ideas-toolbox-dlc

Table of Contents

• Toolbox Description
• How to Get Help
• Navigating the Project Repository

Toolbox Description

A toolbox for running DeepLabCut-based tools on the IDEAS platform.

This toolbox is designed to run as a Docker image, which can be run on the IDEAS platform. This toolbox consists of the following tools:

• Train and Evaluate DeepLabCut Model: Train and evaluate a new DeepLabCut model using a set of labeled images and videos.
• DeepLabCut Pose Estimation: Analyze behavioral videos using any pre-trained DeepLabCut model
• Inscopix Bottom View Mouse Pose Estimation: Analyze Inscopix bottom-view mouse videos using a pre-trained DeepLabCut model, trained internally at Inscopix.

How to Get Help

• IDEAS documentation contains detailed information on how to use the toolbox within the IDEAS platform, the parameters that can be used, and the expected output.
• If you have found a bug, we reccomend searching the issues page to see if it has already been reported. If not, please open a new issue.
• If you have a feature request, please open a new issue with the label enhancement

Executing the Toolbox

To run the toolbox, you can use the following command:

make run TOOL=<tool_name>

Available tools are:

• inscopix_bottom_up_model__run
• training__train_model
• workflow__run_workflow

The command will excute the tool with inputs specified in the inputs folder. The output will be saved in the outputs folder.

Navigating the Project Repository

├── commands                # Standardized scripts to execute tools on the cloud
├── data                    # Small data files used for testing
├── info                    # Information about the toolbox and its tools
├── inputs                  # Example input files for testing the tools
│── toolbox                 # Contains all code for running and testing the tools
│   ├── tools               # Contains the individual analysis tools
│   ├── utils               # General utilities used by the tools
│   ├── tests               # Unit tests for the individual tools
│── Makefile                 # To automated and standardize toolbox usage
│── Dockerfile               # Commands to assemble the Docker imageons
|── LICENSE                 # License file
|── user_deps.txt           # Python dependencies for the toolbox
└── .gitignore              # Tells Git which files & folders to ignore

License

This project is licensed under CC BY-NC 4.0. Note that the software is provided "as is", without warrantee of any kind.