allendpoints/example.py

#!/usr/bin/env python3
"""
Example script demonstrating how to use allendpoints as a Python module.
"""

from allendpoints import (
    run_inference,
    check_available_apis,
    get_ollama_models,
    InferenceHandler,
    CONFIG,
    check_provider_key_available
)

def basic_example():
    """Basic usage example of allendpoints."""
    print("\n=== BASIC EXAMPLE ===")
    
    # Inference with default model for Ollama
    default_model = CONFIG["defaults"]["ollama"]
    response = run_inference(
        prompt="What is the capital of France?",
        provider="ollama",
        model=default_model
    )
    
    print(f"Response from Ollama (model: {default_model}): {response}")
    
    # Inference with specific model and system prompt
    response = run_inference(
        prompt="Write a haiku about AI",
        provider="ollama",
        model="llama3.2:3b",
        system_content="You are a poetic assistant that only writes in haiku."
    )
    
    print(f"\nHaiku from Ollama (llama3.2:3b):\n{response}")

def provider_availability_example():
    """Example showing how to check provider availability."""
    print("\n=== PROVIDER AVAILABILITY EXAMPLE ===")
    
    # Check which providers are available (have valid API keys)
    available_providers = check_available_apis()
    print(f"Available providers: {', '.join(available_providers)}")
    
    # Check for specific providers
    providers_to_check = ["ollama", "gemini", "github", "hf", "together", "aiql", "groq", "nvidia"]
    
    for provider in providers_to_check:
        is_available = check_provider_key_available(provider)
        status = "✅ Available" if is_available else "❌ Not available"
        print(f"{provider}: {status}")

def model_listing_example():
    """Example showing how to list available models."""
    print("\n=== MODEL LISTING EXAMPLE ===")
    
    # Get available Ollama models
    try:
        ollama_models = get_ollama_models()
        print(f"Available Ollama models: {', '.join(ollama_models[:5])}...")
        print(f"Total Ollama models: {len(ollama_models)}")
    except Exception as e:
        print(f"Error getting Ollama models: {str(e)}")
    
    # Show configured models for each provider
    print("\nConfigured models per provider:")
    for provider, models in CONFIG["models"].items():
        model_count = len(models)
        print(f"{provider}: {model_count} models configured")
        
        # Handle both list and dictionary model configurations
        if isinstance(models, dict):
            # For dictionary-based configurations (most providers)
            sample_models = list(models.keys())[:3]
            if sample_models:
                print(f"  Sample models: {', '.join(sample_models)}")
        elif isinstance(models, list):
            # For list-based configurations (ollama)
            sample_models = models[:3]
            if sample_models:
                print(f"  Sample models: {', '.join(sample_models)}")

def direct_provider_example():
    """Example showing how to use provider handlers directly."""
    print("\n=== DIRECT PROVIDER EXAMPLE ===")
    
    # Check if Ollama is available
    if check_provider_key_available("ollama"):
        try:
            # Use the Ollama handler directly
            response = InferenceHandler.ollama(
                prompt="Explain how a computer works in one paragraph",
                model="llama3.2:3b"
            )
            print(f"Direct Ollama response:\n{response}")
        except Exception as e:
            print(f"Error with direct Ollama call: {str(e)}")
    
    # Check if Gemini is available
    if check_provider_key_available("gemini"):
        try:
            # Use the Gemini handler directly
            response = InferenceHandler.gemini(
                prompt="What is quantum computing?",
                model="gemini-1.5-pro"
            )
            print(f"\nDirect Gemini response:\n{response[:150]}...")
        except Exception as e:
            print(f"Error with direct Gemini call: {str(e)}")

def batch_processing_example():
    """Example showing how to process multiple prompts with multiple providers."""
    print("\n=== BATCH PROCESSING EXAMPLE ===")
    
    # Define a list of prompts
    prompts = [
        "What is machine learning?",
        "Explain the theory of relativity briefly"
    ]
    
    # Get available providers (only use the first 2 for this example)
    available_providers = check_available_apis()[:2]
    
    if not available_providers:
        print("No providers available for batch processing")
        return
    
    print(f"Processing {len(prompts)} prompts with {len(available_providers)} providers: {', '.join(available_providers)}")
    
    # Process each prompt with each provider
    for prompt in prompts:
        print(f"\nPrompt: {prompt}")
        
        for provider in available_providers:
            try:
                # Get default model for this provider
                default_model = CONFIG["defaults"][provider]
                
                # Run inference with explicit model parameter
                response = run_inference(prompt, provider, model=default_model)
                
                # Print truncated response
                print(f"  {provider} ({default_model}): {response[:100]}...")
            except Exception as e:
                print(f"  Error with {provider}: {str(e)}")

def main():
    """Run all examples."""
    print("AllEndpoints Python Module Examples")
    print("==================================")
    
    # Run examples
    basic_example()
    provider_availability_example()
    model_listing_example()
    direct_provider_example()
    batch_processing_example()
    
    print("\nExamples completed!")

if __name__ == "__main__":
    main()