Spaces:
Running
Running
Update src/bin/function_predictor.py
Browse files
src/bin/function_predictor.py
CHANGED
@@ -1,4 +1,7 @@
|
|
1 |
# -*- coding: utf-8 -*-
|
|
|
|
|
|
|
2 |
import pandas as pd
|
3 |
import numpy as np
|
4 |
from datetime import datetime
|
@@ -63,7 +66,8 @@ def MultiLabelSVC_cross_val_predict(representation_name, dataset, X, y, classifi
|
|
63 |
y_pred = cross_val_predict(clf, Xn, y, cv=kf)
|
64 |
|
65 |
if detailed_output:
|
66 |
-
|
|
|
67 |
pickle.dump(clf,file)
|
68 |
|
69 |
acc_cv = []
|
@@ -114,7 +118,7 @@ def MultiLabelSVC_cross_val_predict(representation_name, dataset, X, y, classifi
|
|
114 |
|
115 |
def ProtDescModel():
|
116 |
#desc_file = pd.read_csv(r"protein_representations\final\{0}_dim{1}.tsv".format(representation_name,desc_dim),sep="\t")
|
117 |
-
datasets = os.listdir(r"../data/auxilary_input/GO_datasets")
|
118 |
if dataset_type == "All_Data_Sets" and aspect_type == "All_Aspects":
|
119 |
filtered_datasets = datasets
|
120 |
elif dataset_type == "All_Data_Sets":
|
@@ -129,7 +133,7 @@ def ProtDescModel():
|
|
129 |
|
130 |
for dt in tqdm(filtered_datasets,total=len(filtered_datasets)):
|
131 |
print(r"Protein function prediction is started for the dataset: {0}".format(dt.split(".")[0]))
|
132 |
-
dt_file = pd.read_csv(r"../data/auxilary_input/GO_datasets/{0}".format(dt),sep="\t")
|
133 |
dt_merge = dt_file.merge(representation_dataframe,left_on="Protein_Id",right_on="Entry")
|
134 |
|
135 |
dt_X = dt_merge['Vector']
|
@@ -149,7 +153,7 @@ def ProtDescModel():
|
|
149 |
predictions = dt_merge.iloc[:,:6]
|
150 |
predictions["predicted_values"] = list(model[3].toarray())
|
151 |
if detailed_output:
|
152 |
-
predictions.to_csv(r"../results/Ontology_based_function_prediction_{1}_{0}_predictions.tsv".format(representation_name,dt.split(".")[0]),sep="\t",index=None)
|
153 |
|
154 |
return (cv_results, cv_mean_results,cv_std_results)
|
155 |
|
@@ -164,7 +168,7 @@ def pred_output():
|
|
164 |
for i in cv_result:
|
165 |
df_cv_result.loc[len(df_cv_result)] = i
|
166 |
if detailed_output:
|
167 |
-
df_cv_result.to_csv(r"../results/Ontology_based_function_prediction_5cv_{0}.tsv".format(representation_name),sep="\t",index=None)
|
168 |
|
169 |
cv_mean_result = model[1]
|
170 |
df_cv_mean_result = pd.DataFrame({"Model": pd.Series([], dtype='str') ,"Accuracy": pd.Series([], dtype='float'),"F1_Micro": pd.Series([], dtype='float'),\
|
@@ -178,7 +182,7 @@ def pred_output():
|
|
178 |
|
179 |
for j in cv_mean_result:
|
180 |
df_cv_mean_result.loc[len(df_cv_mean_result)] = j
|
181 |
-
df_cv_mean_result.to_csv(r"../results/Ontology_based_function_prediction_5cv_mean_{0}.tsv".format(representation_name),sep="\t",index=None)
|
182 |
|
183 |
#save std deviation of scores to file
|
184 |
cv_std_result = model[2]
|
@@ -193,7 +197,7 @@ def pred_output():
|
|
193 |
|
194 |
for k in cv_std_result:
|
195 |
df_cv_std_result.loc[len(df_cv_std_result)] = k
|
196 |
-
df_cv_std_result.to_csv(r"../results/Ontology_based_function_prediction_5cv_std_{0}.tsv".format(representation_name),sep="\t",index=None)
|
197 |
|
198 |
print(datetime.now())
|
199 |
|
|
|
1 |
# -*- coding: utf-8 -*-
|
2 |
+
import os
|
3 |
+
script_dir = os.path.dirname(os.path.abspath(__file__))
|
4 |
+
|
5 |
import pandas as pd
|
6 |
import numpy as np
|
7 |
from datetime import datetime
|
|
|
66 |
y_pred = cross_val_predict(clf, Xn, y, cv=kf)
|
67 |
|
68 |
if detailed_output:
|
69 |
+
ont_path = r"../results/Ontology_based_function_prediction_{1}_{0}_model.pkl".format(representation_name,dataset.split(".")[0])
|
70 |
+
with open(os.path.join(script_dir, ont_path),"wb") as file:
|
71 |
pickle.dump(clf,file)
|
72 |
|
73 |
acc_cv = []
|
|
|
118 |
|
119 |
def ProtDescModel():
|
120 |
#desc_file = pd.read_csv(r"protein_representations\final\{0}_dim{1}.tsv".format(representation_name,desc_dim),sep="\t")
|
121 |
+
datasets = os.listdir(os.path.join(script_dir, r"../data/auxilary_input/GO_datasets"))
|
122 |
if dataset_type == "All_Data_Sets" and aspect_type == "All_Aspects":
|
123 |
filtered_datasets = datasets
|
124 |
elif dataset_type == "All_Data_Sets":
|
|
|
133 |
|
134 |
for dt in tqdm(filtered_datasets,total=len(filtered_datasets)):
|
135 |
print(r"Protein function prediction is started for the dataset: {0}".format(dt.split(".")[0]))
|
136 |
+
dt_file = pd.read_csv(os.path.join(script_dir, r"../data/auxilary_input/GO_datasets/{0}".format(dt)),sep="\t")
|
137 |
dt_merge = dt_file.merge(representation_dataframe,left_on="Protein_Id",right_on="Entry")
|
138 |
|
139 |
dt_X = dt_merge['Vector']
|
|
|
153 |
predictions = dt_merge.iloc[:,:6]
|
154 |
predictions["predicted_values"] = list(model[3].toarray())
|
155 |
if detailed_output:
|
156 |
+
predictions.to_csv(os.path.join(script_dir, r"../results/Ontology_based_function_prediction_{1}_{0}_predictions.tsv".format(representation_name,dt.split(".")[0])),sep="\t",index=None)
|
157 |
|
158 |
return (cv_results, cv_mean_results,cv_std_results)
|
159 |
|
|
|
168 |
for i in cv_result:
|
169 |
df_cv_result.loc[len(df_cv_result)] = i
|
170 |
if detailed_output:
|
171 |
+
df_cv_result.to_csv(os.path.join(script_dir, r"../results/Ontology_based_function_prediction_5cv_{0}.tsv".format(representation_name)),sep="\t",index=None)
|
172 |
|
173 |
cv_mean_result = model[1]
|
174 |
df_cv_mean_result = pd.DataFrame({"Model": pd.Series([], dtype='str') ,"Accuracy": pd.Series([], dtype='float'),"F1_Micro": pd.Series([], dtype='float'),\
|
|
|
182 |
|
183 |
for j in cv_mean_result:
|
184 |
df_cv_mean_result.loc[len(df_cv_mean_result)] = j
|
185 |
+
df_cv_mean_result.to_csv(os.path.join(script_dir, r"../results/Ontology_based_function_prediction_5cv_mean_{0}.tsv".format(representation_name)),sep="\t",index=None)
|
186 |
|
187 |
#save std deviation of scores to file
|
188 |
cv_std_result = model[2]
|
|
|
197 |
|
198 |
for k in cv_std_result:
|
199 |
df_cv_std_result.loc[len(df_cv_std_result)] = k
|
200 |
+
df_cv_std_result.to_csv(os.path.join(script_dir, r"../results/Ontology_based_function_prediction_5cv_std_{0}.tsv".format(representation_name)),sep="\t",index=None)
|
201 |
|
202 |
print(datetime.now())
|
203 |
|