updated app structure

.gitignore

@@ -1,4 +1,23 @@
-__pycache__
+# Python
+__pycache__/
+*.py[cod]
+*.pyo
+*.pyd
+.Python
+env/
+venv/
+ENV/
+
+# Streamlit
+.streamlit/
+
+# Environment variables
 .env
-venv
-.venv
+
+# IDEs
+.vscode/
+.idea/
+
+# OS generated files
+.DS_Store
+Thumbs.db

README.md

@@ -14,9 +14,9 @@
 - [ ] Add more providers
 - [ ] Use something like LiteLLM to unify models code and avoid repeating code for each provider
 
-## Original G1 Example video
+## Example
 
-[Video Demo](https://github.com/user-attachments/assets/db2a221f-f8eb-48c3-b5a7-8399c6300243)
+![Simple Math](examples/maths.png)
 
 ## Description
 
@@ -36,32 +36,12 @@ At each step, the LLM can choose to continue to another reasoning step, or provi
 The reasoning ability of the LLM is therefore improved through combining Chain-of-Thought with the requirement to try multiple methods, explore alternative answers, question previous draft solutions, and consider the LLM’s limitations. This alone, without any training, is sufficient to achieve ~70% accuracy on the Strawberry problem (n=10, "How many Rs are in strawberry?"). Without prompting, Llama-3.1-70b had 0% accuracy and ChatGPT-4o had 30% accuracy.
 
 
-### Examples
+### Disclaimer
 
 > [!IMPORTANT]
 > multi1 is not perfect, but it can perform significantly better than LLMs out-of-the-box. From initial testing, multi1 accurately solves simple logic problems 60-80% of the time that usually stump LLMs. However, accuracy has yet to be formally evaluated. See examples below.
 
 
-##### How many Rs are in strawberry?
-Prompt: How many Rs are in strawberry?
-
-Result:
-
-![Strawberry example](examples/strawberry.png)
-
----
-
-Prompt: Which is larger, .9 or .11?
-
-Result:
-
-![0.9 or 0.11 example](examples/math.png)
-
-
-Prompt: In the context of Lie Group and Lie Algebra, let $R \in E$ be an irreducible root system. Show that then $E$ is an irreducible representation of the Weyl group $W$.
-
-![](examples/lie.1.png)
-
 
 ### Quickstart
 

app/config_menu.py

@@ -0,0 +1,49 @@
+import streamlit as st
+import os
+from dotenv import load_dotenv, set_key
+
+def load_env_vars():
+    load_dotenv(os.path.join(os.path.dirname(__file__), "..", ".env"))
+    return {
+        'OLLAMA_URL': os.getenv('OLLAMA_URL', 'http://localhost:11434'),
+        'OLLAMA_MODEL': os.getenv('OLLAMA_MODEL', 'mistral'),
+        'PERPLEXITY_API_KEY': os.getenv('PERPLEXITY_API_KEY', ''),
+        'PERPLEXITY_MODEL': os.getenv('PERPLEXITY_MODEL', 'mistral-7b-instruct'),
+        'GROQ_API_KEY': os.getenv('GROQ_API_KEY', ''),
+        'GROQ_MODEL': os.getenv('GROQ_MODEL', 'mixtral-8x7b-32768')
+    }
+
+def save_env_vars(config):
+    env_path = os.path.join(os.path.dirname(__file__), "..", ".env")
+    for key, value in config.items():
+        set_key(env_path, key, value)
+
+def config_menu():
+    st.sidebar.markdown("## 🛠️ Configuration")
+    
+    config = load_env_vars()
+    
+    with st.sidebar.expander("Edit Configuration"):
+        new_config = {}
+        new_config['OLLAMA_URL'] = st.text_input("Ollama URL", value=config['OLLAMA_URL'])
+        new_config['OLLAMA_MODEL'] = st.text_input("Ollama Model", value=config['OLLAMA_MODEL'])
+        new_config['PERPLEXITY_API_KEY'] = st.text_input("Perplexity API Key", value=config['PERPLEXITY_API_KEY'], type="password")
+        new_config['PERPLEXITY_MODEL'] = st.text_input("Perplexity Model", value=config['PERPLEXITY_MODEL'])
+        new_config['GROQ_API_KEY'] = st.text_input("Groq API Key", value=config['GROQ_API_KEY'], type="password")
+        new_config['GROQ_MODEL'] = st.text_input("Groq Model", value=config['GROQ_MODEL'])
+        
+        if st.button("Save Configuration"):
+            save_env_vars(new_config)
+            st.success("Configuration saved successfully!")
+    
+    return config
+
+def display_config(backend, config):
+    st.sidebar.markdown("## 🛠️ Current Configuration")
+    if backend == "Ollama":
+        st.sidebar.markdown(f"- 🖥️ Ollama URL: `{config['OLLAMA_URL']}`")
+        st.sidebar.markdown(f"- 🤖 Ollama Model: `{config['OLLAMA_MODEL']}`")
+    elif backend == "Perplexity AI":
+        st.sidebar.markdown(f"- 🧠 Perplexity AI Model: `{config['PERPLEXITY_MODEL']}`")
+    else:  # Groq
+        st.sidebar.markdown(f"- ⚡ Groq Model: `{config['GROQ_MODEL']}`")

app/main.py

@@ -0,0 +1,69 @@
+import streamlit as st
+from dotenv import load_dotenv
+from api_handlers import OllamaHandler, PerplexityHandler, GroqHandler
+from utils import generate_response
+from config_menu import config_menu, display_config
+import os
+
+# Load environment variables
+load_dotenv()
+
+def load_css():
+    with open(os.path.join(os.path.dirname(__file__), "..", "static", "styles.css")) as f:
+        st.markdown(f'<style>{f.read()}</style>', unsafe_allow_html=True)
+
+def setup_page():
+    st.set_page_config(page_title="multi1 - Unified AI Reasoning Chains", page_icon="🧠", layout="wide")
+    load_css()
+    st.markdown("""
+    <h1 class="main-title">
+        🧠 multi1 - Unified AI Reasoning Chains
+    </h1>
+    """, unsafe_allow_html=True)
+    st.markdown("""
+    <p class="main-description">
+        This app demonstrates AI reasoning chains using different backends: Ollama, Perplexity AI, and Groq.
+        Choose a backend and enter your query to see the step-by-step reasoning process.
+    </p>
+    """, unsafe_allow_html=True)
+
+def get_api_handler(backend, config):
+    if backend == "Ollama":
+        return OllamaHandler(config['OLLAMA_URL'], config['OLLAMA_MODEL'])
+    elif backend == "Perplexity AI":
+        return PerplexityHandler(config['PERPLEXITY_API_KEY'], config['PERPLEXITY_MODEL'])
+    else:  # Groq
+        return GroqHandler(config['GROQ_API_KEY'], config['GROQ_MODEL'])
+
+def main():
+    setup_page()
+
+    st.sidebar.markdown('<h3 class="sidebar-title">⚙️ Settings</h3>', unsafe_allow_html=True)
+    config = config_menu()
+    
+    backend = st.sidebar.selectbox("Choose AI Backend", ["Ollama", "Perplexity AI", "Groq"])
+    display_config(backend, config)
+    api_handler = get_api_handler(backend, config)
+
+    user_query = st.text_input("💬 Enter your query:", placeholder="e.g., How many 'R's are in the word strawberry?")
+
+    if user_query:
+        st.write("🔍 Generating response...")
+        response_container = st.empty()
+        time_container = st.empty()
+
+        for steps, total_thinking_time in generate_response(user_query, api_handler):
+            with response_container.container():
+                for title, content, _ in steps:
+                    if title.startswith("Final Answer"):
+                        st.markdown(f'<h3 class="expander-title">🎯 {title}</h3>', unsafe_allow_html=True)
+                        st.markdown(f'<div>{content}</div>', unsafe_allow_html=True)
+                    else:
+                        with st.expander(f"📝 {title}", expanded=True):
+                            st.markdown(f'<div>{content}</div>', unsafe_allow_html=True)
+
+            if total_thinking_time is not None:
+                time_container.markdown(f'<p class="thinking-time">⏱️ Total thinking time: {total_thinking_time:.2f} seconds</p>', unsafe_allow_html=True)
+
+if __name__ == "__main__":
+    main()

main.py

@@ -1,71 +0,0 @@
-import streamlit as st
-from dotenv import load_dotenv
-from api_handlers import OllamaHandler, PerplexityHandler, GroqHandler
-from utils import generate_response, load_env_vars
-
-# Load environment variables and configuration
-load_dotenv()
-config = load_env_vars()
-
-def setup_page():
-    st.set_page_config(page_title="multi1 - Unified AI Reasoning Chains", page_icon="🧠", layout="wide")
-    st.markdown("""
-    <h1 style='text-align: center; font-family: -apple-system, BlinkMacSystemFont, sans-serif;'>
-        🧠 multi1 - Unified AI Reasoning Chains
-    </h1>
-    """, unsafe_allow_html=True)
-    st.markdown("""
-    <p style='text-align: center; font-family: -apple-system, BlinkMacSystemFont, sans-serif; font-size: 1.1em;'>
-        This app demonstrates AI reasoning chains using different backends: Ollama, Perplexity AI, and Groq.
-        Choose a backend and enter your query to see the step-by-step reasoning process.
-    </p>
-    """, unsafe_allow_html=True)
-
-def get_api_handler(backend):
-    if backend == "Ollama":
-        return OllamaHandler(config['OLLAMA_URL'], config['OLLAMA_MODEL'])
-    elif backend == "Perplexity AI":
-        return PerplexityHandler(config['PERPLEXITY_API_KEY'], config['PERPLEXITY_MODEL'])
-    else:  # Groq
-        return GroqHandler()
-
-def display_config(backend):
-    st.sidebar.markdown("## 🛠️ Current Configuration")
-    if backend == "Ollama":
-        st.sidebar.markdown(f"- 🖥️ Ollama URL: `{config['OLLAMA_URL']}`")
-        st.sidebar.markdown(f"- 🤖 Ollama Model: `{config['OLLAMA_MODEL']}`")
-    elif backend == "Perplexity AI":
-        st.sidebar.markdown(f"- 🧠 Perplexity AI Model: `{config['PERPLEXITY_MODEL']}`")
-    else:  # Groq
-        st.sidebar.markdown("- ⚡ Using Groq API")
-
-def main():
-    setup_page()
-
-    st.sidebar.markdown("<h3 style='font-family: -apple-system, BlinkMacSystemFont, sans-serif;'>⚙️ Settings</h3>", unsafe_allow_html=True)
-    backend = st.sidebar.selectbox("Choose AI Backend", ["Ollama", "Perplexity AI", "Groq"])
-    display_config(backend)
-    api_handler = get_api_handler(backend)
-
-    user_query = st.text_input("💬 Enter your query:", placeholder="e.g., How many 'R's are in the word strawberry?")
-
-    if user_query:
-        st.write("🔍 Generating response...")
-        response_container = st.empty()
-        time_container = st.empty()
-
-        for steps, total_thinking_time in generate_response(user_query, api_handler):
-            with response_container.container():
-                for title, content, _ in steps:
-                    if title.startswith("Final Answer"):
-                        st.markdown(f"<h3 style='font-family: -apple-system, BlinkMacSystemFont, sans-serif;'>🎯 {title}</h3>", unsafe_allow_html=True)
-                        st.markdown(f"<div style='font-family: -apple-system, BlinkMacSystemFont, sans-serif;'>{content}</div>", unsafe_allow_html=True)
-                    else:
-                        with st.expander(f"📝 {title}", expanded=True):
-                            st.markdown(f"<div style='font-family: -apple-system, BlinkMacSystemFont, sans-serif;'>{content}</div>", unsafe_allow_html=True)
-
-            if total_thinking_time is not None:
-                time_container.markdown(f"<p style='font-family: -apple-system, BlinkMacSystemFont, sans-serif;'><strong>⏱️ Total thinking time: {total_thinking_time:.2f} seconds</strong></p>", unsafe_allow_html=True)
-
-if __name__ == "__main__":
-    main()

static/styles.css

@@ -0,0 +1,29 @@
+body {
+    font-family: -apple-system, BlinkMacSystemFont, sans-serif;
+}
+
+h1, h2, h3, h4, h5, h6 {
+    font-family: -apple-system, BlinkMacSystemFont, sans-serif;
+}
+
+.main-title {
+    text-align: center;
+}
+
+.main-description {
+    text-align: center;
+    font-size: 1.1em;
+}
+
+.sidebar-title {
+    font-family: -apple-system, BlinkMacSystemFont, sans-serif;
+}
+
+.expander-title {
+    font-family: -apple-system, BlinkMacSystemFont, sans-serif;
+}
+
+.thinking-time {
+    font-family: -apple-system, BlinkMacSystemFont, sans-serif;
+    font-weight: bold;
+}