Skip to content

matchms/notebooks

Repository files navigation

Binder

Jupyter notebooks with matchms tutorial.

Prerequisites

For running locally on Ubuntu:

sudo apt install jupyter-core
sudo apt install jupyter-notebook

Install Miniconda or Anaconda from https://docs.conda.io/projects/conda/en/latest/user-guide/install/linux.html

Create a conda environment:

conda env create

Activate the conda environment:

conda activate matchms-notebooks

While matchms is being developed, we cannot yet install it from conda-forge, therefore matchms is not yet listed as a dependency in environment.yml. To circumvent this issue, we will check out a development copy of matchms, then use pip install with the --editable argument to add it to the matchms conda environment.

# (first, change to where you want to store the copy of matchms)
git clone https://github.com/matchms/matchms.git
cd matchms
pip install --editable .

Check to see if it all worked by listing the packages that are present in the conda environment named matchms:

conda list

This should yield a table of packages alongside where they were installed from, like so (excerp):

libsodium                 1.0.17               h516909a_0    conda-forge
libstdcxx-ng              9.2.0                hdf63c60_2    conda-forge
markupsafe                1.1.1            py38h1e0a361_1    conda-forge
matchms                   0.1.0                     dev_0    <develop>
mistune                   0.8.4           py38h516909a_1000    conda-forge
nb_conda                  2.2.1                    py38_2    conda-forge
nb_conda_kernels          2.2.3                    py38_0    conda-forge
nbconvert                 5.6.1                    py38_0    conda-forge
nbformat                  5.0.4                      py_0    conda-forge
ncurses                   6.1               hf484d3e_1002    conda-forge
notebook                  6.0.3                    py38_0    conda-forge

You can tell that matchms was not retrieved from conda-forge (the normal source of packages), but instead was installed from local disk as indicated by the <develop> label.

If everything looks OK, start the notebook server:

jupyter notebook

And open a browser to http://localhost:8888 to interact with the notebook.