Skip to content

The datalab user guide

🚧 This guide is under construction 🚧

Almost everything here is placeholder material. The structure below is real — it is the outline we intend to fill — but most pages currently contain section headings, notes to the authors, and gaps where screenshots will go. Nothing on this site should be relied on yet, and anything it does say may be wrong.

  • Finished documentation lives at docs.datalab-org.io.
  • If you would like to help write a page, see contributing.
  • Progress: 0 of 25 screenshots taken, 91 sections still to write.

This guide is for people using a datalab instance that someone else runs: recording samples and cells, attaching raw data, keeping track of what came from where, and getting your data back out again at the end.

If you are standing up an instance for your group, or writing code against the server, you probably want the main datalab documentation instead — though the API and Extending sections here bridge the two.

Every datalab is different

datalab is self-hosted and plugin-extensible, so no two deployments are alike. Which item types exist, which data blocks are installed, which fields your group actually fills in, and how items are named all vary between instances.

This guide describes the upstream project. Where behaviour depends on configuration, we say so — but if something here is missing from your instance, ask whoever administers it before assuming it is a bug.

  • New here?

    Start at the beginning: log in, find your way around, and record your first sample with files and a plot attached to it.

      Getting started

  • Day-to-day work

    Collections, relationships, instrument sync, the inventory, permissions, sharing and export — one page per task.

      Guides

  • Python Automate it

    Do the same things from a script or a notebook with the datalab-api Python client.

      Using the API

  • Make it yours

    Write a data block for your own technique, or package it up as a plugin for your group.

      Extending datalab

Try it without installing anything

A public demo instance runs at demo.datalab-org.io. It is wiped periodically, so it is a good place to follow this guide through without worrying about making a mess.

The shape of the data

Everything in datalab is an item with a permanent refcode. Items link to each other, carry files, and render those files through data blocks.

graph LR
    SM[Starting materials] --> S[Sample]
    S --> S2[Sample]
    S2 --> C[Cell]
    F[Uploaded files] --> S2
    I[Instrument data] --> F
    S2 --> B[Data blocks]
    C --> B