CCR downtime is scheduled for Tuesday 25 January 2022. Details: https://ubccr.freshdesk.com/support/discussions/topics/13000030322
Using geos on GHub
First: Start the GHub Workspace 10 tool.
1. On the command line, enable the current geospatial conda env:
conda activate geospatial-2021-09
2. Start geos by typing:
Note the URL that geos indicates.
3. Use the Workspace menu (bottom left) to open the Firefox browser, and point it at:
(or whatever was indicated when geos started)
Voila, geos on GHub.
New updates to geospatial-plus kernel on theGHub.org:
The geospatial-plus kernel has been updated with packages to support:
The geospatial-plus kernel now utilizes the geospatial-2021-09 conda environment. Any notebook marked to use this kernel will now run this updated environment.
To see all packages in the current geospatial kernel, run this command in your Deb10 jupyter notebook:
!conda list -n geospatial-2021-09
Or, you can view the partial package list here:
Are you using the pyproj package in your notebooks? Since we have several conda installations of pyproj, you may see a SQLite error when using the package.
Here's how to circumvent it, right inside the notebook:
# Check the database we are using:
# we want to use this most current pyproj database:
# so we can set it as follows:
thedir = '/apps/share64/debian10/anaconda/anaconda-6/envs/geospatial_plus/share/proj'
A note about (Jupyter) Notebooks on GHub:
The Jupyter on Deb10 tool should be your tool of choice for running Jupyter Notebooks on theGHub.org. (Note that the older "Jupyter Notebook" tool is unmaintained, and is retained only for backwards compatability. Please avoid using it!)
For more information about GHub's Jupyter Notebooks capabilities, please try out the Jupyter Notebooks Examples and Documentation (Deb10) tool.
Welcome GHub users!
Feel free to look around and bookmark our direct URL, theghub.org.
You are welcome to register--use your academic or official email address and indicate your affiliation!
Any questions or concerns, feel free to enter a ticket.
Introduction to Tool Development
Any registered member of GHub can develop and share tools on the HUBzero platform--this is one of its greatest strengths. Tools can be based on Jupyter notebooks running Python 3, which makes them both easy to assemble and computationally powerful.
For a recently-deployed example of GHub tools, check out Denis Felikson's modgrnld. Load the page and click "Launch Tool".
If you're not familiar with this kind of thing, just select the first block of code ("cell") and click the "Run" button. You can page through the whole notebook, running each cell, and making the computations.
The power of this tool is: a user can see the code; can run the code and tinker with it--but any changes they make will not be permanent.
The overall workflow you use when building a Jupyter notebook tool is outlined here:
If you need additional Python packages, refer here (feel free to ask me if you need package installs):
For an introduction to Jupyter Notebooks, creating tools, running jobs at CCR, and packages available on GHub:
Jupyter Notebooks Examples and Documentation (Debian10 and Python3)
For development and testing, use these GHub tools, which give you the most recent container environment, Debian 10:
User Interface Elements
Hubzero has created a package with user interface elements for interacting with data, downloading and uploading, etc. This is available to use from your Jupyter notebooks and tools. Documentation is here:
Note: don't use the Rappture module, that's old school. You want the UI module here.
This page presents a general guide to developing tools on GHub:
Additional articles about environment variables, submitting jobs to the CCR cluster, and so forth, are here:
To cite GHub please use the following, published 10 December 2020:
Please note that GHub now enjoys priority access to CCR compute nodes. This change is transparent to you (but should result in shorter wait times for resources, both regular 'submit' jobs and Pegasus jobs). Any issues, problems, or questions, please enter a ticket.
3 November 2020
Send me an email (jsperhac-at-buffalo-dot-edu) to join (or with any questions):
9 October 2020
UI and site appearance:
We are excited to announce new ISMIP6 data access!
ISMIP6 collaborators can now access the data store at CCR and UB using Globus for efficient, easy file transfer.
Most datasets stored on 1km grids can be found on UB Geology's Thule server (endpoint name: Thule_ISMIP6); larger grids are stored on CCR's own project space.
Access Globus using your own institutional credentials. This utility makes data access fast, secure, and easy. Just contact us by creating a ticket with your user details to get set up.
Please check out our newly published tools, which make use of the CCR high-performance computing cluster:
A few GHub updates:
All GHub content pages are now browsable by any visitor. Resources, tools, data access, and project materials remain secured to group membership.
Feel free to post on the wiki, forum, blog, calendar, etc. Play around--this space is for the collaboration.
Documentation has been expanded--including tips about using and developing tools, accessing the Dashboard, and registering for membership.
Please comment or inquire if you have questions or remarks.