Spaces:
				
			
			
	
			
			
		Paused
		
	
	
	
			
			
	
	
	
	
		
		
		Paused
		
	Commit 
							
							·
						
						6cf0820
	
0
								Parent(s):
							
							
:gem: [Feature] New GoogleSearcher: Enable google search with query
Browse files- networks/__init__.py +0 -0
 - networks/google_searcher.py +55 -0
 
    	
        networks/__init__.py
    ADDED
    
    | 
         
            File without changes
         
     | 
    	
        networks/google_searcher.py
    ADDED
    
    | 
         @@ -0,0 +1,55 @@ 
     | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
| 
         | 
|
| 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:
         
     | 
| 17 | 
         
            +
                # https://github.com/Nv7-GitHub/googlesearch/blob/master/googlesearch/__init__.py
         
     | 
| 18 | 
         
            +
                def __init__(self):
         
     | 
| 19 | 
         
            +
                    self.url = "https://www.google.com/search"
         
     | 
| 20 | 
         
            +
                    self.enver = enver
         
     | 
| 21 | 
         
            +
                    self.enver.set_envs(proxies=True)
         
     | 
| 22 | 
         
            +
                    self.output_root = Path(__file__).parents[1] / "files"
         
     | 
| 23 | 
         
            +
             
     | 
| 24 | 
         
            +
                def send_request(self, query, result_num=10):
         
     | 
| 25 | 
         
            +
                    res = requests.get(
         
     | 
| 26 | 
         
            +
                        url=self.url,
         
     | 
| 27 | 
         
            +
                        headers=RequestHeaderConstructor().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, res, query):
         
     | 
| 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(res.content)
         
     | 
| 45 | 
         
            +
             
     | 
| 46 | 
         
            +
                def search(self, query):
         
     | 
| 47 | 
         
            +
                    self.query = query
         
     | 
| 48 | 
         
            +
                    res = self.send_request(query)
         
     | 
| 49 | 
         
            +
                    self.save_response(res, query)
         
     | 
| 50 | 
         
            +
             
     | 
| 51 | 
         
            +
             
     | 
| 52 | 
         
            +
            if __name__ == "__main__":
         
     | 
| 53 | 
         
            +
                searcher = GoogleSearcher()
         
     | 
| 54 | 
         
            +
                # searcher.search("python tutorials")
         
     | 
| 55 | 
         
            +
                searcher.search("python教程")
         
     |