|
1
|
|
|
#! /usr/bin/env python |
|
|
|
|
|
|
2
|
|
|
# |
|
3
|
|
|
# Copyright (C) 2007-2009 Rich Lewis <[email protected]> |
|
4
|
|
|
# License: 3-clause BSD |
|
5
|
|
|
|
|
6
|
|
|
""" |
|
7
|
|
|
# skchem.io.smiles |
|
8
|
|
|
|
|
9
|
|
|
Defining input and output operations for smiles files. |
|
10
|
|
|
""" |
|
11
|
|
|
|
|
12
|
|
|
import warnings |
|
13
|
|
|
from functools import wraps |
|
14
|
|
|
|
|
15
|
|
|
import pandas as pd |
|
|
|
|
|
|
16
|
|
|
|
|
17
|
|
|
from ..utils import Suppressor |
|
18
|
|
|
from ..core import Mol |
|
19
|
|
|
|
|
20
|
|
|
def read_smiles(smiles_file, smiles_column=0, name_column=None, delimiter='\t', |
|
|
|
|
|
|
21
|
|
|
title_line=False, error_bad_mol=False, warn_bad_mol=True, |
|
22
|
|
|
drop_bad_mol=True, *args, **kwargs): |
|
23
|
|
|
|
|
24
|
|
|
"""Read a smiles file into a pandas dataframe. |
|
25
|
|
|
|
|
26
|
|
|
The class wraps the pandas read_csv function. |
|
27
|
|
|
|
|
28
|
|
|
smiles_file (str, file-like): |
|
29
|
|
|
Location of data to load, specified as a string or passed directly as a |
|
30
|
|
|
file-like object. URLs may also be used, see the pandas.read_csv |
|
31
|
|
|
documentation. |
|
32
|
|
|
smiles_column (int): |
|
33
|
|
|
The column index at which SMILES are provided. |
|
34
|
|
|
Defaults to `0`. |
|
35
|
|
|
name_column (int): |
|
36
|
|
|
The column index at which compound names are provided, for use as the |
|
37
|
|
|
index in the dataframe. If None, use the default index. |
|
38
|
|
|
Defaults to `None`. |
|
39
|
|
|
delimiter (str): |
|
40
|
|
|
The delimiter used. |
|
41
|
|
|
Defaults to `\t`. |
|
42
|
|
|
title_line (bool): |
|
43
|
|
|
Whether a title line is provided, to use as column titles. |
|
44
|
|
|
Defaults to `False`. |
|
45
|
|
|
error_bad_mol (bool): |
|
46
|
|
|
Whether an error should be raised when a molecule fails to parse. |
|
47
|
|
|
Defaults to `False`. |
|
48
|
|
|
warn_bad_mol (bool): |
|
49
|
|
|
Whether a warning should be raised when a molecule fails to parse. |
|
50
|
|
|
Defaults to `True`. |
|
51
|
|
|
drop_bad_mol (bool): |
|
52
|
|
|
If true, drop any column with smiles that failed to parse. Otherwise, |
|
53
|
|
|
the field is None. Defaults to `True`. |
|
54
|
|
|
*args, **kwargs: |
|
55
|
|
|
Arguments will be passed to pandas read_csv arguments. |
|
56
|
|
|
|
|
57
|
|
|
Returns: |
|
58
|
|
|
pandas.DataFrame: |
|
59
|
|
|
The loaded data frame, with Mols supplied in the `structure` field. |
|
60
|
|
|
|
|
61
|
|
|
See Also: |
|
62
|
|
|
pandas.read_csv |
|
63
|
|
|
skchem.Mol.from_smiles |
|
64
|
|
|
skchem.io.sdf |
|
65
|
|
|
|
|
66
|
|
|
""" |
|
67
|
|
|
|
|
68
|
|
|
with Suppressor(): |
|
69
|
|
|
|
|
70
|
|
|
# set the header line to pass to the pandas parser |
|
71
|
|
|
# we accept True as being line zero, as is usual for smiles |
|
72
|
|
|
# if user specifies a header already, then do nothing |
|
73
|
|
|
|
|
74
|
|
|
header = kwargs.pop('header', None) |
|
75
|
|
|
if title_line is True: |
|
76
|
|
|
header = 0 |
|
77
|
|
|
elif header is not None: |
|
78
|
|
|
pass #remove from the kwargs to not pass it twice |
|
79
|
|
|
else: |
|
80
|
|
|
header = None |
|
81
|
|
|
|
|
82
|
|
|
# read the smiles file |
|
83
|
|
|
data = pd.read_csv(smiles_file, delimiter=delimiter, header=header, |
|
84
|
|
|
*args, **kwargs) |
|
85
|
|
|
|
|
86
|
|
|
# replace the smiles column with the structure column |
|
87
|
|
|
lst = list(data.columns) |
|
88
|
|
|
lst[smiles_column] = 'structure' |
|
89
|
|
|
data.columns = lst |
|
90
|
|
|
|
|
91
|
|
|
def parse(row): |
|
92
|
|
|
""" Parse smiles for row """ |
|
93
|
|
|
try: |
|
94
|
|
|
return Mol.from_smiles(row.structure) |
|
|
|
|
|
|
95
|
|
|
except ValueError: |
|
96
|
|
|
msg = 'Molecule {} could not be decoded.'.format(row.name) |
|
97
|
|
|
if error_bad_mol: |
|
98
|
|
|
raise ValueError(msg) |
|
99
|
|
|
elif warn_bad_mol: |
|
100
|
|
|
warnings.warn(msg) |
|
101
|
|
|
|
|
102
|
|
|
return None |
|
103
|
|
|
|
|
104
|
|
|
data['structure'] = data['structure'].apply(str) |
|
105
|
|
|
data['structure'] = data.apply(parse, axis=1) |
|
106
|
|
|
|
|
107
|
|
|
if drop_bad_mol: |
|
108
|
|
|
data = data[data['structure'].notnull()] |
|
109
|
|
|
|
|
110
|
|
|
# set index if passed |
|
111
|
|
|
if name_column is not None: |
|
112
|
|
|
data = data.set_index(data.columns[name_column]) |
|
113
|
|
|
|
|
114
|
|
|
return data |
|
115
|
|
|
|
|
116
|
|
|
@classmethod |
|
117
|
|
|
@wraps(read_smiles) |
|
118
|
|
|
def _from_smiles(_, *args, **kwargs): |
|
119
|
|
|
return read_smiles(*args, **kwargs) |
|
120
|
|
|
|
|
121
|
|
|
#set on pandas dataframe |
|
122
|
|
|
pd.DataFrame.from_smiles = _from_smiles |
|
123
|
|
|
|
richlewis42 /
scikit-chem
Loading history...
Cyclic imports may cause partly loaded modules to be returned. This might lead to unexpected runtime behavior which is hard to debug.