Data Analysis - jeanollion/bacmman GitHub Wiki

Introduction

In BACMMAN, the term data analysis refers to the analysis of measurements. The Data Analysis tab allows to manage data analysis notebooks through an online library and provides a user-friendly way to execute a Jupyter Notebook. A Jupyter Notebook is an interactive document that displays code cells along with their output below (graphs, tables etc..). To open the data analysis tab, select Run > Data Analysis from the menu.

Start by setting a working directory, either by typing the path manually or using right-click menu, which allows you to choose a new directory or recall a previously selected one. Click the Set + Load button to confirm your choice and set the directory. The working directory serves as the root directory for both the Local Notebooks browser (see Notebook library section) and the Notebook server (see Notebook server section).

Notebook server

The notebook server is a program that allows you to edit and execute notebooks interactively.

This functionality requires the installation of Docker; however, the GPU-related components are not necessary. Docker is a platform that packages applications and their dependencies into portable containers, ensuring they run consistently across different environments. A Docker image contains a list of softwares and includes everything needed to run them: code, runtime, libraries, and settings. A Docker container is a running instance of a Docker image, providing an isolated and portable environment for executing the software.

The Run Notebook panel allows you to manage the notebook server on which a notebook will be executed. Select the Docker image (typically the latest version). When clicking the Start button, a container will be created and associated to the selected port and working directory. The notebook server associated to this container is accessible from a web browser at the address 127.0.0.1:8888. The default port is 8888, but it can be changed if it is already in use (for instance if another container is running on the same port). If the Docker image is not already installed, BACMMAN will download and install it (this may take a few minutes, depending on the internet connection speed).

See this page for an example of data analysis using python and the PyBacmman library.

Notebook library

Similarly to configuration and dl model libraries, BACMMAN allows you to store and exchange notebooks via a Github account.

The left panel, Local Notebooks, lets you browse Notebooks located in the working directory. Double-click a notebook to execute it in the selected container (if one is running) or to launch a new container. Right-click a notebook to upload it to the server (authentication is required).

The right panel, Remote Notebooks, lets you browse Notebooks stored on the server. Some notebooks are private and require authentication to be accessed. Right-click a notebook to download it.

For authentication instructions, see this page.