Datasets:
haneulpark
commited on
Commit
•
0613fc4
1
Parent(s):
26b5a5e
Update Preprocessing Script.py
Browse files- Preprocessing Script.py +7 -5
Preprocessing Script.py
CHANGED
@@ -75,6 +75,8 @@ from scipy.cluster import hierarchy
|
|
75 |
|
76 |
#7. Split the dataset into test and train
|
77 |
|
|
|
|
|
78 |
class MolecularFingerprint:
|
79 |
def __init__(self, fingerprint):
|
80 |
self.fingerprint = fingerprint
|
@@ -140,7 +142,7 @@ def butina_cluster(fingerprints, num_points, distance_threshold, reordering=Fals
|
|
140 |
|
141 |
def hierarchal_cluster(fingerprints):
|
142 |
|
143 |
-
|
144 |
|
145 |
av_cluster_size = 8
|
146 |
dists = []
|
@@ -175,11 +177,9 @@ def cluster_fingerprints(fingerprints, method="Auto"):
|
|
175 |
print("Butina clustering is selected. Dataset size is:", num_fingerprints)
|
176 |
clusters = butina_cluster(fingerprints, num_fingerprints, cutoff)
|
177 |
|
178 |
-
return clusters
|
179 |
-
|
180 |
elif method == "Hierarchy":
|
181 |
print("Hierarchical clustering is selected. Dataset size is:", num_fingerprints)
|
182 |
-
clusters = hierarchal_cluster(fingerprints
|
183 |
|
184 |
return clusters
|
185 |
|
@@ -255,11 +255,13 @@ def split_df_into_train_and_test_sets(df):
|
|
255 |
test = df[df['group'] == 'testing']
|
256 |
return train, test
|
257 |
|
|
|
|
|
258 |
smiles_index = 0 # Because smiles is in the first column
|
259 |
realistic = realistic_split(newAA.copy(), smiles_index, 0.8, split_for_exact_frac=True, cluster_method="Auto")
|
260 |
realistic_train, realistic_test = split_df_into_train_and_test_sets(realistic)
|
261 |
|
262 |
-
#
|
263 |
|
264 |
selected_columns = realistic_train[['new SMILES', 'substance_id', 'aggref_index', 'logP', 'reference']]
|
265 |
selected_columns.to_csv("AggregatorAdvisor_train.csv", index=False)
|
|
|
75 |
|
76 |
#7. Split the dataset into test and train
|
77 |
|
78 |
+
#7. Split the dataset into test and train
|
79 |
+
|
80 |
class MolecularFingerprint:
|
81 |
def __init__(self, fingerprint):
|
82 |
self.fingerprint = fingerprint
|
|
|
142 |
|
143 |
def hierarchal_cluster(fingerprints):
|
144 |
|
145 |
+
num_fingerprints = len(fingerprints)
|
146 |
|
147 |
av_cluster_size = 8
|
148 |
dists = []
|
|
|
177 |
print("Butina clustering is selected. Dataset size is:", num_fingerprints)
|
178 |
clusters = butina_cluster(fingerprints, num_fingerprints, cutoff)
|
179 |
|
|
|
|
|
180 |
elif method == "Hierarchy":
|
181 |
print("Hierarchical clustering is selected. Dataset size is:", num_fingerprints)
|
182 |
+
clusters = hierarchal_cluster(fingerprints)
|
183 |
|
184 |
return clusters
|
185 |
|
|
|
255 |
test = df[df['group'] == 'testing']
|
256 |
return train, test
|
257 |
|
258 |
+
# 8. Test and train datasets have been made
|
259 |
+
|
260 |
smiles_index = 0 # Because smiles is in the first column
|
261 |
realistic = realistic_split(newAA.copy(), smiles_index, 0.8, split_for_exact_frac=True, cluster_method="Auto")
|
262 |
realistic_train, realistic_test = split_df_into_train_and_test_sets(realistic)
|
263 |
|
264 |
+
#9. Select columns and name the datasets
|
265 |
|
266 |
selected_columns = realistic_train[['new SMILES', 'substance_id', 'aggref_index', 'logP', 'reference']]
|
267 |
selected_columns.to_csv("AggregatorAdvisor_train.csv", index=False)
|