curl --noproxy“ *”在Python的请求模块中等效

Question

I'm trying to access an API on a server within my network using Python's request module, but am getting a 502 Bad Gateway error. When I try to do something equivalent using curl , I get the same error and it's because the request is trying to run through the proxy server. Here is the output from curl :

$ curl -v -k -u username:password https://hostname:port/some/api/endpoint              * Uses proxy env variable no_proxy == 'hostname'
* Uses proxy env variable https_proxy == 'http://10.3.0.40:10256'
*   Trying 10.3.0.40...
* TCP_NODELAY set
  % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current
                                 Dload  Upload   Total   Spent    Left  Speed
  0     0    0     0    0     0      0      0 --:--:-- --:--:-- --:--:--     0* Connected to 10.3.0.40 (10.3.0.40) port 10256 (#0)
* allocate connect buffer!
* Establish HTTP proxy tunnel to hostname:port
* Server auth using Basic with user 'user'
> CONNECT hostname:port HTTP/1.1
> Host: hostname:port
> User-Agent: curl/7.61.1
> Proxy-Connection: Keep-Alive
>
< HTTP/1.0 502 Bad Gateway
< Server: Zscaler/5.6
< Content-Type: text/html
< Connection: close
<
* Received HTTP code 502 from proxy after CONNECT
* CONNECT phase completed!
* Closing connection 0
curl: (56) Received HTTP code 502 from proxy after CONNECT

As the output states at the top, curl is automatically implementing the HTTP_PROXY variable I have in my environment variables, which is needed for other Python tasks. In order to avoid this, I can pass --noproxy "*" to curl and the request no longer tries to run through the proxy.

I'd like to try to get the equivalent working with Python's request module. I know that requests will automatically read from the environment variables as well, so I tried to bypass that specifying proxies=None in the code:

response = requests.get('https://hostname:port/some/api/endpoint', proxies=None, auth=(USER, PASS))

But I still get the 502 Bad Gateway. In looking at some questions on SO, it looks as if you can specify the NO_PROXY environment variable and provide URLs that should not use the proxy. I tried populating that variable with https://hostname , https://hostname:port , https://hostname/* , https://hostname:port/* , and hostname . None of which seemed to work. (Are those even the right formats?)

How do I get requests to use the equivalent of curl's --noproxy ?

Answer 1

Setting proxies to None is like omitting it. You want to specify that the http and https proxies are None by passing the value:

proxies={'http': None, 'https': None}

to the requests.get function.

Regarding NO_PROXY (or no_proxy ), its value is a domain name , so you shouldn't add http:// or https:// in front of its value, just use the domain name hostname in your case.