-
Notifications
You must be signed in to change notification settings - Fork 4
/
frontend.py
271 lines (216 loc) · 11.4 KB
/
frontend.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
""" This module includes all streamlit function to display interactive widgets
"""
import logging
import streamlit as st
import base64
from matching import get_confusion_matrix_data
import pandas as pd
_phonetic_encoding = ['metaphone','double_metaphone','soundex','nysiis','match_rating']
_blocking = ['Full Indexing','Standard','SortedNeighbourhood']
def show_ui_import_data():
st.header("Step 1- Let's **begin** with the dataset you want to deduplicate :")
uploaded_file = st.file_uploader("Browse and select the file (only accept CSV file):",type = "csv")
return uploaded_file
def show_ui_set_index(cols):
st.write("## Select the record identifier in your dataset")
# UI set index
if cols is not None:
index_name = st.selectbox("Record Identifier:",cols)
return index_name
else:
return None
def show_ui_set_index_true_links(cols_true,is_gold_standard):
if ( (is_gold_standard) & (cols_true is not None)):
index_name_1 = st.selectbox("Record Identifier 1:",cols_true)
index_name_2 = st.selectbox("Record Identifier 2:",cols_true)
return index_name_1,index_name_2
else:
return None, None
def show_ui_import_gold_standard():
st.markdown("---")
st.header("Step 2- Do you have a **gold standard** for this dataset ?")
st.info("""
In order to asses the quality of the different matching strategies ground-thruth data ,
known as gold standard or labled data are required
""")
is_gold_standard = False if st.radio("Check Yes or No if you have a gold standard ?",("No","Yes")) == "No" else True
if(is_gold_standard) :
uploaded_true = st.file_uploader("Browse and select the gold standard (Review instructions for the correct format of the gold standard):",type = "csv")
if uploaded_true is not None :
return uploaded_true,is_gold_standard
else:
return None,is_gold_standard
else:
return None, is_gold_standard
def show_ui_phonetic_encoding(df_a,index_name):
st.markdown("---")
select_encoding = {}
cols = df_a.columns.to_list()
encoding_cols = list(cols)
st.header("Step 3- Tell us the Pre-processing or Encoding you want on each field")
st.info("""
The main task of data cleaning and standardization is the conversion of the raw input data into well defined, consistent forms, as well as the resolution of inconsistencies
in the way information is represented and encoded.(Christen 2012)
""")
for i,c in enumerate(encoding_cols):
if st.checkbox(c,key="ppp_check"+c) :
ph_selected = st.selectbox("phonetic encoding:",_phonetic_encoding,key='phonetic_'+c)
select_encoding[c] = ph_selected
return select_encoding
def show_ui_blocking(cols):
st.markdown("---")
st.header('Step 4 -Now, please configure the **blocking strategy** you want ')
st.info("""
The step of the process called blocking or indexing try to reduce the number of records we need to compare.
The idea is instead of comparing all records of the dataset between themselves we want
to compare only the records that are most likely to be matched.
""")
blocking_selected = st.selectbox("Select blocking algorithm:",_blocking,key="_blocking_alg")
#FIXME Catch error when blokcing fields empty
if (blocking_selected == "Standard"):
number_blocks = st.number_input("Number of predicates:",1) + 1
options = [st.empty() for i in range(1,number_blocks)]
options_cols =[[] for i in enumerate(options)]
for i,opt in enumerate(options):
options_cols[i] = st.multiselect('Select blocking fields',cols,cols[:2],key="options"+str(i))
else:
options_cols = []
return options_cols,blocking_selected
def show_ui_comparison(cols,_compare_vartype,_compare_string_method):
st.markdown("---")
st.header('Step 5 - Next, please configure the **comparison strategy** you want to apply')
st.info("""
In this step you select the similarity algorihtm and calculate
the similarity score between records pairs to create a comparison vectors
""")
list_checkbox_fields = [st.empty for i in range(1,len(cols))]
select_fields ={}
comparison = []
for i, col in enumerate(cols):
item = {}
if st.checkbox(cols[i]):
vartype = st.selectbox("vartype",_compare_vartype,key="vartype_"+col)
if(vartype =='string'):
method = st.selectbox('method',_compare_string_method,key="method"+col)
if(st.checkbox('set threshold',key='threshold_check'+col)):
threshold = st.number_input('threshold',key='threshold_'+col)
item['threshold'] = threshold
item['method'] = method
item['vartype'] = vartype
item['field'] = col
item['code'] = col
comparison.append(item)
st.markdown("---")
if len(comparison) == 0 :
comparison.append({'vartype':'exact','field': cols[0],'code':cols[0]})
comparison.append({'vartype':'exact','field': cols[1],'code':cols[1]})
return comparison
def show_ui_classification(features,df_a,list_classifiers):
st.markdown("---")
st.header('Step 6 - Classification')
st.info("""
Based on comparison vector, this step uses a classification algorithm to
classify candidate records pairs in: matches and non-matches.
""")
option_classifiers = []
threshold = 1
threshold_wavg = 1
select_wf= {}
for i , classifier in enumerate(list_classifiers):
if st.checkbox(classifier) :
option_classifiers.append(classifier)
if classifier == 'SimSum' :
threshold = st.slider("Threshold :",1,len(features.columns)+1)
elif classifier == "Weighted Average":
threshold_wavg = st.number_input("Threshold",min_value=0.1,max_value=0.999,key="threshold_wf")
cols = features.columns.to_list()
st.write("Select attributes weights:")
for i,c in enumerate(cols):
if st.checkbox(c,key="wf_check_"+c+str(i)) :
wf_selected = st.number_input("weight :",value =1,key="wf_"+c)
select_wf[c] = wf_selected
else:
select_wf[c] = 1
return option_classifiers, threshold,threshold_wavg,select_wf
def get_table_download_link(df,data_name):
"""Generates a link allowing the data in a given panda dataframe to be downloaded
in: dataframe
out: href string
"""
csv = df.to_csv(index=False)
b64 = base64.b64encode(csv.encode()).decode() # some strings <-> bytes conversions necessary here
href = f'<a href="data:file/csv;base64,{b64}">Download {data_name} as csv file</a>'
return href
def show_ui_evaluation():
st.markdown("---")
st.header('Step 7 - Evaluation')
st.info("""
Comparing match results with the known ground truth or gold standard
to measure the performance of the matching process.
""")
def show_ui_leaderboard(app_results,results_dict,option_classifiers,is_gold_standard):
st.markdown('## Leaderboard')
df_a = app_results['data']
features = app_results['comparison_vector']
index_name = app_results['index_name']
if is_gold_standard :
index_name_1 = app_results['index_name_1']
index_name_2 = app_results['index_name_2']
df_true_links = app_results['df_true_links']
# For each classifier display results
df_leaderboard = pd.DataFrame()
for classifier, results in results_dict.items():
if not is_gold_standard :
st.markdown("### "+classifier)
unique_data = results['unique']['data_unique']
nunique = results['metrics']['nunique']
st.write("Number of unique records : "+ str(nunique))
st.write(unique_data)
st.markdown(get_table_download_link(unique_data,""),unsafe_allow_html=True)
else:
df_leaderboard = pd.concat([df_leaderboard,pd.DataFrame(results_dict[classifier]['metrics'],index=[classifier])])
if is_gold_standard :
st.table(df_leaderboard)
def show_ui_dashboard(app_results,results_dict,option_classifiers,is_gold_standard):
df_a = app_results['data']
features = app_results['comparison_vector']
index_name = app_results['index_name']
if is_gold_standard :
index_name_1 = app_results['index_name_1']
index_name_2 = app_results['index_name_2']
df_true_links = app_results['df_true_links']
# For each classifier display results
for classifier, results in results_dict.items():
#st.markdown("### "+classifier)
if not is_gold_standard :
pass
else:
if st.checkbox(classifier,key="leader_details"+classifier):
st.markdown("### Confusion matrix")
matrix = pd.DataFrame(results['matrix'], columns=['Predicted Positives','Predicted Negatives'],index = ['True Positives','True Negatives'])
st.table(matrix)
st.markdown("### Visuals")
st.altair_chart(results['roc'])
st.altair_chart(results['pr'])
if st.checkbox('Show confusion matrix data',key='show_data'+classifier):
selected_data = st.selectbox("select dataset",['matches',
'non-matches',
'true positives',
'false positives',
'true negatives',
'false negatives',
],key='data_'+classifier)
nb_data = st.empty()
data = get_confusion_matrix_data(selected_data,df_a,
df_true_links,
results['matches'],
features,
index_name_1,
index_name_2)
st.write(selected_data +": "+str(len(data))+" records pairs")
st.write(data)
st.markdown(get_table_download_link(data,selected_data),unsafe_allow_html=True)
if st.checkbox("Show unique records",key="show_unique_"+classifier):
unique_data = results['unique']['data_unique']
st.write(unique_data[:1000])
st.markdown(get_table_download_link(unique_data,""),unsafe_allow_html=True)