-
Notifications
You must be signed in to change notification settings - Fork 5
/
test_scoring.py
177 lines (160 loc) · 8.31 KB
/
test_scoring.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
from pathlib import Path
from numpy.core.fromnumeric import mean
from numpy.lib.shape_base import _put_along_axis_dispatcher
from numpy.lib.ufunclike import _fix_and_maybe_deprecate_out_named_y
from pandas.io.pytables import AppendableMultiSeriesTable
from sklearn import metrics
import pandas as pd
import numpy as np
from github import Github
import os
from math import floor, log10
def main():
ak = os.environ.get("SOMEVAR", None)
g = Github(ak)
pr_id = os.environ.get("TRAVIS_PULL_REQUEST", None)
if pr_id == "false" or pr_id == None:
pr_id = 1
repo = g.get_repo(350697371)
pr = repo.get_pull(int(pr_id))
issue_str = ""
points_total = 0
if Path("task_1_predictions.csv").exists():
df = pd.read_csv("hidden_test_data/test_crystals.csv")
y_true = np.array(df["calculated_density"])
y_pred_df = pd.read_csv("task_1_predictions.csv", header=None)
y_pred = [j for i in y_pred_df.to_numpy() for j in i]
mae = np.around(metrics.mean_absolute_error(y_true, y_pred), 3)
r2 = np.around(metrics.r2_score(y_true, y_pred), 3)
nMAE = np.around(calc_nMAE(y_true, y_pred), 3)
npoints = points(1 - calc_nMAE(y_true, y_pred))
points_total += npoints
issue_str += "Task 1 Prediction - Density\n-----------------\n"
issue_str += f"Mean Absolute Error: {mae}\n"
issue_str += f"R<sup>2</sup>: {r2}\n"
issue_str += f"Normalised Mean Absolute Error (Assessed): {nMAE}\n"
issue_str += f"__Points: {npoints}__\n\n"
else:
issue_str += "Task 1 Prediction - Density\n-----------------\n"
issue_str += "No results submitted for task 1\n\n"
if Path("task_2_predictions.csv").exists():
df = pd.read_csv("hidden_test_data/test_crystals.csv")
y_true = np.array(df["is_centrosymmetric"])
y_pred_df = pd.read_csv("task_2_predictions.csv", header=None)
y_pred = [j for i in y_pred_df.to_numpy() for j in i]
acc = np.around(metrics.accuracy_score(y_true, y_pred), 3)
f1mac = np.around(metrics.f1_score(y_true, y_pred, average="macro"), 3)
f1wei = np.around(metrics.f1_score(y_true, y_pred, average="weighted"), 3)
npoints = points(metrics.f1_score(y_true, y_pred, average="macro"))
points_total += npoints
issue_str += "Task 2 Prediction - Centrosymmetric\n-----------------\n"
issue_str += f"Accuracy: {acc}\n"
issue_str += f"Macro F1-score (Assessed): {f1mac}\n"
issue_str += f"Weighted F1-score: {f1wei}\n"
issue_str += f"__Points: {npoints}__\n\n"
else:
issue_str += "Task 2 Prediction - Centrosymmetric\n-----------------\n"
issue_str += "No results submitted for task 2\n\n"
if Path("task_3_predictions.csv").exists():
df = pd.read_csv("hidden_test_data/test_centroid_distances.csv")
y_true = np.array(df["mean"])
y_pred_df = pd.read_csv("task_3_predictions.csv", header=None)
y_pred = [j for i in y_pred_df.to_numpy() for j in i]
mae = np.around(metrics.mean_absolute_error(y_true, y_pred), 3)
r2 = np.around(metrics.r2_score(y_true, y_pred), 3)
nMAE = np.around(calc_nMAE(y_true, y_pred), 3)
npoints = points(1 - calc_nMAE(y_true, y_pred))
points_total += npoints
issue_str += "Task 3 Prediction - Mean Distances\n-----------------\n"
issue_str += f"Mean Absolute Error: {mae}\n"
issue_str += f"R<sup>2</sup>: {r2}\n"
issue_str += f"Normalised Mean Absolute Error (Assessed): {nMAE}\n"
issue_str += f"__Points: {npoints}__\n\n"
else:
issue_str += "Task 3 Prediction - Mean Distances\n-----------------\n"
issue_str += "No results submitted for task 3\n\n"
if Path("task_4_predictions.csv").exists():
df = pd.read_csv("hidden_test_data/test_distances.csv")
y_true = np.array(df["n_vdw_contacts"])
y_pred_df = pd.read_csv("task_4_predictions.csv", header=None)
y_pred = [j for i in y_pred_df.to_numpy() for j in i]
mae = np.around(metrics.mean_absolute_error(y_true, y_pred), 3)
r2 = np.around(metrics.r2_score(y_true, y_pred), 3)
nMAE = np.around(calc_nMAE(y_true, y_pred), 3)
npoints = points(1 - calc_nMAE(y_true, y_pred))
points_total += npoints
issue_str += "Task 4 Prediction - Van der Waals Contacts\n-----------------\n"
issue_str += f"Mean Absolute Error: {mae}\n"
issue_str += f"R<sup>2</sup>: {r2}\n"
issue_str += f"Normalised Mean Absolute Error (Assessed): {nMAE}\n"
issue_str += f"__Points: {npoints}__\n\n"
else:
issue_str += "Task 4 Prediction - Van der Waals Contacts\n-----------------\n"
issue_str += "No results submitted for task 4\n\n"
if Path("bonus_1_predictions.csv").exists():
df = pd.read_csv("hidden_test_data/test_crystals.csv")
y_true = np.array(df["packing_coefficient"])
y_pred_df = pd.read_csv("bonus_1_predictions.csv", header=None)
y_pred = [j for i in y_pred_df.to_numpy() for j in i]
mae = np.around(metrics.mean_absolute_error(y_true, y_pred), 3)
r2 = np.around(metrics.r2_score(y_true, y_pred), 3)
nMAE = np.around(calc_nMAE(y_true, y_pred), 3)
npoints = points(1 - calc_nMAE(y_true, y_pred), max_points=10)
points_total += npoints
issue_str += "Bonus Task 1 Prediction - Packing Coefficient\n-----------------\n"
issue_str += f"Mean Absolute Error: {mae}\n"
issue_str += f"R<sup>2</sup>: {r2}\n"
issue_str += f"Normalised Mean Absolute Error (Assessed): {nMAE}\n"
issue_str += f"__Points: {npoints}__\n\n"
else:
issue_str += "Bonus Task 1 Prediction - Packing Coefficient\n-----------------\n"
issue_str += "No results submitted for bonus task 1\n\n"
if Path("bonus_2_predictions.csv").exists():
df = pd.read_csv("hidden_test_data/test_crystals.csv")
y_true = np.array(df["cell_volume"])
y_pred_df = pd.read_csv("bonus_2_predictions.csv", header=None)
y_pred = [j for i in y_pred_df.to_numpy() for j in i]
mae = np.around(metrics.mean_absolute_error(y_true, y_pred), 3)
r2 = np.around(metrics.r2_score(y_true, y_pred), 3)
nMAE = np.around(calc_nMAE(y_true, y_pred), 3)
npoints = points(1 - calc_nMAE(y_true, y_pred), max_points=10)
points_total += npoints
issue_str += "Bonus Task 2 Prediction - Cell Volume\n-----------------\n"
issue_str += f"Mean Absolute Error: {mae}\n"
issue_str += f"R<sup>2</sup>: {r2}\n"
issue_str += f"Normalised Mean Absolute Error (Assessed): {nMAE}\n"
issue_str += f"__Points: {npoints}__\n\n"
else:
issue_str += "Bonus Task 2 Prediction - Cell Volume\n-----------------\n"
issue_str += "No results submitted for bonus task 2\n\n"
if Path("bonus_3_predictions.csv").exists():
df = pd.read_csv("hidden_test_data/test_crystals.csv")
y_true = np.array(df["spacegroup_symbol"])
y_pred_df = pd.read_csv("bonus_3_predictions.csv", header=None)
y_pred = [j for i in y_pred_df.to_numpy() for j in i]
acc = np.around(metrics.accuracy_score(y_true, y_pred), 3)
f1mac = np.around(metrics.f1_score(y_true, y_pred, average="macro"), 3)
f1wei = np.around(metrics.f1_score(y_true, y_pred, average="weighted"), 3)
npoints = points(metrics.f1_score(y_true, y_pred, average="macro"), max_points=10)
points_total += npoints
issue_str += "Bonus Task 3 Prediction - Space Group Symbol\n-----------------\n"
issue_str += f"Accuracy: {acc}\n"
issue_str += f"Macro F1-score (Assessed): {f1mac}\n"
issue_str += f"Weighted F1-score: {f1wei}\n"
issue_str += f"__Points: {npoints}__\n\n"
else:
issue_str += "Bonus Task 3 Prediction - Space Group Symbol\n-----------------\n"
issue_str += "No results submitted for bonus task 3\n\n"
issue_str += "__Total Points__\n-----------------\n"
issue_str += f"{points_total}/130\n\n"
pr.create_issue_comment(issue_str)
def calc_nMAE(true, pred):
return sum(abs(true - pred)) / sum(abs(true))
def points(score, alpha=1.2, max_points=25):
return np.minimum(
(max_points + 2) ** (np.maximum(score, np.finfo(float).eps) ** alpha)
- 1,
max_points,
).astype(int)
if __name__ == "__main__":
main()