class Classifier:
def __init__(self):
self.bag_of_words = [
'modi', 'pm', 'visit', 'narendra', 'prime minister'
]
self.db_con = DBConnection().create_connection()
self.cursor = self.db_con.cursor()
def classify_title(self):
sql = "SELECT id, title FROM news_url WHERE id > %s ORDER BY id ASC "
sql_update = "UPDATE news_url SET is_valid = True WHERE id = %s"
news_url_id = self.get_last_news_url_id()
counter = 0
while True:
self.cursor.execute(sql, (news_url_id, ))
rs_tuple = self.cursor.fetchone()
if not rs_tuple:
break
(news_url_id, title) = rs_tuple
print(news_url_id)
for key_word in self.bag_of_words:
if re.search(key_word, title.strip(), re.I):
self.cursor.execute(sql_update, (news_url_id, ))
self.db_con.commit()
counter += 1
if counter % 1000 == 0:
print('Sleeping for 5 seconds')
time.sleep(5)
self.update_last_news_url_id(news_url_id)
def update_last_news_url_id(self, news_url_id):
sql = "UPDATE scraper_info SET news_url_id = %s WHERE id = 1"
self.cursor.execute(sql, (news_url_id, ))
self.db_con.commit()
def get_last_news_url_id(self):
sql = "SELECT news_url_id FROM scraper_info"
self.cursor.execute(sql)
rs_tuple = self.cursor.fetchone()
(news_url_id, ) = rs_tuple
return news_url_id
class Classifier:
def __init__(self):
self.bag_of_words = [
'modi', 'pm', 'visit', 'narendra', 'prime minister'
]
self.bag_of_words_set = set(self.bag_of_words)
self.db_con = DBConnection().create_connection()
self.cursor = self.db_con.cursor()
def classify_title(self):
sql = "SELECT id, title FROM news_url WHERE id > %s ORDER BY id ASC "
sql_update = "UPDATE news_url SET is_valid = True WHERE id = %s"
news_url_id = self.get_last_news_url_id()
counter = 0
while True:
self.cursor.execute(sql, (news_url_id, ))
rs_tuple = self.cursor.fetchone()
if not rs_tuple:
break
(news_url_id, title) = rs_tuple
print(news_url_id)
stemmer = PorterStemmer()
result_set = None
title = title.lower()
title_set = set([stemmer.stem(item) for item in title.split()])
#print (title)
result_set = (title_set & self.bag_of_words_set)
print("result_set", result_set)
'''if result_set:
self.cursor.execute(sql_update, (news_url_id, ))
self.db_con.commit()'''
#for item in title_set:
# if item.lower() in COUNTRY:
# # print(item)
# if result_set:
# self.cursor.execute(sql_update, (news_url_id, ))
# self.db_con.commit()
# else:
# if self.get_place(item) and result_set:
# self.cursor.execute(sql_update, (news_url_id, ))
# self.db_con.commit()
counter += 1
if counter % 1000 == 0:
print('Sleeping for 5 seconds')
time.sleep(5)
#self.update_last_news_url_id(news_url_id)
def update_last_news_url_id(self, news_url_id):
sql = "UPDATE scraper_info SET news_url_id = %s WHERE id = 1"
self.cursor.execute(sql, (news_url_id, ))
self.db_con.commit()
def get_last_news_url_id(self):
sql = "SELECT news_url_id FROM scraper_info"
self.cursor.execute(sql)
rs_tuple = self.cursor.fetchone()
(news_url_id, ) = rs_tuple
return news_url_id
def get_place(self, text):
chunked = ne_chunk(pos_tag(word_tokenize(text)))
for i in chunked:
if type(i) == Tree:
if i.label() == "GPE":
for token, pos in i.leaves():
#print("found country %s " %token)
return True
return False
class Xtractor:
def __init__(self):
self.year = 2015
self.base_url = "http://timesofindia.indiatimes.com/%s"
self.url_to_scrape = None
self.db_con = DBConnection().create_connection()
self.cursor = self.db_con.cursor()
self.request_headers = {
"User-Agent":
"Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/55.0.2883.87 Safari/537.36"
}
def save_last_month_date_url(self, month_date_url_id):
sql_scraper_info = "UPDATE scraper_info SET last_month_date_url_id = %s WHERE id=1"
self.cursor.execute(sql_scraper_info, (month_date_url_id, ))
self.db_con.commit()
def get_url_to_scrape(self):
month_date_url_id = 0
sql_last_month_date_url = "SELECT last_month_date_url_id FROM scraper_info WHERE id=1"
sql_get_month_date_url = """SELECT month_date_url ,id
FROM month_date_url
%s """
self.cursor = self.db_con.cursor()
self.cursor.execute(sql_last_month_date_url)
rs_tuple = self.cursor.fetchone()
(month_date_url_id, ) = rs_tuple
if rs_tuple:
(month_date_url_id, ) = rs_tuple
condition = 'WHERE id > %s ORDER BY id asc' % (month_date_url_id)
self.cursor.execute(sql_get_month_date_url % condition)
rs_tuple = self.cursor.fetchone()
if not rs_tuple:
raise NoMoreUrlFoundException(
"No URL left to scrape...you are done scraping all the urls present in DB"
)
(month_date_url, month_date_url_id) = rs_tuple
else:
self.cursor.execute(sql_get_month_date_url % ('ORDER BY id asc'))
rs_tuple = self.cursor.fetchone()
(month_date_url, month_date_url_id) = rs_tuple
return (month_date_url_id, month_date_url)
def start(self):
while True:
try:
time.sleep(5)
(month_date_url_id, month_date_url) = self.get_url_to_scrape()
print(month_date_url)
url_to_open = self.base_url % month_date_url
print(url_to_open)
html = requests.get(url_to_open,
headers=self.request_headers).text
self.parse_titles(html, month_date_url_id)
except NoMoreUrlFoundException:
print("------------------DONE------------------------")
sys.exit(0)
return
def save_tile_and_href(self, month_date_url_id, title, href):
sql = "INSERT INTO news_url (month_date_url_id, title, url) VALUES (%s, %s, %s) ON CONFLICT (title) DO NOTHING"
self.cursor.execute(sql, (month_date_url_id, title, href))
self.db_con.commit()
return
def parse_titles(self, html, month_date_url_id):
soup = BeautifulSoup(html, 'lxml')
div = soup.find(
'div', {
'style':
'font-family:arial ;font-size:12;font-weight:bold; color: #006699'
})
table = div.find('table')
atag_list = table.find_all('a')
with open("/home/aish/Desktop/toi/toi.html", 'w') as f:
f.write(str(table))
for a in atag_list:
title = a.text.strip()
href = a['href']
self.save_tile_and_href(month_date_url_id, title, href)
self.save_last_month_date_url(month_date_url_id)
return
def initiate(self):
sql_insert_month_date_url = "INSERT INTO month_date_url (url_month, url_date, year, month_date_url) VALUES (%s, %s, %s, %s)"
toi_starttime = 42005
month_counter = 1
date_counter = 1
# Add all url of all month date to database
while True:
# Get number of days in a month depending on the year
num_of_days_in_month = calendar.monthrange(self.year,
month_counter)[1]
while True:
toi_url_month_date = "/%s/1/1/archivelist/year-%s,month-%s,starttime-%s.cms" % (
self.year, self.year, month_counter, toi_starttime)
self.cursor.execute(sql_insert_month_date_url,
(month_counter, date_counter, self.year,
toi_url_month_date))
self.db_con.commit()
toi_starttime += 1
date_counter += 1
print(toi_url_month_date)
if date_counter > num_of_days_in_month:
# Reset the date counter to 1 before break for each month
date_counter = 1
break
month_counter += 1
# Break when December is reached
if month_counter > 12:
break
'''
#Stage 4
# keyword specific seach
if re.search("arrives|attend|embarks|reaches", content,re.DOTALL):
filter_count = filter_count + 1
cursor.execute(sql, (filter_count, id))
print (">>>>>>>>>>>>>>>>>",id)
'''
#dbcon.commit()
requests
sql_news_title = "SELECT n.title, n.id, n.url FROM news_url n INNER JOIN classified_article c ON c.news_url_id = n.id WHERE c.counter > 3 "
sql_insert = "INSERT INTO visit_info(place, visit_date) values(%s, %s) ON CONFLICT (place, visit_date) DO NOTHING"
cursor.execute(sql_news_title)
rs_tuple_list = cursor.fetchall()
for rs_tuple in rs_tuple_list:
(title, id, url) = rs_tuple
lst_ner = get_continuous_chunks(title)
country_list = []
for item in lst_ner:
if re.search(regex, item.strip().lower()):
country_list.append(item)
html = requests.get(url).text
soup = BeautifulSoup(html, 'lxml')
if country_list:
for country in country_list:
date = soup.find('span', {'class': "time_cptn"}).text
date_reg = re.search(r'(.*\|)(.*2015)(.*)', date)
date_str = date_reg.group(2)
cursor.execute(sql_insert, (country, date_str))
dbcon.commit()
