feng-lab/OBMI-Platform

Installation

Setup:

• install miniconda from https://docs.conda.io/en/latest/miniconda.html

• For windows, go to start menu, open Anaconda Prompt (Miniconda3), conda base enviroment should be activated. For other systems, open terminal

  # cmd prompt should be something like (base) C:\Users\..., indicating the base conda enviroment is activated now
  
  # install mamba to base enviroment
  conda install -c conda-forge mamba
  
  # create seprate enviroment with caiman by manba (this installation from caiman official guide)
  mamba create -n bmi -c conda-forge caiman
  # switch to bmi enviroment
  conda deactivate   # exit base
  conda activate bmi  # enter bmi
  
  # some other conda stuff
  # to list current installed packages:
  conda list
  # to update conda
  conda update --all -y
  # to remove outdated conda packages
  conda clean --all -y

• install other enviroment for python

  # make sure we are in bmi enviroment (conda activate bmi)
  # this will install packages for python
  conda install -c conda-forge pyside2 opencv pandas scipy comtypes
  # to show that PySide2 is successfully installed
  conda list    # will show something like  pyside2                   5.13.1           py38ha8f7116_6    conda-forge
  # install hnccorr package (it requires Micorsoft Visual C++ build tools)
  pip install hnccorr

• install pytorch with CUDA, you can get command from https://pytorch.org/get-started/locally/

• install your ide, I am using pycharm, it's free.

• download source code

• open the obmi folder with pycharm, In Preferences of pycharm, set the Project Interpreter to our bmi enviroment (add python interpreter -> conda enviroment -> existing enviroment, our bmi should be there). open src/main.py, open context menu (right click), click 'Run main'.