-
Notifications
You must be signed in to change notification settings - Fork 0
/
app_labeler.py
457 lines (358 loc) · 15.8 KB
/
app_labeler.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
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
# -*- coding: utf-8 -*-
import dash
import dash_core_components as dcc
import dash_html_components as html
import plotly.graph_objs as go
import pandas as pd
import numpy as np
import datetime
from datetime import datetime as dt
import re
import os
from dash.dependencies import Output
from dash.dependencies import Input
from dash.dependencies import State
import json
external_stylesheets = ['bWLwgP.css']
app = dash.Dash(__name__, external_stylesheets=external_stylesheets)
##############################################################################
# Define preprocessor functions
##############################################################################
def index_to_time(df_new):
def time_stamp(val):
try:
return pd.to_datetime(val,format='%d-%m-%Y')
except:
#try:
# return pd.to_datetime(val,format='%d-%m-%Y')
print('error with ',val)
return np.nan
df_new['time'] = df_new['date'].apply(time_stamp)
df_new = df_new.dropna()
#df_new.index = pd.DatetimeIndex(df_new['time'])
df_new['time_new'] = df_new["time"].apply( lambda df : datetime.datetime(year=df.year, month=df.month, day=df.day))
df_new.set_index(df_new["time_new"],inplace=True)
return df_new
def render_document(title='', snippet='', body='', art='', id='',date='',publisher='', lines=[], colors=[], labels=[]):
'''
Renders news articles and colors substrings within lines
'''
document = ''
line_br = "<br>"
if len(title) > 0:
document += ("<h3> <center>" + title + " </center></h3>" + line_br)
document += line_br
if len(id)>0:
document += ('ID: '+id)
if len(date)>0:
if len(id) > 0:
document+=', '
document += (date)
if len(publisher)>0:
if len(id)>0 or len(date)>0:
document+=', '
document += publisher
if len(id)>0 or len(date)>0 or len(pubisher)>0:
document += line_br
if len(body) > 0:
document += ('<p>' + snippet + ' ' + body + '</p>')
if len(art) > 0:
document += ('<p>' + art + '</p>')
for i, line in enumerate(lines):
#document = re.sub(line, "<span style='color:" + colors[i] + "'><b>" + line + "</b>[" + labels[i] + "]</span>",
# document)
document = re.sub(line, "<span style='color:" + colors[i] + "'>" + line + "<b>[" + labels[i] + "]</b></span>",
document)
return document
def render_doc_summary(title='', summary='', lines=[], colors=['red', 'blue'], labels=[]):
'''
Renders news articles and colors substrings within lines
'''
document = '<b>'+title+'</b>'+'<br>' + summary
#document = "<h3> <center>" + title + " </center></h3>" + summary
for i, line in enumerate(lines):
document = re.sub(line, "<span style='background-color:" + colors[i] + "'><b>" + line + "<b>[" + labels[i] + "]</span>",
document)
return document
def create_fake_summary(text):
'''
Here we would call a more sophisticaed function... TextRank
'''
return '.'.join(text.split('.')[:2])
##############################################################################
# Define data sources & properties of the sources
##############################################################################
# assumes ['title'],,['text'],'date','file_id' as input fields
input_file = 'jacobs_corpus_body_labeled.csv'
df = pd.read_csv(input_file)
df_pred=pd.read_csv('jacobs_corpus.csv')
df = df.dropna()
df = index_to_time(df)
df = df.sort_index()
# for modified labels
df_mod = df.copy()
# events as they appear in the dataframe columns
event_labels = ['Profit', 'Dividend', 'MergerAcquisition', 'SalesVolume', 'BuyRating', 'QuarterlyResults',
'TargetPrice', 'ShareRepurchase', 'Turnover', 'Debt']
event_colors_tmp = ['r','b','g','y','m','k','gray','blueviolet','firebrick','bisque','orange']
event_colors_tmp = ['red','gray','yellow','blue','navy','silver','aqua','Purple','olive','lime']
event_colors={}
for label,color in zip(event_labels,event_colors_tmp):
event_colors[label] = color
current_file_id = None
##############################################################################
# Build layout
##############################################################################
app.layout = html.Div(children=[
html.H1(children='News Label Dashboard',style={'textAlign': 'center'}),
html.Div(className='row',children=[
html.Div(className='row', children=[dcc.Graph(id="numdocs_graph")], style={'float': 'right', 'clear': 'both',
'padding': 20, 'margin': 5,
'borderRadius': 5,
'border': 'thin lightgrey solid'}),
# initial selection block
html.Div(className='row',children=[
html.Div(children='Select file'), # , style={'float': 'left', 'clear': 'both', 'text-align': 'center'}),
html.Div([
dcc.Dropdown(
id='file_dropdown',
options=[{'label':f,'value':f} for f in sorted([f for f in os.listdir('.') if (os.path.isfile(f) and '.csv' in f)],key=lambda x: os.path.getmtime(x))],
value='jacobs_corpus_body_labeled.csv' # sorted([f for f in os.listdir('.') if os.path.isfile(f)],key=lambda x: os.path.getmtime(x))[-1]
)]),#, className='six columns'),
html.Div(children='Select Event Type'), #, style={'float': 'left', 'clear': 'both', 'text-align': 'center'}),
html.Div([
dcc.Dropdown(
id='event_dropdown',
options=[{'label':ev,'value':ev} for ev in event_labels],
value='MergerAcquisition'
)],className='six columns'),
html.Div(children='Pick your date range'),#, style={'float': 'left', 'clear': 'both', 'text-align': 'center'}),
html.Div([
dcc.DatePickerRange(id='date-picker-range',
start_date=df.index[0],
end_date = df.index[-1]),
],className='six columns'),
# show number of articles in the time range selected
],style={'padding': 20,'margin': 5,'borderRadius': 5,'border': 'thin lightgrey solid','width' : '600'}),
]),
# full document view
html.Div(className="twelve columns",children=[
html.Div(id='full_document',children=[
html.Div(children=[
html.Div(children=[
html.H4(children='Labels'),
dcc.Checklist(
id='articlechecklist',
options=[{'label': label, 'value': label} for label in event_labels],
values=[]
),
html.Button('Update labels', id='updatelabelsbutton'),
html.Button('Previous article', id='prevarticlebutton'),
html.Button('Next article', id='nextarticlebutton'),
html.Button('Save to file', id='savetofilebutton'),
]),
html.Div(children=[
html.Iframe(id='fullarticleiframe',
style={'overflowY': 'scroll', 'height': 600, 'width': 1000,
'overflow': '-moz-scrollbars-vertical',
'padding': 5, 'margin': 0, 'borderRadius': 0, 'border': 'thin lightgrey solid'
})
])
],className="twelve columns",style={'padding': 20, 'margin': 5,
'borderRadius': 5,
'border': 'thin lightgrey solid'})
])
]),
# full document view
html.Div(className="twelve columns", children=[
html.Button('Retrain and predict', id='retrainbutton'),
],style={'padding': 20, 'margin': 5,'borderRadius': 5,'border': 'thin lightgrey solid'}),
html.P(id='placeholder'),
html.Div(id='intermediate-value', style = {'display':'none'}),
html.P(id='updated_df', children=[],style={'display': 'none'}),
html.P(id='updated_df2', children=[],style={'display': 'none'}),
html.P(id='updated_df3', children=[], style={'display': 'none'}),
html.P(id='updated_df4', children=[], style={'display': 'none'}),
html.P(id='updated_df5', children=[], style={'display': 'none'})
## end of layout
])
##############################################################################
# Define plotting functions
##############################################################################
def figure_num_articles(xlimrange=None,event_type=None):
#select event
df_sub = df[df[event_type]==1]
print('done')
traces = [go.Scatter(
x=df_sub['text'].resample('7D', how='count').index,
y=df_sub['text'].resample('7D', how='count'),
line={"color": "rgb(53, 83, 255)"},
mode="lines", # lines,markers
name="500 Index Fund Inv"
)]
layout = go.Layout(
xaxis={'title': 'publication date','range':xlimrange},
yaxis={'title': '# of articles'}, # , 'range': [20, 90]},
margin={'l': 40, 'b': 40, 't': 10, 'r': 10},
height= 170,
width= 600,
hovermode='closest'
)
return {'data': traces, 'layout': layout}
# uses df_pred to get tagged sentences and provide them with color
def get_tagged_sentences(file_id):
global df_pred,event_colors
import ast
df_selection = df_pred[file_id == df_pred['file_id']]
sent_tmp = df_selection['sentences'].values
lab_tmp = df_selection['label'].apply(ast.literal_eval).values
#print('found df selection ',df_selection.shape)
colors=[]
sentences=[]
labels=[]
for i,lab in enumerate(lab_tmp):
#print('search ',lab,event_colors)
#print(type(lab))
if type(lab) == int:
continue
if lab[0] in event_colors:
#print('####!!!!!!!!!!')
colors.append(event_colors[lab[0]])
sentences.append(sent_tmp[i])
labels.append(lab[0])
return sentences,colors,labels
##############################################################################
# Updates (callbacks)
##############################################################################
# update num docs figure based on parameters
@app.callback(
Output(component_id='numdocs_graph', component_property='figure'),
[Input('date-picker-range', 'start_date'),
Input('date-picker-range', 'end_date'),
Input('event_dropdown','value')])
def update_numdocs_figure(start_date,end_date,event_type):
return figure_num_articles(xlimrange=[start_date,end_date],event_type=event_type)
# update the article text view
@app.callback(
Output('fullarticleiframe', 'srcDoc'),
[Input('prevarticlebutton','n_clicks'),
Input('nextarticlebutton','n_clicks'),
Input('event_dropdown', 'value')],
[State('updatelabelsbutton','n_clicks')])
def update_shown_articletext(n_clicks_prev,n_clicks_next,event_type,n_clicks_updatelabels):
global current_file_id,df_mod
# check if need to use modified df
if n_clicks_updatelabels is None:
df_show = df
print('click updates alreadY?')
else:
df_show = df_mod
print('take modified dataset')
if not event_type is None:
df_show = df_show[df_show[event_type]==1.]
# select df based on clicks
if n_clicks_prev is None:
n_clicks_prev=0
if n_clicks_next is None:
n_clicks_next=0
n_clicks = n_clicks_next-n_clicks_prev
print(n_clicks_prev,n_clicks_next,n_clicks)
len_data_entries = len(df_show)
if n_clicks>=len_data_entries:
n_clicks = len_data_entries
idx = n_clicks - 1
current_file_id = df_show.iloc[idx]['file_id']
print('set current_file_id',current_file_id)
lines,colors,labels=get_tagged_sentences(current_file_id)
print('lines1',lines,len(lines))
print('colors1',colors,len(colors))
return render_document(title=df_show.iloc[idx]['title'], body=df_show.iloc[idx]['text'],
id=str(current_file_id),date=df_show.iloc[idx]['date'],
lines=lines, colors=colors,labels=labels)
# show article labels for article considered
@app.callback(
dash.dependencies.Output('articlechecklist', 'values'),
[Input('fullarticleiframe','srcDoc')],
#[Input('prevarticlebutton', 'n_clicks'),
#Input('nextarticlebutton', 'n_clicks')],
[State('updatelabelsbutton', 'n_clicks')])
def update_shown_labels(srcDocs,n_clicks_updatelabels):
global current_file_id,df_mod
print('enter update_shown_labels')
# check if need to use modified df
if n_clicks_updatelabels is None:
print('click updates alreadY?')
df_show = df
else:
print('take modified dataset')
df_show = df_mod
row = df_show.loc[df_show['file_id']==current_file_id]
if row.empty:
print('error row is empty for id ',current_file_id)
print('check row ',row[event_labels].values, 'for id ',current_file_id
)
active_labels=[]
for label in event_labels:
if row[label].values[0] == 1.:
active_labels.append(label)
print('found labels ',active_labels)
return active_labels
# update df-mod with labels submitted when update button is clicked
@app.callback(
Output('updated_df', 'children'), #updated_df, placeholder
[Input('updatelabelsbutton', 'n_clicks')],
[State('articlechecklist','values')])
def update_article_label(n_clicks,checklistvals):
global df_mod
global current_file_id
#print('executed!! ')
#print('number clicks:',n_clicks)
#print('checklinkvals',checklistvals)
# modify df_mod
df_mod[df_mod['file_id'] == current_file_id][event_labels] = 0 # set all 0 first
for label in checklistvals:
print('set label ',label)
df_mod.loc[df_mod['file_id']==current_file_id,label] = 1
print('labels ',checklistvals,' were updated for id ',current_file_id)
# save df_mod to file
@app.callback(
Output('updated_df2', 'children'), #updated_df, placeholder
[Input('savetofilebutton', 'n_clicks')])
def save_updated_labelsfile(n_clicks):
global df_mod,input_file
import time
filename=input_file[:-3]+str(int(time.time()))+'.csv'
print('save to ',filename)
df_mod.to_csv(filename)
# update file list once df_mod is saved
@app.callback(
Output('file_dropdown', 'options'), # updated_df, placeholder
[Input('updated_df2', 'children')])
def save_updated_labelsfile(vals):
return [{'label':f,'value':f} for f in sorted([f for f in os.listdir('.') if (os.path.isfile(f) and '.csv' in f)],key=lambda x: os.path.getmtime(x))]
#return [{'label': file, 'value': file} for file in [f for f in os.listdir('.') if (os.path.isfile(f) and 'csv' in f)]]
# read file selected in dropdown menu
@app.callback(
Output('updated_df4', 'children'), # updated_df, placeholder
[Input('file_dropdown', 'value')])
def load_file(file_to_load):
global df,df_mod
print('read file into df: ',file_to_load)
df = pd.read_csv(file_to_load)
df = df.dropna()
df = index_to_time(df)
df = df.sort_index()
# for modified labels
df_mod = df.copy()
# read file selected in dropdown menu
#@app.callback(
# Output('updated_df5', 'children'), # updated_df, placeholder
# [Input('retrainbutton', 'value'),
# Input('file_dropdown','file_to_load')])
#def retraining(value,filename):
# start training based on filename and
##############################################################################
# Execute main
##############################################################################
if __name__ == '__main__':
app.run_server(debug=True)