Skip to content

Meeting_minutes_181105.md

Latest commit

 

History

History
87 lines (46 loc) · 3.02 KB

Meeting_minutes_181105.md

File metadata and controls

87 lines (46 loc) · 3.02 KB

05/11/2018's meeting minutes with DR, YC, EC, JW, BL, OM, DH

Generalities

  • Capsul, soma-base and soma-workflow have been added to the Populse GitHub project

    • Quick reminder: soma-base contains some generic tools that are used in Capsul and Brainvisa

  • Should we change the name of the several repositories to maintain a certain coherence ?

    • Either rename populse_mia and populse_db to MIA and DB

      • Could be too generic (hard to find the database API by googling only "db")
      • These names already exist on Pypi

    • Or rename Capsul, soma-base and soma-workflow to populse_capsul, populse_base and populse_workflow

      • There could be too many dependencies to change

    • This point will be on the agenda of the next meeting

Populse_MIA's deployment

  • Bring together MIA's current config with Capsul v3's platform

    • There should be three levels of config

      • system config
      • user config
      • project config

    • The user config could be initialized when the software is launched for the first time

      • The user could set the different default values (e.g. to use SPM, FSL)

    • The project config would inherit the user config and would be specific for each project

      • The project config could specify where are the data, the modules to use etc.

    • The user/project config corresponds to the Platform object developed in Capsul v3

      • We should rapidly agree on the development of this API and its concepts (e.g. what do we need to run SPM)
      • @sapetnioc proposed to create a feature branch on Capsul's GitHub repository with the Platform class

  • Populse_MIA's architecture

    • Two choices:

      • 1: keep the same architecture and create a name_of_package.py file in "populse_mia/python/populse_mia" importing all classes in the package

        • e.g. in data_browser.py: "from src.modules.data_browser.count_table import CountTable" etc.
          • So we could write "from populse_mia.data_browser import CountTable"
        • Even if this could avoid to change all the architectures, this could lead to circular imports (not sure)

      • 2: change the architecture to put everything that is under "src/modules" in "populse_mia/python/populse_mia"

    • Option n°2 has been chosen

  • How to deploy/install Populse_MIA?

    • This is not a simple question

      • If the Data Viewer uses compiled files, the deployment on Pypi will be problematic
      • This point will be discussed more precisely in a further meeting

  • Where should be the main.py file?

    • It could be in the python modules and has to be found by the python path

Data Viewer

  • Some problems have been encountered by using PyAnatomist on Windows (PyAnatomist being ensured to work on BrainVisa environment)

    • Using a virtual machine could be a solution

      • @sapetnioc proposed to send a "tutorial" about how to set it up

  • The specification of the Data Viewer should be added to populse.github.io repository

    • Adding a canvas could be great to start the development of the widget

Next meeting

  • On Monday 19/11/2018 at 2PM