Fetch image/video from cloud bucket and send it+prompt to Gemini vision in Go

[Y-Noor]

Hello hello

I am trying to get an image from my cloud bucket and send it + a user prompt to gemini vision. I managed to do it for locally stored images but cannot see how to do it when bickets are involved.

Below is the local version:

client, err := genai.NewClient(ctx, option.WithAPIKey(string(API_KEY)))
if err != nil {
log.Fatal(err)
}
defer client.Close()
modelfile, err := client.UploadFileFromPath(ctx, filepath.Join(uploadPath, dtf+".jpg"), nil)
if err != nil {
log.Fatal(err)
}
defer client.DeleteFile(ctx, modelfile.Name)
// ----------------------------- THIS ONE
model := client.GenerativeModel("gemini-1.5-flash")
resp, err := model.GenerateContent(ctx,
genai.FileData{URI: modelfile.URI},
genai.Text(prompt))

Thanks in advance for your help!

[ruthseki]

Hi @Y-Noor,

Welcome to Google Cloud Community!

Here's an insight of how to fetch an image from a cloud bucket and send it to Gemini Vision in Go.

1. Authentication and Setup:
• You'll need a Google Cloud Platform project with an API key for Gemini Vision.
• Configure your Go environment to use this API key.
• Ensure your GCP project has access to your cloud storage bucket.
2. Download Image from Bucket:
• Use the Google Cloud Storage client library in Go to interact with your cloud bucket.
• Specify the bucket name and image file name.
• Download the image file contents into memory.
3. Upload Image to Gemini Vision:
• Use the Gemini Vision API client library in Go.
• Upload the downloaded image file as a data stream.
• Retrieve a unique URI (identifier) for the uploaded image.
4. Send Prompt to Gemini Vision:
• Construct a request to the Gemini Vision API.
• Include the image URI from the previous step.
• Add your user prompt (text description) about what you want Gemini Vision to analyze.
5. Process Gemini Vision Response:
• The Gemini Vision API will respond with an analysis of the image based on your prompt.
• Parse the response (usually in JSON format) to extract relevant data.

Additionally, here are some useful documentation that you may check as well:

• Setting up your development environment - Go
• Explore vision capabilities with the Gemini API
• Send requests to the Vertex AI API for Gemini - Go

I hope the above information is helpful.