Registering a dataset using the API

#!/bin/bash -eu
#
# This is an outline Bash script to show updating a set of datasets registered with GBIF.org.
#
# The script isn't recommended for production use; it shows the steps involved, but does not have any error checking.
#
# Using a process like this, you should make sure you store the UUID GBIF assigns to your dataset, so you don't accidentally
# re-register existing datasets as new ones.

# Starting point:
# - One or more datasets as DarwinCore Archives or EML files in the current directory.
#   - This directory has an EML file, which is sufficient for a metadata-only dataset.
# - A web server exposing these files
#   - We can use the GitHub view of this directory for that:
ACCESS_ENDPOINT=https://techdocs.gbif.org/en/data-publishing/_attachments
# - An organization registered in GBIF
# - An installation registered in GBIF (represents the server this script runs on)
# - A GBIF.org user account with publishing rights for the registered organization.
#   - These are the test values available for use on GBIF-UAT.org
ORGANIZATION=0a16da09-7719-40de-8d4f-56a15ed52fb6
INSTALLATION=92d76df5-3de1-4c89-be03-7a17abad962a
GBIF_USER=ws_client_demo
GBIF_PASSWORD=Demo123

# Loop through all the DWCA and EML files:

shopt -s extglob
for dataset_file in *.@(zip|eml) ; do

    # Guess dataset type (script doesn't handle checklists or sampling event datasets)
    case $dataset_file in
        *.eml)
            dataset_type=METADATA
            endpoint_type=EML
            ;;
        *)
            dataset_type=OCCURRENCE
            endpoint_type=DWC_ARCHIVE
            ;;
    esac

    # Check if the dataset is already registered -- we have a local file recording the UUID if that is the case.
    if [[ -e $dataset_file.registration ]]; then
        dataset=$(cat $dataset_file.registration)
        echo "Dataset $dataset_file is already registered at $dataset"
    else

        echo "Registering dataset $dataset_file"

        # Make a JSON object representing the minimum metadata necessary to register a dataset.  The rest of the metadata will
        # be added when GBIF.org retrieves the dataset for indexing.

        # The license isn't essential, but is very helpful to us if you can provide it (correctly) in the initial registration.

        # If using your own DOIs, add "doi": "10.xxxx/xxxx" to this JSON object.

        cat > $dataset_file.registration_json <<-EOF
        {
            "publishingOrganizationKey": "$ORGANIZATION",
            "installationKey": "$INSTALLATION",
            "type": "$dataset_type",
            "title": "Example dataset registration",
            "description": "The dataset is registered with minimal metadata, which is overwritten once GBIF can access the file.",
            "language": "eng",
            "license": "http://creativecommons.org/publicdomain/zero/1.0/legalcode"
        }
        EOF

        # Send the request by HTTP:
        curl -Ssf --user $GBIF_USER:$GBIF_PASSWORD -H "Content-Type: application/json" -X POST --data @$dataset_file.registration_json https://api.gbif-uat.org/v1/dataset | tr -d '"' > $dataset_file.registration
        dataset=$(cat $dataset_file.registration)
    fi

    if [[ -e $dataset_file.endpoint ]]; then
        echo "  Endpoint is already set"
    else

        # Add an endpoint, the location GBIF.org will retrieve the archive (or EML) file from:
        cat > $dataset_file.endpoint_json <<-EOF
        {
            "type": "$endpoint_type",
            "url": "$ACCESS_ENDPOINT/$dataset_file"
        }
        EOF

        curl -Ssf --user $GBIF_USER:$GBIF_PASSWORD -H "Content-Type: application/json" -X POST --data @$dataset_file.endpoint_json https://api.gbif-uat.org/v1/dataset/$dataset/endpoint > $dataset_file.endpoint

        echo "Dataset registered, see https://registry.gbif-uat.org/dataset/$dataset or https://api.gbif-uat.org/v1/dataset/$dataset"
    fi

done