|
1
|
|
|
""" |
|
2
|
|
|
Command-line interface for mandos. |
|
3
|
|
|
""" |
|
4
|
|
|
|
|
5
|
|
|
from __future__ import annotations |
|
6
|
|
|
|
|
7
|
|
|
import inspect |
|
8
|
|
|
import logging |
|
9
|
|
|
from pathlib import Path, PurePath |
|
10
|
|
|
from typing import Sequence, Set, Optional, Mapping, Any |
|
|
|
|
|
|
11
|
|
|
from typing import Tuple as Tup |
|
12
|
|
|
from typing import Type, Union |
|
|
|
|
|
|
13
|
|
|
|
|
14
|
|
|
import pandas as pd |
|
|
|
|
|
|
15
|
|
|
import typer |
|
|
|
|
|
|
16
|
|
|
from chembl_webresource_client.new_client import new_client as Chembl |
|
|
|
|
|
|
17
|
|
|
from pocketutils.core.dot_dict import NestedDotDict |
|
|
|
|
|
|
18
|
|
|
|
|
19
|
|
|
from mandos.model.chembl_api import ChemblApi |
|
20
|
|
|
from mandos.model.chembl_support.chembl_targets import TargetType |
|
21
|
|
|
from mandos.model.hits import Triple |
|
22
|
|
|
from mandos.model.pubchem_api import ( |
|
23
|
|
|
CachingPubchemApi, |
|
24
|
|
|
QueryingPubchemApi, |
|
25
|
|
|
PubchemCompoundLookupError, |
|
26
|
|
|
) |
|
27
|
|
|
from mandos.model.searches import Search |
|
28
|
|
|
from mandos.model.taxonomy import Taxonomy |
|
29
|
|
|
from mandos.model.taxonomy_caches import TaxonomyFactories |
|
30
|
|
|
from mandos.model.settings import MANDOS_SETTINGS |
|
31
|
|
|
from mandos.search.chembl.binding_search import BindingSearch |
|
32
|
|
|
from mandos.search.chembl.atc_search import AtcSearch |
|
33
|
|
|
from mandos.search.chembl.go_search import GoType, GoSearch |
|
34
|
|
|
from mandos.search.chembl.indication_search import IndicationSearch |
|
35
|
|
|
from mandos.search.chembl.mechanism_search import MechanismSearch |
|
36
|
|
|
|
|
37
|
|
|
logger = logging.getLogger(__package__) |
|
38
|
|
|
cli = typer.Typer() |
|
39
|
|
|
|
|
40
|
|
|
|
|
41
|
|
|
class Utils: |
|
|
|
|
|
|
42
|
|
|
@staticmethod |
|
43
|
|
|
def split(st: str) -> Set[str]: |
|
|
|
|
|
|
44
|
|
|
return {s.strip() for s in st.split(",")} |
|
45
|
|
|
|
|
46
|
|
|
@staticmethod |
|
47
|
|
|
def get_taxon(taxon: int) -> Taxonomy: |
|
|
|
|
|
|
48
|
|
|
return TaxonomyFactories.from_uniprot(MANDOS_SETTINGS.taxonomy_cache_path).load(taxon) |
|
49
|
|
|
|
|
50
|
|
|
@staticmethod |
|
51
|
|
|
def get_target_types(st: str) -> Set[str]: |
|
|
|
|
|
|
52
|
|
|
st = st.strip() |
|
53
|
|
|
if st == "all": |
|
54
|
|
|
return {str(s) for s in TargetType.all_types()} |
|
55
|
|
|
if st == "known": |
|
56
|
|
|
return {str(s) for s in TargetType.all_types() if not s.is_unknown} |
|
57
|
|
|
if st == "protein": |
|
58
|
|
|
return {str(s) for s in TargetType.protein_types()} |
|
59
|
|
|
return Utils.split(st) |
|
60
|
|
|
|
|
61
|
|
|
@staticmethod |
|
62
|
|
|
def get_params() -> Mapping[str, Any]: |
|
|
|
|
|
|
63
|
|
|
frame = inspect.getouterframes(inspect.currentframe())[1][3] |
|
64
|
|
|
# assume there aren't any varargs or kwargs -- that would be strange in a CLI |
|
65
|
|
|
xx = inspect.getargvalues(frame) |
|
|
|
|
|
|
66
|
|
|
return {a: xx.locals[a] for a in xx.args} |
|
67
|
|
|
|
|
68
|
|
|
|
|
69
|
|
|
class Searcher: |
|
|
|
|
|
|
70
|
|
|
def __init__(self, search: Search): |
|
71
|
|
|
self.what = search |
|
72
|
|
|
self._params = Utils.get_params() |
|
73
|
|
|
|
|
74
|
|
|
def search( |
|
75
|
|
|
self, |
|
|
|
|
|
|
76
|
|
|
path: Path, |
|
|
|
|
|
|
77
|
|
|
) -> Tup[pd.DataFrame, Sequence[Triple]]: |
|
78
|
|
|
""" |
|
79
|
|
|
Performs the search, and writes data. |
|
80
|
|
|
|
|
81
|
|
|
Args: |
|
82
|
|
|
path: Path to the input file of one of the formats: |
|
83
|
|
|
- .txt containing one key (InChI / CHEMBL ID) per line |
|
84
|
|
|
- .csv/.tsv/.tab containing one key per row |
|
85
|
|
|
- .csv/.tsv/.tab of a symmetric affinity matrix, with a row header and column header with the keys |
|
|
|
|
|
|
86
|
|
|
|
|
87
|
|
|
Returns: |
|
88
|
|
|
|
|
89
|
|
|
""" |
|
90
|
|
|
df, triples = self.search_for(path) |
|
|
|
|
|
|
91
|
|
|
df_out = Path(str(path.with_suffix("")) + "-" + self.what.search_name.lower() + ".csv") |
|
92
|
|
|
df.to_csv(df_out) |
|
93
|
|
|
triples_out = df_out.with_suffix(".triples.txt") |
|
94
|
|
|
triples_out.write_text("\n".join([t.statement for t in triples]), encoding="utf8") |
|
95
|
|
|
triples_out.write_text( |
|
96
|
|
|
"\n".join([Triple.tab_header(), *[t.tabs for t in triples]]), encoding="utf8" |
|
97
|
|
|
) |
|
98
|
|
|
NestedDotDict(dict(args=self._params)).write_json(df_out.with_suffix(".metadata.json")) |
|
99
|
|
|
return df, triples |
|
100
|
|
|
|
|
101
|
|
|
def search_for( |
|
102
|
|
|
self, |
|
|
|
|
|
|
103
|
|
|
compounds: Union[Sequence[str], PurePath], |
|
|
|
|
|
|
104
|
|
|
) -> Tup[pd.DataFrame, Sequence[Triple]]: |
|
105
|
|
|
""" |
|
106
|
|
|
Performs the search. Does not write any files. |
|
107
|
|
|
|
|
108
|
|
|
Args: |
|
109
|
|
|
compounds: |
|
110
|
|
|
|
|
111
|
|
|
Returns: |
|
112
|
|
|
|
|
113
|
|
|
""" |
|
114
|
|
|
if isinstance(compounds, (PurePath, str)): |
|
115
|
|
|
compounds = Path(compounds).read_text(encoding="utf8").splitlines() |
|
116
|
|
|
compounds = [c.strip() for c in compounds if len(c.strip()) > 0] |
|
117
|
|
|
cache = CachingPubchemApi(MANDOS_SETTINGS.pubchem_cache_path, QueryingPubchemApi()) |
|
118
|
|
|
compounds = list(compounds) |
|
119
|
|
|
# TODO |
|
|
|
|
|
|
120
|
|
|
for compound in compounds: |
|
121
|
|
|
try: |
|
122
|
|
|
cache.fetch_data(compound) |
|
123
|
|
|
except PubchemCompoundLookupError: |
|
124
|
|
|
logger.error(f"Did not find compound {compound}") |
|
|
|
|
|
|
125
|
|
|
logger.debug(f"Did not find compound {compound}", exc_info=True) |
|
|
|
|
|
|
126
|
|
|
hits = self.what.find_all(compounds) |
|
127
|
|
|
# collapse over and sort the triples |
|
128
|
|
|
triples = sorted(list({hit.to_triple() for hit in hits})) |
|
129
|
|
|
df = pd.DataFrame( |
|
|
|
|
|
|
130
|
|
|
[pd.Series({f: getattr(h, f) for f in self.what.hit_fields()}) for h in hits] |
|
131
|
|
|
) |
|
132
|
|
|
return df, triples |
|
133
|
|
|
|
|
134
|
|
|
|
|
135
|
|
|
class Commands: |
|
136
|
|
|
""" |
|
137
|
|
|
Entry points for mandos. |
|
138
|
|
|
""" |
|
139
|
|
|
|
|
140
|
|
|
@staticmethod |
|
141
|
|
|
@cli.command("chembl:binding") |
|
142
|
|
|
def binding( |
|
|
|
|
|
|
143
|
|
|
path: Path, |
|
|
|
|
|
|
144
|
|
|
taxon: int = 7742, |
|
|
|
|
|
|
145
|
|
|
traversal_strategy: str = "strategy0", |
|
|
|
|
|
|
146
|
|
|
target_types: str = "single_protein,protein_family,protein_complex,protein_complex_group,selectivity_group", |
|
|
|
|
|
|
147
|
|
|
min_confidence: int = 3, |
|
|
|
|
|
|
148
|
|
|
relations: str = "<,<=,=", |
|
|
|
|
|
|
149
|
|
|
min_pchembl: float = 6.0, |
|
|
|
|
|
|
150
|
|
|
banned_flags: str = "potential missing data,potential transcription error,outside typical range", |
|
|
|
|
|
|
151
|
|
|
) -> Tup[pd.DataFrame, Sequence[Triple]]: |
|
152
|
|
|
""" |
|
153
|
|
|
Process data. |
|
154
|
|
|
""" |
|
155
|
|
|
api = ChemblApi.wrap(Chembl) |
|
156
|
|
|
search = BindingSearch( |
|
157
|
|
|
chembl_api=api, |
|
158
|
|
|
tax=Utils.get_taxon(taxon), |
|
159
|
|
|
traversal_strategy=traversal_strategy, |
|
160
|
|
|
allowed_target_types=Utils.get_target_types(target_types), |
|
161
|
|
|
min_confidence_score=min_confidence, |
|
162
|
|
|
allowed_relations=Utils.split(relations), |
|
163
|
|
|
min_pchembl=min_pchembl, |
|
164
|
|
|
banned_flags=Utils.split(banned_flags), |
|
165
|
|
|
) |
|
166
|
|
|
return Searcher(search).search(path) |
|
167
|
|
|
|
|
168
|
|
|
@staticmethod |
|
169
|
|
|
@cli.command("chembl:mechanism") |
|
170
|
|
|
def mechanism( |
|
171
|
|
|
path: Path, |
|
|
|
|
|
|
172
|
|
|
taxon: int = 7742, |
|
|
|
|
|
|
173
|
|
|
traversal_strategy: str = "strategy0", |
|
|
|
|
|
|
174
|
|
|
target_types: str = "single_protein,protein_family,protein_complex,protein_complex_group,selectivity_group", |
|
|
|
|
|
|
175
|
|
|
min_confidence: int = 3, |
|
|
|
|
|
|
176
|
|
|
) -> Tup[pd.DataFrame, Sequence[Triple]]: |
|
177
|
|
|
""" |
|
178
|
|
|
Process data. |
|
179
|
|
|
""" |
|
180
|
|
|
api = ChemblApi.wrap(Chembl) |
|
181
|
|
|
search = MechanismSearch( |
|
182
|
|
|
chembl_api=api, |
|
183
|
|
|
tax=Utils.get_taxon(taxon), |
|
184
|
|
|
traversal_strategy=traversal_strategy, |
|
185
|
|
|
allowed_target_types=Utils.get_target_types(target_types), |
|
186
|
|
|
min_confidence_score=min_confidence, |
|
187
|
|
|
) |
|
188
|
|
|
return Searcher(search).search(path) |
|
189
|
|
|
|
|
190
|
|
|
@staticmethod |
|
191
|
|
|
@cli.command("chembl:trials") |
|
192
|
|
|
def trials( |
|
193
|
|
|
path: Path, |
|
|
|
|
|
|
194
|
|
|
min_phase: int = 3, |
|
|
|
|
|
|
195
|
|
|
) -> Tup[pd.DataFrame, Sequence[Triple]]: |
|
196
|
|
|
""" |
|
197
|
|
|
Process data. |
|
198
|
|
|
""" |
|
199
|
|
|
api = ChemblApi.wrap(Chembl) |
|
200
|
|
|
search = IndicationSearch( |
|
201
|
|
|
chembl_api=api, |
|
202
|
|
|
min_phase=min_phase, |
|
203
|
|
|
) |
|
204
|
|
|
return Searcher(search).search(path) |
|
205
|
|
|
|
|
206
|
|
|
@staticmethod |
|
207
|
|
|
@cli.command("chembl:atc") |
|
208
|
|
|
def atc( |
|
209
|
|
|
path: Path, |
|
|
|
|
|
|
210
|
|
|
) -> Tup[pd.DataFrame, Sequence[Triple]]: |
|
211
|
|
|
""" |
|
212
|
|
|
Process data. |
|
213
|
|
|
""" |
|
214
|
|
|
api = ChemblApi.wrap(Chembl) |
|
215
|
|
|
search = AtcSearch( |
|
216
|
|
|
chembl_api=api, |
|
217
|
|
|
) |
|
218
|
|
|
return Searcher(search).search(path) |
|
219
|
|
|
|
|
220
|
|
|
@staticmethod |
|
221
|
|
|
@cli.command("chembl:go") |
|
222
|
|
|
def go_search( |
|
|
|
|
|
|
223
|
|
|
path: Path, |
|
|
|
|
|
|
224
|
|
|
kind: GoType, |
|
|
|
|
|
|
225
|
|
|
taxon: int = 7742, |
|
|
|
|
|
|
226
|
|
|
traversal_strategy: str = "strategy0", |
|
|
|
|
|
|
227
|
|
|
target_types: str = "single_protein,protein_family,protein_complex,protein_complex_group,selectivity_group", |
|
|
|
|
|
|
228
|
|
|
min_confidence: int = 3, |
|
|
|
|
|
|
229
|
|
|
relations: str = "<,<=,=", |
|
|
|
|
|
|
230
|
|
|
min_pchembl: float = 6.0, |
|
|
|
|
|
|
231
|
|
|
banned_flags: str = "potential missing data,potential transcription error,outside typical range", |
|
|
|
|
|
|
232
|
|
|
) -> Tup[pd.DataFrame, Sequence[Triple]]: |
|
233
|
|
|
""" |
|
234
|
|
|
Process data. |
|
235
|
|
|
""" |
|
236
|
|
|
api = ChemblApi.wrap(Chembl) |
|
237
|
|
|
binding_search = BindingSearch( |
|
238
|
|
|
chembl_api=api, |
|
239
|
|
|
tax=Utils.get_taxon(taxon), |
|
240
|
|
|
traversal_strategy=traversal_strategy, |
|
241
|
|
|
allowed_target_types=Utils.get_target_types(target_types), |
|
242
|
|
|
min_confidence_score=min_confidence, |
|
243
|
|
|
allowed_relations=Utils.split(relations), |
|
244
|
|
|
min_pchembl=min_pchembl, |
|
245
|
|
|
banned_flags=Utils.split(banned_flags), |
|
246
|
|
|
) |
|
247
|
|
|
search = GoSearch(api, kind, binding_search) |
|
248
|
|
|
return Searcher(search).search(path) |
|
249
|
|
|
|
|
250
|
|
|
@staticmethod |
|
251
|
|
|
@cli.command(hidden=True) |
|
252
|
|
|
def process_tax( |
|
253
|
|
|
taxon: int, |
|
|
|
|
|
|
254
|
|
|
) -> None: |
|
255
|
|
|
""" |
|
256
|
|
|
Preps a new taxonomy file for use in mandos. |
|
257
|
|
|
Just returns if a corresponding file already exists in the resources dir or mandos cache (``~/.mandos``). |
|
|
|
|
|
|
258
|
|
|
Otherwise, downloads a tab-separated file from UniProt. |
|
259
|
|
|
(To find manually, follow the ``All lower taxonomy nodes`` link and click ``Download``.) |
|
260
|
|
|
Then applies fixes and reduces the file size, creating a new file alongside. |
|
261
|
|
|
Puts both the raw data and fixed data in the cache under ``~/.mandos/taxonomy/``. |
|
262
|
|
|
|
|
263
|
|
|
Args: |
|
264
|
|
|
taxon: The **ID** of the UniProt taxon |
|
265
|
|
|
""" |
|
266
|
|
|
TaxonomyFactories.from_uniprot(MANDOS_SETTINGS.taxonomy_cache_path).load(taxon) |
|
267
|
|
|
|
|
268
|
|
|
|
|
269
|
|
|
if __name__ == "__main__": |
|
270
|
|
|
cli() |
|
271
|
|
|
|
|
272
|
|
|
|
|
273
|
|
|
__all__ = ["Commands", "Searcher"] |
|
274
|
|
|
|
dmyersturnbull /
mandos
Loading history...