edited app.py and Dockerfile

Dockerfile

@@ -1,33 +1,17 @@
-# Use a slim Python base image
-FROM python:3.13.5-slim
+# Use the official Ollama image as the base
+FROM ollama/ollama
 
-# Set the working directory
-WORKDIR /app
+# Switch to the root user to install system and Python packages
+USER root
 
-# Install system dependencies
-# DEBIAN_FRONTEND is set to noninteractive to prevent prompts
-ENV DEBIAN_FRONTEND=noninteractive
-RUN apt-get update && apt-get install -y --no-install-recommends \
-    build-essential \
-    curl \
-    git \
-    ca-certificates \
+# Install Python and pip (Ollama image doesn't have it by default)
+RUN apt-get update && apt-get install -y \
+    python3 \
+    python3-pip \
     && rm -rf /var/lib/apt/lists/*
 
-# --- Manual Ollama Installation ---
-# This is the new, robust method.
-# Download the latest Linux binary from Ollama's GitHub releases
-# The URL below is a generic way to get the latest release
-RUN curl -L https://github.com/ollama/ollama/releases/latest/download/ollama-linux-amd64.tgz \
-    -o /tmp/ollama-linux-amd64.tgz
-
-# Extract the tarball and place the binary directly in /usr/local/bin/
-RUN tar -xzf /tmp/ollama-linux-amd64.tgz -C /usr/local/bin/
-
-# Remove the temporary file
-RUN rm /tmp/ollama-linux-amd64.tgz
-
 # Copy your application files
+WORKDIR /app
 COPY requirements.txt ./
 COPY app.py ./
 
@@ -37,8 +21,6 @@ RUN pip3 install -r requirements.txt
 # Expose the Streamlit port
 EXPOSE 8501
 
-# Healthcheck for Streamlit (useful for Hugging Face Spaces)
-HEALTHCHECK CMD curl --fail http://localhost:8501/_stcore/health
-
-# Entrypoint to run your Streamlit app
-ENTRYPOINT ["streamlit", "run", "app.py", "--server.port=8501", "--server.address=0.0.0.0"]
+# --- Ollama will be running on container startup ---
+# The entrypoint will now start the Streamlit app
+ENTRYPOINT ["python3", "-m", "streamlit", "run", "app.py", "--server.port=8501", "--server.address=0.0.0.0"]

app.py

@@ -4,27 +4,21 @@ from langchain.memory import ConversationBufferMemory
 from langchain.memory.chat_message_histories import ChatMessageHistory
 from langchain.prompts import PromptTemplate
 from langchain_core.runnables import RunnableSequence
-import subprocess
-import time
+import os
 
 # --- Ollama Setup ---
-def start_ollama_and_pull_models():
-    """Starts the Ollama server and pulls all required models."""
-    st.write("Starting Ollama server...")
-    
-    # Use the absolute path to the ollama binary
-    subprocess.Popen(['/usr/local/bin/ollama', 'serve'])
-    time.sleep(5)  # Give the server time to start
-    
-    # Pull all models at once
+def pull_models():
+    """Pulls all required models."""
     models_to_pull = ["phi3:mini", "deepseek-coder:1.3b"]
     st.write(f"Pulling models: {', '.join(models_to_pull)}...")
+    
+    # Use the official ollama client to pull models.
+    # The `ollama` executable is now in the PATH, so the call works.
     for model in models_to_pull:
         try:
-            # Use the absolute path for pulling the model too
-            subprocess.run(['/usr/local/bin/ollama', 'pull', model], check=True)
+            os.system(f"ollama pull {model}")
             st.success(f"Model '{model}' pulled successfully.")
-        except subprocess.CalledProcessError as e:
+        except Exception as e:
             st.error(f"Failed to pull model '{model}': {e}")
 
 
@@ -32,10 +26,10 @@ def start_ollama_and_pull_models():
 st.set_page_config(layout="wide")
 st.title("My Local Chatbot")
 
-if "ollama_started" not in st.session_state:
+if "ollama_pulled" not in st.session_state:
     with st.spinner("Setting up the local LLM server... this may take a moment."):
-        start_ollama_and_pull_models()
-    st.session_state.ollama_started = True
+        pull_models()
+    st.session_state.ollama_pulled = True
 
 # --- Sidebar Inputs ---
 st.sidebar.header("Settings")
@@ -50,8 +44,7 @@ TOP_P = st.sidebar.slider("Top-p (nucleus sampling)", 0.0, 1.0, 0.9, 0.05)
 TOP_K = st.sidebar.slider("Top-k", 0, 100, 40, 5)
 MAX_TOKENS = st.sidebar.number_input("Max Tokens", min_value=256, max_value=16384, value=2048, step=256)
 
-# ... (rest of the code is the same) ...
-
+# ... (remaining code from your app.py) ...
 if "chat_history" not in st.session_state:
     st.session_state.chat_history = []
 if "memory" not in st.session_state: