basesPerSite.sh

#!/bin/sh
# Get ID-total sequences from Quanto data set and calculate total

RDF_PORTAL_EP="http://integbio.jp/rdf/sparql"

QUERY=$(
CAT <<'EOF'
PREFIX rdfs: <http://www.w3.org/2000/01/rdf-schema#>
PREFIX sos: <http://purl.jp/bio/01/quanto/ontology/sos#>

SELECT ?insdcId ?total_sequences
FROM <http://quanto.dbcls.jp>
WHERE {
  ?quanto rdfs:seeAlso ?insdcId;
    sos:totalSequences ?x1 .
  ?x1 rdf:value ?total_sequences .
}
EOF
)

curl -H 'Accept: text/tab-separated-values' --data-urlencode "query=${QUERY}" "${RDF_PORTAL_EP}" | \
awk '
  /SRR/ {
    ncbi += $2
  } /ERR/ {
    ebi += $2
  } /DRR/ {
    ddbj += $2
  } END {
    total=ncbi+ebi+ddbj;
    print "total " NR " records";
    print "ncbi: " ncbi " (" ncbi / total * 100 "%)";
    print "ebi: " ebi " (" ebi / total * 100 "%)";
    print "ddbj: " ddbj " (" ddbj / total * 100 "%)";
  }
'