def get_source_code_submissions(proxies,
proxy_index,
contest_id,
submission_ids,
pbar,
submission_index=0,
iterations=1000):
if iterations > 0:
# Specify a random user agent
headers = {'User-Agent': 'anything'}
# if all proxies have been exhausted, get new list of proxies
if proxy_index >= len(proxies):
proxies = Proxies.get_proxies()
proxy_index = 0
# Set the proxy for the requests
proxy_dict = {
'http': f'{proxies[proxy_index]}',
'https': f'{proxies[proxy_index]}',
}
try:
# Go over all submissions and make a request to the submission page
for submission in submission_ids[submission_index:]:
response = requests.get(
f'https://codeforces.com/contest/{contest_id}/submission/{submission}',
timeout=10,
headers=headers,
proxies=proxy_dict,
allow_redirects=False)
# Transform to soup object with html parser
soup = BeautifulSoup(response.content, 'html.parser')
# Find the actual program code on page
solution = soup.find("pre", {
"id": "program-source-text"
}).text.strip()
os.makedirs(f'data/contests_solutions_code/{contest_id}/',
exist_ok=True)
text_file = open(
f'data/contests_solutions_code/{contest_id}/{submission}.txt',
"w")
text_file.write(solution)
text_file.close()
# update where we are for if we need to update the proxy and update the progress bar
submission_index += 1
pbar.update()
except Exception:
proxy_index += 1
iterations -= 1
Scraper.get_source_code_submissions(proxies, proxy_index,
contest_id, submission_ids,
pbar, submission_index,
iterations)
def baiduzhidaosearch(keyword, page):
ret = {
'code': 1002,
'msg': 'failure',
'data': []
}
try:
page = int(page) * 10
print 111
keyword_u = keyword.encode('utf-8')
print chardet.detect(keyword_u)
# url = 'http://zhidao.baidu.com/search?word=%s&ie=gbk&site=-1&sites=0&date=0&pn=%s' % (keyword.encode('utf-8 ').decode('gbk','ignore'), page)
url = u'http://zhidao.baidu.com/search?ct=17&pn=%s&tn=ikaslist&rn=10&word=%s' % (page, keyword)
#print(url)
print 222
headers = Headers.getHeaders()
proxies = Proxies.get_proxies()
req = requests.get(url, headers=headers, timeout=60, proxies=proxies)
if req.status_code == 200:
ret['code'] = 1001
ret['msg'] = 'success'
id = []
title = []
req.encoding = 'gbk'
html = req.text.encode(encoding="utf-8", errors="ignore").decode("utf-8", errors="ignore")
selector = etree.HTML(html)
urls = selector.xpath('//div[@class="list"]/dl/dt[1]/a/@href')
for u in urls:
match_obj = re.search(r'question/(.*?).html', u, re.M | re.I)
id.append(match_obj.group(1))
titles = selector.xpath('//div[@class="list"]/dl/dt[1]/a')
for t in titles:
title.append(etree.tostring(t, encoding='utf8', method="html"))
max_n = len(id)
n = 0
while True:
if n >= max_n:
break
# print(title[n])
ret['data'].append(
{'cid': id[n], 'title': re.search(r'"ti">(.*?)</a>', title[n], re.M | re.I).group(1)})
n = n + 1
except Exception as e :
print e
return simplejson.dumps(ret)
from headers import Headers
from config import Config
from mysqldao import MysqlDao
import time
from proxies import Proxies
url_main = 'http://www.bttiantang.com/movie.php?/order/id/'
page_count = 10
# 1367为总页数,需要定期修改
for one in xrange(1, page_count):
url = url_main + str(one)
print(url)
headers = Headers.getHeaders()
try:
proxies = Proxies.get_proxies()
req = requests.get(url, headers=headers, timeout=30, proxies=proxies)
code = req.status_code
print(code)
if code == 200:
html = req.content
selector = etree.HTML(html)
content_urls = selector.xpath('//*[contains(@class,"litpic")]/a[1]/@href')
for content_url in content_urls:
content_url = Config.url_main + content_url
created_at = time.strftime('%Y-%m-%d %H:%M:%S')
insert_value = '"' + content_url + '",0,"' + created_at + '"'
mysqlDao = MysqlDao()
sql = 'select url from bttiantang_url ORDER BY id desc limit 0,1 '
ret = mysqlDao.execute(sql)
for r in ret:
def get_solutions(amt_threads=40):
# Get all contests that have already been scraped
scraped_contests = []
for _, dirs, _ in os.walk('data/contests_solutions_code'):
scraped_contests += [int(dir) for dir in dirs]
# Get the metadata files to find submission ids to scrape solutions for
metadata_files = [
file for file in os.listdir('data/contests_solutions_metadata')
if os.path.isfile(f'data/contests_solutions_metadata/{file}')
]
if len(metadata_files) == 0:
print(
'Run "get_solutions_metadata" first to fetch the solutions to scrape before running this function'
)
for metadata_file in metadata_files:
# Open the solutions metadata file where solutions will be coupled to
solutions_df = pd.read_csv(
f'data/contests_solutions_metadata/{metadata_file}')
# Get all unique contests in this dataset that are not scraped yet
contests = list(
solutions_df[~solutions_df['contestId'].isin(scraped_contests)]
['contestId'].unique())
for contest in contests:
print(f'starting with contest {contest}...')
contest_submissions = solutions_df[solutions_df['contestId'] ==
contest]['solutionId']
# Work with amt of threads to parallelize the requests
threads = []
submissions_per_thread = math.ceil(
len(contest_submissions) / amt_threads)
proxies = Proxies.get_proxies()
# Progress bar to indicate current progress and speed
pbar = tqdm(total=len(contest_submissions))
for index in range(0, len(contest_submissions),
submissions_per_thread):
# Let every thread start with a random proxy to spread the search space
proxy_index = random.randrange(0, len(proxies))
# Get solutions for the contests from the scraper
threads.append(
threading.Thread(
target=Scraper.get_source_code_submissions,
args=(
proxies,
proxy_index,
contest,
contest_submissions[index:index +
submissions_per_thread],
pbar,
)))
for t in threads:
t.start()
for t in threads:
t.join()
