Spaces:
Paused
Paused
:recycle: [Refactor] Move header constructor, and prettier logging
Browse files- networks/google_searcher.py +15 -23
networks/google_searcher.py
CHANGED
|
@@ -1,16 +1,7 @@
|
|
| 1 |
import requests
|
| 2 |
from pathlib import Path
|
| 3 |
from utils.enver import enver
|
| 4 |
-
|
| 5 |
-
|
| 6 |
-
class RequestHeaderConstructor:
|
| 7 |
-
def __init__(self):
|
| 8 |
-
self.construct()
|
| 9 |
-
|
| 10 |
-
def construct(self):
|
| 11 |
-
self.headers = {
|
| 12 |
-
"User-Agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/111.0.0.0 Safari/537.36 Edg/111.0.1661.62",
|
| 13 |
-
}
|
| 14 |
|
| 15 |
|
| 16 |
class GoogleSearcher:
|
|
@@ -21,35 +12,36 @@ class GoogleSearcher:
|
|
| 21 |
self.enver.set_envs(proxies=True)
|
| 22 |
self.output_root = Path(__file__).parents[1] / "files"
|
| 23 |
|
| 24 |
-
def send_request(self,
|
| 25 |
-
|
|
|
|
| 26 |
url=self.url,
|
| 27 |
-
headers=
|
|
|
|
|
|
|
| 28 |
params={
|
| 29 |
"q": self.query,
|
| 30 |
"num": result_num,
|
| 31 |
-
# "hl": "en",
|
| 32 |
-
# "start": 0,
|
| 33 |
},
|
| 34 |
proxies=self.enver.requests_proxies,
|
| 35 |
)
|
| 36 |
-
return res
|
| 37 |
|
| 38 |
-
def save_response(self
|
| 39 |
-
output_filename = query.replace(" ", "_") + ".html"
|
| 40 |
if not self.output_root.exists():
|
| 41 |
self.output_root.mkdir(parents=True, exist_ok=True)
|
| 42 |
output_path = self.output_root / output_filename
|
|
|
|
| 43 |
with open(output_path, "wb") as wf:
|
| 44 |
-
wf.write(
|
| 45 |
|
| 46 |
def search(self, query):
|
| 47 |
self.query = query
|
| 48 |
-
|
| 49 |
-
self.save_response(
|
| 50 |
|
| 51 |
|
| 52 |
if __name__ == "__main__":
|
| 53 |
searcher = GoogleSearcher()
|
| 54 |
-
# searcher.search("python
|
| 55 |
-
searcher.search("python
|
|
|
|
| 1 |
import requests
|
| 2 |
from pathlib import Path
|
| 3 |
from utils.enver import enver
|
| 4 |
+
from utils.logger import logger
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 5 |
|
| 6 |
|
| 7 |
class GoogleSearcher:
|
|
|
|
| 12 |
self.enver.set_envs(proxies=True)
|
| 13 |
self.output_root = Path(__file__).parents[1] / "files"
|
| 14 |
|
| 15 |
+
def send_request(self, result_num=10):
|
| 16 |
+
logger.note(f"Searching: [{self.query}]")
|
| 17 |
+
self.request_response = requests.get(
|
| 18 |
url=self.url,
|
| 19 |
+
headers={
|
| 20 |
+
"User-Agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/111.0.0.0 Safari/537.36 Edg/111.0.1661.62",
|
| 21 |
+
},
|
| 22 |
params={
|
| 23 |
"q": self.query,
|
| 24 |
"num": result_num,
|
|
|
|
|
|
|
| 25 |
},
|
| 26 |
proxies=self.enver.requests_proxies,
|
| 27 |
)
|
|
|
|
| 28 |
|
| 29 |
+
def save_response(self):
|
| 30 |
+
output_filename = self.query.replace(" ", "_") + ".html"
|
| 31 |
if not self.output_root.exists():
|
| 32 |
self.output_root.mkdir(parents=True, exist_ok=True)
|
| 33 |
output_path = self.output_root / output_filename
|
| 34 |
+
logger.note(f"Saving to: [{output_path}]")
|
| 35 |
with open(output_path, "wb") as wf:
|
| 36 |
+
wf.write(self.request_response.content)
|
| 37 |
|
| 38 |
def search(self, query):
|
| 39 |
self.query = query
|
| 40 |
+
self.send_request()
|
| 41 |
+
self.save_response()
|
| 42 |
|
| 43 |
|
| 44 |
if __name__ == "__main__":
|
| 45 |
searcher = GoogleSearcher()
|
| 46 |
+
# searcher.search("python教程")
|
| 47 |
+
searcher.search("python tutorials")
|