3

Author: thisaltennakoon

docker/Dockerfile

@@ -1,51 +1,36 @@
-# Multi-stage Dockerfile that works on macOS for development and Ubuntu for GPU deployment
-ARG BASE_IMAGE=python:3.11-slim
+# CPU-only Dockerfile for macOS development and Ubuntu CPU deployment
+FROM python:3.11.13-slim
 
-# Use CUDA runtime for GPU support when building on Linux/Ubuntu
-# Use regular Python when building on macOS
-FROM ${BASE_IMAGE}
+# Set working directory
+WORKDIR /app
 
 # Set environment variables
+ENV PYTHONDONTWRITEBYTECODE=1
 ENV PYTHONUNBUFFERED=1
-ENV DEBIAN_FRONTEND=noninteractive
 
 # Install system dependencies
 RUN apt-get update && apt-get install -y \
     build-essential \
-    curl \
+    wget \
+    ca-certificates \
     && rm -rf /var/lib/apt/lists/*
 
-# Create a non-root user
-RUN useradd --create-home --shell /bin/bash app
-USER app
-WORKDIR /home/app
-
-# Set up Python environment
-ENV PATH="/home/app/.local/bin:$PATH"
-
 # Copy requirements first for better caching
-COPY --chown=app:app docker/requirements.txt ./requirements.txt
+COPY docker/requirements.txt .
 
 # Install Python dependencies
-RUN pip install --user --no-cache-dir -r requirements.txt
-
-# Pre-install spaCy models to avoid runtime downloads
-RUN python -c "import spacy; spacy.cli.download('en_core_web_lg')"
-
-# Optionally pre-download GLiNER model (this will reduce first-run time)
-# Comment out if you want to download at runtime instead
-RUN python -c "from gliner import GLiNER; GLiNER.from_pretrained('urchade/gliner_small-v2.1')"
+RUN pip install --no-cache-dir -r requirements.txt
 
-# Copy only necessary source files
-COPY --chown=app:app validator/ ./validator/
-COPY --chown=app:app docker/server_docker.py ./server.py
+# Copy application code
+COPY validator/ ./validator/
+COPY docker/server.py ./server.py
 
-# Set default environment variables
-ENV PII_ENTITIES='["EMAIL_ADDRESS", "PHONE_NUMBER", "PERSON", "DATE_TIME"]'
-ENV USE_GPU=false
+# Pre-download and cache model files during build to avoid runtime delays
+# This runs the exact same guard initialization as the server to cache models
+RUN python -c "from validator import GuardrailsPII; from guardrails import Guard; print('Caching PII detection models...'); guard = Guard().use(GuardrailsPII(entities=['EMAIL_ADDRESS', 'PHONE_NUMBER', 'PERSON', 'DATE_TIME'], on_fail='fix', use_gpu=False)); result = guard.validate('Test email john@example.com and phone 555-123-4567'); print('Models cached successfully!')"
 
-# Expose port
+# Expose port 8000
 EXPOSE 8000
 
-# Run the application
+# Start the Flask application
 CMD ["python", "server.py"]

docker/Dockerfile.gpu

@@ -1,49 +1,43 @@
 # GPU-enabled Dockerfile for Ubuntu deployment with NVIDIA CUDA support
 FROM nvidia/cuda:11.8-runtime-ubuntu22.04
 
+# Set working directory
+WORKDIR /app
+
 # Set environment variables
+ENV PYTHONDONTWRITEBYTECODE=1
 ENV PYTHONUNBUFFERED=1
 ENV DEBIAN_FRONTEND=noninteractive
 
-# Install Python and essential packages
+# Install Python and system dependencies
 RUN apt-get update && apt-get install -y \
     python3 \
     python3-pip \
     python3-dev \
     build-essential \
+    wget \
+    ca-certificates \
     && rm -rf /var/lib/apt/lists/*
 
-# Create a non-root user
-RUN useradd --create-home --shell /bin/bash app
-USER app
-WORKDIR /home/app
-
-# Set up Python environment
-ENV PATH="/home/app/.local/bin:$PATH"
+# Create symlink for python command
+RUN ln -s /usr/bin/python3 /usr/bin/python
 
 # Copy requirements first for better caching
-COPY --chown=app:app docker/requirements.txt ./requirements.txt
+COPY docker/requirements.txt .
 
 # Install Python dependencies
-RUN pip3 install --user --no-cache-dir -r requirements.txt
-
-# Pre-install spaCy models to avoid runtime downloads
-RUN python3 -c "import spacy; spacy.cli.download('en_core_web_lg')"
-
-# Optionally pre-download GLiNER model (this will reduce first-run time)
-# Comment out if you want to download at runtime instead
-RUN python3 -c "from gliner import GLiNER; GLiNER.from_pretrained('urchade/gliner_small-v2.1')"
+RUN pip install --no-cache-dir -r requirements.txt
 
-# Copy only necessary source files
-COPY --chown=app:app validator/ ./validator/
-COPY --chown=app:app docker/server_docker.py ./server.py
+# Copy application code
+COPY validator/ ./validator/
+COPY docker/server.py ./server.py
 
-# Set default environment variables for GPU
-ENV PII_ENTITIES='["EMAIL_ADDRESS", "PHONE_NUMBER", "PERSON", "DATE_TIME"]'
-ENV USE_GPU=true
+# Pre-download and cache model files during build to avoid runtime delays
+# This runs the exact same guard initialization as the server to cache models with GPU support
+RUN python -c "from validator import GuardrailsPII; from guardrails import Guard; print('Caching PII detection models with GPU support...'); guard = Guard().use(GuardrailsPII(entities=['EMAIL_ADDRESS', 'PHONE_NUMBER', 'PERSON', 'DATE_TIME'], on_fail='fix', use_gpu=True)); result = guard.validate('Test email john@example.com and phone 555-123-4567'); print('Models cached successfully!')"
 
-# Expose port
+# Expose port 8000
 EXPOSE 8000
 
-# Run the application
-CMD ["python3", "server.py"]
+# Start the Flask application
+CMD ["python", "server.py"]

docker/docker-README.md

@@ -1,16 +1,3 @@
-# PII Detection Docker Container
-
-This directory contains Docker configuration for the PII detection and masking service that works on both macOS (development) and Ubuntu (GPU deployment).
-
-## Features
-
-- **Cross-Platform**: Works on macOS for development and Ubuntu for GPU deployment
-- **GPU Support**: NVIDIA GPU acceleration for improved performance (Ubuntu only)
-- **Configurable Entities**: Dynamic PII entity types via environment variables
-- **Lightweight**: Optimized image size with minimal dependencies
-
-## Development on macOS
-
 ### Build and Run (CPU-only)
 
 ```bash
@@ -39,99 +26,4 @@ docker build -f docker/Dockerfile.gpu -t pii-detector-gpu:latest .
 docker run --gpus all -p 8000:8000 \
   -e USE_GPU=true \
   -e PII_ENTITIES='["EMAIL_ADDRESS", "PHONE_NUMBER", "PERSON", "DATE_TIME"]' \
-  pii-detector-gpu:latest
-
-# Or use docker-compose
-docker-compose up pii-detector-gpu
-```
-
-## Docker Files
-
-- **`Dockerfile`**: CPU-only version, works on macOS and Ubuntu
-- **`Dockerfile.gpu`**: GPU-enabled version for Ubuntu deployment
-- **`docker-compose.yml`**: Multi-service setup for different scenarios
-
-## Environment Variables
-
-| Variable | Description | Default | Example |
-|----------|-------------|---------|---------|
-| `USE_GPU` | Enable/disable GPU acceleration | `false` (CPU), `true` (GPU) | `false` |
-| `PII_ENTITIES` | JSON array of PII entity types | `["EMAIL_ADDRESS", "PHONE_NUMBER", "PERSON", "DATE_TIME"]` | `["EMAIL_ADDRESS", "PHONE_NUMBER"]` |
-
-## Docker Compose Services
-
-| Service | Purpose | Port | GPU |
-|---------|---------|------|-----|
-| `pii-detector-gpu` | Production with GPU | 8000 | ✅ |
-| `pii-detector-cpu` | Production without GPU | 8001 | ❌ |
-| `pii-detector-dev` | Development/testing | 8002 | ❌ |
-
-## API Endpoints
-
-### POST /validate
-Validates text for PII and returns anonymized version.
-
-**Request:**
-```json
-{
-  "text": "Contact John Doe at john.doe@example.com"
-}
-```
-
-**Response:**
-```json
-{
-  "verdict": false,
-  "assessment": [
-    {
-      "piiEntity": "PERSON",
-      "piiValue": "John Doe"
-    },
-    {
-      "piiEntity": "EMAIL_ADDRESS", 
-      "piiValue": "john.doe@example.com"
-    }
-  ],
-  "anonymizedText": "Contact [PERSON] at [EMAIL_ADDRESS]"
-}
-```
-
-## Requirements
-
-### For macOS Development
-- Docker Desktop
-
-### For Ubuntu GPU Deployment
-- Docker with NVIDIA Container Toolkit
-- NVIDIA GPU with CUDA 11.8+
-
-## Supported PII Entities
-
-Common entity types include:
-- `EMAIL_ADDRESS`
-- `PHONE_NUMBER`
-- `PERSON`
-- `DATE_TIME`
-- `CREDIT_CARD`
-- `SSN`
-- `IBAN_CODE`
-- `IP_ADDRESS`
-- `LOCATION`
-- `ORGANIZATION`
-
-## Quick Commands
-
-```bash
-# Development on macOS
-docker build -f docker/Dockerfile -t pii-detector:latest .
-docker run -p 8000:8000 -e USE_GPU=false pii-detector:latest
-
-# Production on Ubuntu
-docker build -f docker/Dockerfile.gpu -t pii-detector-gpu:latest .
-docker run --gpus all -p 8000:8000 -e USE_GPU=true pii-detector-gpu:latest
-
-# Using docker-compose
-docker-compose up pii-detector-cpu    # CPU only
-docker-compose up pii-detector-gpu    # GPU enabled
-docker-compose up pii-detector-dev    # Development
-```
+  pii-detector-gpu:latest

docker/docker-compose.yml

@@ -1,42 +1,34 @@
 version: '3.8'
 
 services:
-  # For GPU deployment on Ubuntu/Linux
-  pii-detector-gpu:
+  # CPU-only deployment (works on macOS and Ubuntu)
+  pii-detection-cpu:
+    build:
+      context: ..
+      dockerfile: docker/Dockerfile
+    ports:
+      - "8000:8000"
+    environment:
+      - PYTHONUNBUFFERED=1
+    restart: unless-stopped
+    container_name: pii-detection-cpu
+
+  # GPU-enabled deployment (Ubuntu with NVIDIA GPU)
+  pii-detection-gpu:
     build:
       context: ..
       dockerfile: docker/Dockerfile.gpu
     ports:
       - "8000:8000"
     environment:
-      - USE_GPU=true
-      - PII_ENTITIES=["EMAIL_ADDRESS", "PHONE_NUMBER", "PERSON", "DATE_TIME"]
+      - PYTHONUNBUFFERED=1
+    restart: unless-stopped
+    container_name: pii-detection-gpu
+    # Enable GPU support - requires Docker with NVIDIA Container Toolkit
     deploy:
       resources:
         reservations:
           devices:
             - driver: nvidia
-              count: 1
+              count: all
               capabilities: [gpu]
-
-  # For CPU-only deployment (works on macOS and Ubuntu)
-  pii-detector-cpu:
-    build:
-      context: ..
-      dockerfile: docker/Dockerfile
-    ports:
-      - "8001:8000"
-    environment:
-      - USE_GPU=false
-      - PII_ENTITIES=["EMAIL_ADDRESS", "PHONE_NUMBER", "PERSON", "DATE_TIME"]
-
-  # Minimal configuration for development
-  pii-detector-dev:
-    build:
-      context: ..
-      dockerfile: docker/Dockerfile
-    ports:
-      - "8002:8000"
-    environment:
-      - USE_GPU=false
-      - PII_ENTITIES=["EMAIL_ADDRESS", "PHONE_NUMBER"]