Count class instances from LOD Laundromat

count_class_instances_from_lod_laundromat.sh

#!/usr/bin/env bash
# 
# A demo script that retrieves counts of class instances from LOD Laundromat datasets
# Usage:
#   Reads a list of newline separated class IRIs either from its first argument or standard input.
#   $ count_class_instances_from_lod_laundromat.sh < classes.txt
#
# Caveat: 
#   Since we use an old-school shell script, class IRIs that contain special characters,
#   like %-encoding or commas will break the CSV output.
#   In general, using <https://github.com/LOD-Laundromat/Frank> might be a saner option.

set -e

die () {
  echo >&2 "$@"
  exit 1
}

# Test if JQ is installed.
command -v jq >/dev/null 2>&1 || die "Missing JQ!"

RDF_TYPE="http://www.w3.org/1999/02/22-rdf-syntax-ns#type"
VOID_TRIPLES="http://rdfs.org/ns/void#triples"

while IFS=$' \t\n\r' read -r CLASS; do
  # Get datasets in which $CLASS occurs
  curl \
    --silent \
    --data-urlencode "uri=${CLASS}" \
    --get http://index.lodlaundromat.org/r2d |
  jq --raw-output '.results | .[]' | # Output dataset IDs separated by newlines
  while read -r DATASET; do 
    COUNT=$(curl \
              --silent \
              --header "Accept:application/ld+json" \
              --data-urlencode "predicate=${RDF_TYPE}" \
              --data-urlencode "object=${CLASS}" \
              --get http://ldf.lodlaundromat.org/${DATASET} |
            jq '..? | .["'${VOID_TRIPLES}'"]? | select (. != null)')

    # Only output non-zero results
    [ ${COUNT} -eq 0 ] || echo ${CLASS}","${DATASET}","${COUNT}
  done
done < "${1:-/dev/stdin}" | # Read classes from the first argument or standard input
(echo "class,dataset,count" && cat) # Prefix output with a CSV header