def main():
search_url = "https://russian.alibaba.com/trade/search?fsb=y&IndexArea=product_en&SearchText="
# Retrieve command line args
pages = sys.argv[1].rstrip()
search_str = ""
if pages.isdigit():
pages = int(pages)
search_str = sys.argv[2].rstrip()
else:
pages = 2
search_str = sys.argv[1].rstrip()
merged_result = []
sessionManager = SessionManager()
search_url = search_url + parse.quote_plus(search_str) + '&page='
# Iterate over pages
for i in range(1, pages + 1):
url = search_url + str(i)
scrapper = PageScrapper(url)
scrapper.sessionManager = sessionManager
scrapper.get_page()
scrapper.get_items()
merged_result.append(scrapper.result)
del sessionManager
merged_result = [i for i in itertools.chain(*merged_result)]
print(
json.dumps(merged_result, indent=4, sort_keys=True,
ensure_ascii=False))
def main(secondsAgo=None, year=None, exclude=None, reinit=False):
manager = SessionManager()
manager.generateEngine()
manager.initializeDatabase(reinit)
manager.createSession()
loadFromTime = None
startTime = datetime.now()
if secondsAgo is not None:
loadFromTime = startTime - timedelta(seconds=secondsAgo)
if exclude != 'cce':
loadCCE(manager, loadFromTime, year)
if exclude != 'ccr':
loadCCR(manager, loadFromTime, year)
indexUpdates(manager, loadFromTime)
manager.closeConnection()
# Session config
app.config['SESSION_TYPE'] = 'mongodb'
app.config['SESSION_MONGODB'] = mongoClient
app.config['SESSION_MONGODB_DB'] = "citations-app2"
app.config['SESSION_MONGODB_COLLECT'] = "sessions"
app.config['SESSION_COOKIE_SECURE'] = True
app.config['SESSION_COOKIE_HTTPONLY'] = True
app.config['SESSION_PERMANENT'] = True
# For a 1 minute validity session, set minutes=121 (date is not the same format before database & this API)
app.config['PERMANENT_SESSION_LIFETIME'] = timedelta(minutes=122)
# if SESSION_USE_SIGNER set to True, you have to set flask.Flask.secret_key
app.config['SESSION_USE_SIGNER'] = True
Session(app)
sessMgr = SessionManager(mongo)
mailMgr = MailManager()
ph = PasswordHasher()
''' HEADERS
X-Frame-Options :
éviter les attaques de clickjacking pour s'assurer que le contenu ne soit pas embarqué dans d'autres sites
X-Content-Type-Options
est un marqueur utilisé par le serveur pour indiquer que les types MIME annoncés dans les en-têtes Content-Type ne doivent pas être modifiés ou et suivis
'''
httpResponseHeaderOptions = {
'X-Frame-Options': 'sameorigin',
'X-Content-Type-Options': 'nosniff'
}
@app.route('/', methods=['GET', 'POST'])
