The api_key client option must be set either by passing api_key to the client or by setting the OPENAI_API_KEY environment variable

[bigdataer01]

在本地使用如下命令启动了api server
MODEL=../chatglm-ggml.bin uvicorn chatglm_cpp.openai_api:app --host 127.0.0.1 --port 8000

使用openai_client访问的时候报错
OPENAI_API_BASE=http://127.0.0.1:8000/v1 python3 examples/openai_client.py --stream --prompt 你好

报错如下：
openai.OpenAIError: The api_key client option must be set either by passing api_key to the client or by setting the OPENAI_API_KEY environment variable

[chibinjiang]

随便给 OPENAI_API_KEY 赋值就行:
OPENAI_API_KEY=hello OPENAI_API_BASE=http://127.0.0.1:8000/v1 python openai_client.py --stream --prompt 你好

[beiwei30]

openai 1.5.0 中，指定 OPENAI_API_KEY=hello 后 openai_client.py 超时

openai.APITimeoutError: Request timed out.

openai_client.py 执行了 openai migrate 做了升级：

diff --git a/examples/openai_client.py b/examples/openai_client.py
index 372a541..e1ddfa5 100644
--- a/examples/openai_client.py
+++ b/examples/openai_client.py
@@ -1,6 +1,8 @@
 import argparse

-import openai
+from openai import OpenAI
+
+client = OpenAI()

 parser = argparse.ArgumentParser()
 parser.add_argument("--stream", action="store_true")
@@ -9,11 +11,11 @@ args = parser.parse_args()

 messages = [{"role": "user", "content": args.prompt}]
 if args.stream:
-    response = openai.ChatCompletion.create(model="default-model", messages=messages, stream=True)
+    response = client.chat.completions.create(model="default-model", messages=messages, stream=True)
     for chunk in response:
         content = chunk["choices"][0]["delta"].get("content", "")
         print(content, end="", flush=True)
     print()
 else:
-    response = openai.ChatCompletion.create(model="default-model", messages=messages)
+    response = client.chat.completions.create(model="default-model", messages=messages)
     print(response["choices"][0]["message"]["content"])

[beiwei30]

补充一下，安装 0.28.1 是工作的

pip install openai==0.28.1