Proxy settings configuration for Vertex AI SDK

Hello !

I'm using the Vertex AI Python SDK with a generative model in my application as follows:

```python
import vertexai
vertexai.init(...)
model = GenerativeModel("...")
generation_config = GenerationConfig("...")
model.generate_content(..., stream=True)
```

I want the `generate_content` method to route its requests through a proxy.

From my research, I found that, with this library, proxies can be configured using the `https_proxy` and `http_proxy` environment variables or programmatically within the Python code, for example:

```python
import os
os.environ["https_proxy"] = "http://my_proxy:8080"
```

However, I would like to know if there is another way to configure a proxy for this call without relying on `os.environ` or setting environment variables.

This is important because modifying `os.environ` or system-level environment variables could unintentionally affect other parts of the application.
I'd appreciate any suggestions on alternative approaches.
Thanks!

Start with the Vertex AI Python SDK's generate_content and initialization paths, and compare their documented proxy configuration options with the current http_proxy and https_proxy environment-variable behavior. Done would require an agreed, scoped alternative that isolates proxy settings to this client or call, with coverage or documentation showing it does not alter unrelated application behavior.

由索引模型根据 Issue 内容生成。