Upload 2 files

app.py

@@ -8,6 +8,7 @@ from dotenv import load_dotenv
 import asyncio
 import httpx
 import time
+import subprocess
 
 # Load environment variables
 load_dotenv()
@@ -16,6 +17,18 @@ load_dotenv()
 st.title("Math Reasoning Chatbot")
 st.write("Select a provider and chat with the bot to solve math problems!")
 
+# Function to start Ollama server if not already running
+def start_ollama_server():
+    try:
+        # Check if Ollama is already running
+        response = httpx.get("http://localhost:11434/v1")
+        if response.status_code == 200:
+            return True
+    except httpx.RequestError:
+        # Start Ollama server in the background
+        subprocess.Popen(["ollama", "serve"], stdout=subprocess.PIPE, stderr=subprocess.PIPE)
+        return False
+
 # Function to check if the Ollama server is running with retries
 async def check_ollama_health(max_retries=5, retry_delay=5):
     for attempt in range(max_retries):
@@ -44,14 +57,16 @@ def setup_client(provider):
         display_model = "OpenAI (gpt-4o-mini)"
     elif provider == "ollama":
         from openai import AsyncOpenAI as OllamaClient
-        # Check if Ollama server is running with retries
-        if not asyncio.run(check_ollama_health()):
-            st.error("Ollama server is not running or not accessible at http://localhost:11434 after multiple attempts. Please try again later or select a different provider.")
-            return None, None, None
+        # Start Ollama server if not running
+        if not start_ollama_server():
+            # Wait and check health
+            if not asyncio.run(check_ollama_health()):
+                st.error("Failed to start Ollama server or connect to it at http://localhost:11434 after multiple attempts.")
+                return None, None, None
         client = instructor.from_openai(
             OllamaClient(base_url="http://localhost:11434/v1", api_key="ollama"), mode=instructor.Mode.JSON
         )
-        model = "llama3.2:1b"  # Updated to use the smaller model
+        model = "llama3.2:1b"
         display_model = "Ollama (llama3.2:1b)"
     else:
         st.error(f"Unsupported provider: {provider}")
@@ -91,7 +106,9 @@ if "agent" not in st.session_state or st.session_state.get("current_model") != m
         client=client,
         model=model,
         system_prompt_generator=system_prompt_generator,
-        memory=st.session_state.memory,
+        memory ~
+
+=st.session_state.memory,
         system_role="developer",
     ))
     st.session_state.current_model = model  # Track the current model to detect changes

dockerfile

@@ -4,18 +4,18 @@ FROM python:3.10-slim
 # Set working directory
 WORKDIR /app
 
-# Copy the requirements file
-COPY requirements.txt .
-
-# Install dependencies
-RUN pip install --no-cache-dir -r requirements.txt
+# Install system dependencies
+RUN apt-get update && apt-get install -y curl && rm -rf /var/lib/apt/lists/*
 
 # Install Ollama
-RUN apt-get update && apt-get install -y curl
 RUN curl -fsSL https://ollama.com/install.sh | sh
 
-# Pre-download the llama3.2:1b model and debug
-RUN ollama pull llama3.2:1b && ollama list
+# Pre-download the llama3.2:1b model
+RUN ollama pull llama3.2:1b
+
+# Copy the requirements file and install Python dependencies
+COPY requirements.txt .
+RUN pip install --no-cache-dir -r requirements.txt
 
 # Copy the app code
 COPY app.py .
@@ -23,5 +23,5 @@ COPY app.py .
 # Expose the Streamlit port
 EXPOSE 8501
 
-# Start Ollama in the background, log output, wait 10 seconds, then run Streamlit
-CMD ollama serve > ollama.log 2>&1 & sleep 10 && streamlit run app.py --server.port 8501 --server.address 0.0.0.0
+# Run Streamlit (Ollama will be started by the app code)
+CMD ["streamlit", "run", "app.py", "--server.port", "8501", "--server.address", "0.0.0.0"]