|
1
|
|
|
""" |
|
2
|
|
|
Command-line interface for mandos. |
|
3
|
|
|
""" |
|
4
|
|
|
|
|
5
|
|
|
from __future__ import annotations |
|
6
|
|
|
|
|
7
|
|
|
import enum |
|
8
|
|
|
import logging |
|
9
|
|
|
from pathlib import Path, PurePath |
|
10
|
|
|
from typing import Any, Mapping, Optional, Sequence |
|
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.chembl_api import ChemblApi |
|
20
|
|
|
from mandos.model import Search, Triple |
|
21
|
|
|
from mandos.model.caches import TaxonomyFactories |
|
22
|
|
|
from mandos.model.settings import DEFAULT_TAXONOMY_CACHE, Settings |
|
23
|
|
|
from mandos.search.chembl.activity_search import ActivitySearch |
|
24
|
|
|
from mandos.search.chembl.atc_search import AtcSearch |
|
25
|
|
|
from mandos.search.chembl.go_search import GoSearchFactory, GoType |
|
26
|
|
|
from mandos.search.chembl.indication_search import IndicationSearch |
|
27
|
|
|
from mandos.search.chembl.mechanism_search import MechanismSearch |
|
28
|
|
|
|
|
29
|
|
|
logger = logging.getLogger(__package__) |
|
30
|
|
|
cli = typer.Typer() |
|
31
|
|
|
|
|
32
|
|
|
|
|
33
|
|
|
class What(enum.Enum): |
|
34
|
|
|
""" |
|
35
|
|
|
List of search items. |
|
36
|
|
|
""" |
|
37
|
|
|
|
|
38
|
|
|
activity = enum.auto(), ActivitySearch |
|
39
|
|
|
mechanism = enum.auto(), MechanismSearch |
|
40
|
|
|
atc = enum.auto(), AtcSearch |
|
41
|
|
|
trial = enum.auto(), IndicationSearch |
|
42
|
|
|
go_proc_moa = enum.auto(), GoSearchFactory.create(GoType.process, MechanismSearch) |
|
43
|
|
|
go_fn_moa = enum.auto(), GoSearchFactory.create(GoType.function, MechanismSearch) |
|
44
|
|
|
go_comp_moa = enum.auto(), GoSearchFactory.create(GoType.component, MechanismSearch) |
|
45
|
|
|
go_proc_act = enum.auto(), GoSearchFactory.create(GoType.process, ActivitySearch) |
|
46
|
|
|
go_fn_act = enum.auto(), GoSearchFactory.create(GoType.function, ActivitySearch) |
|
47
|
|
|
go_comp_act = enum.auto(), GoSearchFactory.create(GoType.component, ActivitySearch) |
|
48
|
|
|
|
|
49
|
|
|
def __new__(cls, *args, **kwargs): |
|
|
|
|
|
|
50
|
|
|
obj = object.__new__(cls) |
|
51
|
|
|
obj._value_ = args[0] |
|
52
|
|
|
return obj |
|
53
|
|
|
|
|
54
|
|
|
# ignore the first param since it's already set by __new__ |
|
55
|
|
|
def __init__(self, _: str, clazz: Type[Search]): |
|
56
|
|
|
self._clazz_ = clazz |
|
57
|
|
|
|
|
58
|
|
|
@property |
|
59
|
|
|
def clazz(self) -> Type[Search]: |
|
|
|
|
|
|
60
|
|
|
return self._clazz_ |
|
61
|
|
|
|
|
62
|
|
|
|
|
63
|
|
|
class Commands: |
|
64
|
|
|
""" |
|
65
|
|
|
Entry points for mandos. |
|
66
|
|
|
""" |
|
67
|
|
|
|
|
68
|
|
|
@staticmethod |
|
69
|
|
|
@cli.command() |
|
70
|
|
|
def search( |
|
71
|
|
|
what: str, |
|
|
|
|
|
|
72
|
|
|
path: Path, |
|
|
|
|
|
|
73
|
|
|
config: Optional[Path] = None, |
|
|
|
|
|
|
74
|
|
|
) -> None: |
|
75
|
|
|
""" |
|
76
|
|
|
Process data. |
|
77
|
|
|
|
|
78
|
|
|
Args: |
|
79
|
|
|
what: Comma-separated list of ``activity``, ``mechanism``, ``atc``, and ``indication``. |
|
80
|
|
|
path: Path to the input file of one of the formats: |
|
81
|
|
|
- .txt containing one key (InChI / CHEMBL ID) per line |
|
82
|
|
|
- .csv/.tsv/.tab containing one key per row |
|
83
|
|
|
- .csv/.tsv/.tab of a symmetric affinity matrix, with a row header and column header with the keys |
|
|
|
|
|
|
84
|
|
|
config: Path to a TOML config file |
|
85
|
|
|
""" |
|
86
|
|
|
for w in what.split(","): |
|
|
|
|
|
|
87
|
|
|
w = What[w.lower()] |
|
|
|
|
|
|
88
|
|
|
df, triples = Commands.search_for(w, path, config=config) |
|
|
|
|
|
|
89
|
|
|
df_out = Path(str(path.with_suffix("")) + "-" + w.name.lower() + ".csv") |
|
90
|
|
|
df.to_csv(df_out) |
|
91
|
|
|
triples_out = df_out.with_suffix(".triples.txt") |
|
92
|
|
|
triples_out.write_text("\n".join([t.statement for t in triples]), encoding="utf8") |
|
93
|
|
|
triples_out.write_text( |
|
94
|
|
|
"\n".join([Triple.tab_header(), *[t.tabs for t in triples]]), encoding="utf8" |
|
95
|
|
|
) |
|
96
|
|
|
|
|
97
|
|
|
@staticmethod |
|
98
|
|
|
@cli.command(hidden=True) |
|
99
|
|
|
def process_tax( |
|
100
|
|
|
taxon: int, |
|
|
|
|
|
|
101
|
|
|
cache_path: Optional[Path] = None, |
|
|
|
|
|
|
102
|
|
|
) -> None: |
|
103
|
|
|
""" |
|
104
|
|
|
Preps a new taxonomy file for use in mandos. |
|
105
|
|
|
Just returns if a corresponding file already exists in the resources dir or mandos cache (``~/.mandos``). |
|
|
|
|
|
|
106
|
|
|
Otherwise, downloads a tab-separated file from UniProt. |
|
107
|
|
|
(To find manually, follow the ``All lower taxonomy nodes`` link and click ``Download``.) |
|
108
|
|
|
Then applies fixes and reduces the file size, creating a new file alongside. |
|
109
|
|
|
Puts both the raw data and fixed data in the cache under ``~/.mandos/taxonomy/``. |
|
110
|
|
|
|
|
111
|
|
|
Args: |
|
112
|
|
|
taxon: The **ID** of the UniProt taxon |
|
113
|
|
|
cache_path: |
|
114
|
|
|
""" |
|
115
|
|
|
if cache_path is None: |
|
116
|
|
|
cache_path = DEFAULT_TAXONOMY_CACHE |
|
117
|
|
|
TaxonomyFactories.from_uniprot(cache_path).load(taxon) |
|
118
|
|
|
|
|
119
|
|
|
@staticmethod |
|
120
|
|
|
def search_for( |
|
121
|
|
|
what: What, |
|
|
|
|
|
|
122
|
|
|
compounds: Union[Sequence[str], PurePath], |
|
|
|
|
|
|
123
|
|
|
config: Union[None, Mapping[str, Any], Path], |
|
|
|
|
|
|
124
|
|
|
) -> Tup[pd.DataFrame, Sequence[Triple]]: |
|
125
|
|
|
""" |
|
126
|
|
|
|
|
127
|
|
|
Args: |
|
128
|
|
|
what: |
|
129
|
|
|
compounds: |
|
130
|
|
|
config: |
|
131
|
|
|
|
|
132
|
|
|
Returns: |
|
133
|
|
|
|
|
134
|
|
|
""" |
|
135
|
|
|
if isinstance(compounds, (PurePath, str)): |
|
136
|
|
|
compounds = Path(compounds).read_text(encoding="utf8").splitlines() |
|
137
|
|
|
compounds = [c.strip() for c in compounds if len(c.strip()) > 0] |
|
138
|
|
|
if config is None: |
|
139
|
|
|
settings = Settings.load(NestedDotDict({})) |
|
140
|
|
|
elif isinstance(config, PurePath): |
|
141
|
|
|
settings = Settings.load(NestedDotDict.read_toml(config)) |
|
142
|
|
|
elif isinstance(config, NestedDotDict): |
|
143
|
|
|
settings = config |
|
144
|
|
|
else: |
|
145
|
|
|
settings = Settings.load(NestedDotDict(config)) |
|
146
|
|
|
settings.set() |
|
147
|
|
|
compounds = list(compounds) |
|
148
|
|
|
api = ChemblApi.wrap(Chembl) |
|
149
|
|
|
taxonomy = TaxonomyFactories.from_uniprot(settings.taxonomy_cache_path).load(settings.taxon) |
|
150
|
|
|
hits = what.clazz(api, settings, taxonomy).find_all(compounds) |
|
151
|
|
|
# collapse over and sort the triples |
|
152
|
|
|
triples = sorted(list({hit.to_triple() for hit in hits})) |
|
153
|
|
|
df = pd.DataFrame( |
|
|
|
|
|
|
154
|
|
|
[pd.Series({f: getattr(h, f) for f in what.clazz.hit_fields()}) for h in hits] |
|
155
|
|
|
) |
|
156
|
|
|
return df, triples |
|
157
|
|
|
|
|
158
|
|
|
|
|
159
|
|
|
if __name__ == "__main__": |
|
160
|
|
|
cli() |
|
161
|
|
|
|
|
162
|
|
|
|
|
163
|
|
|
__all__ = ["Commands", "What"] |
|
164
|
|
|
|
dmyersturnbull /
mandos
Loading history...