README.md

PositronicPython

PositronicPython is an innovative Python framework designed to dynamically augment Python code with decision-making capabilities using Large Language Models (LLMs). It utilizes Python's inspect module to analyze code at runtime and augment it with LLM-generated code or data, enhancing the Python language with human-like decision-making capabilities.

The vision

To augment the python language with a human-like resourcefulness and decision making capabilities, making the development of gen AI agents to be more natural, repeatable, and testable.

Roadmap features:

• Dynamic Code Augmentation: Automatically enhances Python code by injecting intelligent, context-aware snippets generated by LLMs.
• Real-Time Analysis: Uses runtime inspection to understand and modify code behavior dynamically.
• Seamless Integration: Easy to integrate with existing Python projects, requiring minimal changes to the codebase.
• Customizable AI Models: Supports customization of the AI model to better fit specific use cases and requirements.

Available features:

• Automatic population: of python data classes by analysing the source code of the class and following the user instruction

Quick Start

Here’s a simple example to get you started with PositronicPython:

Example: Enhancing a Python Class

pythonCopy code

import dataclasses
import json
import logging
from typing import List

from framework.ai_interface import ai_enhanced


@ai_enhanced
class EnglishTeacher:
    # A class representing an assignment given by an English teacher to a Hebrew-speaking student.
    # The teacher is very friendly and supportive and has a good sense of humor. He wants to make sure the student will succeed in the test
    # The teacher is very creative, and he wants to make sure the students will enjoy the learning process
    class Word:
        word_str: str
        definition_english: str
        explanation_hebrew: str
        examples: List[str]
        synonyms: List[str]

    introduction: str # in English
    words: List[Word]  # Use HTML to make it more engaging
    funny_story_containing_the_words: str # Mark important words in bold (<b>)
    exercises: List[str]  # List of exercises to practice the words structure as HTML. Make it easy for the student to test himself. Use colors to make it engaging
    solutions: List[str]


words = "Folloed by, Praced, Suited, Informed, Detected, Perspective"

logging.info("Generating the exercise")
exercise = EnglishTeacher.generate(words)
logging.info("Done generating the exercise")
exercise.to_html("Structure the page in a form of cards. Hebrew explanations should be next to the English ones (same line). It should be easy to read")
print("Done")