Check LOP Install

Overview

The Check LOP Install component verifies that all Python libraries required by LOPs operators are properly installed in the LOPs virtual environment. It provides a dashboard of availability toggles and install buttons so you can quickly see what is missing and install it without leaving TouchDesigner.

Key Features

• Detects whether the LOPs Python virtual environment is set up and has a valid Python executable
• Checks availability of five commonly used libraries: PyTorch, OpenAI, Tiktoken, Llama-Index, and Sentence-Transformers
• One-click installation for each library through ChatTD’s Python manager
• Automatic CUDA version detection for PyTorch installation on Windows, with manual fallback
• Dependency enforcement: Llama-Index and Sentence-Transformers prompt you to install PyTorch first if it is missing on Windows

How It Works

On the Check Libraries page, each library has a read-only toggle showing whether it is currently available and an Install button that is enabled only when the library is missing. On the Check Environment page, a toggle shows whether the LOPs virtual environment exists, and a Setup button lets you create it if needed. Pulse Refresh All Checks to re-scan everything at once.

All installations are routed through ChatTD’s built-in Python manager, so packages land in the correct LOPs virtual environment automatically.

Usage

First-Time Setup

1. Open the Check Environment page.
2. If LOPS Venv Available is off, pulse Setup LOPS Venv. This calls ChatTD to create the shared virtual environment.
3. Once the venv is ready, pulse Refresh All Checks to scan all libraries.

Installing a Library

1. Open the Check Libraries page.
2. Find the library you need — its toggle will be off if it is missing.
3. Pulse the corresponding Install button (e.g., Install PyTorch).
4. For PyTorch on Windows, a dialog will detect your CUDA version or let you choose manually (CUDA 11.8 or 12.1). On macOS, the CPU-only version is installed automatically.
5. After installation completes, the component reinitializes and the availability toggle updates.

Re-installing PyTorch

If PyTorch is already detected but you suspect a broken install, pulse Install PyTorch again. A dialog will offer a forced clean reinstall with cache clearing.

Troubleshooting

• All library toggles are off after refreshing: The LOPs virtual environment is not set up. Install it first from the Check Environment page, then refresh.
• “ChatTD reference not found” errors: ChatTD must be present in the project. This is standard LOPs infrastructure and should always be available.
• CUDA detection fails on Windows: If nvcc --version is not on your system PATH, the installer will prompt you to manually select CUDA 11.8 or 12.1.
• Llama-Index or Sentence-Transformers install is blocked: These libraries depend on PyTorch on Windows. Install PyTorch first when prompted.