-
Notifications
You must be signed in to change notification settings - Fork 0
/
Main.py
316 lines (259 loc) · 12.9 KB
/
Main.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
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
import numpy as np
import pandas as pd
from data_prep_sup import cleanDataset, separation
from model_sup import logistic, randomForest, xgBoost, lightGbm, naiveBayes
from dashboard_sup import dashViz
from Check import checksPassed
from diagnostic import function_diagnostic
from model_non_sup import Kmeans, predictions_tous_les_clusters_separe_KM, Meanshift, \
predictions_tous_les_clusters_separe_MS, Mixturegaussian, \
prediction_tous_les_clusters_separes_MG, Bayesian_mixture_gaussian, prediction_tous_les_clusters_separes_BMG
from data_prep_non_sup import scale, pca2, pca, fig_pca, tsne, isomap, local_lin, mds
from dashboard_non_sup import dashviznsup, server
from flask import request, render_template, Response
import io
from matplotlib.backends.backend_agg import FigureCanvasAgg as FigureCanvas
# Launch home page
@server.route('/')
def home():
return render_template("home.html")
# When a dataset is selected, we call the page 2 to select the kind of learning
@server.route('/uploader', methods=['POST'])
def upload_file():
if request.method == 'POST':
f = request.files['file']
global dataframe
dataframe = pd.read_csv(f)
return render_template("page2.html")
# If the supervised learning is selected, we call the page 3 with the name of all the column of the dataframe
# Otherwise, we call the page 4
@server.route('/Page2', methods=['POST'])
def page2():
if request.method == 'POST':
global learning
if request.form['learning'] == "supervised":
learning = "supervised"
return render_template("page3.html", column_name=dataframe.columns)
elif request.form['learning'] == "unsupervised":
learning = "unsupervised"
return render_template("page4.html")
##Tabs
if request.form['learning'] == "Dataset":
return render_template("home.html")
elif request.form['learning'] == "Type of dataset":
return render_template("page2.html")
# If it's in supervised learning mode, the user need to choose the feature to predict
# Then the page 4 is launch
@server.route('/Page3', methods=['POST'])
def page3():
if request.method == 'POST':
global feature_target
feature_target = request.form['feature_target']
# If the check are passed, no dataprep
if request.form['tabs'] == "Submit" and checksPassed(dataframe) == True:
X_train, X_test, Y_train, Y_test = separation(dataframe, feature_target)
Y_test = pd.Series(Y_test)
lin = logistic(X_train, X_test, Y_train)
lin_res = [Y_test, pd.Series(lin), "Linear"]
nb = naiveBayes(X_train, X_test, Y_train)
nb_res = [Y_test, pd.Series(nb), "Naive Bayes"]
xgb = xgBoost(X_train, X_test, Y_train)
xgb_res = [Y_test, pd.Series(xgb), "XGBoost"]
rf = randomForest(X_train, X_test, Y_train, grid_search=False)
rf_res = [Y_test, pd.Series(rf), "Random forest"]
lgb = lightGbm(X_train, X_test, Y_train, grid_search=False)
lgb_res = [Y_test, pd.Series(lgb), "LightGbm"]
dash_appS = dashViz([lin_res, nb_res, xgb_res, rf_res, lgb_res])
return dash_appS.index()
elif request.form['tabs'] == "Submit" and checksPassed(dataframe) == False:
return render_template("page4.html")
## Tabs
if request.form['tabs'] == "Dataset":
return render_template("home.html")
elif request.form['tabs'] == "Type of dataset":
return render_template("page2.html")
elif request.form['tabs'] == "Feature to predict":
return render_template("page3.html")
# Choose of data processing
@server.route('/Page4', methods=['POST'])
def page4():
if request.method == 'POST':
global dataset_cleaned
if request.form['Processing'] == "automatic" and learning == "supervised":
dataset_cleaned = cleanDataset(dataframe, feature_target, manual=False, supervised=True)
X_train, X_test, Y_train, Y_test = separation(dataset_cleaned, feature_target)
Y_test = pd.Series(Y_test)
lin = logistic(X_train, X_test, Y_train)
lin_res = [Y_test, pd.Series(lin), "Linear"]
nb = naiveBayes(X_train, X_test, Y_train)
nb_res = [Y_test, pd.Series(nb), "Naive Bayes"]
xgb = xgBoost(X_train, X_test, Y_train)
xgb_res = [Y_test, pd.Series(xgb), "XGBoost"]
rf = randomForest(X_train, X_test, Y_train, grid_search=False)
rf_res = [Y_test, pd.Series(rf), "Random forest"]
lgb = lightGbm(X_train, X_test, Y_train, grid_search=False)
lgb_res = [Y_test, pd.Series(lgb), "LightGbm"]
dash_appS = dashViz([lin_res, nb_res, xgb_res, rf_res, lgb_res])
return dash_appS.index()
elif request.form['Processing'] == "diagnostic" and learning == "supervised":
messages = function_diagnostic(dataframe, feature_target, supervised=True)
return render_template("page_manuelle.html", messages=messages)
elif request.form['Processing'] == "automatic" and learning == "unsupervised":
dataset_cleaned = cleanDataset(dataframe, feature_target=None, manual=False, supervised=False)
return render_template("page5.html", column_number=range(2, len(dataset_cleaned.columns)))
elif request.form['Processing'] == "diagnostic" and learning == "unsupervised":
messages = function_diagnostic(dataframe, feature_target=None, supervised=False)
return render_template("page_manuelle.html", messages=messages)
##Tabs
elif request.form['Processing'] == "Dataset" and (
learning == "supervised" or learning == "unsupervised"):
return render_template("home.html")
elif request.form['Processing'] == "Type of dataset" and (
learning == "supervised" or learning == "unsupervised"):
return render_template("page2.html")
elif request.form['Processing'] == "Feature to predict" and (learning == "supervised"):
return render_template("page3.html", column_name=dataframe.columns)
elif request.form['Processing'] == "Feature to predict" and (learning == "unsupervised"):
return "impossible"
elif request.form['Processing'] == "Data prep" and (
learning == "supervised" or learning == "unsupervised"):
return render_template("page4.html")
# Diagnostic
@server.route('/page_manuelle', methods=['POST'])
def function_return():
if request.method == 'POST':
if request.form['return'] == "yes":
return render_template("page4.html")
## Tabs
elif request.form['return'] == "Dataset" and (learning == "supervised" or learning == "unsupervised"):
return render_template("home.html")
elif request.form['return'] == "Type of dataset" and (
learning == "supervised" or learning == "unsupervise"):
return render_template("page2.html")
elif request.form['return'] == "Feature to predict" and (learning == "supervised"):
return render_template("page3.html", column_name=dataframe.columns)
elif request.form['return'] == "Feature to predict" and (learning == "unsupervised"):
return "impossible"
elif request.form['return'] == "Data prep" and (
learning == "supervised" or learning == "unsupervised"):
return render_template("page4.html")
elif request.form['return'] == "Diagnostic" and learning == "unsupervised":
messages = function_diagnostic(dataframe, feature_target=None, supervised=False)
return render_template("page_manuelle.html", messages=messages)
elif request.form['return'] == "Diagnostic" and learning == "unsupervised":
messages = function_diagnostic(dataframe, feature_target, supervised=True)
return render_template("page_manuelle.html", messages=messages)
# image pca
@server.route('/plot.png')
def plot_png():
global X_scaled
X_scaled = scale(dataset_cleaned)
fig = fig_pca(X_scaled)
output = io.BytesIO()
FigureCanvas(fig).print_png(output)
return Response(output.getvalue(), mimetype='image/png')
# We train the differents models and we display the dashboard if the number of dimensions chosen by the user is equal to 2.
# Otherwise, we go to the next page which will ask for the size reduction method
@server.route('/Page5', methods=['POST'])
def page5():
if request.method == 'POST':
if request.form['tabs'] == "Submit":
global nb_pca, X_pca, X2, tab2
nb_pca = request.form['nb_pca']
nb_pca = int(nb_pca)
X_pca = pca(X_scaled, nb_pca)
X2 = dataset_cleaned
pourcentage = np.arange(0.9, 1.0, 0.01)
deb = 2
fin = 8
tab2 = []
# Training of the models
Km = Kmeans(X_pca, deb, fin)
Km = list(Km)
Ms = Meanshift(X_pca)
Ms = list(Ms)
Mg = Mixturegaussian(X_pca, deb, fin)
Mg = list(Mg)
Bmg = Bayesian_mixture_gaussian(X_pca)
Bmg = list(Bmg)
# Detection of outliers for every percentage
tab1 = []
for i in pourcentage:
# KMeans
Km2 = Km.copy()
Km_predict_anomaly = predictions_tous_les_clusters_separe_KM(X_pca, i, Km[4], Km[0])
Km2.insert(7, Km_predict_anomaly)
tab1.append(Km2)
tab2.append(tab1)
tab1 = []
for i in pourcentage:
# MShift
Ms2 = Ms.copy()
Ms_predict_anomaly = predictions_tous_les_clusters_separe_MS(X_pca, i, Ms[4], Ms[0])
Ms2.insert(7, Ms_predict_anomaly)
tab1.append(Ms2)
tab2.append(tab1)
tab1 = []
for i in pourcentage:
# MixtureGaussian
Mg2 = Mg.copy()
Mg_predict_anomaly = prediction_tous_les_clusters_separes_MG(i, Mg[0], Mg[4])
Mg2.insert(7, Mg_predict_anomaly)
tab1.append(Mg2)
tab2.append(tab1)
tab1 = []
for i in pourcentage:
# BayesianMixtureGaussian
Bmg2 = Bmg.copy()
Bmg_predict_anomaly = prediction_tous_les_clusters_separes_BMG(i, Bmg[0], Bmg[4])
Bmg2.insert(7, Bmg_predict_anomaly)
tab1.append(Bmg2)
tab2.append(tab1)
if nb_pca == 2:
dashboard_non_sup = dashviznsup(X_pca, X2, tab2)
return dashboard_non_sup.index()
elif nb_pca != 2:
return render_template("page6.html", reduc_name=['tsne', 'isomap', 'local lin', 'pca', 'mds'])
elif request.form['tabs'] == "Dataset":
return render_template("home.html")
elif request.form['tabs'] == "Type of dataset":
return render_template("page2.html")
elif request.form['tabs'] == "Feature to predict":
return "impossible"
elif request.form['tabs'] == "Data prep":
return render_template("page4.html")
elif request.form['tabs'] == "PCA":
return render_template("page5.html", column_number=range(2, len(dataset_cleaned.columns)))
# Choice of the 2D dimension reduction. Then we display the dashboard
@server.route('/Page6', methods=['POST'])
def choix_reduc():
if request.method == 'POST':
global type_of_dimension_reduction
type_of_dimension_reduction = request.form['type_of_dimension_reduction']
if request.form['tabs'] == "Submit":
if type_of_dimension_reduction == "tsne":
X = tsne(X_pca)
if type_of_dimension_reduction == "isomap":
X = isomap(X_pca)
if type_of_dimension_reduction == "local lin":
X = local_lin(X_pca)
if type_of_dimension_reduction == "pca":
X = pca2(X_pca)
if type_of_dimension_reduction == "mds":
X = mds(X_pca)
dashboard_non_sup = dashviznsup(X, X2, tab2)
return dashboard_non_sup.index()
elif request.form['tabs'] == "Dataset":
return render_template("home.html")
elif request.form['tabs'] == "Type of dataset":
return render_template("page2.html")
elif request.form['tabs'] == "Feature to predict":
return "impossible"
elif request.form['tabs'] == "Data prep":
return render_template("page4.html")
elif request.form['tabs'] == "PCA":
return render_template("page5.html", column_number=range(2, len(dataset_cleaned.columns)))
elif request.form['tabs'] == "Other reduction":
return render_template("page6.html", reduc_name=['tsne', 'isomap', 'local lin', 'pca', 'mds'])
if __name__ == '__main__':
server.run()