class scrappers:
data_path = "././data/raw"
def __init__(self):
__dir_path = os.path.dirname(os.path.realpath(__file__))
credentials = get_credidentials()
self.twitter_premium_api = load_credentials(
filename="{}/{}".format(__dir_path, "twitter_keys.yaml"),
yaml_key="search_tweets_api_30day")
self.twitter_api = Twitter(auth=OAuth(
consumer_key=credentials['twitter']['consumer_key'],
consumer_secret=credentials['twitter']['consumer_secret'],
token=credentials['twitter']['access_token_key'],
token_secret=credentials['twitter']['access_token_secret']))
self.yelp_api = YelpAPI(credentials['yelp']['api_key'])
self.__data_path = "../data/raw"
logger.info("initiation started.")
def tw_verify_credentials(self):
obj = self.twitter_api.VerifyCredentials()
print(json.dumps(obj._json, indent=4, sort_keys=True))
def tw_get_statuses(self, user_list):
for username in user_list:
with open(f'datasets/tw_{username}_statuses.json', 'w') as f:
try:
f.write('{"statuses": [')
max_id = 0
while (True):
# status scheme available at: https://developer.twitter.com/en/docs/tweets/timelines/api-reference/get-statuses-user_timeline.html
statuses = self.twitter_api.GetUserTimeline(
screen_name=username, count=100, max_id=max_id)
if len(statuses) == 1 and statuses[0].id == max_id:
break
else:
for status in statuses:
if status.id != max_id:
f.write("%s," % json.dumps(status._json))
max_id = statuses[-1].id
finally:
max_id != 0 and f.seek(f.tell() - 1, os.SEEK_SET)
f.write("]}")
def tw_get_search(self, user_list):
for user_name, keyword_list in user_list.items():
with open(f'datasets/tw_{user_name}_searches.json', 'w') as f:
try:
f.write('{"statuses": [')
max_id = 0
user = self.twitter_api.GetUser(screen_name=user_name)
keyword_list.append(f'{user.name}')
keyword_list.append(f'{user_name}')
keyword_list.append(f'#{user_name}')
keyword_list.append(f'@{user_name}')
term = '{}'.format(' OR '.join(keyword_list))
while (True):
# status scheme available at: https://developer.twitter.com/en/docs/tweets/search/api-reference/get-search-tweets.html
statuses = self.twitter_api.GetSearch(
term=term.encode('utf-8'),
geocode=None,
count=100,
max_id=max_id)
if (len(statuses) == 1 and statuses[0].id
== max_id) or statuses == []:
break
else:
for status in statuses:
if status.id != max_id:
"""status_text = json.dumps(status._json)
status_json = json.loads(status_text)
status_json['keyword'] = keyword"""
f.write("%s," % json.dumps(status._json))
max_id = statuses[-1].id
finally:
max_id != 0 and f.seek(f.tell() - 1, os.SEEK_SET)
f.write("]}")
def tw_get_premium_search(self, keyword: str):
with open(f'datasets/tw_{keyword.lower()}_searches_premium.json',
'w') as f:
try:
f.write('{"statuses": [')
rule = gen_rule_payload(
pt_rule="near:\"New York, NY\" within:50mi".format(),
results_per_call=100,
from_date="2018-07-01",
to_date="2018-10-01")
rule = gen_rule_payload(
pt_rule="place:\"New York, NY\"".format(),
results_per_call=100,
from_date=(datetime.date.today() -
datetime.timedelta(31)).isoformat(),
to_date=datetime.date.today().isoformat())
next_token = None
while True:
results = ResultStream(rule_payload=rule,
**self.twitter_premium_api)
results.next_token = next_token
tweets = []
try:
tweets = list(results.stream())
except Exception as ex:
print(str(ex))
for tweet in tweets:
f.write("%s," % json.dumps(tweet))
if results.next_token is None:
break
else:
next_token = results.next_token
next_token is not None and f.seek(f.tell() - 1, os.SEEK_SET)
f.write("]}")
except Exception as ex:
print("Error:\n" + str(ex))
def yp_get_businesses(self, business_list):
"""
Get reviews for each business in the business_list and creates separate data files.
File Type: JSON
"""
for business in business_list:
with open(f'{self.data_path}/yp_{business}_competitors.json',
'w') as f:
try:
f.write('{"businesses": [')
branch = self.yelp_api.business_query(business)
offset = 0
while (True):
try:
# status scheme available at: # https://www.yelp.com/developers/documentation/v3/business_search
competitors = self.yelp_api.search_query(
longitude=branch['coordinates']['longitude'],
latitude=branch['coordinates']['latitude'],
radius=40000,
# categories='bars,french'
sort_by='distance',
limit=50,
offset=offset)
f.write("%s," %
json.dumps(competitors['businesses']))
offset = offset + 50
except self.yelp_api.YelpAPIError:
break
finally:
offset != 0 and f.seek(f.tell() - 1, os.SEEK_SET)
f.write("]}")
def yp_get_competitors(self, business_list):
"""
Gets business list in consideration to the existing business list file. Adds any additional business, if it is not recorded yet.
"""
file_path = fp.yp_raw_competitors(self.data_path)
index_list = []
existing_list = []
"""
if os.path.exists(file_path):
with open(file_path, 'r') as f:
current_file = f.readlines()
if len(current_file) > 0:
existing_list = json.loads(current_file[0])
index_list = [_business["alias"] for _business in existing_list]
logger.info(f"existing file found: {len(index_list)} total entries")
"""
with open(file_path, 'w') as f:
# find businesses
for business in business_list:
new_list = []
try:
logger.info(f"import started for : {business}")
branch = self.yelp_api.business_query(business)
offset = 0
while (True):
try:
# status scheme available at: # https://www.yelp.com/developers/documentation/v3/business_search
competitors = self.yelp_api.search_query(
longitude=branch['coordinates']['longitude'],
latitude=branch['coordinates']['latitude'],
radius=40000,
# categories='bars,french'
sort_by='distance',
limit=50,
offset=offset)
# add alias name for distance measurement as dist_to_alias
businesses = competitors["businesses"]
[
i.update({"dist_to_alias": business})
for i in businesses
]
for i in businesses:
if i['alias'] not in index_list:
new_list.append(i)
index_list.append(i['alias'])
offset = offset + 50
except self.yelp_api.YelpAPIError:
break
finally:
existing_list.extend(new_list)
logger.info(
f"import completed. existing: {len(existing_list)} new: {len(new_list)}"
)
# saving into file
json.dump(existing_list, f)
def yp_get_business_reviews(self, business_list):
"""
Gets three reviews from the yelp api.
"""
for business in business_list:
with open(f'{self.data_path}/yp_{business}_rws.json', 'w') as f:
try:
f.write('{"reviews": [')
offset = 0
while (True):
reviews_set = self.yelp_api.reviews_query(
business, limit=5, offset=offset)
reviews = reviews_set['reviews']
if len(reviews) > 0:
for review in reviews:
f.write("%s,\n" % review)
offset = offset + 5
else:
break
finally:
offset != 0 and f.seek(f.tell() - 1, os.SEEK_SET)
f.write("]}")
def yp_get_competitor_reviews(self,
business_list=None,
start_index=0,
end_index=5):
"""
Gets reviews by scraping through the site. Reviews are saved by business name and reviews. Uses Competitors reviews file as default file. Given index controls regions of Competitors.
business_list: None or List
start_index: int, interested region's starting index
end_index: int, interested region's ending index
File Type: CSV
"""
file_path = fp.yp_raw_competitors_reviews(self.data_path)
headers = {
'User-Agent':
'Mozilla/5.0 (Macintosh; Intel Mac OS X 10_10_1) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/39.0.2171.95 Safari/537.36'
}
columns = [
'alias', 'ratingValue', 'dataPublished', 'description', 'author'
]
df: pd.DataFrame
# getting competitors list
businesses_file_path = fp.yp_raw_competitors(self.data_path)
businesses_index_list = []
if os.path.exists(businesses_file_path):
with open(businesses_file_path, 'r') as f:
current_file = f.readlines()
if len(current_file) > 0:
businesses_index_list = [
_business["alias"]
for _business in json.loads(current_file[0])
]
# needed every time
if os.path.exists(file_path):
with open(file_path, 'r') as f:
df = pd.read_csv(file_path)
logger.info(
f"existing file found. total reviews count: {len(df)}")
# need only once, if file doesn't exists
if os.path.exists(file_path) is False:
with open(file_path, 'w') as f:
writer = csv.writer(f)
writer.writerow(columns)
logger.info("file created at: {}".format(file_path))
# ops
with open(file_path, 'a', newline='') as f:
if business_list is None:
business_list = businesses_index_list
current_index = start_index - 1
for business in business_list[start_index:end_index]:
cnt_imported = 0
current_index = current_index + 1
logger.info(f"index: {current_index} of {end_index - 1}")
try:
writer = csv.writer(f)
logger.info(f"import started for : {business}")
start = 0
cnt_requests = 0
while (True):
url = '{}/{}?sort_by=date_desc&start={}'.format(
'https://www.yelp.com/biz', business, start)
response = requests.get(url, headers)
soup = BeautifulSoup(response.text, 'html.parser')
html_script = soup.findAll(
'script', {'type': 'application/ld+json'})[-1]
obj = json.loads(html_script.string)
reviews = obj['review']
if len(reviews) > 0:
for review in reviews:
data = [
business,
review['reviewRating']['ratingValue'],
review['datePublished'],
review['description'], review['author']
]
check = np.array(data, dtype='O')
if not (df.values == check).all(1).any():
writer.writerow(data)
cnt_imported = cnt_imported + 1
start = start + 20
cnt_requests = cnt_requests + 1
else:
logger.info(
f"import completed. total reviews cnt: {cnt_imported} total request cnt: {cnt_requests}"
)
break
except Exception as ex:
logger.warning(
f"error: alias: {business} index: {current_index} total reviews cnt: {cnt_imported}"
)
logger.warning(f"error message: {ex}")
logger.warning("Let me sleep for some time..")
second = int(round(random.expovariate(1) * 100))
time.sleep(second)
logger.warning(
f"{second} seconds slept, now back on scrapping..")
continue
def yp_get_business_reviews2(self, business_list):
"""
Gets reviews by scraping through the site.
"""
for business in business_list:
headers = {
'User-Agent':
'Mozilla/5.0 (Macintosh; Intel Mac OS X 10_10_1) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/39.0.2171.95 Safari/537.36'
}
with open(f'{self.data_path}/yp_{business}_rws.json', 'w') as f:
try:
f.write('{"reviews": [')
start = 0
while (True):
url = '{}/{}?sort_by=date_desc&start={}'.format(
'https://www.yelp.com/biz', business, start)
response = requests.get(url, headers)
soup = BeautifulSoup(response.text, 'html.parser')
html_script = soup.find(
'script', {'type': 'application/ld+json'})
obj = json.loads(html_script.string)
reviews = obj['review']
if len(reviews) > 0:
for review in reviews:
data = {
'ratingValue':
review['reviewRating']['ratingValue'],
'datePublished':
review['datePublished'],
'description':
review['description'],
'author':
review['author']
}
f.write("%s," % json.dumps(data))
start = start + 20
else:
break
finally:
start != 0 and f.seek(f.tell() - 1, os.SEEK_SET)
f.write("]}")
with open(f'datasets/yp_businesses.json', 'a') as f:
obj['review'] = []
