Google

pip install datapizza-ai-clients-google

datapizza.clients.google.GoogleClient

Bases: Client

A client for interacting with Google's Generative AI APIs.

This class provides methods for invoking the Google GenAI API to generate responses based on given input data. It extends the Client class.

__init__

__init__(
    api_key=None,
    model="gemini-2.0-flash",
    system_prompt="",
    temperature=None,
    cache=None,
    project_id=None,
    location=None,
    credentials_path=None,
    use_vertexai=False,
)

Parameters:

Name	Type	Description	Default
api_key	str | None	The API key for the Google API.	None
model	str	The model to use for the Google API.	'gemini-2.0-flash'
system_prompt	str	The system prompt to use for the Google API.	''
temperature	float | None	The temperature to use for the Google API.	None
cache	Cache | None	The cache to use for the Google API.	None
project_id	str | None	The project ID for the Google API.	None
location	str | None	The location for the Google API.	None
credentials_path	str | None	The path to the credentials for the Google API.	None
use_vertexai	bool	Whether to use Vertex AI for the Google API.	False

Usage example

from datapizza.clients.google import GoogleClient

client = GoogleClient(
    api_key="YOUR_API_KEY",
    model="gemini-2.5-flash",
    system_prompt="You are a helpful assistant.",
    temperature=0.7,
)

response = client.invoke("What is the photoelectric effect?")
print(response.text)

Thinking Configuration

Gemini models support explicit reasoning via thinking_config. Use include_thoughts=True to receive ThoughtBlock in the response.

Gemini 2.5 – Budget-based thinking

response = client.invoke(
    input="Explain step by step why the sky is blue.",
    thinking_config={
        "thinking_budget": 1024,
        "include_thoughts": True,
    },
)

print("Thoughts:", response.thoughts)
print("Answer:", response.text)

Gemini 3 – Level-based thinking

response = client_3.invoke(
    input="Design a simple movie recommendation algorithm.",
    thinking_config={
        "thinking_level": "high",  # "low" | "high"
        "include_thoughts": True,
    },
)

print("Thoughts:", response.thoughts)
print("Answer:", response.text)

Native Tools

The GoogleClient supports Gemini's native tools: Google Search, URL Context, Google Maps and Code Execution.

Google Search

from google.genai import types

response = client.invoke(
    input="What are the latest news about the James Webb telescope?",
    tools=[{"google_search": types.GoogleSearch()}],
)
print(response.text)

URL Context

response = client.invoke(
    input="Summarize this page: https://example.com/article",
    tools=[{"url_context": {}}],
)
print(response.text)

Google Maps

response = client.invoke(
    input="List the best restaurants in Rome",
    tools=[{"google_maps": types.GoogleMaps()}],
)
print(response.text)

Code Execution

from google.genai import types

response = client.invoke(
    input="Calculate the sum of the first 50 prime numbers.",
    tools=[{"code_execution": types.ToolCodeExecution()}],
)
print(response.text)

Media Resolution

Control the quality/cost/latency trade-off for multimedia inputs (images, PDFs) with media_resolution (Gemini 3 models only).

from google.genai import types

response = client.invoke(
    input="Analyze this technical diagram in detail.",
    media_resolution=types.MediaResolution.MEDIA_RESOLUTION_HIGH,
)

Available values: MEDIA_RESOLUTION_LOW, MEDIA_RESOLUTION_MEDIUM, MEDIA_RESOLUTION_HIGH.

Image Generation

from google.genai import types
from datapizza.type import MediaBlock
import base64

client_img = GoogleClient(
    api_key="YOUR_API_KEY",
    model="gemini-2.5-flash-image",
)

response = client_img.invoke(
    input="Create an image of a cat baking a pizza",
    image_config=types.ImageConfig(aspect_ratio="4:3", image_size="1K"), # Gemini 3 only
)
print(response.text)

# response.content could contain both TextBlock and MediaBlock
for block in response.content:
    if isinstance(block, MediaBlock):
        image_bytes = base64.b64decode(block.media.source)
        with open(f"generated.{block.media.extension}", "wb") as f:
            f.write(image_bytes)