fractal-tictactoe-analyzer

A lightweight Python package that parses natural‑language descriptions of fractal Tic‑Tac‑Toe game states or strategies and returns a structured analysis. It can identify optimal moves, detect winning patterns, suggest game variations, and more—using a pattern‑matched response format that is easy to consume by downstream tools or AI agents.

Installation

pip install fractal_tictactoe_analyzer

Quick Start

from fractal_tictactoe_analyzer import fractal_tictactoe_analyzer

# Minimal usage – the function will create a default ChatLLM7 instance
response = fractal_tictactoe_analyzer(
    user_input="I have a 3‑level fractal board, the top‑left corner is X, the centre is O..."
)

print(response)   # -> List of extracted analysis strings

Advanced usage – providing your own LLM

You can pass any LangChain‑compatible chat model. The default is ChatLLM7 from the langchain_llm7 package.

OpenAI

from langchain_openai import ChatOpenAI
from fractal_tictactoe_analyzer import fractal_tictactoe_analyzer

my_llm = ChatOpenAI(model="gpt-4o")
response = fractal_tictactoe_analyzer(
    user_input="Explain the best move on this fractal board...",
    llm=my_llm
)

Anthropic

from langchain_anthropic import ChatAnthropic
from fractal_tictactoe_analyzer import fractal_tictactoe_analyzer

my_llm = ChatAnthropic(model="claude-3-5-sonnet")
response = fractal_tictactoe_analyzer(
    user_input="Find a winning pattern in the nested 2×2 grid.",
    llm=my_llm
)

Google Gemini

from langchain_google_genai import ChatGoogleGenerativeAI
from fractal_tictactoe_analyzer import fractal_tictactoe_analyzer

my_llm = ChatGoogleGenerativeAI(model="gemini-1.5-flash")
response = fractal_tictactoe_analyzer(
    user_input="Generate a new variation of fractal Tic‑Tac‑Toe.",
    llm=my_llm
)

Parameters

How It Works

1. Prompt Construction – The package builds a system prompt and a human prompt (defined in prompts.py) that guide the LLM to produce output matching a strict regular‑expression pattern.
2. Pattern Matching – The response from the LLM is validated against pattern (also defined in prompts.py). Only data that conforms to the pattern is returned.
3. Extraction – The llmatch helper extracts the structured pieces of information, returning them as a list of strings.

Rate Limits & API Keys

• The free tier of LLM7 provides generous rate limits that are sufficient for typical development and small‑scale usage.
• For higher throughput, obtain a personal API key from LLM7:
  • Register at: https://token.llm7.io/
  • Export it in your environment: export LLM7_API_KEY="your_key_here"
  • Or pass it directly to the function: api_key="your_key_here".

Support & Contributions

• GitHub Issues: https://github.com/chigwell/fractal-tictactoe-analyzer/issues
• Feel free to open an issue for bugs, feature requests, or general questions.

License

This project is licensed under the MIT License.

Author

Eugene Evstafev
Email: hi@euegne.plus
GitHub: https://github.com/chigwell