Update Dockerfile

Dockerfile

@@ -1,20 +1,15 @@
-# Dockerfile for Hugging Face Spaces - Non-root setup
+# Minimal Dockerfile for Hugging Face Spaces - Non-root setup
 FROM node:18-slim
 
 # Create non-root user
-RUN groupadd -r firecrawl && useradd -r -g firecrawl firecrawl
+RUN groupadd -r firecrawl && useradd -r -g firecrawl -m firecrawl
 
-# Install system dependencies as root
+# Install only essential system dependencies
 RUN apt-get update && apt-get install -y \
     git \
     curl \
-    wget \
     bash \
     ca-certificates \
-    redis-server \
-    postgresql \
-    postgresql-contrib \
-    sudo \
     && rm -rf /var/lib/apt/lists/*
 
 # Create app directory and set permissions
@@ -25,153 +20,12 @@ RUN mkdir -p /home/firecrawl/app && \
 USER firecrawl
 WORKDIR /home/firecrawl/app
 
-# Clone Firecrawl repository
-RUN git clone https://github.com/mendableai/firecrawl.git .
-
-# Install Node.js dependencies for API
-WORKDIR /home/firecrawl/app/apps/api
-RUN npm ci --only=production
-
-# Install Node.js dependencies for Playwright service (lightweight version)
-WORKDIR /home/firecrawl/app/apps/playwright-service-ts
-RUN npm ci --only=production
-
-# Build the applications
-WORKDIR /home/firecrawl/app/apps/api
-RUN npm run build
-
-WORKDIR /home/firecrawl/app/apps/playwright-service-ts
-RUN npm run build
-
-# Create directories for runtime
-WORKDIR /home/firecrawl/app
-RUN mkdir -p logs tmp data redis-data postgres-data
-
-# Copy environment template and create .env
-RUN cp apps/api/.env.example .env
-
-# Create a simple startup script for HF Spaces
-RUN cat > start.sh << 'EOF' && chmod +x start.sh
-#!/bin/bash
-
-# Set environment variables for HF Spaces
-export PORT=7860
-export HOST=0.0.0.0
-export REDIS_URL=redis://localhost:6379
-export PLAYWRIGHT_MICROSERVICE_URL=http://localhost:3000/scrape
-export USE_DB_AUTHENTICATION=false
-export ENV=production
-export LOGGING_LEVEL=info
-
-# Start Redis in background (if we can)
-redis-server --daemonize yes --port 6379 --bind 127.0.0.1 --dir /home/firecrawl/app/redis-data || echo "Redis start failed, continuing..."
-
-# Start Playwright service in background
-cd /home/firecrawl/app/apps/playwright-service-ts
-npm start &
-PLAYWRIGHT_PID=$!
-
-# Wait a bit for Playwright to start
-sleep 5
-
-# Start API service
-cd /home/firecrawl/app/apps/api
-exec node dist/src/harness.js --start-docker
-EOF
-
-# Create a simple health check script
-RUN cat > health.sh << 'EOF' && chmod +x health.sh
-#!/bin/bash
-curl -f http://localhost:${PORT:-7860}/health 2>/dev/null || exit 1
-EOF
-
-# Create minimal configuration files
-RUN cat > /home/firecrawl/app/.env << 'EOF'
-# HF Spaces Configuration
-PORT=7860
-HOST=0.0.0.0
-ENV=production
-LOGGING_LEVEL=info
-
-# Disable database authentication for simplicity
-USE_DB_AUTHENTICATION=false
-
-# Redis (will try to use local instance)
-REDIS_URL=redis://localhost:6379
-REDIS_RATE_LIMIT_URL=redis://localhost:6379
-
-# Playwright service
-PLAYWRIGHT_MICROSERVICE_URL=http://localhost:3000/scrape
-BLOCK_MEDIA=true
-
-# Disable optional services that might not work in HF Spaces
-BULL_AUTH_KEY=fc-demo
-TEST_API_KEY=fc-demo-key
-
-# Model configuration (can be overridden with HF Spaces secrets)
-MODEL_NAME=gpt-3.5-turbo
-MODEL_EMBEDDING_NAME=text-embedding-ada-002
-EOF
-
-# Create a simple API-only version for HF Spaces constraints
-RUN cat > start-api-only.sh << 'EOF' && chmod +x start-api-only.sh
-#!/bin/bash
-
-# Simplified startup for HF Spaces - API only
-export PORT=${PORT:-7860}
-export HOST=0.0.0.0
-export ENV=production
-export LOGGING_LEVEL=info
-export USE_DB_AUTHENTICATION=false
-
-# Disable services that might not work
-export PLAYWRIGHT_MICROSERVICE_URL=http://localhost:3000/scrape
-export REDIS_URL=redis://localhost:6379
-
-echo "🔥 Starting Firecrawl API on port $PORT"
-echo "⚠️  Note: This is a simplified version for Hugging Face Spaces"
-echo "📝 Some features may be limited due to platform constraints"
-
-cd /home/firecrawl/app/apps/api
-
-# Create a simple in-memory fallback if Redis isn't available
-cat > dist/src/lib/redis-fallback.js << 'JSEOF'
-// Simple in-memory fallback for Redis when not available
-class MemoryStore {
-  constructor() {
-    this.store = new Map();
-  }
-  
-  async get(key) {
-    return this.store.get(key);
-  }
-  
-  async set(key, value, options = {}) {
-    this.store.set(key, value);
-    if (options.EX) {
-      setTimeout(() => this.store.delete(key), options.EX * 1000);
-    }
-    return 'OK';
-  }
-  
-  async del(key) {
-    return this.store.delete(key);
-  }
-}
-
-module.exports = { MemoryStore };
-JSEOF
-
-# Start the API service
-exec node dist/src/harness.js --start-docker
-EOF
+# Copy the setup script
+COPY --chown=firecrawl:firecrawl setup-firecrawl.sh /home/firecrawl/app/
+RUN chmod +x setup-firecrawl.sh
 
 # Expose the port that HF Spaces expects
 EXPOSE 7860
 
-# Health check
-HEALTHCHECK --interval=30s --timeout=10s --start-period=60s --retries=3 \
-    CMD ./health.sh
-
-# Default command - use the API-only version for HF Spaces
-CMD ["./start-api-only.sh"]
+# Run the setup script which handles everything
+CMD ["./setup-firecrawl.sh"]