Refactor Dockerfile and start script: remove dos2unix dependency, streamline installation commands, and enhance startup process

Dockerfile

@@ -1,10 +1,13 @@
 # Use official Python base image
 FROM python:3.10-slim
 
-# Install curl, git, and dependencies for Ollama
-RUN apt-get update && apt-get install -y curl git && \
-    curl -fsSL https://ollama.com/install.sh | sh && \
-    rm -rf /var/lib/apt/lists/*
+# Install curl, git, dos2unix and dependencies for Ollama
+RUN apt-get update && apt-get install -y \
+    curl \
+    git \
+    dos2unix \
+    && curl -fsSL https://ollama.com/install.sh | sh \
+    && rm -rf /var/lib/apt/lists/*
 
 # Set working directory
 WORKDIR /app
@@ -12,13 +15,13 @@ WORKDIR /app
 # Copy app files
 COPY requirements.txt ./requirements.txt
 COPY fast.py ./fast.py
-COPY startup.py ./startup.py
+COPY start.sh ./start.sh
 
 # Install Python dependencies
 RUN pip install --no-cache-dir -r requirements.txt
 
-# Make startup script executable
-RUN chmod +x startup.py
+# Fix line endings and make start script executable
+RUN dos2unix start.sh && chmod +x start.sh
 
 # Create ollama home directory
 RUN mkdir -p /tmp/ollama_home/.ollama
@@ -31,5 +34,5 @@ ENV OLLAMA_ORIGINS="*"
 # Expose the port (required by Hugging Face Spaces)
 EXPOSE 7860
 
-# Start your app via the Python startup script
-CMD ["python3", "startup.py"]
+# Start your app via the startup script
+CMD ["./start.sh"]

start.sh

@@ -1,114 +1,51 @@
-#!/usr/bin/env python3
-import os
-import subprocess
-import time
-import requests
-import sys
+#!/bin/bash
 
-def setup_environment():
-    """Setup environment variables and directories"""
-    os.environ['HOME'] = '/tmp/ollama_home'
-    os.environ['OLLAMA_HOST'] = '0.0.0.0:11434'
-    os.environ['OLLAMA_ORIGINS'] = '*'
-    
-    # Create directories
-    ollama_dir = '/tmp/ollama_home/.ollama'
-    os.makedirs(ollama_dir, exist_ok=True)
-    os.makedirs(f'{ollama_dir}/models', exist_ok=True)
-    
-    print(f"Environment setup complete. HOME={os.environ['HOME']}")
+# Set Ollama environment variables
+export OLLAMA_HOST=0.0.0.0:11434
+export OLLAMA_ORIGINS="*"
+export OLLAMA_DATA_DIR=/tmp/.ollama
+export OLLAMA_MODELS=/tmp/.ollama/models
 
-def start_ollama():
-    """Start Ollama server"""
-    print("Starting Ollama server...")
-    
-    # Start Ollama in background
-    process = subprocess.Popen(
-        ['ollama', 'serve'],
-        env=os.environ.copy(),
-        stdout=subprocess.PIPE,
-        stderr=subprocess.PIPE
-    )
-    
-    # Wait for Ollama to start
-    for attempt in range(30):
-        try:
-            response = requests.get('http://localhost:11434/api/version', timeout=2)
-            if response.status_code == 200:
-                print(f"Ollama server started successfully! (attempt {attempt + 1})")
-                return process
-        except requests.exceptions.RequestException:
-            pass
-        
-        print(f"Attempt {attempt + 1}/30: Waiting for Ollama...")
-        time.sleep(2)
-        
-        # Check if process is still alive
-        if process.poll() is not None:
-            stdout, stderr = process.communicate()
-            print(f"Ollama process died. STDOUT: {stdout.decode()}")
-            print(f"STDERR: {stderr.decode()}")
-            return None
-    
-    print("Failed to start Ollama server")
-    return None
+# Create directories with proper permissions
+mkdir -p /tmp/.ollama/models
+chmod 755 /tmp/.ollama
+chmod 755 /tmp/.ollama/models
 
-def pull_model():
-    """Pull the LLaVA model"""
-    print("Pulling LLaVA model...")
-    try:
-        result = subprocess.run(
-            ['ollama', 'pull', 'llava:7b'],
-            env=os.environ.copy(),
-            capture_output=True,
-            text=True,
-            timeout=600  # 10 minutes timeout
-        )
-        
-        if result.returncode == 0:
-            print("Model pulled successfully!")
-            return True
-        else:
-            print(f"Failed to pull model: {result.stderr}")
-            return False
-    except subprocess.TimeoutExpired:
-        print("Model pull timed out")
-        return False
-    except Exception as e:
-        print(f"Error pulling model: {e}")
-        return False
+# Start Ollama server in background with explicit data directory
+echo "Starting Ollama server..."
+OLLAMA_DATA_DIR=/tmp/.ollama ollama serve &
 
-def start_fastapi():
-    """Start FastAPI server"""
-    print("Starting FastAPI server on port 7860...")
-    os.chdir('/app')
-    
-    # Start FastAPI
-    subprocess.run([
-        'python3', '-m', 'uvicorn', 
-        'fast:app', 
-        '--host', '0.0.0.0', 
-        '--port', '7860'
-    ])
+# Store the PID
+OLLAMA_PID=$!
 
-def main():
-    """Main startup sequence"""
-    print("===== Application Startup =====")
-    
-    # Setup environment
-    setup_environment()
-    
-    # Start Ollama
-    ollama_process = start_ollama()
-    if not ollama_process:
-        print("Failed to start Ollama. Exiting.")
-        sys.exit(1)
-    
-    # Pull model
-    pull_model()  # Continue even if this fails
-    
-    # Start FastAPI
-    start_fastapi()
+# Wait for Ollama to start
+echo "Waiting for Ollama server to start..."
+for i in {1..30}; do
+  if curl -s http://localhost:11434/api/version > /dev/null 2>&1; then
+    echo "Ollama server started successfully!"
+    break
+  fi
+  echo "Attempt $i/30: Waiting for Ollama..."
+  sleep 2
+done
 
-if __name__ == "__main__":
-    main()
+# Check if Ollama actually started
+if ! curl -s http://localhost:11434/api/version > /dev/null 2>&1; then
+  echo "Error: Ollama server failed to start"
+  exit 1
+fi
+
+# Pull the LLaVA model for vision analysis
+echo "Pulling LLaVA model for vision analysis..."
+OLLAMA_DATA_DIR=/tmp/.ollama ollama pull llava:7b
+
+if [ $? -eq 0 ]; then
+    echo "Model pulled successfully!"
+else
+    echo "Warning: Failed to pull model, but continuing..."
+fi
+
+# Start FastAPI on port 7860 (HF Spaces requirement)
+echo "Starting FastAPI server on port 7860..."
+cd /app
+python3 -m uvicorn fast:app --host 0.0.0.0 --port 7860