Working with Gemini in Pixeltable

Pixeltable's Gemini integration enables you to access the Gemini LLM via the Google Gemini API.

Prerequisites

• A Google AI Studio account with an API key (https://aistudio.google.com/app/apikey)

Important Notes

• Google AI Studio usage may incur costs based on your plan.
• Be mindful of sensitive data and consider security measures when integrating with external services.

First you'll need to install required libraries and enter a Gemini API key obtained via Google AI Studio.

%pip install -qU pixeltable google-generativeai

Note: you may need to restart the kernel to use updated packages.

import os
import getpass

if 'GEMINI_API_KEY' not in os.environ:
    os.environ['GEMINI_API_KEY'] = getpass.getpass('Google AI Studio API Key:')

Now let's create a Pixeltable directory to hold the tables for our demo.

import pixeltable as pxt

# Remove the `gemini_demo` directory and its contents, if it exists
pxt.drop_dir('gemini_demo', force=True)
pxt.create_dir('gemini_demo')

Connected to Pixeltable database at: postgresql+psycopg
://postgres:@/pixeltable?host=/Users/anushas-pxt/.pixeltable/pgdata
Created directory `gemini_demo`.
<pixeltable.catalog.dir.Dir at 0x3425c73a0>

Generate content

Create a Table: In Pixeltable, create a table with columns to represent your input data and the columns where you want to store the results from Gemini.

from pixeltable.functions import gemini

# Create a table in Pixeltable and pick a model hosted on Google AI Studio with some parameters

t = pxt.create_table('gemini_demo.story', {'input': pxt.String})

t.add_computed_column(output=gemini.generate_content(
    t.input,
    model_name='gemini-1.5-flash',
    candidate_count=3,
    stop_sequences=['\n'],
    max_output_tokens=300,
    temperature=1.0,
    top_p=0.95,
    top_k=40,
))

Created table `story`.
Added 0 column values with 0 errors.
UpdateStatus(num_rows=0, num_computed_values=0, num_excs=0, updated_cols=[], cols_with_excs=[])

# Ask Gemini to generate some content based on the input
t.insert([
    {'input': 'Write a story about a magic backpack.'}, 
    {'input': 'Tell me a science joke.'}
])

Inserting rows into `story`: 2 rows [00:00, 361.73 rows/s]████████| 4/4 [00:03<00:00,  1.10 cells/s]
Computing cells: 100%|████████████████████████████████████████████| 4/4 [00:03<00:00,  1.10 cells/s]
Inserted 2 rows with 0 errors.
UpdateStatus(num_rows=2, num_computed_values=4, num_excs=0, updated_cols=[], cols_with_excs=[])

# Parse the response into a new column
t.add_computed_column(response=t.output['candidates'][0]['content']['parts'][0]['text'])
t.select(t.input, t.response).show()

Computing cells: 100%|████████████████████████████████████████████| 2/2 [00:00<00:00, 25.38 cells/s]
Added 2 column values with 0 errors.

Learn More

To learn more about advanced techniques like RAG operations in Pixeltable, check out the RAG Operations in Pixeltable tutorial.

If you have any questions, don't hesitate to reach out.