Extending video is not working while using proxy

[WilliamChen-luckbob]

Environment details

• Programming language: Python
• OS: Windows
• Language runtime version: Python3.12
• Package version: google-genai==1.57.0 and it's dependencies

Steps to reproduce

just run the code

import os
import time

import httpx
import requests
from dotenv import load_dotenv
from google import genai
from google.genai import types
from google.genai.types import Video

if __name__ == '__main__':
    load_dotenv()
    api_key = os.getenv("GENAI_API_KEY")
    proxy = os.getenv("GRPC_PROXY")
    genai_client = genai.Client(
        api_key=api_key,
        http_options=types.HttpOptions(
            httpx_client=httpx.Client(proxy=proxy)
        )
    )

    # Generate a video via Veo
    source = types.GenerateVideosSource(
        prompt="Upbeat and adorable background music throughout the video. A close-up shot focusing on the face of a cute lynx-point Ragdoll cat, with its big, bright blue eyes staring at the camera curiously, then tilting its head adorably toward the lens."
    )
    config = types.GenerateVideosConfig(
        duration_seconds=4,
        aspect_ratio="16:9",
        resolution="720p",
        negative_prompt="nsfw",
    )

    operation: types.GenerateVideosOperation = genai_client.models.generate_videos(
        model="veo-3.1-fast-generate-preview",
        source=source,
        config=config
    )
    while not operation.done:
        print("Waiting for video generation to complete...")
        time.sleep(10)
        operation = genai_client.operations.get(operation)

    generated_video = operation.response.generated_videos[0] if operation.response.generated_videos else None
    # download the video directly via proxy and save
    if generated_video:
        video = generated_video.video
        download_url = video.uri
        actual_download_url = f'{download_url}&key={api_key}'
        rsp = requests.get(actual_download_url, proxies={
            'http': proxy,
            'https': proxy
        })
        video_bytes = rsp.content
        video = Video(uri=download_url, video_bytes=video_bytes)
        video.save("cat-gen.mp4")

    # extend the video
    # re-create the config
    config = types.GenerateVideosConfig(
        duration_seconds=4,
        aspect_ratio="16:9",
        resolution="720p",
        negative_prompt="nsfw",
    )

    # request for extend (will crash here)
    operation: types.GenerateVideosOperation = genai_client.models.generate_videos(
        model="veo-3.1-fast-generate-preview",
        video=operation.response.generated_videos[0].video,
        prompt="Upbeat and adorable background music throughout the video. A close-up shot focusing on the face of a cute lynx-point Ragdoll cat, with its big, bright blue eyes staring at the camera curiously, then tilting its head adorably toward the lens. and eyes blinking.",
        config=config
    )

    while not operation.done:
        print("Waiting for video generation to complete...")
        time.sleep(10)
        operation = genai_client.operations.get(operation)

    generated_video = operation.response.generated_videos[0] if operation.response.generated_videos else None

    # download the video directly via proxy and save
    if generated_video:
        video = generated_video.video
        download_url = video.uri
        actual_download_url = f'{download_url}&key={api_key}'
        rsp = requests.get(actual_download_url, proxies={
            'http': proxy,
            'https': proxy
        })
        video_bytes = rsp.content
        video = Video(uri=download_url, video_bytes=video_bytes)
        video.save("cat-gen-extend.mp4")

Here is the .env look like:

# this is my proxy location, which work well in video generation
GRPC_PROXY='http://user:password@192.168.24.100:6666'
# gemini api token
GENAI_API_KEY=xxxxxxxxxxxxxxxxxxxxxxxxxx

the first generation is downloaded successfully, but there is an error while trying to extend the video passing the first object via genai client.

Traceback (most recent call last):
  File "C:\Program Files\JetBrains\PyCharm Community Edition 2023.3.4\plugins\python-ce\helpers\pydev\pydevd.py", line 1534, in _exec
    pydev_imports.execfile(file, globals, locals)  # execute the script
    ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Program Files\JetBrains\PyCharm Community Edition 2023.3.4\plugins\python-ce\helpers\pydev\_pydev_imps\_pydev_execfile.py", line 18, in execfile
    exec(compile(contents+"\n", file, 'exec'), glob, loc)
  File "C:\code\repository-python\dify-official-plugins\tools\gemini_video\test\test_gen.py", line 69, in <module>
    operation: types.GenerateVideosOperation = genai_client.models.generate_videos(
                                               ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\code\repository-python\dify-official-plugins\tools\gemini_video\venv\Lib\site-packages\google\genai\models.py", line 5725, in generate_videos
    return self._generate_videos(
           ^^^^^^^^^^^^^^^^^^^^^^
  File "C:\code\repository-python\dify-official-plugins\tools\gemini_video\venv\Lib\site-packages\google\genai\models.py", line 5082, in _generate_videos
    response = self._api_client.request(
               ^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\code\repository-python\dify-official-plugins\tools\gemini_video\venv\Lib\site-packages\google\genai\_api_client.py", line 1375, in request
    response = self._request(http_request, http_options, stream=False)
               ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\code\repository-python\dify-official-plugins\tools\gemini_video\venv\Lib\site-packages\google\genai\_api_client.py", line 1211, in _request
    return self._retry(self._request_once, http_request, stream)  # type: ignore[no-any-return]
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\code\repository-python\dify-official-plugins\tools\gemini_video\venv\Lib\site-packages\tenacity\__init__.py", line 477, in __call__
    do = self.iter(retry_state=retry_state)
         ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\code\repository-python\dify-official-plugins\tools\gemini_video\venv\Lib\site-packages\tenacity\__init__.py", line 378, in iter
    result = action(retry_state)
             ^^^^^^^^^^^^^^^^^^^
  File "C:\code\repository-python\dify-official-plugins\tools\gemini_video\venv\Lib\site-packages\tenacity\__init__.py", line 420, in exc_check
    raise retry_exc.reraise()
          ^^^^^^^^^^^^^^^^^^^
  File "C:\code\repository-python\dify-official-plugins\tools\gemini_video\venv\Lib\site-packages\tenacity\__init__.py", line 187, in reraise
    raise self.last_attempt.result()
          ^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Users\william\AppData\Local\Programs\Python\Python312\Lib\concurrent\futures\_base.py", line 449, in result
    return self.__get_result()
           ^^^^^^^^^^^^^^^^^^^
  File "C:\Users\william\AppData\Local\Programs\Python\Python312\Lib\concurrent\futures\_base.py", line 401, in __get_result
    raise self._exception
  File "C:\code\repository-python\dify-official-plugins\tools\gemini_video\venv\Lib\site-packages\tenacity\__init__.py", line 480, in __call__
    result = fn(*args, **kwargs)
             ^^^^^^^^^^^^^^^^^^^
  File "C:\code\repository-python\dify-official-plugins\tools\gemini_video\venv\Lib\site-packages\google\genai\_api_client.py", line 1188, in _request_once
    errors.APIError.raise_for_response(response)
  File "C:\code\repository-python\dify-official-plugins\tools\gemini_video\venv\Lib\site-packages\google\genai\errors.py", line 121, in raise_for_response
    cls.raise_error(response.status_code, response_json, response)
  File "C:\code\repository-python\dify-official-plugins\tools\gemini_video\venv\Lib\site-packages\google\genai\errors.py", line 146, in raise_error
    raise ClientError(status_code, response_json, response)
google.genai.errors.ClientError: 400 INVALID_ARGUMENT. {'error': {'code': 400, 'message': 'Your use case is currently not supported.  Please refer to Gemini API documentation for current model offering.', 'status': 'INVALID_ARGUMENT'}}

I’m using direct download via the requests library because of issue #1946, and I haven’t been able to resolve this problem yet.

According to the official documentation, I used the Video instance returned from the first generation as the video parameter (matching the approach in the official code example), but it still doesn’t work.
I also investigated the root cause by checking the google/genai/models.py file and found that the code simply creates a new Video object with uri and mime_type like this:

Unfortunately, the genai tool will not return mime_type in the first generation, and even I put video/mp4 in it, the error still remains.