extract_accession_numbers

#!/usr/bin/env bash
# Extract accession numbers from database search results

# Author: Helge Knüttel
# Date: 2020-2022

# Some sensible settings. Taken from https://kvz.io/bash-best-practices.html
set -o errexit  # exit when a command fails
set -o pipefail # catch failing commands in a pipe
set -o nounset  # exit when script tries to use undeclared variables
# set -o xtrace # Trace what gets executed. Useful for debugging.

# Help text
PrintHelp() {

  cat << "EOF"

Extract accession numbers from database search results.

  Input is read from STDIN. Output is written to STDOUT.


Options

  --format          Export format. Mandatory option. 
  

Export formats

  Known export formats are:

  pubmed             Export from PubMed as format 'PubMed' (formerly 'MEDLINE')

  ovid_medline       Export from Ovid MEDLINE as:
                      - Format: Citavi, Endnote or ReferenceManager
		      - Fields: Complete Reference

  ovid_embase       Export from Ovid Embase as:
                      - Format: Citavi, Endnote or ReferenceManager
		      - Fields: Complete Reference

  ovid_ris          Export from Ovid Embase or Ovid MEDLINE as:
                      - Format: RIS
		      - Fields: Complete Reference

  cochrane_reviews_endnote_ris
                     Export from Cochrane Library (Reviews) as format 'RIS (EndNote)'

  cochrane_reviews_refman_ris
                     Export from Cochrane Library (Reviews) as format 'RIS (Reference Manager)'

  cochrane_reviews_plain_text
                     Export from Cochrane Library (Reviews) as format 'Plain text'

  cochrane_protocols_endnote_ris
                     Export from Cochrane Library (Protocols) as format 'RIS (EndNote)'

  cochrane_protocols_refman_ris
                     Export from Cochrane Library (Protocols) as format 'RIS (Reference Manager)'

  cochrane_protocols_plain_text
                     Export from Cochrane Library (Protocols) as format 'Plain text'

  cochrane_trials_endnote_ris
                     Export from Cochrane Library (Trials) as format 'RIS (EndNote)'

  cochrane_trials_refman_ris
                     Export from Cochrane Library (Trials) as format 'RIS (Reference Manager)'

  cochrane_trials_plain_text
                     Export from Cochrane Library (Trials) as format 'Plain text'

  cochrane_clinical_answers_endnote_ris
                     Export from Cochrane Library (Clinical Answers) as format 'RIS (EndNote)'

  wos_endnote        Export from Web of Science Core Collection as: 
                       Endnote Desktop --> Record content: Full Record (.ciw-file)

  wos_other          Export from Web of Science Core Collection as: 
                       Other file format --> Record content: Full Record; 
                       File Format: Other reference software (.txt-file)

  ebscohost_psycinfo_ris
		     Export from EBSCOhost PsycInfo as format 'RIS (e.g. CITAVI, EasyBib, EndNote, ProCite, Reference Manager, Zotero)'

  ebscohost_cinahl_ris
		     Export from EBSCOhost CINAHL as format 'RIS (e.g. CITAVI, EasyBib, EndNote, ProCite, Reference Manager, Zotero)'

  ictrp
		     Export from WHO ICTRP as format 'XML'

  clinicaltrials_gov_txt
		     Export from ClinicalTrials.gov as format 'txt'

  clinicaltrials_gov_tsv
		     Export from ClinicalTrials.gov as format 'TSV'

  clinicaltrials_gov_xml
		     Export from ClinicalTrials.gov as format 'XML'


Examples

  cat ovid_embase_export.cgi | extract_accession_numbers --format ovid_embase | wc -l
  
EOF
}


# Read command line arguments
if [ $# -eq 0 ]
then
    echo "ERROR: Mandatory argument(s) missing. Try '${0} --help'" >&2
    exit 1
fi

while [ $# -gt 0 ]
do
    case "$1" in
	--format )
	    shift
	    if [ $# -gt 0 ]
	    then
		export_format="$1"
		shift
	    else
		echo "ERROR: Missing --format argument" >&2
		exit 1
	    fi
	    ;;
	-h|--help )
	    PrintHelp
	    exit 0
	    ;;    
	* )
	    # unknown argument
	    unknown_arg=$1
	    echo "ERROR: Unknown argument ${unknown_arg}" >&2
	    exit 1
	    ;;
    esac
done


case "$export_format" in
    pubmed )
	grep "^PMID- " - | sed -e 's/^PMID- //'
	;;
    ovid_medline|ovid_embase )
	grep "^UI  - " - | sed -e 's/^UI  - //' -e 's/\r//g'
	;;
    ovid_ris )
	grep "^ID  - " - | sed -e 's/^ID  - //' -e 's/\r//g'
	;;
    cochrane_reviews_endnote_ris|cochrane_reviews_refman_ris )
	grep "^AN  -  " - | sed -e 's/^AN  -  //' -e 's/\r//g'
	;;
    cochrane_reviews_plain_text )
	grep "^ID: " - | sed -e 's/^ID: //' -e 's/\r//g'
	;;
    cochrane_protocols_endnote_ris|cochrane_protocols_refman_ris )
	grep "^AN  -  " - | sed -e 's/^AN  -  //' -e 's/\r//g'
	;;
    cochrane_protocols_plain_text )
	grep "^ID: " - | sed -e 's/^ID: //' -e 's/\r//g'
	;;
    cochrane_trials_endnote_ris|cochrane_trials_refman_ris )
	grep "^AN  -  " - | sed -e 's/^AN  -  //' -e 's/\r//g'
	;;
    cochrane_trials_plain_text )
	grep "^ID: " - | sed -e 's/^ID: //' -e 's/\r//g'
	;;
    cochrane_clinical_answers_endnote_ris )
	grep "^DO  -  " - | sed -e 's/^DO  -  //' -e 's/\r$//g'
	;;
    wos_endnote|wos_other )
	grep "^UT " - | sed -e 's/^UT //' -e 's/\r//g'
	;;
    ebscohost_psycinfo_ris )
	grep "^AN  - " - | sed -e 's/^AN  - //' -e 's/\r//g'
	;;
    ebscohost_cinahl_ris )
	grep "^ID  - " - | sed -e 's/^ID  - //' -e 's/\r//g'
	;;
    ictrp )
	grep "^    <TrialID>" - | sed -e 's/^    <TrialID>//' -e 's/\s*$//'
	;;
    clinicaltrials_gov_txt )
	grep "^  NCT Number:" - | sed -e 's/^  NCT Number:\s\+//' -e 's/\s*$//'
	;;
    clinicaltrials_gov_tsv )
	cut -f2 - | tail -n+2
	;;
    clinicaltrials_gov_xml )
	grep "^    <nct_id>" - | sed -e 's/^    <nct_id>//' -e 's/<\/nct_id>\s*$//'
	;;
    * )	
	echo "ERROR: Unknown export format ${export_format}" >&2
	exit 1
	;;
esac