2PyiT04G

1def processing(request, id=None):
2    template = 'kmeans/processing.html'
3    q = get_object_or_404(FileDetails, id = id, user = request.user)
4    print(q)
5    c = q.data_columns.replace(',', '\n').replace('[', '').replace(']', '').replace("'", '')  # cleaning data columns
6    
7    import os
8    media_path = settings.MEDIA_ROOT  # reading media path
9    file_path = os.path.join(media_path, q.file_name) # makeing file url by joining media_path & filtered file name
10    data = pd.read_csv(file_path)  # reading selected file
11    data['id'] = range(0, len(data))  # create unique id for csv file 
12    head = data.head()
13    km = KMeans(n_clusters=3) # defining no. of cluster
14    
15    if request.method == 'POST':    
16        v = request.POST.getlist('sel')  # taking file column as input for clustering
17        print(v)
18        X_std = StandardScaler().fit_transform(data[v])  # data scaling
19        pca = PCA(n_components=2)  # applying pca
20        x_pca = pca.fit_transform(X_std)  # fitting & tramsformation
21        cl_data = pd.DataFrame(columns=['id','PCA1', 'PCA2', 'cluster'])  # create new dataframe 
22        cl_data['id'] = data['id']  # adding id 
23        cl_data['PCA1'] = [xx[0] for xx in x_pca]  # adding pca1
24        cl_data['PCA2'] = [xx[1] for xx in x_pca]  # adding pca2
25        y_predicted = km.fit_predict(x_pca)  # making predicting
26        cl_data['cluster'] = y_predicted     # adding prediction to cluster
27        data['cluster'] = y_predicted
28        # creating a table dynamically
29        from django.db import connection
30        cur = connection.cursor()
31
32        cl = list(data.columns)  # saving data columns in a list
33        clms = [re.sub('\W+', '',i) for i in cl]  # cleaning column names
34        import datetime
35        fl = re.sub('\W+', '',q.file_name+str(datetime.datetime.now().strftime("%X")))
36        # using session to store table name
37        request.session['fl'] = fl
38        
39        # creating query dynamically
40        def cr(s):
41            return s + ' varchar(255)'
42        zz = "CREATE TABLE IF NOT EXISTS "+fl+" (" + ', '.join(map(cr, clms)) + ')'  # creating query
43        cur.execute(zz)  # executing query
44
45        # setting up database
46        database_url = 'mysql+pymysql://{user}:{password}@127.0.0.1/{database_name}'.format(
47            user='root',
48            password='user1234',
49            database_name="clustered",
50        )
51        engine = create_engine(database_url, echo=False)
52        data.to_sql(name=fl, con=engine, if_exists='append', index=False)  # saving data into a table
53
54        # inserting data into database table
55        for uid, pca1, pca2, cluster in zip(cl_data['id'], cl_data['PCA1'], cl_data['PCA2'], cl_data['cluster']):
56            ClusterData.objects.create(file_id=q, u_id=uid, pca1=pca1, pca2=pca2, cluster=cluster)
57
58        FileDetails.objects.filter(user= request.user).filter(file_name=q.file_name).update(status = True, sel_columns = v, table_name = fl)
59
60        return HttpResponse('<h2>Your file has been successfully proceed.</h2>')
61
62    context = {'q':q, 'columns':c.split(), 'head':head.to_html}
63    return render(request, template, context)
64
65# dashboard function
66def dashboard(request, id=None):
67    template = 'kmeans/dashboard.html'
68    request.session['fid'] = id
69    q = FileDetails.objects.filter(id = id)
70    
71    for qq in q:
72        aa = qq.sel_columns
73        xx = aa.replace(',', '\n').replace('[', '').replace(']', '').replace("'", '')
74    context = {'q':q, 'col':xx.split()}
75    return render(request, template, context)