1. import pandas as pd
2. import numpy as np
3. from itertools import combinations
4. from itertools import permutations
5.  
6. from prompt_toolkit.layout import Dimension
7.  
8. # Data
9. # DODAC wpisywanie min_sup !!!!
10. min_sup = 0.25
11. custID = [1, 1, 2, 2, 2, 3, 4, 4, 4, 5]
12. dateOrder = [1, 2, 1, 2, 3, 1, 1, 2, 3, 1]
13. items = [['a'], ['a'], ['a'], ['b'], ['c', 'e'], ['a', 'e'], ['a'], ['c', 'd', 'e'], ['a'], ['a']]
14. singleItems = ['a', 'b', 'c', 'd', 'e']
15. d = {'custID': custID, 'dateOrder': dateOrder, 'items': items}
16. dataDF = pd.DataFrame(d)
17.  
18. print(dataDF)
19. dataDF = dataDF.set_index(pd.Index(items))
20.  
21. groupedData = dataDF.groupby('custID').groups
22.  
23. def isSubset(subset, set):
24.     if len(subset) > len(set):
25.         return False
26.     counter = 0
27.     for sub_item in subset:
28.         if sub_item in set:
29.             counter += 1
30.  
31.     return counter == len(subset)
32.  
33.  
34. def SupportCounter(pattern):
35.     support = 0
36.     for key in groupedData:
37.         shouldOccure = len(pattern)
38.         occurance = 0
39.         for item in pattern:
40.             occurance += groupedData[key].count(item)
41.         if (occurance == shouldOccure):
42.             support = support + 1
43.         occurance = 0
44.     return support
45.  
46.  
47. def FindFrequentSetInDataSet(itemToCheckIfFrequent):
48.     inHowManyGroupsItemOccured = 0
49.     itemCounter = 0
50.     for group in groupedData:
51.         itemCounterPerGroup = 0
52.         for items in groupedData[group]:
53.             if isSubset(itemToCheckIfFrequent, items):
54.                 itemCounter += 1
55.                 itemCounterPerGroup += 1
56.         if (itemCounterPerGroup > 0):
57.             inHowManyGroupsItemOccured += 1
58.     return inHowManyGroupsItemOccured
59.  
60.  
61. for group in groupedData:
62.     groupedData[group] = pd.array(groupedData[group])
63.  
64. L1 = []
65. itemTable = singleItems[:]
66. all_frequent_sets = []
67. DSPower = len(groupedData)
68. for item in singleItems:
69.     occurancesInAllEvents = FindFrequentSetInDataSet(item)
70.     if (occurancesInAllEvents / DSPower >= min_sup):
71.         L1.append(item)
72.         all_frequent_sets.append([item])
73.  
74. print("L1:", all_frequent_sets)
75. flag = True
76. counter_length = 2
77. frequent_set = []
78. for i in range(0, len(items)):
79.     dlugosc = len(items[i])
80.     j = 0
81.     while j < dlugosc:
82.         if items[i][j] not in L1:
83.             items[i].pop(j)
84.             dlugosc -= 1
85.         j += 1
86.  
87. d2 = {'custID': custID, 'items': items}
88. dataDF2 = pd.DataFrame(d2)
89. newDFArray = []
90. for index, row in dataDF2.iterrows():
91.     emptyArrayCheck = row["items"]
92.     if len(emptyArrayCheck) > 0:
93.         newDFArray.append(row)
94. newDF = pd.DataFrame(newDFArray)
95.  
96. while flag:
97.  
98.     current_candidates = combinations(L1, counter_length)
99.     current_candidates = np.array(list(current_candidates))
100.     for item in current_candidates:
101.         occurrences = FindFrequentSetInDataSet(item)
102.         if occurrences / DSPower >= min_sup:
103.             frequent_set.append(item)
104.             all_frequent_sets.append(list(item))
105.     print("L" + str(counter_length) + ":", np.array(list(frequent_set)))
106.     if len(frequent_set) == 0 or len(L1) - 1 <= counter_length:
107.         break
108.     frequent_set = []
109.     counter_length += 1
110.  
111. frequent_dict = dict()
112. number = 1
113. for i in all_frequent_sets:
114.     d = {number: i}
115.     frequent_dict.update(d)
116.     number += 1
117.  
118. columns = {'sets': all_frequent_sets, 'values': range(1, len(all_frequent_sets) + 1)}
119. df_frequent = pd.DataFrame(columns)
120. print("Odwzorowanie:")
121. print(df_frequent)
122.  
123. tmpTable = []
124. new_list = [[x] for x in newDF['items']]
125.  
126. for item in new_list:
127.     if len(item[0]) > 1:
128.         tmp = item[:]
129.         z = [[x] for x in tmp[0]]
130.         y = []
131.         z.append(item[0])
132.         for i in range(0, len(z)):
133.             if (z[i] in all_frequent_sets):
134.                 y.append(z[i])
135.  
136.         item[0] = y[:]
137.  
138. for i in range(0, len(new_list)):
139.     if len(new_list[i][0]) > 1:
140.         new_list[i] = new_list[i][0]
141.  
142.  
143. for i in range(0, len(new_list)):
144.     for j in range(0, len(new_list[i])):
145.         iterator = 0
146.         for mySet in df_frequent["sets"]:
147.             tmpValue = df_frequent.at[iterator, "values"]
148.             if new_list[i][j] == mySet:
149.                 new_list[i][j] = tmpValue
150.                 break
151.             iterator += 1
152.  
153. newDF["items"] = new_list
154. print("Po Transfromacji i odwzorowaniu:")
155. print(newDF)
156.  
157. newDF = newDF.set_index(pd.Index(newDF["items"]))
158. groupedData2 = newDF.groupby('custID').groups
159.  
160. allCand = []
161. for i in range(1, len(df_frequent["values"]) + 1):
162.     for j in range(1, len(df_frequent["values"]) + 1):
163.         tmp = [i, j]
164.         allCand.append(tmp)
165.  
166.  
167. frequent_formatted = []
168. for cand in allCand:
169.     occurances = 0
170.     for i in range(1, max(newDF["custID"]) + 1):
171.         semi_occurance = 0
172.         counter = 0
173.         for tran in groupedData2[i]:
174.             if cand[counter] in tran:
175.                 semi_occurance += 1
176.                 counter += 1
177.                 if counter > 1:
178.                     break
179.         if semi_occurance > 1:
180.             occurances += 1
181.     support = occurances/max(newDF["custID"])
182.     if support > min_sup:
183.         frequent_formatted.append((cand , support))
184.  
185. for i in frequent_formatted:
186.     print(i)