Initial Implementation of PydanticPrompt Library

[abrichr]

Initial Implementation of PydanticPrompt Library

This PR introduces PydanticPrompt, a lightweight Python library for documenting Pydantic models for LLM interactions using standard Python docstrings.

Overview

PydanticPrompt allows developers to document their Pydantic models with standard Python docstrings and then easily format that documentation for inclusion in LLM prompts, providing a clean, consistent way to describe expected outputs for large language models.

Key Features

• Simple decorator interface (@prompt_schema)
• Uses standard Python docstrings for field documentation
• Automatic type inference including generics and nested models
• Optional validation rule display
• Warnings for undocumented fields
• Seamless integration with existing Pydantic models

Implementation Details

• Core functionality built around field inspection and docstring extraction
• Uses Pydantic's JSON schema for validation rule discovery
• Modern type annotations with mypy validation
• Automated CI setup with GitHub Actions
• Comprehensive linting with Ruff

Example Usage

from pydantic import BaseModel, Field
from pydantic_prompt import prompt_schema

@prompt_schema
class UserProfile(BaseModel):
    name: str
    """The full name of the user"""
    
    age: int = Field(ge=0)
    """User's age in years - must be a positive integer"""
    
    email: str = Field(pattern=r"[^@]+@[^@]+\.[^@]+")
    """Valid email address that will be used for communication"""

# In an LLM prompt
prompt = f"""
Extract a UserProfile from this text according to this schema:

{UserProfile.format_for_llm(include_validation=True)}

Text: John Doe is 32 years old, contact him at [email protected]
"""

Test Results

All tests are passing:

====================================================================== test session starts ======================================================================
platform linux -- Python 3.12.1, pytest-8.3.5, pluggy-1.5.0 -- /workspaces/PydanticPrompt/.venv/bin/python
cachedir: .pytest_cache
rootdir: /workspaces/PydanticPrompt
configfile: pyproject.toml
testpaths: tests
collected 4 items                                                                                                                                               

tests/test_pydantic_prompt.py::test_basic_docstring_extraction PASSED                                                                                     [ 25%]
tests/test_pydantic_prompt.py::test_optional_fields PASSED                                                                                                [ 50%]
tests/test_pydantic_prompt.py::test_validation_rules PASSED                                                                                               [ 75%]
tests/test_pydantic_prompt.py::test_nested_models PASSED                                                                                                  [100%]

======================================================================= 4 passed in 0.13s =======================================================================

Changes Made

• Created core implementation with docstring extraction
• Added comprehensive type hints with mypy validation
• Set up CI with GitHub Actions for automated testing
• Added linting script for consistent code style
• Implemented warnings for undocumented fields
• Created detailed documentation and examples

[codecov]

Welcome to Codecov 🎉

Once you merge this PR into your default branch, you're all set! Codecov will compare coverage reports and display results in all future pull requests.

ℹ️ You can also turn on project coverage checks and project coverage reporting on Pull Request comment

Thanks for integrating Codecov - We've got you covered ☂️