Learn more about it from Hannah Clark, Enterprise Client Services at LabArchives, LLC
- Why your company decided to work on the LA and Jupyter integration?
Nearly all LabArchives integrations start with feedback from our users. We regularly interview researchers to learn more about their workflows. We found that many of our researchers were using Jupyter for some of their computational work. Jupyter is one of the most popular tools for working with code. For our team it was clear that an integration with Jupyter would be very useful for our researchers and educators.
We interviewed Jupyter users to understand how the integration would be useful. We heard that researchers still want to use Jupyter but would like to export files from Jupyter and view them in LabArchives. This way, a lab manager or PI could access the data without having to install Jupyter on a home computer.
- When this integration became available?
It became available in mid-November. We hope that more and more researchers will learn about it and use it.
- What are the main benefits and primary users impacted by it?
I often think of Jupyter as a LabArchives notebook for code. It allows you to view your code, explore visualizations, and add text to describe your code. The benefit of this integration is that you can still use Jupyter, but upload Jupyter files to LabArchives so that you have all your research documented in one place – including data, revisions history, code, equations, and visualizations. With the Jupyter integration a LabArchives notebook can really tell the full story of your research, from the moment the data is collected, to the experiments performed, and finally to the analysis that is performed.
- Only Jupyter files can be uploaded to LA and not the other way, correct?
You can upload Jupyter notebook files (.ipynb) to LabArchives. Within Jupyter you can always link to or reference information stored in LabArchives. LabArchives is discipline agnostic; we see more and more researchers from social sciences, natural sciences and even humanities using LabArchives. This integration provides additional options for dry labs to benefit from the revisions history, search, and other great tools in LabArchives.
- What were the challenges your company needed to overcome in order to integrate these two notebooks?
Before development started, we wanted to interview a current Jupyter expert to better understand how they would like the integration to work. Our hardest challenge was finding someone to interview! We were very lucky that one of our site administrators also manages Jupyter for the university and they were able to provide some great insight! Some integrations can take weeks or months to negotiate, design, and develop. Fortunately, the Jupyter integration was a fairly seamless process. If there is someone in your institution who is interested to give us more feedback about the integration, we would love to hear from them.
- Are there any tradeoffs/disadvantages of the integration?
Currently, the integration is simply a viewer for Jupyter Files (.pynb file type), so you still need Jupyter to edit the files. You would need to upload the Jupyter files to LabArchives in the browser or using a tool like Folder Monitor. It is not as detailed as some of our integrations like SnapGene. With SnapGene, you can enable an option to "Save to LabArchives" directly within SnapGene and there are some additional options while viewing a SnapGene file in LabArchives. In the future, there may be further integration with Jupyter.
- What are other future projects? Other integrations in the works?
We have several ideas for 2021 and those are in the works now. Almost all of our integrations (GraphPad Prism, Vernier Logger Pro, Typeset, SnapGene, Qeios and more) start from user feedback. We ask our researchers, research administrators, educators, and research integrity experts what tools they use and how they use that tool as part of their workflow. That feedback is so valuable for our product team to investigate future integration partners. If you have any feature requests or ideas for future integrations, please let us know!
Also, the LabArchives API is available for Duke University researchers who have experience working with an API. The API allows programmers to develop programs that access notebook data or to send data from external applications and software to a notebook, or to provision a notebook on behalf of a user.
Learn more about the Jupyter Integration: https://labarchives.kayako.com/Knowledgebase/Article/View/1008-jupyter-i...
Watch a demo about the LabArchives and Jupyter integration: https://youtu.be/Tlukm65DcD8