ds_zagres_ai/LOCAL_TESTING_README.md

Local Testing with Remote Server Resources

This guide explains how to set up local testing that connects to remote server resources for the chatbot project.

Overview

These scripts allow you to:

1. Test the connection to the remote Ollama server
2. Create SSH tunnels to access remote resources locally
3. Run tests against the remote Ollama models from your local machine

Setup Instructions

1. Initial Setup

Run the setup script to create a virtual environment and install dependencies:

chmod +x setup_local_test_env.sh
./setup_local_test_env.sh

This will:

• Create a Python virtual environment
• Install required dependencies
• Create a local_test.env file from the template

2. Configure Environment Variables

Edit the local_test.env file with your server credentials:

# Remote server configuration
SERVER_IP=104.225.217.215
SERVER_PORT=22
SERVER_USER=root
SERVER_PASSWORD=your_password_here  # Add your actual password

# Ollama configuration
OLLAMA_API_URL=http://104.225.217.215:11434

# OpenWebUI configuration
OPENWEBUI_URL=http://104.225.217.215:8080
OPENWEBUI_API_KEY=your_openwebui_api_key_here

3. Create SSH Tunnels (Optional)

If you want to access the remote resources through localhost (recommended for security and to avoid firewall issues), create SSH tunnels:

chmod +x create_ssh_tunnel.sh
./create_ssh_tunnel.sh

This will create SSH tunnels for:

• Ollama API (localhost:11434 → remote:11434)
• OpenWebUI (localhost:8080 → remote:8080)

If you use SSH tunnels, update your local_test.env file to use localhost URLs:

# Ollama configuration
OLLAMA_API_URL=http://localhost:11434

# OpenWebUI configuration
OPENWEBUI_URL=http://localhost:8080

4. Run Tests

Activate the virtual environment and run the test script:

source venv/bin/activate
python test_remote_ollama.py

Test Options

You can customize the test with command-line arguments:

# Test with a specific model
python test_remote_ollama.py --model llama3.3

# Test with a custom prompt
python test_remote_ollama.py --prompt "Explain quantum computing"

# Test with a different Ollama URL
python test_remote_ollama.py --ollama-url http://your-server-ip:11434

# Test with a different timeout
python test_remote_ollama.py --timeout 600

5. Stop SSH Tunnels

When you're done testing, stop the SSH tunnels:

chmod +x stop_ssh_tunnels.sh
./stop_ssh_tunnels.sh

Troubleshooting

Connection Issues

If you can't connect to the remote server:

1. Check if the server is reachable:

   ping 104.225.217.215
   

2. Verify that Ollama is running on the server:

   ssh root@104.225.217.215 "curl http://localhost:11434/api/tags"
   

3. Check if there are firewall rules blocking the connection:

   ssh root@104.225.217.215 "iptables -L"
   

SSH Tunnel Issues

If the SSH tunnels aren't working:

1. Make sure you have SSH access to the server:

   ssh root@104.225.217.215
   

2. Check if the ports are already in use:

   netstat -tuln | grep 11434
   netstat -tuln | grep 8080
   

3. Try creating the tunnels manually:

   ssh -N -L 11434:localhost:11434 root@104.225.217.215
   

Advanced Usage

Testing with OpenWebUI's RAG Capabilities

To test document-based question answering using OpenWebUI's knowledge database:

python test_remote_ollama.py --prompt "What information do you have about our project?" --model llama3.1

Integrating with Your Local Development

You can use these scripts as a foundation for local development that connects to the remote resources. This allows you to develop and test locally while using the remote server's models and data.