-
Notifications
You must be signed in to change notification settings - Fork 31
/
Copy pathMatBenchDataset2020.py
190 lines (165 loc) · 11 KB
/
MatBenchDataset2020.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
import os
import pandas as pd
from kgcnn.data.crystal import CrystalDataset
from kgcnn.data.download import DownloadDataset
from kgcnn.data.utils import load_json_file, save_json_file
class MatBenchDataset2020(CrystalDataset, DownloadDataset):
r"""Base class for loading graph datasets from `MatBench <https://matbench.materialsproject.org/>`__ , collection
of materials datasets. For graph learning only those with structure are relevant.
Process and loads from serialized :obj:`pymatgen` structures.
.. note::
This class does not follow the interface of `MatBench <https://github.com/materialsproject/matbench>`__
and therefore also not the original splits required for submission of benchmark values.
Matbench is an automated leaderboard for benchmarking state of the art ML algorithms predicting a diverse range
of solid materials' properties. It is hosted and maintained by the
`Materials Project <https://materialsproject.org/>`_ .
`Matbench <https://www.nature.com/articles/s41524-020-00406-3>`__ is an `ImageNet <https://image-net.org/>`__
for materials science; a curated set of 13 supervised, pre-cleaned, ready-to-use ML tasks for benchmarking
and fair comparison.
The tasks span a wide domain of inorganic materials science applications including electronic, thermodynamic,
mechanical, and thermal properties among crystals, 2D materials, disordered metals, and more.
References:
(1) Dunn, A., Wang, Q., Ganose, A. et al. Benchmarking materials property prediction methods: the Matbench
test set and Automatminer reference algorithm. npj Comput Mater 6, 138 (2020).
`<https://doi.org/10.1038/s41524-020-00406-3>`_ .
"""
datasets_download_info = {
"matbench_steels": {"dataset_name": "matbench_steels",
"download_file_name": 'matbench_steels.json.gz',
"data_directory_name": "matbench_steels", "extract_gz": True,
"extract_file_name": 'matbench_steels.json'},
"matbench_jdft2d": {"dataset_name": "matbench_jdft2d",
"download_file_name": 'matbench_jdft2d.json.gz',
"data_directory_name": "matbench_jdft2d", "extract_gz": True,
"extract_file_name": 'matbench_jdft2d.json'},
"matbench_phonons": {"dataset_name": "matbench_phonons",
"download_file_name": 'matbench_phonons.json.gz',
"data_directory_name": "matbench_phonons", "extract_gz": True,
"extract_file_name": 'matbench_phonons.json'},
"matbench_expt_gap": {"dataset_name": "matbench_expt_gap",
"download_file_name": 'matbench_expt_gap.json.gz',
"data_directory_name": "matbench_expt_gap", "extract_gz": True,
"extract_file_name": 'matbench_expt_gap.json'},
"matbench_dielectric": {"dataset_name": "matbench_dielectric",
"download_file_name": 'matbench_dielectric.json.gz',
"data_directory_name": "matbench_dielectric", "extract_gz": True,
"extract_file_name": 'matbench_dielectric.json'},
"matbench_expt_is_metal": {"dataset_name": "matbench_expt_is_metal",
"download_file_name": 'matbench_expt_is_metal.json.gz',
"data_directory_name": "matbench_expt_is_metal", "extract_gz": True,
"extract_file_name": 'matbench_expt_is_metal.json'},
"matbench_glass": {"dataset_name": "matbench_glass",
"download_file_name": 'matbench_glass.json.gz',
"data_directory_name": "matbench_glass", "extract_gz": True,
"extract_file_name": 'matbench_glass.json'},
"matbench_log_gvrh": {"dataset_name": "matbench_log_gvrh",
"download_file_name": 'matbench_log_gvrh.json.gz',
"data_directory_name": "matbench_log_gvrh", "extract_gz": True,
"extract_file_name": 'matbench_log_gvrh.json'},
"matbench_log_kvrh": {"dataset_name": "matbench_log_kvrh",
"download_file_name": 'matbench_log_kvrh.json.gz',
"data_directory_name": "matbench_log_kvrh", "extract_gz": True,
"extract_file_name": 'matbench_log_kvrh.json'},
"matbench_perovskites": {"dataset_name": "matbench_perovskites",
"download_file_name": 'matbench_perovskites.json.gz',
"data_directory_name": "matbench_perovskites", "extract_gz": True,
"extract_file_name": 'matbench_perovskites.json'},
"matbench_mp_gap": {"dataset_name": "matbench_mp_gap",
"download_file_name": 'matbench_mp_gap.json.gz',
"data_directory_name": "matbench_mp_gap", "extract_gz": True,
"extract_file_name": 'matbench_mp_gap.json'},
"matbench_mp_is_metal": {"dataset_name": "matbench_mp_is_metal",
"download_file_name": 'matbench_mp_is_metal.json.gz',
"data_directory_name": "matbench_mp_is_metal", "extract_gz": True,
"extract_file_name": 'matbench_mp_is_metal.json'},
"matbench_mp_e_form": {"dataset_name": "matbench_mp_e_form",
"download_file_name": 'matbench_mp_e_form.json.gz',
"data_directory_name": "matbench_mp_e_form", "extract_gz": True,
"extract_file_name": 'matbench_mp_e_form.json'},
}
datasets_prepare_data_info = {
"matbench_steels": {"file_column_name": "composition"},
"matbench_jdft2d": {"file_column_name": "structure"},
"matbench_phonons": {"file_column_name": "structure"},
"matbench_expt_gap": {"file_column_name": "composition"},
"matbench_dielectric": {"file_column_name": "structure"},
"matbench_expt_is_metal": {"file_column_name": "composition"},
"matbench_glass": {"file_column_name": "composition"},
"matbench_log_gvrh": {"file_column_name": "structure"},
"matbench_log_kvrh": {"file_column_name": "structure"},
"matbench_perovskites": {"file_column_name": "structure"},
"matbench_mp_gap": {"file_column_name": "structure"},
"matbench_mp_is_metal": {"file_column_name": "structure"},
"matbench_mp_e_form": {"file_column_name": "structure"},
}
datasets_read_in_memory_info = {
"matbench_steels": {"label_column_name": "yield strength"},
"matbench_jdft2d": {"label_column_name": "exfoliation_en"},
"matbench_phonons": {"label_column_name": "last phdos peak"},
"matbench_expt_gap": {"label_column_name": "gap expt"},
"matbench_dielectric": {"label_column_name": "n"},
"matbench_expt_is_metal": {"label_column_name": "is_metal"},
"matbench_glass": {"label_column_name": "gfa"},
"matbench_log_gvrh": {"label_column_name": "log10(G_VRH)"},
"matbench_log_kvrh": {"label_column_name": "log10(K_VRH)"},
"matbench_perovskites": {"label_column_name": "e_form"},
"matbench_mp_gap": {"label_column_name": "gap pbe"},
"matbench_mp_is_metal": {"label_column_name": "is_metal"},
"matbench_mp_e_form": {"label_column_name": "e_form"},
}
def __init__(self, dataset_name: str, reload: bool = False, verbose: int = 10):
"""Initialize a `GraphTUDataset` instance from string identifier.
Args:
dataset_name (str): Name of a dataset.
reload (bool): Download the dataset again and prepare data on disk.
verbose (int): Print progress or info for processing where 60=silent. Default is 10.
"""
if not isinstance(dataset_name, str):
raise ValueError("Please provide string identifier for TUDataset.")
CrystalDataset.__init__(self, verbose=verbose, dataset_name=dataset_name)
# Prepare download
if dataset_name in self.datasets_download_info:
self.download_info = self.datasets_download_info[dataset_name]
self.download_info.update({"download_url": "https://ml.materialsproject.org/projects/" +
self.download_info["download_file_name"]})
else:
raise ValueError(
"Can not resolve %s as a MatBench dataset. Pick " % dataset_name, self.datasets_download_info.keys(),
"For new dataset, add to `datasets_download_info` list manually.")
DownloadDataset.__init__(self, **self.download_info, reload=reload, verbose=verbose)
self.data_directory = os.path.join(self.data_main_dir, self.data_directory_name)
file_name_download = self.download_file_name if self.extract_file_name is None else self.extract_file_name
self.file_name = "%s.csv" % os.path.splitext(file_name_download)[0]
self.dataset_name = dataset_name
self.require_prepare_data = True
self.fits_in_memory = True
if self.require_prepare_data:
self.prepare_data(overwrite=reload, **self.datasets_prepare_data_info[self.dataset_name])
if self.fits_in_memory:
self.read_in_memory(**self.datasets_read_in_memory_info[self.dataset_name])
def prepare_data(self, file_column_name: str = None, overwrite: bool = False):
file_name_download = self.download_file_name if self.extract_file_name is None else self.extract_file_name
# file_name_base = os.path.splitext(self.file_name)[0]
if all([os.path.exists(self.pymatgen_json_file_path), os.path.exists(self.file_path), not overwrite]):
self.info("Found '%s' of structures." % self.pymatgen_json_file_path)
return self
self.info("Load dataset '%s' to memory..." % self.dataset_name)
data = load_json_file(os.path.join(self.data_directory, file_name_download))
self.info("Process database with %s and columns %s" % (data.keys(), data["columns"]))
data_columns = data["columns"]
index_structure = 0
for i, col in enumerate(data_columns):
if col == file_column_name:
index_structure = i
break
py_mat_list = [x[index_structure] for x in data["data"]]
self.info("Write structures or compositions '%s' to file." % self.pymatgen_json_file_path)
save_json_file(py_mat_list, self.pymatgen_json_file_path)
df_dict = {"index": data["index"]}
for i, col in enumerate(data_columns):
if i != index_structure:
df_dict[col] = [x[i] for x in data["data"]]
df = pd.DataFrame(df_dict)
self.info("Write dataset table '%s' to file." % self.file_name)
df.to_csv(self.file_path)
return self