[Python] How to pass fps metadata to gemini_model.generate_content?

I currently have code that uses the Google Gemini developer API to build a request and send that request to a gemini-2.5-flash model. The request includes a video with metadata specifying a Frames Per Second (fps) value for Gemini to analyze the video at. The documentation for this can be found here.

contents = types.Content(parts=[
types.Part(
inline_data=types.Blob(
data=video_bytes,
mime_type=mime_type
),
video_metadata=types.VideoMetadata(fps=8)
),
types.Part(text=PROMPT)
])

resp = client.models.generate_content(
model="gemini-2.5-flash",
contents=contents,
config=types.GenerateContentConfig(
temperature=0.2,
max_output_tokens=60000
)
)



I am now trying to switch from the developer API to Google VertexAI as this will allow me to upload videos to Google cloud and simply send Gemini a Google cloud uri.

I have rebuilt a request similar to the one above that uses Google Developer API, but I cannot figure out how to pass Frames Per Second (fps) metadata with this new request using VertexAI.

contents = [
Part.from_uri(uri=gcs_uri, mime_type=mime_type),
Part.from_text(prompt)
]
resp = gemini_model.generate_content(contents)

Continue reading...