Skip to content

Latest commit

 

History

History
49 lines (36 loc) · 2.73 KB

instructions.md

File metadata and controls

49 lines (36 loc) · 2.73 KB
Raw

Detailed Submission Instructions

Fill out and submit and issue

  • Click on Issues from the menu bar
  • Click on New issue
  • Select the Dataset submission/request template by clicking the Get started button
  • Fill out the issue template, remembering to enter your dataset name in the issue title
  • Submit the issue

Prepare your dataset

  • Create a folder to contain all your dataset files
  • Give this folder the name of your dataset, replacing spaces with underscores → _
  • Move your dataset files (and ONLY your dataset files) into this folder

Fork the colabfit-data repository

  • Click on the Fork dropdown menu near the upper right-hand side of the page
  • Click on Create new fork
  • Leave Copy the main branch only selected
  • Click on Create fork

Upload your dataset

  • Make sure you are on your fork of the repository
  • Select Add file → Upload files from the right side of the activity bar
  • Using your mouse, drag your entire dataset folder (containing all dataset files) onto the GitHub pane
    • Note: You must drag your files from your file explorer application onto the GitHub pane. The choose your files button will not allow you to upload a folder.
  • Click on Commit changes. Leave Commit directly to the main branch selected.

Submit a pull request

  • Click on Compare & pull request
    • You may need to find the Contribute button and click `Open pull request
  • Select the submit branch as your PR base
    • The bar above the submission template should look something like: base repository: colabfit/data-lake base: main ← head repository: your-GitHub/data-lake compare: main
  • Read through the pull request checklist and ensure you are prepared to submit
  • Click Create pull request

File expectations

Content

Files should contain the data necessary to model atomic systems: atom types, atom coordinates, unit cell/lattice. They should also contain the values for calculated properties of the systems. For example properties, see the current list of properties included in ColabFit. We are most interested in files describing the relaxation trajectories of systems calculated at the quantum (e.g., CCSD) or DFT levels of theory.

Format

Files in the .extxyz format (e.g., as implemented by ASE) are the most convenient to ingest, though we are able to work with other formats, including CIF, npy/npz and JSON.

Files may be compressed using a common compression format (tar, zip). Please do not use unusual or proprietary compression formats. If compressing a directory, keep in mind that GitHub has an upper limit on file sizes. It may be better to upload a larger number of smaller files.