Spaces:
Runtime error
Runtime error
Commit
·
9ad659c
1
Parent(s):
b970341
Upload _base.py
Browse files- finnlp/data_sources/_base.py +80 -0
finnlp/data_sources/_base.py
ADDED
|
@@ -0,0 +1,80 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
from finnlp.utils.get_proxy import get_china_free_proxy, get_us_free_proxy, Kuaidaili
|
| 2 |
+
import requests
|
| 3 |
+
|
| 4 |
+
class FinNLP_Downloader:
|
| 5 |
+
def __init__(self, args = {}):
|
| 6 |
+
self.use_proxy = True if "use_proxy" in args.keys() else False
|
| 7 |
+
if self.use_proxy:
|
| 8 |
+
self.country = args["use_proxy"]
|
| 9 |
+
else:
|
| 10 |
+
self.country = None
|
| 11 |
+
self.max_retry = args["max_retry"] if "max_retry" in args.keys() else 1
|
| 12 |
+
self.proxy_pages = args["proxy_pages"] if "proxy_pages" in args.keys() else 5
|
| 13 |
+
if self.use_proxy:
|
| 14 |
+
if "kuaidaili" in self.country:
|
| 15 |
+
# tunnel, username, password
|
| 16 |
+
assert "tunnel" in args.keys(), "Please make sure \'tunnel\' in your keys"
|
| 17 |
+
assert "username" in args.keys(), "Please make sure \'username\' in your keys"
|
| 18 |
+
assert "password" in args.keys(), "Please make sure \'password\' in your keys"
|
| 19 |
+
self.proxy_list = Kuaidaili(args["tunnel"], args["username"], args["password"])
|
| 20 |
+
else:
|
| 21 |
+
self.proxy_id = 0
|
| 22 |
+
self.proxy_list = self._update_proxy()
|
| 23 |
+
else:
|
| 24 |
+
self.proxy_list = []
|
| 25 |
+
|
| 26 |
+
def _get_proxy(self):
|
| 27 |
+
if self.use_proxy:
|
| 28 |
+
if "kuaidaili" in self.country:
|
| 29 |
+
proxy = self.proxy_list.get_kuaidaili_tunnel_proxy()
|
| 30 |
+
return proxy
|
| 31 |
+
elif len(self.proxy_list) >0:
|
| 32 |
+
proxy = self.proxy_list[self.proxy_id]
|
| 33 |
+
self.proxy_id += 1
|
| 34 |
+
if self.proxy_id == len(self.proxy_list):
|
| 35 |
+
self.proxy_id = 0
|
| 36 |
+
return proxy
|
| 37 |
+
else:
|
| 38 |
+
return None
|
| 39 |
+
|
| 40 |
+
def _update_proxy(self):
|
| 41 |
+
if "china" in self.country or "China" in self.country:
|
| 42 |
+
return get_china_free_proxy(self.proxy_pages)
|
| 43 |
+
else:
|
| 44 |
+
return get_us_free_proxy(self.proxy_pages)
|
| 45 |
+
|
| 46 |
+
def _request_get(self, url, headers = None, verify = None, params = None):
|
| 47 |
+
if headers is None:
|
| 48 |
+
headers = {
|
| 49 |
+
"User-Agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:109.0) Gecko/20100101 Firefox/112.0"
|
| 50 |
+
}
|
| 51 |
+
max_retry = self.max_retry
|
| 52 |
+
proxies = self._get_proxy()
|
| 53 |
+
for _ in range(max_retry):
|
| 54 |
+
try:
|
| 55 |
+
response = requests.get(url = url, proxies = proxies, headers = headers, verify = verify, params = params)
|
| 56 |
+
if response.status_code == 200:
|
| 57 |
+
break
|
| 58 |
+
except:
|
| 59 |
+
response = None
|
| 60 |
+
|
| 61 |
+
if response is not None and response.status_code != 200:
|
| 62 |
+
response = None
|
| 63 |
+
|
| 64 |
+
return response
|
| 65 |
+
|
| 66 |
+
def _request_post(self, url, headers, json):
|
| 67 |
+
max_retry = self.max_retry
|
| 68 |
+
proxies = self._get_proxy()
|
| 69 |
+
for _ in range(max_retry):
|
| 70 |
+
try:
|
| 71 |
+
response = requests.post(url = url, headers = headers, json = json, proxies = proxies)
|
| 72 |
+
if response.status_code == 200:
|
| 73 |
+
break
|
| 74 |
+
except:
|
| 75 |
+
response = None
|
| 76 |
+
|
| 77 |
+
if response is not None and response.status_code != 200:
|
| 78 |
+
response = None
|
| 79 |
+
|
| 80 |
+
return response
|