forked from je-suis-tm/web-scraping
-
Notifications
You must be signed in to change notification settings - Fork 0
/
MENA Newsletter.py
565 lines (427 loc) · 16.9 KB
/
MENA Newsletter.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
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
# coding: utf-8
#this script is about the latest news of MENA region
#we scrape different influential media websites, or so-called fake news, lol
#and send only updates to the mailbox for daily newsletter
#in order to do that, we need a db to store all the historical content of websites
#and all the scraping techniques from html parse tree to regular expression
#over time, i also discovered the issue of information overload in daily newsletter
#hence, i invented a graph theory based algorithm to extract key information
#a part of this algo will also be featured in this script to solve info redundancy
#as u can see, this is the most advanced script in web scraping repository
#it contains almost every technique we have introduced so far
#make sure you have gone through all the other scripts before moving onto this one
import pandas as pd
from bs4 import BeautifulSoup as bs
import requests
import datetime as dt
import win32com.client as win32
import sqlite3
import os
import re
import copy
import time
os.chdir('d:/')
#this is a home made special package for text mining
#it is designed to extract key information and remove similar contents
#for details of this graph traversal algorithm plz refer to the following link
# https://github.com/je-suis-tm/graph-theory/blob/master/Text%20Mining%20project/alternative%20bfs.py
import graph
#main stuff
def main():
ec=scrape('https://www.economist.com/middle-east-and-africa/',economist)
aj=scrape('https://www.aljazeera.com/topics/regions/middleeast.html',aljazeera)
tr=scrape('https://www.reuters.com/news/archive/middle-east',reuters)
bc=scrape('https://www.bbc.co.uk/news/world/middle_east',bbc)
ws=scrape('https://www.wsj.com/news/types/middle-east-news',wsj)
ft=scrape('https://www.ft.com/world/mideast',financialtimes)
bb=scrape('https://www.bloomberg.com/view/topics/middle-east',bloomberg)
cn=scrape('https://edition.cnn.com/middle-east',cnn)
fo=scrape('https://fortune.com/tag/middle-east/',fortune)
#concat scraped data via append, can use pd.concat as an alternative
#unlike the previous version, current version does not sort information by source
#the purpose of blending data together is to go through text mining pipeline
df=ft
for i in [aj,tr,bc,ws,cn,fo,ec,bb]:
df=df.append(i)
#CRUCIAL!!!
#as we append dataframe together, we need to reset the index
#otherwise, we would not be able to use reindex in database function call
df.reset_index(inplace=True,drop=True)
#first round, insert into database and remove outdated information
df=database(df)
#second round, use home made package to remove similar contents
output=graph.remove_similar(df,graph.stopword)
#if the link is not correctly captured
#remove anything before www and add https://
for i in range(len(output)):
if 'https://' not in output['link'][i]:
temp=re.search('www',output['link'][i]).start()
output.at[i,'link']='http://'+output['link'][i][temp:]
print(output)
#using html email template
#check stripo for different templates
# https://stripo.email/templates/
html="""
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"
"http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
<html>
<head>
<meta charset="UTF-8">
<meta content="width=device-width, initial-scale=1" name="viewport">
<meta name="x-apple-disable-message-reformatting">
<meta http-equiv="X-UA-Compatible" content="IE=edge">
<meta content="telephone=no" name="format-detection">
<title></title>
<!--[if (mso 16)]>
<style type="text/css">
a {text-decoration: none;}
</style>
<![endif]-->
<!--[if gte mso 9]><style>sup
{ font-size: 100% !important; }</style><![endif]-->
</head>
<body>
<div class="es-wrapper-color">
<!--[if gte mso 9]>
<v:background xmlns:v="urn:schemas-microsoft-com:vml"
fill="t">
<v:fill type="tile" color="#333333"></v:fill>
</v:background>
<![endif]-->
<table class="es-content-body" width="600"
cellspacing="15" cellpadding="15" bgcolor="#ffffff"
align="center">
<tr>
<td class="esd-block-text" align="center">
<h2>Middle East</h2></td>
</tr></table>
<div><br></div>
"""
#there are a few ways for embed image in html email
#here, we use the link of the image
#it may be a lil bit slow to load the image but its the most efficient way
#alternatively, we can use mail.Attachments.add()
#we attach all images, and set <img src='cid: imagename.jpg'>
#the downside is that we have to scrape the website repeatedly to get images
#or we can use < img src='data:image/jpg; base64, [remove the brackets and paste base64]'/>
#but this is blocked by most email clients including outlook 2016
for i in range(len(output)):
html+="""<table class="es-content-body" width="600"
cellspacing="10" cellpadding="5" bgcolor="#ffffff"
align="center">"""
html+="""<tr><td class="esd-block-text es-p10t es-p10b"
align="center"><p><a href="%s">
<font color="#6F6F6F">%s<font><a></p></td></tr>
<tr><td align="center">
<img src="%s" width="200" height="150"/></td></tr>
<tr>"""%(output['link'][i],output['title'][i],output['image'][i])
html+="""</tr></table><div><br></div>"""
html+="""
</div>
</body>
</html>
"""
send(html)
#i use win32 to control outlook and send emails
#when you have a win 10 pro, it is the easiest way to do it
#cuz windows pro automatically launches outlook at startup
#otherwise, there is a library called smtp for pop3/imap server
#supposedly authentication of corporate email would kill u
#i definitely recommend folks to use win32 library
#note that using win32.email requires outlook to stay active
#do not close the app until u actually send out the email
#win32 library uses COM api to control windows
#go to microsoft developer network
#check mailitem object model to learn how to manipulate outlook emails
#the website below is the home page of outlook vba reference
# https://msdn.microsoft.com/en-us/vba/vba-outlook
def send(html):
#create an email with recipient, subject, context and attachment
outlook = win32.Dispatch('outlook.application')
mail = outlook.CreateItem(0)
#these email addresses are fabricated, PLZ DO NOT HARASS OUR GODDESS
#just some random pornstar i love
receivers = ['lana.rhodes@brazzers.com',
'tori.black@brazzers.com',
'naomi.woods@brazzers.com']
#use ';' to separate receipients
#this is a requirement of outlook
mail.To = ';'.join(receivers)
mail.Subject ='Mid East Newsletter %s'%(dt.datetime.now())
mail.BodyFormat=2
#use html to make email looks more elegant
#html is very simple
#use br for line break, b for bold fonts
#font for color and size, a href for hyperlink
#check the website below to see more html tutorials
# https://www.w3schools.com/html/
#Alternatively, we can use plain text email
#remember to use '\r\n' to jump line
#assuming html is a list of str
#the code should be mail.Body = '\r\n'.join(html)
mail.HTMLBody=html
#i usually print out everything
#need to check carefully before sending to stakeholders
#we can use mail.Display() to see the draft instead
condition=str(input('0/1 for no/yes:'))
if condition=='1':
mail.Send()
print('\nSENT')
else:
print('\nABORT')
return
#database insertion and output the latest feeds
#i assume you are familiar with sqlite3
#if not, plz check the following link
# https://github.com/je-suis-tm/web-scraping/blob/master/LME.py
def database(df):
temp=[]
conn = sqlite3.connect('mideast_news.db')
c = conn.cursor()
#the table structure is simple
#the table name is new
#there are three columns, title, link and image
#the data types of all of them are TEXT
#title is the primary key which forbids duplicates
for i in range(len(df)):
try:
c.execute("""INSERT INTO news VALUES (?,?,?)""",df.iloc[i,:])
conn.commit()
print('Updating...')
#the idea is very simple
#insert each line from our scraped result into database
#as the primary key has been set up
#we have non-duplicate title constraint
#insert what has already been in database would raise an error
#if so, just ignore the error and pass to the next iteration
#we can utilize the nature of database to pick out the latest information
#every successful insertion into the database also goes to the output
#at the end, output contains nothing but latest updates of websites
#that is what we call newsletter
temp.append(i)
except Exception as e:
print(e)
conn.close()
#check if the output contains no updates
if temp:
output=df.loc[[i for i in temp]]
output.reset_index(inplace=True,drop=True)
else:
output=pd.DataFrame()
output['title']=['No updates yet.']
output['link']=output['image']=['']
return output
#scraping webpages and do some etl
def scrape(url,method):
print('scraping webpage effortlessly')
time.sleep(5)
session=requests.Session()
response = session.get(url,headers={'User-Agent': 'Mozilla/5.0'})
page=bs(response.content,'html.parser',from_encoding='utf_8_sig')
df=method(page)
out=database(df)
return out
"""
the functions below are data etl of different media sources
"""
#the economist etl
def economist(page):
title,link,image=[],[],[]
df=pd.DataFrame()
prefix='https://www.economist.com'
a=page.find_all('div',class_="topic-item-container")
for i in a:
link.append(prefix+i.find('a').get('href'))
title.append(i.find('a').text)
image.append(i.parent.find('img').get('src'))
df['title']=title
df['link']=link
df['image']=image
return df
#fortune etl
def fortune(page):
title,link,image=[],[],[]
df=pd.DataFrame()
prefix='https://fortune.com'
a=page.find_all('article')
for i in a:
link.append(prefix+i.find('a').get('href'))
if 'http' in i.find('img').get('src'):
image.append(i.find('img').get('src'))
else:
image.append('')
temp=re.split('\s*',i.find_all('a')[1].text)
temp.pop()
temp.pop(0)
title.append(' '.join(temp))
df['title']=title
df['link']=link
df['image']=image
return df
#cnn etl
def cnn(page):
title,link,image=[],[],[]
df=pd.DataFrame()
prefix='https://edition.cnn.com'
a=page.find_all('div', class_='cd__wrapper')
for i in a:
title.append(i.find('span').text)
link.append(prefix+i.find('a').get('href'))
try:
image.append('https:'+i.find('img').get('data-src-medium'))
except:
image.append('')
df['title']=title
df['link']=link
df['image']=image
return df
#bloomberg etl
def bloomberg(page):
title,link,image=[],[],[]
df=pd.DataFrame()
prefix='https://www.bloomberg.com'
a=page.find_all('h1')
for i in a:
try:
link.append(prefix+i.find('a').get('href'))
title.append(i.find('a').text.replace('’','\''))
except:
pass
b=page.find_all('li')
for j in b:
try:
temp=j.find('article').get('style')
image.append( \
re.search('(?<=url\()\S*(?=\))', \
temp).group() \
)
except:
temp=j.find('article')
try:
temp2=temp.get('id')
if not temp2:
image.append('')
except:
pass
df['title']=title
df['link']=link
df['image']=image
return df
#financial times etl
def financialtimes(page):
title,link,image=[],[],[]
df=pd.DataFrame()
prefix='https://www.ft.com'
a=page.find_all('a',class_='js-teaser-heading-link')
for i in a:
link.append(prefix+i.get('href'))
temp=i.text.replace('’','\'').replace('‘','\'')
title.append(temp.replace('\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t',''))
for j in a:
temp=j.parent.parent.parent
try:
text=re.search('(?<=")\S*(?=next)',str(temp)).group()
image.append(text+'next&fit=scale-down&compression=best&width=210')
except:
image.append('')
df['title']=title
df['link']=link
df['image']=image
return df
#wall street journal etl
def wsj(page):
df=pd.DataFrame()
text=str(page)
link=re.findall('(?<=headline"> <a href=")\S*(?=">)',text)
image=re.findall('(?<=img data-src=")\S*(?=")',text)
title=[]
for i in link:
try:
temp=re.search('(?<={}")>(.*?)<'.format(i),text).group()
title.append(temp)
except:
pass
for i in range(len(title)):
title[i]=title[i].replace('’',"'").replace('<','').replace('>','')
df['title']=title
df['link']=link[:len(title)]
df['image']=image+[''] if (len(image)!=len(title)) else image
return df
#bbc etl
def bbc(page):
title,link,image=[],[],[]
df=pd.DataFrame()
prefix='https://www.bbc.co.uk'
a=page.find_all('span',class_='title-link__title-text')
for i in a:
temp=i.parent.parent.parent.parent
b=(re.findall('(?<=src=")\S*(?=jpg)',str(temp)))
if len(b)>0:
b=copy.deepcopy(b[0])+'jpg'
else:
b=''
image.append(b)
for j in a:
title.append(j.text)
for k in a:
temp=k.parent.parent
c=re.findall('(?<=href=")\S*(?=">)',str(temp))
link.append(prefix+c[0])
df['title']=title
df['link']=link
df['image']=image
return df
#thompson reuters etl
def reuters(page):
title,link,image=[],[],[]
df=pd.DataFrame()
prefix='https://www.reuters.com'
for i in page.find('div', class_='news-headline-list').find_all('h3'):
temp=i.text.replace(' ','')
title.append(temp.replace('\n',''))
for j in page.find('div', class_='news-headline-list').find_all('a'):
link.append(prefix+j.get('href'))
link=link[0::2]
for k in page.find('div', class_='news-headline-list').find_all('img'):
if k.get('org-src'):
image.append(k.get('org-src'))
else:
image.append('')
df['title']=title
df['link']=link
df['image']=image
return df
#al jazeera etl
def aljazeera(page):
title,link,image=[],[],[]
df=pd.DataFrame()
prefix='https://www.aljazeera.com'
a=page.find_all('div',class_='frame-container')
for i in a:
title.append(i.find('img').get('title'))
image.append(prefix+i.find('img').get('src'))
temp=i.find('a').get('href')
link.append(temp if 'www' in temp else (prefix+temp))
b=page.find_all('div',class_='col-sm-7 topics-sec-item-cont')
c=page.find_all('div',class_='col-sm-5 topics-sec-item-img')
limit=max(len(b),len(c))
j,k=0,0
while j<limit:
title.append(b[j].find('h2').text)
temp=b[j].find_all('a')[1].get('href')
link.append(temp if 'www' in temp else (prefix+temp))
#when there is an opinion article
#the image tag would change
#terrible website
if 'opinion' in b[j].find('a').get('href'):
image.append(' ')
else:
image.append(prefix+c[k].find_all('img')[1].get('data-src'))
k+=1
j+=1
df['title']=title
df['link']=link
df['image']=image
return df
if __name__ == "__main__":
main()