Skip to content

Latest commit

 

History

History
80 lines (50 loc) · 2.93 KB

user_guide.rst

File metadata and controls

80 lines (50 loc) · 2.93 KB
Raw

User Guide

This website consists of isolated fully described projects, runnable locally and also deployed as public examples using Anaconda Enterprise (AE).

Download a Project

All of the examples on this website have a link to download the project. Once you have downloaded a project you can run it locally or on AE.

Run Locally

To run an example locally, first install anaconda-project.

conda install anaconda-project=0.8.3

Once you unpack the project locally and visit that directory, you can see that each project directory has a text file anaconda-project.yml that defines an environment along with predefined commands that can be run in that environment. To run the default command defined in that project, do:

anaconda-project run

Running this command will install the dependencies for the particular project, then execute whatever the first command is. E.g. for a Bokeh or Panel dashboard, the default command could start a Bokeh server (e.g. it will end with a statement like: Bokeh app running at: http://localhost:5006/attractors_panel ). You can then open the given link to see the running dashboard.

If the default command is a dashboard or app but you want to see or edit the individual steps involved, most projects also provide a predefined "notebook" command:

anaconda-project run notebook

Other commands might be defined in the .yml file as well, e.g. multiple notebooks, multiple dashboards, or other tasks. You can also run any command you like in the provided environment, even if it's not defined in the .yml already. E.g. to launch a Jupyter notebook server for the entire directory, you can ask anaconda-project to run jupyter notebook, jupyter lab, or any other program:

anaconda-project run jupyter notebook

If you don’t want to use anaconda-project at all, you can create a regular conda environment using:

conda env create --file anaconda-project.yml

Activate the environment (be sure to replace env-name with the real name of the environment you created):

conda activate <env-name>

Then start a jupyter notebook as usual:

jupyter notebook

NOTE: If the notebook depends on data files, you will need to download them explicitly if you don’t use anaconda-project, by extracting the URLs defined in anaconda-project.yml and saving the file(s) to this directory.

Run on AE

In addition to running examples locally you can upload and share them using Anaconda Enterprise, which is the platform we use for publishing our public deployments. If you’ve already installed anaconda-project, then for an example named “bears” just do:

cd bears
anaconda-project archive bears.zip

Then in the AE interface select “Create”, “Upload Project” and navigate to the zip file. Once your project has been created, you can deploy it.