Silhouette

Silhouette is an AI-powered Python code transformation tool that uses libcst, instructor and GPT models to enhance your code with PEP-compliant docstrings and type hints.

Features

• Automatically generate comprehensive docstrings for functions, classes, and modules
• Add type hints to function parameters and return values
• Preserve existing code structure and comments
• Use LibCST to validate python code
• Use instructor and pydantic to validate gpt outputs

Example Usage: Add docstring and typing.

• note typing works better if the function already has a docstring

from silhouette.cst_transformers import add_type_hints, add_docstrings
import os

api_key = os.getenv("OPENAI_API_KEY")

source_code = """
def calculate_average(numbers, weights=None):
    total = 0
    count = 0
    for i, num in enumerate(numbers):
        if weights:
            total += num * weights[i]
            count += weights[i]
        else:
            total += num
            count += 1
    return total / count if count else 0
"""

documented_code = add_docstrings(source_code, api_key)

result = add_type_hints(documented_code, api_key)
print(result)

Output

def calculate_average(numbers: List[float], weights: Optional[List[float]]=None) -> float:
    """This function calculates the weighted or unweighted average of a list of numbers.
    
    Args:
        numbers: A list of numbers for which the average is to be calculated.
        weights: An optional list of weights for each number. If provided, the function calculates the weighted average. If not provided, the function calculates the unweighted average.
    
    Returns:
        The weighted or unweighted average of the input numbers. If the list of numbers is empty, the function returns 0.
    """
    total = 0
    count = 0
    for i, num in enumerate(numbers):
        if weights:
            total += num * weights[i]
            count += weights[i]
        else:
            total += num
            count += 1
    return total / count if count else 0

Example Usage: Add New Imports

from silhouette.utils.cst_helpers import add_import

result = add_import(result, "from typing import Optional, List")
print(result)

Output

from typing import Optional, List
def calculate_average(numbers: List[float], weights: Optional[List[float]]=None) -> float:
    """This function calculates the weighted or unweighted average of a list of numbers.
    
    Args:
        numbers: A list of numbers for which the average is to be calculated.
        weights: An optional list of weights for each number. If provided, the function calculates the weighted average. If not provided, the function calculates the unweighted average.
    
    Returns:
        The weighted or unweighted average of the input numbers. If the list of numbers is empty, the function returns 0.
    """
    total = 0
    count = 0
    for i, num in enumerate(numbers):
        if weights:
            total += num * weights[i]
            count += weights[i]
        else:
            total += num
            count += 1
    return total / count if count else 0