Installation¶
Agilepy is available as Anaconda package or into a ready-to-use Docker container (from 1.4.0)
Note
AGILE DATASET DOWNLOAD Now it possible to download all the public AGILE dataset stored on SSDC datacenter through a REST Api. Agilepy automatically handles the data and no actions are required from the user. For more information visit this page. This major release includes many new important features and a general refactoring.
Installation with Anaconda¶
Agilepy (and its dependencies) can be easily installed using Anaconda. You just need to decide the name of the virtual environment that will be created by anaconda.
conda config --add channels conda-forge
conda config --add channels plotly
conda create -n <virtualenv_name> -c agilescience agilepy
Note
If you want to try agilepy’s new features that are not officially released yet, a develpoment environment called agilepy-environment is available into Anaconda cloud. It contains all the dependencies unless agilepy, which must be installed by hand cloning the repository. Check the installation instructions here
Supported platforms:
linux-64
osx-64
Note
An experimental package for IBM POWER architecture(ppc64le) is available on Anaconda cloud. Due to some incompability this package does not contain ROOT and AGILE science tools that need to be installed from source. Check the instructions to install AGILE science tools here
Tested on:
CentOs 7.6
Ubuntu 18.04
Ubuntu 19.10
Ubuntu 20.04
macOs 10.14
macOs 10.15
macOS 12.0.1
In order to use the software you need to activate the virtual environment first:
conda activate <virtualenv_name>
or
source activate <virtualenv_name>
Running jupyter server:
start_agilepy_notebooks.sh
Installation with Docker¶
You can pull the image directly from dockerhub using the following command:
docker pull agilescience/agilepy:release-<version>
Note
Check the installation instructions for Docker here
Note
If you want to try agilepy’s new features that are not officially released yet, you need to pull a develop image available using agilepy:develop-latest tag
Using this command you can launch the container and automatically start jupyter notebook.
docker run --rm -it -p 8888:8888 \
-e DISPLAY=$DISPLAY \
-v /tmp/.X11-unix:/tmp/.X11-unix:rw \
-v $PWD/shared_dir:/shared_dir \
agilescience/agilepy:release-<version> /bin/bash -c \
"source /opt/anaconda3/etc/profile.d/conda.sh && conda activate agilepydev && \
jupyter notebook --ip='*' --port=8888 --no-browser --allow-root --notebook-dir="/Agilepy/agilepy/notebooks" --NotebookApp.token='' --NotebookApp.password=''"
shared_dir must be created before launching the command, it is not necessary, but useful for several cases (exporting analysis outside the container, link another dataset etc.)
Jupyter server is at localhost:8888
Agilepy’s containers can be found at dockerhub page
Supported platforms:
linux-64
osx-64
win-64(see note)
Tested on:
CentOs 7.6
Ubuntu 18.04
Ubuntu 19.10
Ubuntu 20.04
macOs 10.14
macOs 10.15
Windows 10 v2004 (May 2020 Update)
Manual Installation¶
If the isntallation does not work with the instructions above, it is recommended to install agilepy and its dependencies from scratch. The dependencies required by Agilepy are:
Root 6.26 Cfitsio 4.1 Zlib
AGILE’s Science Tools (the correct tag to install is on sciencetools_version.txt in the repository main directory)
Uninstalling¶
Anaconda
conda env remove --name <virtualenv_name>
Docker
docker rmi agilescience/agilepy:release-<version>
Package distribution structure¶
The virtual environment <virtualenv_name> folder is under the “envs” folder within the root folder of your anaconda installation.
It contains all the dependencies Agilepy requires. Here, there is the “agiletools” directory, containing AGILE’s scientific software.