removed anything webui: let's keep things simple

README.md

@@ -1,6 +1,6 @@
 # DualMind: AI Conversation Simulator
 
-DualMind is an innovative AI conversation simulator that facilitates engaging dialogues between two AI models using the Ollama API. It offers both a command-line interface (CLI) and a Streamlit-based web interface for immersive and customizable AI interactions.
+DualMind is an innovative AI conversation simulator that facilitates engaging dialogues between two AI models using the Ollama API. It offers a command-line interface (CLI) for immersive and customizable AI interactions.
 
 ![Screenshot](imgs/screenshot.png)
 
@@ -8,15 +8,13 @@ DualMind is an innovative AI conversation simulator that facilitates engaging di
 
 - 🤖 Dual-model conversation: Engage two different AI models in a thought-provoking dialogue
 - 🎭 Customizable system prompts: Tailor the behavior and personality of each AI model
-- 🖥️ Multiple interface options:
+- 🖥️ Command-line interface for quick interactions
-  - Command-line interface for quick interactions
-  - Streamlit web interface for a user-friendly experience
 - 🛠️ Conversation customization:
   - Adjust the number of exchanges
   - Modify the initial prompt
   - Select different AI models
-- 💾 Save and download conversation logs
+- 💾 Save conversation logs
-- 🎨 Responsive and visually appealing design
+- 🔢 Token count display and limit: Monitor the token usage for each message and limit the conversation to a specified number of tokens.
 
 ## Prerequisite: Ollama
 
@@ -46,6 +44,7 @@ Please refer to [Ollama](https://ollama.com/download) to install Ollama on your
    MODEL_1=llama2
    MODEL_2=mistral
    INITIAL_PROMPT="Let's discuss the future of AI. What are your thoughts on its potential impact on society?"
+   MAX_TOKENS=8000
    ```
 
    Feel free to use the env.example file as a template.
@@ -54,42 +53,29 @@ Please refer to [Ollama](https://ollama.com/download) to install Ollama on your
 
 ## Usage
 
-### Command-line Interface
-
 To run DualMind in CLI mode:
 
 ```sh
 ./run_cli.sh
 ```
 
-### Streamlit Web Interface
-
-To run DualMind in Streamlit mode:
-
-```sh
-./run_streamlit.sh
-```
-
-Then, open your web browser and navigate to the URL provided in the terminal (usually `http://localhost:8501`).
-
 ## Customization
 
 ### System Prompts
 
 You can customize the system prompts for each AI model by editing the `system_prompt_1.txt` and `system_prompt_2.txt` files in the project root.
 
-### Styling
+### Options
+
+You can customize the options for the conversation by editing the `options.json` file in the project root.
 
-The appearance of the Streamlit interface can be customized by modifying the `style/custom.css` file.
 
 ## Project Structure
 
 - `main.py`: Entry point of the application
 - `ai_conversation.py`: Core logic for AI conversations
-- `streamlit_app.py`: Streamlit web interface implementation
-- `style/custom.css`: Custom styles for the web interface
 - `run_cli.sh`: Shell script to run the CLI version
-- `run_streamlit.sh`: Shell script to run the Streamlit version
+
 
 ## Contributing
 
@@ -102,4 +88,3 @@ This project is open source and available under the [MIT License](LICENSE).
 ## Acknowledgements
 
 - This project uses the [Ollama](https://ollama.ai/) API for AI model interactions.
-- The web interface is built with [Streamlit](https://streamlit.io/).

ai_conversation.py

@@ -41,7 +41,7 @@ class AIConversation:
                 break  # Avoid removing the system message
         return messages
 
-    def start_conversation(self, initial_message, num_exchanges=0):
+    def start_conversation(self, initial_message, num_exchanges=0, options=None):
         # Main conversation loop
         current_message = initial_message
         color_1, color_2 = "cyan", "yellow"
@@ -78,10 +78,7 @@ class AIConversation:
                 response = self.client.chat(
                     model=self.current_model,
                     messages=messages,
-                    options={
+                    options=options,
-                        "temperature": 0.7,  # Control randomness
-                        "repeat_penalty": 1.2,  # Penalize repetition
-                    },
                 )
                 response_content = response["message"]["content"]
 

main.py

@@ -1,5 +1,5 @@
 import os
-import argparse
+import json
 from dotenv import load_dotenv, set_key
 from ai_conversation import AIConversation
 
@@ -8,6 +8,11 @@ def load_system_prompt(filename):
     with open(filename, "r") as file:
         return file.read().strip()
 
+def load_options_from_json(filename):
+    """Load options from a JSON file."""
+    with open(filename, "r") as file:
+        return json.load(file)
+
 def main():
     # Load environment variables
     load_dotenv()
@@ -25,38 +30,24 @@ def main():
     max_tokens = int(os.getenv("MAX_TOKENS", 4000))
     print(f"Max tokens: {max_tokens}")
 
+    # Load options from JSON file
+    options = load_options_from_json("options.json")
+    print(f"Options: {options}")
+
     # Initialize the AI conversation object
     conversation = AIConversation(
         model_1, model_2, system_prompt_1, system_prompt_2, ollama_endpoint, max_tokens
     )
 
-    # Set up command-line argument parser
-    parser = argparse.ArgumentParser(description="AI Conversation")
-    parser.add_argument("--cli", action="store_true", help="Run in CLI mode")
-    parser.add_argument(
-        "--streamlit", action="store_true", help="Run in Streamlit mode"
-    )
-    args = parser.parse_args()
 
     # Run the appropriate interface based on command-line arguments
-    if args.cli:
+    run_cli(conversation, initial_prompt, options)
-        run_cli(conversation, initial_prompt)
-    elif args.streamlit:
-        run_streamlit(conversation, initial_prompt)
-    else:
-        print("Please specify either --cli or --streamlit mode.")
 
-def run_cli(conversation, initial_prompt):
+def run_cli(conversation, initial_prompt, options):
     """Run the conversation in command-line interface mode."""
     load_dotenv()
-    conversation.start_conversation(initial_prompt, num_exchanges=0)
+    conversation.start_conversation(initial_prompt, num_exchanges=0, options=options)
 
-def run_streamlit(conversation, initial_prompt):
-    """Run the conversation in Streamlit interface mode."""
-    import streamlit as st
-    from streamlit_app import streamlit_interface
-
-    streamlit_interface(conversation, initial_prompt)
 
 if __name__ == "__main__":
     main()

options.json

@@ -0,0 +1,8 @@
+{
+    "temperature": 0.8,
+    "top_k": 40,
+    "top_p": 0.9,
+    "repeat_penalty": 1.1,
+    "presence_penalty": 0.5,
+    "frequency_penalty": 0.5
+    }

requirements.txt

@@ -1,6 +1,6 @@
 python-dotenv
 requests
 termcolor
-streamlit
 Pillow
 tiktoken
+ollama

start_gui.sh

@@ -1,2 +0,0 @@
-#!/bin/bash
-streamlit run main.py -- --streamlit

streamlit_app.py

@@ -1,246 +0,0 @@
-import streamlit as st
-import os
-import datetime
-from dotenv import load_dotenv
-
-
-# Function to load and apply custom CSS
-def load_css(file_name):
-    with open(file_name, "r") as f:
-        st.markdown(f"<style>{f.read()}</style>", unsafe_allow_html=True)
-
-
-# Function to set page configuration
-def set_page_config():
-    st.set_page_config(
-        page_title="DualMind",
-        page_icon="🤖",
-        layout="wide",
-        initial_sidebar_state="expanded",
-    )
-
-
-def streamlit_interface(conversation, initial_prompt):
-    set_page_config()
-    load_css("style/custom.css")
-
-    st.markdown(
-        '<h1 class="main-title">🤖 DualMind</h1>', unsafe_allow_html=True
-    )
-
-    # Sidebar for customization
-    st.sidebar.title("🛠️ Settings")
-
-    # Load default values from .env
-    load_dotenv()
-    default_endpoint = os.getenv("OLLAMA_ENDPOINT")
-    default_model_1 = os.getenv("MODEL_1")
-    default_model_2 = os.getenv("MODEL_2")
-
-    # Sidebar for customization
-    ollama_endpoint = st.sidebar.text_input("Ollama Endpoint", value=default_endpoint)
-
-    # Update the OllamaClient endpoint
-    conversation.ollama_client.endpoint = ollama_endpoint
-
-    # Fetch available models
-    try:
-        available_models = conversation.ollama_client.get_available_models()
-    except Exception as e:
-        st.error(f"Error fetching models: {str(e)}")
-        available_models = []
-
-    # Model selection dropdowns
-    model_1 = st.sidebar.selectbox(
-        "Model 1",
-        options=available_models,
-        index=(
-            available_models.index(default_model_1)
-            if default_model_1 in available_models
-            else 0
-        ),
-    )
-    model_2 = st.sidebar.selectbox(
-        "Model 2",
-        options=available_models,
-        index=(
-            available_models.index(default_model_2)
-            if default_model_2 in available_models
-            else 0
-        ),
-    )
-
-    # System prompt customization
-    col1, col2 = st.columns(2)
-
-    with col1:
-        with st.expander("Customize System Prompt 1"):
-            system_prompt_1 = st.text_area(
-                "System Prompt 1", value=conversation.system_prompt_1, height=150
-            )
-            if st.button("Save Prompt 1"):
-                with open("system_prompt_1.txt", "w") as f:
-                    f.write(system_prompt_1)
-                st.success("System Prompt 1 saved!")
-
-    with col2:
-        with st.expander("Customize System Prompt 2"):
-            system_prompt_2 = st.text_area(
-                "System Prompt 2", value=conversation.system_prompt_2, height=150
-            )
-            if st.button("Save Prompt 2"):
-                with open("system_prompt_2.txt", "w") as f:
-                    f.write(system_prompt_2)
-                st.success("System Prompt 2 saved!")
-
-    # Update conversation with new settings
-    conversation.model_1 = model_1
-    conversation.model_2 = model_2
-    conversation.system_prompt_1 = system_prompt_1
-    conversation.system_prompt_2 = system_prompt_2
-
-    if "messages" not in st.session_state:
-        st.session_state.messages = []
-
-    if "current_message" not in st.session_state:
-        st.session_state.current_message = initial_prompt
-
-    # Add this new section for customizing the initial message
-    initial_message = st.text_area(
-        "Customize initial message:", value=st.session_state.current_message
-    )
-    if st.button("Set Initial Message"):
-        st.session_state.current_message = initial_message
-        st.success("Initial message updated!")
-
-    if "exchange_count" not in st.session_state:
-        st.session_state.exchange_count = 0
-
-    # Update the chat message display
-    for message in st.session_state.messages:
-        with st.chat_message(
-            message["role"], avatar="🧑" if message["role"] == "user" else "🤖"
-        ):
-            st.markdown(
-                f'<div class="chat-message {"user-message" if message["role"] == "user" else "assistant-message"}">{message["content"]}</div>',
-                unsafe_allow_html=True,
-            )
-
-    num_exchanges = st.number_input(
-        "Number of exchanges", min_value=1, max_value=10, value=3
-    )
-
-    if st.button("Generate Responses"):
-        with st.spinner("Generating responses..."):
-            for _ in range(num_exchanges):
-                response = conversation.get_conversation_response(
-                    st.session_state.current_message
-                )
-                model_name, content = response.split("\n", 1)
-
-                avatar = "🔵" if model_name == model_1 else "🟢"
-                st.session_state.messages.append(
-                    {
-                        "role": "assistant",
-                        "content": f"**{model_name}**\n\n{content}",
-                        "avatar": avatar,
-                    }
-                )
-                st.session_state.current_message = content
-                st.session_state.exchange_count += 1
-
-                with st.chat_message("assistant", avatar=avatar):
-                    st.markdown(f"**{model_name}**\n\n{content}")
-
-    if st.button("Reset Conversation"):
-        st.session_state.messages = []
-        st.session_state.current_message = (
-            initial_message  # Use the customized initial message here
-        )
-        st.session_state.exchange_count = 0
-        conversation.current_model = conversation.model_1
-        conversation.current_system_prompt = conversation.system_prompt_1
-
-    st.write(f"Total exchanges: {st.session_state.exchange_count}")
-
-    user_input = st.text_input("Your message:", key="user_input")
-    if st.button("Send"):
-        if user_input:
-            st.session_state.messages.append(
-                {"role": "user", "content": user_input, "avatar": "🧑‍💻"}
-            )
-            st.session_state.current_message = user_input
-            with st.spinner("Processing your message..."):
-                st.experimental_rerun()
-
-    col1, col2, col3 = st.columns([1, 1, 1])
-    with col2:
-        if st.button(
-            "💾 Save Conversation",
-            key="save_button",
-            help="Save the current conversation",
-            use_container_width=True,
-        ):
-            log_content = create_conversation_log(
-                conversation, st.session_state.messages
-            )
-            timestamp = datetime.datetime.now().strftime("%Y%m%d_%H%M%S")
-            filename = f"conversation_log_{timestamp}.txt"
-
-            with open(filename, "w") as f:
-                f.write(log_content)
-
-            st.success(f"Conversation saved as {filename}")
-            st.balloons()  # Add a celebratory animation when saving
-            st.experimental_rerun()  # Rerun the app to update the saved conversations list
-
-    # Add collapsible section for saved conversations
-    with st.sidebar.expander("📚 Saved Conversations"):
-        saved_conversations = get_saved_conversations()
-        if saved_conversations:
-            for conv_file in saved_conversations:
-                if st.button(f"📥 {conv_file}", key=f"download_{conv_file}"):
-                    with open(conv_file, "r") as f:
-                        content = f.read()
-                    st.download_button(
-                        label=f"📥 Download {conv_file}",
-                        data=content,
-                        file_name=conv_file,
-                        mime="text/plain",
-                        key=f"download_button_{conv_file}",
-                    )
-        else:
-            st.info("No saved conversations found.")
-
-    # Add a footer
-    st.markdown(
-        """
-        <footer>
-            <p>Made with ❤️ by <a href="https://github.com/tcsenpai">TCSenpai</a></p>
-        </footer>
-        """,
-        unsafe_allow_html=True,
-    )
-
-
-def create_conversation_log(conversation, messages):
-    log = f"Conversation Log - {datetime.datetime.now().strftime('%Y-%m-%d %H:%M:%S')}\n\n"
-    log += f"Ollama Endpoint: {conversation.ollama_client.endpoint}\n"
-    log += f"Model 1: {conversation.model_1}\n"
-    log += f"Model 2: {conversation.model_2}\n"
-    log += f"System Prompt 1:\n{conversation.system_prompt_1}\n\n"
-    log += f"System Prompt 2:\n{conversation.system_prompt_2}\n\n"
-    log += "Conversation:\n\n"
-
-    for message in messages:
-        log += f"{message['role'].capitalize()}: {message['content']}\n\n"
-
-    return log
-
-
-def get_saved_conversations():
-    return [
-        f
-        for f in os.listdir()
-        if f.startswith("conversation_log_") and f.endswith(".txt")
-    ]