keeya 0.1.2

AI-Powered Python Library for Code Generation

Keeya

AI-Powered Python Library for Code Generation

Keeya is a simple Python library that uses AI to generate clean, executable Python code on-demand. Unlike traditional code completion tools, Keeya runs in your Python environment and generates production-ready code based on your requirements.

Installation

pip install keeya

Setup

No setup required! Keeya works out of the box with free AI models.

Optional: If you want to use your own OpenRouter API key:

1. Get your key from OpenRouter
2. Set it: export OPENROUTER_API_KEY="your_key_here"

Quick Start

import keeya

# Generate any Python function
code = keeya.generate("function to add two numbers")
print(code)

# Generate complex algorithms
code = keeya.generate("function to implement quicksort")
print(code)

Examples

Basic Code Generation

import keeya

# Generate any Python function
code = keeya.generate("function to add two numbers")
print(code)
# Output: def add_numbers(a, b): return a + b

# Generate data processing function
code = keeya.generate("function to calculate mean of a list")
print(code)
# Output: def calculate_mean(numbers): return sum(numbers) / len(numbers)

Data Science Operations

import keeya
import pandas as pd

# Load your data
df = pd.read_csv('data.csv')

# AI-powered data cleaning
cleaned_df = keeya.clean(df)

# AI-powered analysis
insights = keeya.analyze(df)

# AI-powered visualization
keeya.visualize(df, plot_type='scatter')

# AI-powered ML training
model = keeya.train(df, target='price')

Features

• Simple API: Just call keeya.generate() or keeya.clean()
• AI-Powered: Uses AI to generate code based on your data
• Context-Aware: Understands your DataFrames and generates appropriate code
• Smart Model Selection: Automatically chooses the best AI model based on task complexity
• Jupyter Ready: Works seamlessly in notebooks and Colab
• Safe Execution: Safely executes generated code and returns results
• Multi-Model Support: GPT-OSS-20B (fast), Qwen2.5-32B (balanced), Qwen3-480B (powerful)

Examples

Basic Functions

# Generate utility functions
code = keeya.generate("function to reverse a string")
code = keeya.generate("function to find duplicates in a list")
code = keeya.generate("function to sort a dictionary by values")

Data Science

# Data cleaning
cleaned_df = keeya.clean(df)

# Data analysis
analysis = keeya.analyze(df)

# Visualizations
keeya.visualize(df, plot_type='histogram')
keeya.visualize(df, plot_type='correlation')

# Machine learning
model = keeya.train(df, target='target_column')
predictions = model.predict(test_df)

Smart Model Selection

Keeya automatically selects the best AI model based on task complexity:

• GPT-OSS-20B (2-4 seconds): Fast fallback for simple tasks
• Qwen2.5-32B (3-6 seconds): Sweet spot for balanced performance
• Qwen3-480B (6-12 seconds): Worth the wait for complex tasks

Manual Model Selection

You can also specify a model manually:

# Use specific model
code = keeya.generate("complex function", model="qwen3-480b")
cleaned_df = keeya.clean(df, model="gpt-oss-20b")

# See available models
models = keeya.get_available_models()
print(models)

API Reference

keeya.generate(prompt, model=None)

Generate Python code from natural language prompt.

keeya.clean(df, model=None)

AI-powered data cleaning. Returns cleaned DataFrame.

keeya.analyze(df, model=None)

AI-powered data analysis. Returns analysis results.

keeya.visualize(df, plot_type=None, model=None)

AI-powered visualization. Creates and displays plots.

keeya.train(df, target, model=None)

AI-powered ML model training. Returns trained model.

keeya.get_available_models()

Get available models and their descriptions.

Requirements

• Python 3.8+
• pandas
• requests

License

MIT License