Merge remote-tracking branch 'Farmerobot/main'

# Conflicts:
#	src/llm_postor/game/game_state.py

README.md

@@ -1,4 +1,4 @@
-# Analysis of Persuasive Capabilities of Large Language Models in Computer Games
+# LLMPostor
 
 [![Tests](https://github.com/Farmerobot/mk-ai-agents/actions/workflows/test.yml/badge.svg)](https://github.com/Farmerobot/mk-ai-agents/actions/workflows/test.yml)
 
@@ -10,58 +10,108 @@ This project aims to test and compare the persuasive abilities of large language
 
 ## Current Features
 
-1. **Web GUI**: A user-friendly interface built using [insert framework used, e.g., Flask, Streamlit] that allows:
-   - Observation of game progress in real-time.
-   - Analysis of specific agent actions at various points in the game, including transcripts of conversations and decision-making processes.
-   - Insight into the thought and decision-making process of LLMs through logging of their internal states (if available from the LLM).
+TODO
 
-2. **Interactive Map**: A visual representation of the game environment, showing locations and movements of AI agents, using [insert library used, e.g., Pygame, a JavaScript library].
+## Project Architecture
 
-3. **Post-Game Analysis**: After the game concludes, the system provides an in-depth analysis of the persuasive abilities demonstrated by AI agents, including metrics such as:
-    * Success rate of persuasive attempts.
-    * Frequency of different persuasive techniques used.
-    * Correlation between LLM model and persuasive success.
+The project is structured into several modules:
 
-4. **"Among Us" Simulation**: The project uses the popular game "Among Us" as a framework to test AI agents' abilities in interaction, deception, and persuasion.  The core mechanics of tasks, meetings, voting, and impostors are replicated.
+* **`agents`**: Contains implementations of different AI agents, each using a specific LLM.
+* **`models`**: Defines data structures for tasks and game state.
+* **`players`**: Defines the players.
 
-5. **Multiple AI Agents**: The system supports multiple AI agents, each of which can be powered by a different LLM (e.g., GPT-3, GPT-4, other open-source models), enabling comparative analysis of their persuasive capabilities.
+## Table of Contents
 
-## Project Architecture
+- Installation
+- Usage
+- Configuration
+- Running Tests
+- Contributing
+- License
 
-The project is structured into several modules:
+## Installation
 
-* **`game_engine`**: Manages the game logic, including task assignment, player interactions, and game state transitions.
-* **`agents`**: Contains implementations of different AI agents, each using a specific LLM.
-* **`models`**: Defines data structures for players, tasks, and game state.
-* **`gui`**: Handles the user interface and visualization.
-* **`utils`**: Contains helper functions.
+This project requires Python 3.11 or higher.
 
-## Technologies Used
+### 1. Install Poetry
 
-* **Python 3.11**: Programming language.
-* **[List Libraries/Frameworks Used, e.g., Pydantic, Flask, Pygame, etc.]**:  Specific libraries and frameworks used for data modeling, web development, game visualization, etc.
-* **OpenAI API**: For accessing large language models.
+If you don't have Poetry installed, you can install it by following these steps:
 
+- For Unix/macOS:
+```bash
+  curl -sSL https://install.python-poetry.org | python3 -
+```
+- For Windows:
 
-## Running Instructions
+  Visit the official [Poetry installation page](https://python-poetry.org/docs/#installation) for Windows-specific instructions.
 
-Required Python version: 3.11
+Once installed, verify Poetry's installation by running:
+```bash
+  poetry --version
+```
+### 2. Clone the Repository and Install Dependencies
 
-The project uses the `poetry` package manager, which can be installed following the instructions at [https://python-poetry.org/docs/](https://python-poetry.org/docs/)
+- Clone the repository:
 
-To install all dependencies using `poetry`, execute the following command:
 ```bash
-poetry install
+  git clone https://github.com/username/llm_poster.git
+  cd llm_poster
 ```
 
-Due to the use of large language models, it is necessary to provide an OpenAI API key. The key can be obtained at [https://beta.openai.com/signup/](https://beta.openai.com/signup/). After obtaining the key, assign it to the `OPENAI_API_KEY` environment variable.
+- Install dependencies using Poetry:
 
-To run the demonstration version, execute the following command:
 ```bash
-cd src
-poetry run python demo.py
+  poetry install
 ```
 
+### 3. Activate the Virtual Environment
+
+Once the dependencies are installed, activate the virtual environment:
+
+```bash
+  poetry shell
+```
+
+## Usage
+
+To run the simulation with the GUI:
+
+```bash
+  poetry run run-gui
+```
+
+## Configuration
+
+The project requires API keys for LLMs such as OpenAI. Set the following environment variables in your shell:
+
+```bash
+  export OPENAI_API_KEY=your-api-key
+```
+
+Alternatively, you can set these in a `.env` file at the project root.
+
+## Running Tests
+
+To run all tests, use the following command:
+
+```bash
+  poetry run pytest
+```
+
+## Contributing
+
+Contributions are welcome! Please follow these steps to contribute:
+
+1. Fork the repository.
+2. Create a new branch (git checkout -b feature-branch).
+3. Commit your changes (git commit -m 'Add new feature').
+4. Push to the branch (git push origin feature-branch).
+5. Open a pull request on GitHub.
+
+## License
+
+This project is licensed under the MIT License. See the LICENSE file for details.
+
 ## Future Enhancements (Roadmap)
 
 We are planning to add the following features to enhance the project:
@@ -76,7 +126,6 @@ We are planning to add the following features to enhance the project:
 - **API Documentation**: Comprehensive API documentation for integrating custom LLM models or analytical tools.
 - **Automated Testing**:  Implementation of automated tests to ensure the robustness and reliability of the game engine and AI agents.
 
-
 These planned features aim to make the project more comprehensive and versatile for researchers and enthusiasts in AI and NLP fields.
 
 This project offers a unique and interactive way to study the persuasive abilities of LLMs in a controlled, game-like environment, providing valuable insights for researchers in the field of artificial intelligence and natural language processing.

pyproject.toml

@@ -1,15 +1,15 @@
 [tool.poetry]
-name = "mk-ai-agent-test"
+name = "llm_postor"
 version = "0.1.0"
-description = ""
-authors = ["Marcin Korcz, Andrii Chmutov, Vasyl Korzavatykh, Mateusz Idziejczak, Mateusz Stawicki"]
+description = "This project tests and compares the persuasive abilities of large language models (LLMs) in a game-like environment simulating 'Among Us', where AI agents powered by LLMs interact, make decisions, and attempt to persuade each other."
+authors = ["Marcin Korcz", "Andrii Chmutov", "Vasyl Korzavatykh", "Mateusz Idziejczak", "Mateusz Stawicki"]
 readme = "README.md"
-package-mode = false
+package-mode = true
 
 [tool.poetry.dependencies]
 python = "^3.11"
 numpy = "^1.26.4"
-openai = "^1.23.1"
+openai = "^1.23.1"  
 pandas = "^2.2.2"
 torch = "^2.2.2"
 torchvision = "^0.17.2"
@@ -22,22 +22,24 @@ llama-index-embeddings-huggingface = "^0.2.0"
 ipykernel = "^6.29.4"
 ipywidgets = "^8.1.2"
 langchain = "^0.3.3"
-langchain-community = "^0.3.2"
+langchain-community = "^0.3.2"  
 langchain-openai = "^0.2.2"
 langchain-google-genai = "^2.0.1"
 streamlit = "^1.39.0"
-pytest = "^8.3.3"
 st-annotated-text = "^4.0.1"
 
-[tool.pytest.ini_options]                                                                                                                                     
-pythonpath = ["src"]
-testpaths = ["tests"]
-
 [tool.poetry.group.dev.dependencies]
 ipykernel = "^6.29.4"
 pytest = "^8.3.3"
 pytest-mock = "^3.14.0"
 
+[tool.pytest.ini_options]
+pythonpath = ["src"]
+testpaths = ["tests"]
+
 [build-system]
 requires = ["poetry-core"]
-build-backend = "poetry.core.masonry.api"
+build-backend = "poetry.core.masonry.api"
+
+[tool.poetry.scripts]
+run-gui = "scripts.run_gui:run_gui"

scripts/run_gui.py

@@ -0,0 +1,19 @@
+import os
+import subprocess
+
+def run_gui():
+    # Define the project root directory (this assumes the script is located in the scripts/ folder)
+    project_root = os.path.abspath(os.path.join(os.path.dirname(__file__), '..'))
+
+    # Navigate to the project root directory
+    os.chdir(project_root)
+
+    # Define the path to the demo_gui.py script relative to the project root
+    demo_gui_path = os.path.join(project_root, "src", "llm_postor", "demo_gui.py")
+
+    # Check if the demo_gui.py file exists
+    if not os.path.exists(demo_gui_path):
+        raise FileNotFoundError(f"File does not exist: {demo_gui_path}")
+
+    # Run the Streamlit app with the correct path
+    subprocess.run(["poetry", "run", "streamlit", "run", demo_gui_path], check=True)

scripts/run_gui.sh

@@ -1,10 +1,10 @@
-#!/bin/bash
+    #!/bin/bash
 
-# Get the directory of the current script
-SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+    # Get the directory of the current script
+    SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
 
-# Navigate to the project root (assuming this script is located inside `scripts/` directory)
-cd "$SCRIPT_DIR/.."
+    # Navigate to the project root (assuming this script is located inside `scripts/` directory)
+    cd "$SCRIPT_DIR/.."
 
-# Run the Streamlit app in the Poetry environment
-poetry run streamlit run src/demo_gui.py
+    # Run the Streamlit app in the Poetry environment
+    poetry run streamlit run src/demo_gui.py

src/annotation.py → src/llm_postor/annotation.py

@@ -1,11 +1,11 @@
 import os
 import json
 from typing import List
-
 from langchain_openai import ChatOpenAI
-from game.llm_prompts import ANNOTATION_TEMPLATE
 from langchain.schema import HumanMessage
 
+from llm_postor.game.llm_prompts import ANNOTATION_TEMPLATE
+
 def annotate_dialogue(dialogue: str, llm_model_name: str = "gpt-4o-mini") -> str:
     """
     Annotates a dialogue with persuasion techniques using OpenAI API.

src/demo_gui.py → src/llm_postor/demo_gui.py

@@ -1,26 +1,22 @@
 import json
 import time
-from openai import OpenAIError
 import streamlit as st
-from game.game_engine import GameEngine
-from game.players.ai import AIPlayer
-from game.gui_handler import GUIHandler
-from game.game_state import GameState
+from openai import OpenAIError
 from types import SimpleNamespace as Namespace
-from game.players.fake_ai import FakeAIPlayer
-from game.models.engine import GamePhase
-from game.chat_analyzer import ChatAnalyzer
-
 
+from llm_postor.game.game_engine import GameEngine
+from llm_postor.game.players.ai import AIPlayer
+from llm_postor.game.gui_handler import GUIHandler
+from llm_postor.game.game_state import GameState
+from llm_postor.game.players.fake_ai import FakeAIPlayer
+from llm_postor.game.models.engine import GamePhase
+from llm_postor.game.chat_analyzer import ChatAnalyzer
 
 # To run this script, you need to poetry install and then run the following command:
 # streamlit run src/demo_gui.py
 
-
-
-
 def main():
-    st.title("Among Us Game - susGPT")
+    st.title("Among Us Game - LLMPostor")
     gui_handler = GUIHandler()
     game_engine = GameEngine()
 

src/game/agents/adventure_agent.py → src/llm_postor/game/agents/adventure_agent.py

@@ -1,13 +1,11 @@
+import re
 from typing import List, Any
-
-from langchain_openai import ChatOpenAI
-
-from game.consts import ASCII_MAP
-from .base_agent import Agent
 from langchain.schema import HumanMessage
-from game.llm_prompts import ADVENTURE_PLAN_TEMPLATE, ADVENTURE_ACTION_TEMPLATE
-import re
+from langchain_openai import ChatOpenAI
 
+from llm_postor.game.agents.base_agent import Agent
+from llm_postor.game.consts import ASCII_MAP
+from llm_postor.game.llm_prompts import ADVENTURE_PLAN_TEMPLATE, ADVENTURE_ACTION_TEMPLATE
 
 class AdventureAgent(Agent):
     def update_state(

src/game/agents/base_agent.py → src/llm_postor/game/agents/base_agent.py

@@ -2,9 +2,8 @@
 from typing import Any, List
 from pydantic import BaseModel, Field
 
-from game.agents.usage_metadata import UsageMetadata
-from game.consts import TOKEN_COSTS
-
+from llm_postor.game.agents.usage_metadata import UsageMetadata
+from llm_postor.game.consts import TOKEN_COSTS
 
 class AgentState(BaseModel):
     history: str = Field(default_factory=str)

src/game/agents/discussion_agent.py → src/llm_postor/game/agents/discussion_agent.py

@@ -1,8 +1,9 @@
 from typing import List, Any
 from .base_agent import Agent
 from langchain.schema import HumanMessage
-from game.llm_prompts import DISCUSSION_TEMPLATE, DISCUSSION_RESPONSE_TEMPLATE
 
+from llm_postor.game.llm_prompts import DISCUSSION_TEMPLATE
+from llm_postor.game.llm_prompts import DISCUSSION_RESPONSE_TEMPLATE
 
 class DiscussionAgent(Agent):
     def update_state(

src/game/agents/voting_agent.py → src/llm_postor/game/agents/voting_agent.py

@@ -1,9 +1,9 @@
+import re
 from typing import List, Any
 from .base_agent import Agent
 from langchain.schema import HumanMessage
-from game.llm_prompts import VOTING_TEMPLATE
-import re
 
+from llm_postor.game.llm_prompts import VOTING_TEMPLATE
 
 class VotingAgent(Agent):
     def update_state(

src/game/chat_analyzer.py → src/llm_postor/game/chat_analyzer.py

@@ -1,10 +1,11 @@
+import re
 from typing import List, Dict
 from pydantic import BaseModel
-from game.players.base_player import Player
 from langchain_openai import ChatOpenAI
-from game.llm_prompts import PERSUASION_TECHNIQUES
 from langchain.schema import HumanMessage
-import re
+
+from llm_postor.game.players.base_player import Player
+from llm_postor.game.llm_prompts import PERSUASION_TECHNIQUES
 
 class ChatAnalyzer(BaseModel):
     players: List[Player]