Skip to content

Information on the Statistical Assessment of the Modeling of Proteins and Ligands (SAMPL) series of blind prediction challenges for computational chemistry

Notifications You must be signed in to change notification settings

samplchallenges/samplchallenges.github.io

Repository files navigation

SAMPL: Crowdsourcing challenges to improve prediction of protein-ligand interactions

This site, and repository, contains information, plans, and announcements relating to the Statistical Assessment of Modeling of Proteins and Ligands (SAMPL) series of blind predictive challenges. This NIH-funded series of challenges tests computational models on predictions of properties related to challenges faced in a drug discovery setting.

If you are reading this file, it means you are visiting our GitHub repository. The contents of this repository are meant to be served instead via our SAMPL website, so please visit there instead.

Site status

These pages are currently under construction and will be updated further in the immediate future with the latest SAMPL news and announcements. (Status date April 30, 2019).

Contacts

If you have questions or concerns, please use our issue tracker here, or contact David L. Mobley

Site administration

Generate navigation

The subpages of livecomsjournal.github.io have a navigation column on the left (see livecomsjournal.github.io/authors/policies/ for an example). This navigation is read from _data/navigation.yml and is static, meaning that it needs to be re-generated whenever changes to the structure of a document are done. Specifically, the navigation consists of links to all first and second title levels (# Title and ## Title in Markdown), and if any of those are changed, the navigation needs to be updated.

There is a script named generate_navigation.py which generates the content of the navigation file. Calling

python3 generate_navigation.py > _data/navigation.yml

regenerates the navigation file with the current state of the Markdown-files. It does so by reading through all .md files (except index.md) in the _about_livecoms, _author_instructions and _editorial_policies and generating a menu item for every first-level or second-level title. It also read the frontmatter of each file and rewrites it in case no permalink or no sidebar tag is found.

The script is helpful, but probably not fool-proof - a visual inspection of the result via git diff, or even better a check of the result in a local jekyll installation is highly recommended before pushing the new navigation file to the master branch.

Updating List of All References _history/02_allreferences.md

The list of all references is automatically updated by a github actions workflow using .github/workflows/zotero_pull.yml. Currently the repository is set to update references by running the workflow weekly every Saturday at 00:00 or 12:00 AM. This can easily by changed by updating the cron string on line 13 of .github/workflows/zotero_pull.yml.

All references are pulled from the Mobley Group Zotero library using the tag SAMPL-website. The ZOTERO_LIBRARY_ID and ZOTERO_API_KEY are stored as secrets for this repository.

The workflow pulls the "header text" the formatting, title, and text that will appear above the references list from a file called _history/allreferences_header.md.

  • To update this text, you will need to edit the text in _history/allreferences_header.md rather than _history/02_allreferences.md otherwise your changes will not persist the next time the workflow is run.
  • When updating _history/allreferences_header.md, please leave the first line "##AllReferencesHeader" as is as this is stops this page from being displayed on the website.

Upon running the workflow, the script generate_references.py pulls all references with the correct tag, and organizes in reverse chronological order. It then overwites _history/02_allreferences.md by adding the header text from _history/allreferences_header.md then the formatted references. Finally, it creates a pull request with the changes to be merged into the repository. A PR is only created when changes to the reference list happens, some workflow runs may result in no changes, and thus no PR.

To test this locally you can run:

python generate_references.py --tag SAMPL-website --lib-id [group ID from Dr. Mobley] --api-key [api key for group from Dr. Mobley] --lib-type group --headerfile _history/02_allreferences.md --outfile _history/02_allreferences.md

Testing Locally

First you will need to install the gem for github-pages. Run the following command:

gem install github-pages

This will install the github-pages gem and all dependencies (including jekyll).

Later, to update the gem, type:

gem update github-pages

To build a local copy, go into the directory and type:

jekyll build

This will create (or modify) a _site/ directory, containing everything from assets/, and then the index.md and all pages/*.md files, converted to html. (So there’ll be _site/index.html and the various _site/pages/*.html.)

Do not commit this directory in a Pull Request!! Github will compile this directory on its own.

Type the following in order to “serve” the site. This will first run build, and so it does not need to be preceded by jekyll build.

jekyll serve

Now open your browser and go to http://localhost:4000

Credits

This adapted from/based on the construction of the LiveCoMS Journal GitHub pages site based on the Jekyll Mininimal Mistakes theme.

About

Information on the Statistical Assessment of the Modeling of Proteins and Ligands (SAMPL) series of blind prediction challenges for computational chemistry

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published