Add : Docker Support

Author: ToastCoder

Dockerfile

@@ -0,0 +1,25 @@
+# Use a lightweight Python base image
+FROM python:3.10-slim
+
+# Set the working directory inside the container
+WORKDIR /app
+
+# Install system dependencies (required for PyMuPDF and other tools)
+RUN apt-get update && apt-get install -y \
+    build-essential \
+    && rm -rf /var/lib/apt/lists/*
+
+# Copy only the requirements first to leverage Docker caching
+COPY requirements.txt .
+
+# Install Python dependencies
+RUN pip install --no-cache-dir --default-timeout=1000 --retries 5 -i https://pypi.org/simple/ --extra-index-url https://download.pytorch.org/whl/cpu -r requirements.txt
+
+# Copy the rest of your application code
+COPY . .
+
+# Expose the port Streamlit runs on
+EXPOSE 8501
+
+# Command to run the application
+CMD ["streamlit", "run", "main.py", "--server.port=8501", "--server.address=0.0.0.0"]

docker-compose.yml

@@ -0,0 +1,23 @@
+services:
+  mongodb:
+    image: mongo:6.0
+    container_name: hr_mongodb
+    ports:
+      - "27017:27017"
+    volumes:
+      - mongo_data:/data/db 
+
+  web:
+      build:
+        context: .
+        network: host
+      container_name: hr_streamlit
+      ports:
+        - "8501:8501"
+      depends_on:
+        - mongodb
+      extra_hosts:
+        - "host.docker.internal:host-gateway" 
+
+volumes:
+  mongo_data:

main.py

@@ -19,8 +19,14 @@
 import streamlit as st
 import hashlib
 
+# Variables
+MONGO_URL = "mongodb://mongodb:27017/"
+OLLAMA_BASE_URL = "http://host.docker.internal:11434"
+DB_NAME = "hr_buddy_db"
+PDF_PATH = "rag_source/nasscom-hr-manual-2016.pdf"
+
 # Streamlit webpage Setup
-st.set_page_config(page_title="HR Buddy", page_icon="🤖")
+st.set_page_config(page_title="HR Buddy")
 st.title("HR Policy Assistant")
 
 # Cached Setup for loading model and the database
@@ -33,7 +39,7 @@ def initialize_backend():
 
     # Connecting to default local MongoDB Port
     print("[INFO] Connecting to local MongoDB...")
-    client = MongoClient("mongodb://localhost:27017/")
+    client = MongoClient(MONGO_URL)
 
     # Retrieve Memory Database
     db = client["hr_assistant"]
@@ -42,7 +48,7 @@ def initialize_backend():
 
     # Load and Chunk the Data
     print("[INFO] Reading HR Policy PDF with PyMuPDF...")
-    loader = PyMuPDFLoader("rag_source/nasscom-hr-manual-2016.pdf") 
+    loader = PyMuPDFLoader(PDF_PATH) 
     pages = loader.load()
 
     # Break the dense PDF into small, 1000-character chunks
@@ -53,7 +59,10 @@ def initialize_backend():
     print("[INFO] Initializing vector embeddings. This might take a second...")
 
     # Embed and Store in Vector DB
-    embeddings = OllamaEmbeddings(model="nomic-embed-text")
+    embeddings = OllamaEmbeddings(model="nomic-embed-text",
+                                  base_url=OLLAMA_BASE_URL
+                                  )
+    
     vectorstore = Chroma.from_documents(documents=chunks, embedding=embeddings)
     retriever = vectorstore.as_retriever(search_kwargs={"k": 2})
 
@@ -176,7 +185,8 @@ def initialize_backend():
     with st.chat_message("assistant"):
 
         # Generate the answer using Llama 3.2
-        response = ollama.chat(model='llama3.2', messages=[{'role': 'user', 'content': rag_prompt}])
+        ollama_client = ollama.Client(host=OLLAMA_BASE_URL)
+        response = ollama_client.chat(model='llama3.2', messages=[{'role': 'user', 'content': rag_prompt}])
         ai_answer = response['message']['content']
         st.markdown(ai_answer)
 

dependencies.txt requirements.txtdependencies.txt renamed to requirements.txt

@@ -10,5 +10,4 @@ pymupdf
 protobuf==3.20.3
 pymongo
 streamlit
-torchvision
-hashlib
+torchvision

run.sh

@@ -1,100 +1,53 @@
-# Variables
-PKG_MGR="pip"
-FILE="dependencies.txt"
-
-# Check for dependency file
-if [ ! -f "$FILE" ]; then
-    echo "[ERROR] '$FILE' not found."
-    exit 1
-fi
-
-echo "[OK] dependencies.txt found"
-
-# Parser for choosing Package Manager (pip or conda)
-while [[ "$#" -gt 0 ]]; do
-    case $1 in
-        -m)
-            PKG_MGR="$2"
-            shift 2
-            ;;
-
-        *)
-            echo "[ERROR] Unknown Option: $1. Usage ./run.sh [-m pip | conda]"
-            exit 1
-            ;;
-    esac
-done
-
-echo "[INFO] Selected Package Manager: '$PKG_MGR'"
-
-# Install Dependencies
-if [ "$PKG_MGR" == "conda" ]; then
-    echo "[OK] Starting installation using Conda..."
-    conda install --file "$FILE" -c conda-forge -y
-
-elif [ "$PKG_MGR" == "pip" ]; then
-    echo "[OK] Starting installation using Pip..."
-    pip install -r "$FILE"
-
+#!/bin/bash
+
+# nstall/Check Docker Engine
+if ! command -v docker &> /dev/null; then
+    echo "[INFO] Docker not found. Installing..."
+    curl -fsSL https://get.docker.com -o get-docker.sh && sudo sh get-docker.sh
+    sudo usermod -aG docker $USER
+    echo "[OK] Docker installed. Note: You may need to reboot for permissions."
 else
-    echo "[Error] Invalid manager '$PKG_MANAGER'. Please use 'pip' or 'conda'."
-    exit 1
+    echo "[OK] Docker is already installed."
 fi
 
-# Check for Ollama
-echo "[INFO] Verifying Python environment binding..."
-if ! python3 -c "import ollama" &> /dev/null; then
-    echo "[WARNING] Python cannot find 'ollama' library. Forcing strict inline installation..."
-    python3 -m pip install ollama
-    
-    if [ $? -ne 0 ]; then
-        echo "[ERROR] Could not force-install Python ollama library."
-        exit 1
-    fi
-    echo "[OK] 'ollama' bound to current Python environment."
-else
-    echo "[OK] Python recognizes 'ollama'."
-fi
+# Ensure Docker service is active
+sudo systemctl start docker
+sudo systemctl enable docker &> /dev/null
 
-# System Ollama Engine Check and Install
-echo "[INFO] Checking system Ollama engine..."
+# Install/Check Ollama Engine
 if ! command -v ollama &> /dev/null; then
-    echo "[WARNING] Ollama engine not found. Installing now (may require password)..."
+    echo "[INFO] Ollama not found. Installing..."
     curl -fsSL https://ollama.com/install.sh | sh
-    
-    if [ $? -ne 0 ]; then
-        echo "[ERROR] Ollama engine installation failed."
-        exit 1
-    fi
-    echo "[OK] Ollama engine installed successfully."
 else
-    echo "[OK] Ollama engine is already installed."
+    echo "[OK] Ollama is already installed."
 fi
 
-# Check for PyMuPDF
-if ! python3 -c "import pymupdf" &> /dev/null; then
-    echo "[WARNING] Python cannot find 'pymupdf' library. Forcing strict inline installation..."
-    python3 -m pip install pymupdf
-    
-    if [ $? -ne 0 ]; then
-        echo "[ERROR] Could not force-install Python pymupdf library."
-        exit 1
-    fi
-    echo "[OK] 'pymupdf' bound to current Python environment."
-else
-    echo "[OK] Python recognizes 'pymupdf'."
+# Patch Ollama for Docker Networking
+# Allows the container to communicate with the host machine
+CONF_DIR="/etc/systemd/system/ollama.service.d"
+CONF_FILE="$CONF_DIR/override.conf"
+
+if [ ! -f "$CONF_FILE" ]; then
+    echo "[INFO] Patching Ollama to allow Docker connections..."
+    sudo mkdir -p "$CONF_DIR"
+    echo -e "[Service]\nEnvironment=\"OLLAMA_HOST=0.0.0.0\"" | sudo tee "$CONF_FILE" > /dev/null
+    sudo systemctl daemon-reload
+    sudo systemctl restart ollama
+    echo "[OK] Ollama network patch applied."
 fi
 
-# Check and configure MongoDB
-chmod +x scripts/mongodb_cfg.sh
-./scripts/mongodb_cfg.sh
-
-# Required Models for RAG
-echo "[INFO] Verifying AI models (Llama3.2 & Nomic)..."
+# Pull AI Models
+echo "[INFO] Syncing AI models..."
 ollama pull llama3.2
 ollama pull nomic-embed-text
-echo "[OK] All models are ready."
+echo "[OK] Models are ready."
+
+# Launch Docker Compose
+echo "[INFO] Building and Launching Containers..."
 
-# Start Application
-echo "[INFO] Starting RAG application..."
-streamlit run main.py
+# Try 'docker compose' first, fallback to 'docker-compose'
+if docker compose version &> /dev/null; then
+    docker compose up --build
+else
+    docker-compose up --build
+fi