.6.0; extra == "dev"
|
|
40
38
|
Provides-Extra: rdkit
|
|
41
|
-
Requires-Dist: rdkit
|
|
42
|
-
Requires-Dist: rdkit>=2022.3.4; python_version >= "3.7" and extra == "rdkit"
|
|
39
|
+
Requires-Dist: rdkit>=2022.3.4; extra == "rdkit"
|
|
43
40
|
Dynamic: license-file
|
|
44
41
|
|
|
45
42
|
# Extraction of organic chemistry grammar from unsupervised learning of chemical reactions
|
|
@@ -51,37 +48,27 @@ Enable robust atom mapping on valid reaction SMILES. The atom-mapping informatio
|
|
|
51
48
|
|
|
52
49
|
## Installation
|
|
53
50
|
|
|
54
|
-
###
|
|
55
|
-
|
|
56
|
-
|
|
57
|
-
|
|
58
|
-
|
|
51
|
+
### Create virtual environment (optional)
|
|
52
|
+
|
|
53
|
+
```bash
|
|
54
|
+
python3 -m venv .venv
|
|
55
|
+
source .venv/bin/activate
|
|
59
56
|
```
|
|
60
57
|
|
|
61
|
-
###
|
|
62
|
-
You can install the package and setup the environment directly from github using:
|
|
58
|
+
### Install from pip
|
|
63
59
|
|
|
64
|
-
```
|
|
65
|
-
|
|
66
|
-
cd rxnmapper
|
|
67
|
-
conda create -n rxnmapper python=3.6 -y
|
|
68
|
-
conda activate rxnmapper
|
|
69
|
-
pip install -e .
|
|
60
|
+
```bash
|
|
61
|
+
pip install "rxnmapper[rdkit]"
|
|
70
62
|
```
|
|
71
63
|
|
|
72
|
-
|
|
64
|
+
You can leave out `[rdkit]` if RDKit is already available in your Python environment.
|
|
73
65
|
|
|
74
|
-
|
|
75
|
-
It can also be installed via Conda or Pypi:
|
|
66
|
+
### From source
|
|
76
67
|
|
|
77
68
|
```bash
|
|
78
|
-
|
|
79
|
-
|
|
80
|
-
|
|
81
|
-
# Install RDKit from Pypi
|
|
82
|
-
pip install rdkit
|
|
83
|
-
# for Python<3.7
|
|
84
|
-
# pip install rdkit-pypi
|
|
69
|
+
git clone https://github.com/rxn4chemistry/rxnmapper.git
|
|
70
|
+
cd rxnmapper
|
|
71
|
+
pip install -e ".[rdkit]"
|
|
85
72
|
```
|
|
86
73
|
|
|
87
74
|
## Usage
|
|
@@ -118,10 +105,12 @@ results = list(rxn_mapper.map_reactions_with_info(rxns)) # results as dictionar
|
|
|
118
105
|
|
|
119
106
|
### Testing
|
|
120
107
|
|
|
121
|
-
You can run the
|
|
108
|
+
You can run the test suite with:
|
|
122
109
|
|
|
123
|
-
|
|
124
|
-
|
|
110
|
+
```bash
|
|
111
|
+
pip install -e .[dev,rdkit]
|
|
112
|
+
pytest tests
|
|
113
|
+
```
|
|
125
114
|
|
|
126
115
|
## Examples
|
|
127
116
|
|
|
@@ -11,20 +11,18 @@ classifiers =
|
|
|
11
11
|
Operating System :: OS Independent
|
|
12
12
|
Intended Audience :: Developers
|
|
13
13
|
Intended Audience :: Science/Research
|
|
14
|
-
License :: OSI Approved :: MIT License
|
|
15
14
|
Programming Language :: Python :: 3
|
|
16
|
-
Programming Language :: Python :: 3.6
|
|
17
|
-
Programming Language :: Python :: 3.7
|
|
18
|
-
Programming Language :: Python :: 3.8
|
|
19
|
-
Programming Language :: Python :: 3.9
|
|
20
15
|
Programming Language :: Python :: 3.10
|
|
21
16
|
Programming Language :: Python :: 3.11
|
|
17
|
+
Programming Language :: Python :: 3.12
|
|
18
|
+
Programming Language :: Python :: 3.13
|
|
19
|
+
Programming Language :: Python :: 3.14
|
|
22
20
|
Topic :: Software Development :: Libraries :: Python Modules
|
|
23
21
|
url = https://github.com/rxn4chemistry/rxnmapper
|
|
24
22
|
|
|
25
23
|
[options]
|
|
26
24
|
packages = find:
|
|
27
|
-
python_requires = >= 3.
|
|
25
|
+
python_requires = >= 3.10
|
|
28
26
|
zip_safe = False
|
|
29
27
|
include_package_data = True
|
|
30
28
|
install_requires =
|
|
@@ -52,8 +50,7 @@ dev =
|
|
|
52
50
|
twine==3.1.1
|
|
53
51
|
types-setuptools>=62.6.0
|
|
54
52
|
rdkit =
|
|
55
|
-
rdkit
|
|
56
|
-
rdkit>=2022.3.4 ; python_version>="3.7"
|
|
53
|
+
rdkit>=2022.3.4
|
|
57
54
|
|
|
58
55
|
[flake8]
|
|
59
56
|
extend-ignore = E203, E501
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|