Skip to content

Latest commit

 

History

History
76 lines (60 loc) · 5.78 KB

add-json-records.md

File metadata and controls

76 lines (60 loc) · 5.78 KB
Raw

Add and validate GeoJSON records in GitHub

Context

We have two repositories where we store metadata records for the SDR:

  • opengeometadata/edu.nyu is where cannonical production records live; we and other institutions will index records from these repos directly to their GeoBlacklight instances.
  • nyu-dataservices/gis-metadata-staging is a fork of opengeometadata/edu.nyu; we use it to work on records that are in process and to index to our NYU SDR staging instance of GeoBlacklight.

All new NYU records should be made in nyu-dataservices/gis-metadata-staging, validated, then merged via PR over to opengeometadata/edu.nyu when ready. Both repositories should be in-sync whenever possible.

Instructions

  1. Clone NYU-DataServices/gis-metadata-staging and cd gis-metadata-staging.

  2. Create a new branch (git checkout -b [[branch name]]) using the following conventions for branch name:

    Action type Branch name
    Edit existing records edit-COLLECTION-SHORT-NAME
    Create new records new-COLLECTION-SHORT-NAME

    The collection short name should be a keyword from the title of the collection. Use lower case and separate words with hyphens.

    Example: git checkout -b edit-egress

  3. Edit the relevant JSON records, replace existing records with the JSON modified in Airtable, or add new records to the metadata-aardvark folder. (Not metadata-1.0 Only aardvark records will get indexed to the NYU Spatial Data Repository as of Fall 2023.) 4. When the records in the branch are ready, make a pull request from your branch to main in the same staging repo, using a standardized commit message:

    • git add JSON-FILES-CHANGED, e.g. git add metadata-aardvark/Datasets/nyu-2451-33876.json
    • git commit <-- this will open a Vim editor window where you can write the title of the commit and a description. Hit ito start editing, entering your commit title on the first line after the comment section, a blank line, then the commit description.

      • Use the following convention for commit titles:

        Commit title keyword Record change type
        Field Error Intended field value not correct
        Bitstream Edit Change to location of download URL for asset
        Update Field Change a field or fields with new information, as in a new version of a description, title, etc.
        Typo Correction Inadvertent error in field value
        Object Type Error Incorrect JSON object type (e.g. array for string in key's value
        General Validation Error Fix to general linting validation fail
        Add New Records Add new record to system 
    - For a description (entered after leaving a blank line after the commit title), offer a short narrative of a 1-2 sentences about what edits are being made.
- Close vim using `esc`, `:`, `wq`
- Push using `git push origin [[local branch name]]`, e.g. `git push origin edit-egress`
- Return to the  [gis-metadata-staging]((https://github.com/NYU-DataServices/gis-metadata-staging)) repo, click on "Pull Requests" on the top menu, and select "New pull request." Set it so that the pull is from your local branch into the repo's `main` branch.

  4. When a PR into main is created, this will trigger linting for the records in GitHub Actions. (See: gis-metadata-staging/actions). When they pass, go ahead and merge into main

  5. *** TO DO / NOT IMPLEMENTED * ** Commits to main (directly or via merged PR) will trigger indexing of the aardvark records found in the main branch of NYU-DataServices/gis-metadata-staging to NYU's staging instance of the SDR for QA testing. (Must be on NYU VPN to view the staging instance).

  6. After QA on staging, a PR can be submitted from NYU-DataServices/gis-metadata-staging over to the canonical repo OpenGeoMetadata/edu.nyu. Any records pulled into this canonical OpenGeoMetadata repo will get indexed to our production instance (and other institutions!)

Local development & validation

Follow the steps below to run linting and other tasks locally.

Prerequisites

  • Git
  • Ruby (via RVM, Chruby, or Rbenv)

Steps

  1. Clone the repository 
    gh repo clone NYU-DataServices/gis-metadata-staging & cd gis-metadata-staging
  2. Install ruby dependencies 
    bundle install
  3. Run linting on one or both groups of records 
    bundle exec rake lint:v1
bundle exec rake lint:aardvark
bundle exec rake lint:all
  4. If the aarvark linters fail, you might try to run and/or update the script in /misc: 
    bundle exec ruby misc/aardvark-schema-patching.rb