Anaconda
Open Data Science Core: Anaconda is a package manager, an environment manager, a Python distribution, and a collection of over 1,000+ open source packages. It is free and easy to install, and it offers free community support. https://anaconda.org/
Context
Anaconda is well documented and has excellent support. All I include here are a few tasks that I have had to repeat.
Jupyter
Activate and deactivate the environment
-
Windows:
activate env_name-
deactivate env_name
-
-
Linux:
source activate env_name-
source deactivate env_name
-
Run a notebook
jupyter notebook
Save the files of a notebook tree into an archive (assumes system with tar)
-
When in a notebook, go to a new cell
-
!tar chzf notebook.tgz --exclude notebook.tgz *(change notebook to your choice)
Download that archive
-
From the Jupyter File menu, select Open
-
Click the archive you created, i.e. notebook.tgz
-
It will download to your location
-
Unarchive to use it
Run a command in a notebook
Prepend the command with !
Build identical environment
The operating systems of the host computers must be similar. This cannot be done to copy a Linux environment to Windows.
First a listing of the environment is needed.
-
Try
conda list --explicit-
If you get an error similar to Error: environment does not exist: ../anaconda/envs/name/envs/name
-
-
Then try
conda list -n root --explicit -
To save the output to a file use a pipe appended to the end of the command, similar to
> spec-file.txt -
Next, on your anaconda installation:
conda create --name myenv --file spec-file.txt