Q & A: Can I bring my own tools and algorithms to BDC?
Nathalie Volkheimer
started a topic
12 days ago
For bringing your own tools and algorithms, there are two options: create a CWL wrapper for the software using a Docker container that has the software installed or use an interactive Data Studio session. Which of these you prefer depends on some specific technical specifications of the software.
1. CWL tool wrapping:
1. If the software has clear, predefined inputs and outputs and parameters, and can be executed from a bash command line, this is the way to go.
4. Seven Bridges staff are available via email (support@velsera.com) and at live drop-in office hours to provide assistance.
2. Data Studio session:
1. For most developers, Data Studio is the “quick start”, because it allows you to upload and run code in a familiar IDE. Python is available via JupyterLab Notebooks, and R is available via R Studio in the cloud. SAS is also available, but not typically used for AI algorithms.
2. Drivers and dependencies can be installed inside the interactive session and saved for later sessions.
A common concern is that by putting the software on BDC, that will release it to the public or affect its licensing. This is not true. If the software is being loaded via Docker container, we provide a private registry that can only be accessed by users that you specify. If the software is installed in a Data Studio (interactive) session, it can still only be accessed by users that you add as members to the project. Even BDC-Seven Bridges staff members will be unable to access the Docker container or the code unless proactively added to the project/container registry.
Nathalie Volkheimer
For bringing your own tools and algorithms, there are two options: create a CWL wrapper for the software using a Docker container that has the software installed or use an interactive Data Studio session. Which of these you prefer depends on some specific technical specifications of the software.
1. CWL tool wrapping:
1. If the software has clear, predefined inputs and outputs and parameters, and can be executed from a bash command line, this is the way to go.
2. Create a Docker container with your software installed, using our guidelines from the BDC-SB Documentation:https://sb-biodatacatalyst.readme.io/docs/sdk-overview
3. Use the CWL Tool Editor to create the workflow execution wrapper.https://sb-biodatacatalyst.readme.io/docs/about-the-tool-editor
4. Seven Bridges staff are available via email (support@velsera.com) and at live drop-in office hours to provide assistance.
2. Data Studio session:
1. For most developers, Data Studio is the “quick start”, because it allows you to upload and run code in a familiar IDE. Python is available via JupyterLab Notebooks, and R is available via R Studio in the cloud. SAS is also available, but not typically used for AI algorithms.
2. Drivers and dependencies can be installed inside the interactive session and saved for later sessions.
A common concern is that by putting the software on BDC, that will release it to the public or affect its licensing. This is not true. If the software is being loaded via Docker container, we provide a private registry that can only be accessed by users that you specify. If the software is installed in a Data Studio (interactive) session, it can still only be accessed by users that you add as members to the project. Even BDC-Seven Bridges staff members will be unable to access the Docker container or the code unless proactively added to the project/container registry.