privacy_preserving_document_classification/client.py at main · tabeagnade/privacy_preserving_document_classification · GitHub

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
import dataManager
import model
import clientApi
import performance
import globals
import time
import preprocessing
import matplotlib.pyplot as plt
from tensorflow.keras.optimizers import Adam


class Client:
    def __init__(self):
        api = clientApi.ClientApi()
        p = performance.Performance()
        self.preprocessor = preprocessing.Preprocessor()
        d = dataManager.dataManager()

        globalModel = api.getArchitecture()
        if globals.szenario == "federatedLearning":
            optimizer = Adam(learning_rate=0.01)
        else:
            optimizer = "adam"
        globalModel.compile(loss=globals.loss,
                    optimizer= optimizer,
                    metrics=["acc"])
        globalModel.set_weights(api.getWeights())

        localModel = model.Model(newModelPath= "localModel", model = globalModel)

        p.printAccuracy(localModel.getModelPath())
        #print(globalModel.get_weights())
        method = input("Please enter the method (train or predict): ")
        if method == "predict":
            file = input("Please enter a file path: ")
            f = open(file, "r")
            text = f.read()
            localModel.predict(text)

        if method == "train":
            file = input("Please enter a file path: ")
            f = open(file, "r")
            text = [f.read()]
            label = input("Please enter the {} (1 or 0): ".format(globals.property))
            start = time.time()
            binaryLabels = [int(label)]
            localModel.trainModel(text, binaryLabels)

            if globals.szenario == "federatedLearning":
                api.sendWeights(localModel.getWeights())
            else:
                api.sendTrainData(text, binaryLabels)

            #p.printAccuracy(localModel.getModelPath())
            localModel.setWeights(api.getWeights())

            end = time.time()
            print ("train time: {}".format(end - start))
            #p.printAccuracy(localModel.getModelPath())

        if method == "trainTest":
            labels = []
            texts = []
            for idx, i in enumerate(d.secondTrainingIndices):
                label = d.produceTestFile(i, globals.property)
                labels.append(label) # returns label
                file = open("TestDokumente/test{}{}{}.txt".format(i, globals.property, labels[idx]), "r")
                texts.append(file.read())
            totalTime = 0
            print(len(texts))
            acc = []
            f1_list = []
            for idx, text in enumerate(texts):
                localModel.trainModel([text], [labels[idx]])
                start = time.time()
                if globals.szenario == "federatedLearning":
                    api.sendWeights(localModel.getWeights())
                else:
                    text = self.preprocessor.vectorize_text([text])
                    api.sendTrainData(text, [labels[idx]])

                #p.printAccuracy(localModel.getModelPath())
                localModel.setWeights(api.getWeights())

                end = time.time()
                totalTime += end - start
                print ("train time: {}".format(end - start))
                ac, f1 = p.printAccuracy(localModel.getModelPath())
                acc.append(ac)
                f1_list.append(f1)
            print("average time: " + str(totalTime/ len(texts)))
            datei = open('f1.txt','a')
            datei.write(str(f1_list))
            dateia = open('acc.txt','a')
            dateia.write(str(acc))
            plt.ylim(0.6,0.9)
            plt.plot(acc)
            plt.plot(f1_list)
            plt.show()

if __name__ == "__main__":
    client = Client()