When calling the API interface of ChatGpt, due to the need to surf the Internet scientifically, some problems will be encountered. According to the purpose of calling the API interface, it can be divided into two categories:
1. Call the API to build a dialogue robot
报错：openai.error.APIConnectionError: Error communicating with OpenAI: HTTPSConnectionPool(host='api.openai.com', port=443): Max retries exceeded with url: /v1/engines/text-davinci-002/completions (Caused by SSLError(SSLEOFError(8, 'EOF occurred in violation of protocol (_ssl.c:1129)')))
Use the interface packaged by others:
| parameter name | type | length | must | Remark |
|---|---|---|---|---|
| apiKey | String | 64 | yes | OpenAI's ApiKey |
| sessionId | String | 64 | yes | Session ID, associated context, it is recommended to use UUID as sessionId |
| content | String | 1000 | yes | sent content |
The code example is as follows, just change the value at 'apikey' to your own apikey.
2. Call the API interface and use the GPT3 or GPT4 algorithm to process related NLP tasks
Take the concept extraction task as an example:
问题：openai.error.APIConnectionError: Error communicating with OpenAI: HTTPSConnectionPool(host='api.openai.com', port=443): Max retries exceeded with url: /v1/engines/text-davinci-002/completions (Caused by SSLError(SSLEOFError(8, 'EOF occurred in violation of protocol (_ssl.c:1129)')))
Solution: Use a proxy to solve this problem
The code example is as follows:
your_proxy_addressHere you need to replace and with your own actual proxy address and port your_proxy_port. This way, your request will be sent through the specified proxy server.
In this case, since the OpenAI library has no api_clientattributes, an error attributeError: module 'openai' has no attribute 'api_client' will be reported. In this case, you need to create a openai.api_resources.completion.Completionnew class inherited from and rewrite _createthe method. In this new _createmethod, proxy settings can be specified. code show as below:
Same as above, just replace the proxy address and port in proxies with your own. If the error is still reported at this time: openai.error.APIConnectionError: Error communicating with OpenAI: HTTPSConnectionPool(host='api.openai.com', port=443): Max retries exceeded with url: /v1/engines/text-davinci-002 /completions (Caused by SSLError(SSLEOFError(8, 'EOF occurred in violation of protocol (_ssl.c:1129)')))
Then the problem is probably the version of urllib3, and the problem will be solved by returning the version of urllib3 to version 1.25.11.
The result of the operation is as follows:
Reference link: