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. dataDF = dataDF.set_index(pd.Index(items))
19. print(dataDF)
20.  
21. groupedData = dataDF.groupby('custID').groups
22. print(groupedData)
23.  
24.  
25. def isSubset(subset, set):
26.     if len(subset) > len(set):
27.         return False
28.     counter = 0
29.     for sub_item in subset:
30.         if sub_item in set:
31.             counter += 1
32.  
33.     return counter == len(subset)
34.  
35.  
36. def SupportCounter(pattern):
37.     support = 0
38.     for key in groupedData:
39.         shouldOccure = len(pattern)
40.         occurance = 0
41.         for item in pattern:
42.             occurance += groupedData[key].count(item)
43.         if (occurance == shouldOccure):
44.             support = support + 1
45.         occurance = 0
46.     return support
47.  
48.  
49. def FindFrequentSetInDataSet(itemToCheckIfFrequent):
50.     inHowManyGroupsItemOccured = 0
51.     itemCounter = 0
52.     for group in groupedData:
53.         itemCounterPerGroup = 0
54.         for items in groupedData[group]:
55.             if isSubset(itemToCheckIfFrequent, items):
56.                 itemCounter += 1
57.                 itemCounterPerGroup += 1
58.         if (itemCounterPerGroup > 0):
59.             inHowManyGroupsItemOccured += 1
60.     return inHowManyGroupsItemOccured
61.  
62.  
63. for group in groupedData:
64.     groupedData[group] = pd.array(groupedData[group])
65.  
66. L1 = []
67. itemTable = singleItems[:]
68. all_frequent_sets = []
69. DSPower = len(groupedData)
70. for item in singleItems:
71.     occurancesInAllEvents = FindFrequentSetInDataSet(item)
72.     if (occurancesInAllEvents / DSPower >= min_sup):
73.         L1.append(item)
74.         all_frequent_sets.append([item])
75.  
76. print("L1:", all_frequent_sets)
77. flag = True
78. counter_length = 2
79. frequent_set = []
80. for i in range(0, len(items)):
81.     dlugosc = len(items[i])
82.     j = 0
83.     while j < dlugosc:
84.         if items[i][j] not in L1:
85.             items[i].pop(j)
86.             dlugosc -= 1
87.         j += 1
88.  
89. d2 = {'custID': custID, 'items': items}
90. dataDF2 = pd.DataFrame(d2)
91. newDFArray = []
92. for index, row in dataDF2.iterrows():
93.     emptyArrayCheck = row["items"]
94.     if len(emptyArrayCheck) > 0:
95.         newDFArray.append(row)
96. newDF = pd.DataFrame(newDFArray)
97. print(newDF)
98.  
99. while flag:
100.  
101.     current_candidates = combinations(L1, counter_length)
102.     current_candidates = np.array(list(current_candidates))
103.     for item in current_candidates:
104.         occurrences = FindFrequentSetInDataSet(item)
105.         if occurrences / DSPower >= min_sup:
106.             frequent_set.append(item)
107.             all_frequent_sets.append(list(item))
108.     print("L" + str(counter_length) + ":", np.array(list(frequent_set)))
109.     if len(frequent_set) == 0 or len(L1) - 1 <= counter_length:
110.         break
111.     frequent_set = []
112.     counter_length += 1
113.  
114. frequent_dict = dict()
115. number = 1
116. for i in all_frequent_sets:
117.     d = {number: i}
118.     frequent_dict.update(d)
119.     number += 1
120.  
121. columns = {'sets': all_frequent_sets, 'values': range(1, len(all_frequent_sets) + 1)}
122. df_frequent = pd.DataFrame(columns)
123. print("Odwzorowanie:")
124. print(frequent_dict)
125.  
126. # ODWZOROWANIE
127. tmpTable = []
128. # for item in newDF['items']:
129.  
130. new_list = [[x] for x in newDF['items']]
131.  
132. for item in new_list:
133.     if len(item[0]) > 1:
134.         tmp = item[:]
135.         z = [[x] for x in tmp[0]]
136.         y = []
137.         print(len(z))
138.         print(all_frequent_sets)
139.         z.append(item[0])
140.         for i in range(0, len(z)):
141.             if (z[i] in all_frequent_sets):
142.                 y.append(z[i])
143.  
144.         print(y)
145.  
146.         item[0] = y[:]
147.         print(item[0])
148.         # x = combinations(item[0])
149. print(new_list)
150. for i in range(0, len(new_list)):
151.     if len(new_list[i][0]) > 1:
152.         new_list[i] = new_list[i][0]
153.  
154. print(new_list)
155. print(df_frequent)
156.  
157. # for item in new_list:
158. for i in range(0, len(new_list)):
159.     for j in range(0, len(new_list[i])):
160.         iterator = 0
161.         for mySet in df_frequent["sets"]:
162.             tmpValue = df_frequent.at[iterator, "values"]
163.             # print(tmpValue)
164.             if new_list[i][j] == mySet:
165.                 new_list[i][j] = tmpValue
166.                 break
167.             iterator += 1
168. print(new_list)
169. # df_frequent = df_frequent.set_index('sets')
170. # print(list(frequent_dict.keys()[frequent_dict.values().index(['a'])]))
171.  
172. newDF["items"] = new_list
173. print(newDF)
174.  
175. newDF = newDF.set_index(pd.Index(newDF["items"]))
176. groupedData2 = newDF.groupby('custID').groups
177. print(groupedData2)
178.  
179. allCand = []
180. for i in range(1, len(df_frequent["values"]) + 1):
181.     for j in range(1, len(df_frequent["values"]) + 1):
182.         tmp = [i, j]
183.         allCand.append(tmp)
184. print(allCand)
185.  
186. # (1,1), konkretny klient transakcje (1: [[1], [1]],).
187. # wez pierwsze elem z (1,1) spr czy zawiera se w pierwszej trazakji jak tak  occur +=1 CONTINUE , jak nie Continue
188.  
189. frequent_formatted = []
190. for cand in allCand:
191.     occurances = 0
192.     for i in range(1, max(newDF["custID"]) + 1):
193.         semi_occurance = 0
194.         counter = 0
195.         for tran in groupedData2[i]:
196.             if cand[counter] in tran:
197.                 semi_occurance += 1
198.                 counter += 1
199.                 if counter > 1:
200.                     break
201.         if semi_occurance > 1:
202.             occurances += 1
203.     support = occurances/max(newDF["custID"])
204.     if support > min_sup:
205.         frequent_set.append((cand , support))
206.  
207. for i in frequent_set:
208.     print(i)