an2query

#!/usr/bin/env bash
# Convert a list of database accession numbers or other IDs into a search string

# Author: Helge Knüttel
# Date: 2020

# 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"

Convert a list of database accession numbers or other IDs into a search string

  Each accession number/ID must be on an individual line.
  
  Input is not checked for correct accession numbers/IDs but whitespace at beginning 
  or end of lines as well as empty lines will be deleted.

  The ID types known are specific for a database syntax.

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


Options

  --syntax          Database syntax. Mandatory option. 
  
  --idtype          Type of acession number/ID. Mandatory option.
  

Database syntax formats and ID types

  Known syntax formats are host and database specific:

  pubmed              ID types: pmid, pmcid, doi, aid, lid
  pmc                 ID types: pmid, pmcid, doi
  ovid_medline        ID types: pmid, pmcid, doi, pii
  ovid_embase         ID types: an, eu, pmid, doi
  cochrane_library    ID types: an, doi, tn, none
  web_of_science      ID types: an, doi, pmid
  citavi              ID types: doi, pmcid, pmid
  ebscohost_psycinfo  ID types: an, pmid, doi
  ebscohost_cinahl    ID types: an, pmid
  dimensions          ID types: doi
  dimensions_api      ID types: doi

Examples

  cat pmid.txt | an2query --syntax ovid_medline --idtype pmid > query.txt
  cat ovid_embase_export.cgi | extract_accession_numbers --format ovid_embase | an2query --syntax ovid_embase --idtype an > query.txt
  
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
	--syntax )
	    shift
	    if [ $# -gt 0 ]
	    then
		syntax="$1"
		shift
	    else
		echo "ERROR: Missing --syntax argument" >&2
		exit 1
	    fi
	    ;;
	--idtype )
	    shift
	    if [ $# -gt 0 ]
	    then
		idtype="$1"
		shift
	    else
		echo "ERROR: Missing --idtype 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 "$syntax" in

    pubmed )
	# Fields to search for in PubMed
	case "$idtype" in
	    pmid )
		PUBMED_FIELD=UID
		QUOTE=""
		;;
	    pmcid )
		PUBMED_FIELD=PMCID
		QUOTE=""
		;;
	    doi|aid )
		PUBMED_FIELD=AID
		QUOTE='"'
		;;
	    lid )
		# Location ID
		PUBMED_FIELD=LID
		QUOTE='"'
		;;
	    * )
	        # unknown argument
		echo "ERROR: Unknown idtype ${idtype}" >&2
		exit 1
		;;
	esac
	# Steps: 
	#  * Delete empty lines or lines containig only whitespace.
	#  * Add ${QUOTE} to beginning of each line of the input.
	#  * Add ${QUOTE} and field specification to end of each line of the input.
	#    Whitespace at beginning or end of line will be deleted.
	#  * Add ' OR ' to end of each line except the last one.
	sed\
	    -e '/^\s*$/d' - | \
	sed\
	    -e "s/^\s*/${QUOTE}/" \
	    -e "s/\s*$/${QUOTE}\[${PUBMED_FIELD}\]/" \
	    -e '$! s/$/ OR /'
	;;

    pmc )
	# Fields to search for in PubMed Central
	case "$idtype" in
	    pmid )
		PMC_FIELD=PMID
		;;
	    pmcid )
		# Apparently, there is no dedicated field for PMCIDs that can be searched.
		PMC_FIELD=ALL
		;;
	    doi )
		PMC_FIELD=AID
		;;
	    * )
	        # unknown argument
		echo "ERROR: Unknown idtype ${idtype}" >&2
		exit 1
		;;
	esac
	# Steps: 
	#  * Delete empty lines or lines containig only whitespace.
	#  * Add '"' to beginning of each line of the input.
	#  * Add '"' and field specification to end of each line of the input.
	#    Whitespace at beginning or end of line will be deleted.
	#  * Add ' OR ' to end of each line except the last one.
	sed\
	    -e '/^\s*$/d' - | \
	sed\
	    -e 's/^\s*/"/' \
	    -e "s/\s*$/\"\[${PMC_FIELD}\]/" \
	    -e '$! s/$/ OR /'
	;;

    ovid_medline )
	# grep "^UI  - " - | sed -e 's/^UI  - //' -e 's/\r//g'
	# Fields to search for in Ovid MEDLINE
	# This is the part in between the periods in a search statement.
	case "$idtype" in
	    pmid )
		OVID_FIELD_LIST=ui
		;;
	    pmcid )
		OVID_FIELD_LIST=pm
		;;
	    doi )
		OVID_FIELD_LIST=do,id
		;;
	    pii )
	        OVID_FIELD_LIST=id
		;;
	    * )
	        # unknown argument
		echo "ERROR: Unknown idtype ${idtype}" >&2
		exit 1
		;;
	esac
	# Steps: 
	#  * Delete empty lines or lines containig only whitespace.
	#  * Add '"' to beginning of each line of the input.
	#  * Add '"' to end of each line of the input.
	#    Whitespace at beginning or end of line will be deleted.
	#  * Add ' OR ' to end of each line except the last one.
	#  * Add "(" to beginning of file
	#  * Add ")" and field specification to end of file
	sed\
	    -e '/^\s*$/d' - | \
	sed\
	    -e 's/^\s*/"/' \
	    -e 's/\s*$/"/' \
	    -e '$! s/$/ OR /' | \
	sed\
	    -e '1 i (' \
	    -e "\$ a ).${OVID_FIELD_LIST}."
	;;

    ovid_embase )
	# Fields to search for in Ovid Embase
	# This is the part in between the periods in a search statement.
	case "$idtype" in
	    an )
		OVID_FIELD_LIST=an
		;;
	    eu )
		OVID_FIELD_LIST=eu
		;;
	    pmid )
		OVID_FIELD_LIST=pm
		;;
	    doi )
		OVID_FIELD_LIST=do
		;;
	    * )
	        # unknown argument
		echo "ERROR: Unknown idtype ${idtype}" >&2
		exit 1
		;;
	esac
	# Steps: 
	#  * Delete empty lines or lines containig only whitespace.
	#  * Add '"' to beginning of each line of the input.
	#  * Add '"' to end of each line of the input.
	#    Whitespace at beginning or end of line will be deleted.
	#  * Add ' OR ' to end of each line except the last one.
	#  * Add "(" to beginning of file
	#  * Add ")" and field specification to end of file
	sed\
	    -e '/^\s*$/d' - | \
	sed\
	    -e 's/^\s*/"/' \
	    -e 's/\s*$/"/' \
	    -e '$! s/$/ OR /' | \
	sed\
	    -e '1 i (' \
	    -e "\$ a ).${OVID_FIELD_LIST}."
	;;

    cochrane_library )
	# Fields to search for in Cochrane Library
	case "$idtype" in
	    an )
		COCHRANE_FIELD_SPEC=":an"
		;;
	    doi )
		COCHRANE_FIELD_SPEC=":doi"
		;;
	    tn )
		COCHRANE_FIELD_SPEC=":tn"
		;;
	    none )
		# No field specification i.a. all fields
		COCHRANE_FIELD_SPEC=""
		;;
	    * )
	        # unknown argument
		echo "ERROR: Unknown idtype ${idtype}" >&2
		exit 1
		;;
	esac
	# Steps: 
	#  * Delete empty lines or lines containig only whitespace.
	#  * Add '(' to beginning of each line of the input.
	#  * Add ')' and field specification to end of each line of the input.
	#    Whitespace at beginning or end of line will be deleted.
	#  * Add ' OR ' to end of each line except the last one.
	#  * Join lines
	sed\
	    -e '/^\s*$/d' - | \
	sed\
	    -e 's/^\s*/(/' \
	    -e "s/\s*$/)${COCHRANE_FIELD_SPEC}/" \
	    -e '$! s/$/ OR /' | \
	tr "\n" " "
	;;

    web_of_science )
        # Fields to search for in Web of Science (Core Collection)
	case "$idtype" in
	    an )
		WOS_FIELD_SPEC="UT"
		;;
	    doi )
		WOS_FIELD_SPEC="DO"
		;;
	    pmid )
		WOS_FIELD_SPEC="PMID"
		;;
	    * )
	        # unknown argument
		echo "ERROR: Unknown idtype ${idtype}" >&2
		exit 1
		;;
	esac
	# Steps: 
	#  * Delete empty lines or lines containig only whitespace.
	#  * Add '"' to beginning of each line of the input.
	#  * Add '"' to end of each line of the input.
	#    Whitespace at beginning or end of line will be deleted.
	#  * Add ' OR ' to end of each line except the last one.
	#  * Add "<field specification>=(" to beginning of file
	#  * Add ")" to end of file
	sed\
	    -e '/^\s*$/d' - | \
	sed\
	    -e 's/^\s*/"/' \
	    -e 's/\s*$/"/' \
	    -e '$! s/$/ OR /' | \
	sed\
	    -e "1 i ${WOS_FIELD_SPEC}=(" \
	    -e '$ a )'
	;;

    citavi )
    	# Fields to search for in Citavi citation manager
	case "$idtype" in
	    doi )
		CITAVI_FIELD_SPEC="DOI"
		;;
	    pmcid )
		CITAVI_FIELD_SPEC="PMCID"
		;;
	    pmid )
		CITAVI_FIELD_SPEC="PubMedId"
		;;
	    * )
		# unknown argument
		echo "ERROR: Unknown idtype ${idtype}" >&2
		exit 1
		;;
	esac
	# Steps: 
	#  * Delete empty lines or lines containig only whitespace.
	#  * Add '(' and field specification to beginning of each line of the input.
	#  * Add ')' to end of each line of the input.
	#    Whitespace at beginning or end of line will be deleted.
	#  * Add ' OR ' to end of each line except the last one.
	#  * Join lines
	sed\
	    -e '/^\s*$/d' - | \
	sed\
    	    -e "s/^\s*/(${CITAVI_FIELD_SPEC}: /" \
	    -e 's/\s*$/)/' \
	    -e '$! s/$/ OR /' | \
	tr "\n" " "
	;;

    ebscohost_psycinfo )
	# Fields to search for in EBSCOhost PsycInfo
	case "$idtype" in
	    an )
		EBSCOHOST_FIELD=AN
		;;
	    pmid )
		EBSCOHOST_FIELD=PM
		;;
	    doi )
		EBSCOHOST_FIELD=DI
		;;
	    * )
	        # unknown argument
		echo "ERROR: Unknown idtype ${idtype}" >&2
		exit 1
		;;
	esac
	# Steps: 
	#  * Delete empty lines or lines containig only whitespace.
	#  * Add '"' and field specification to beginning of each line of the input.
	#  * Add '"' to end of each line of the input.
	#    Whitespace at beginning or end of line will be deleted.
	#  * Add ' OR ' to end of each line except the last one.
	sed\
	    -e '/^\s*$/d' - | \
	sed\
	    -e "s/^\s*/${EBSCOHOST_FIELD} \"/" \
	    -e 's/\s*$/"/' \
	    -e '$! s/$/ OR /'
	;;

    ebscohost_cinahl )
	# Fields to search for in EBSCOhost PsycInfo
	case "$idtype" in
	    an )
		# Steps: 
		#  * Delete empty lines or lines containig only whitespace.
		#  * Add '"' and field specification to beginning of each line of the input.
		#  * Add '"' to end of each line of the input.
		#    Whitespace at beginning or end of line will be deleted.
		#  * Add ' OR ' to end of each line except the last one.
		sed\
		    -e '/^\s*$/d' - | \
		sed\
		    -e "s/^\s*/AN \"/" \
		    -e 's/\s*$/"/' \
		    -e '$! s/$/ OR /'
		;;
	    pmid )
		# Steps: 
		#  * Delete empty lines or lines containig only whitespace.
		#  * Add '"', field specification and prefix to PMID to beginning of each line of the input.
		#  * Add '"' to end of each line of the input.
		#    Whitespace at beginning or end of line will be deleted.
		#  * Add ' OR ' to end of each line except the last one.
		sed\
		    -e '/^\s*$/d' - | \
		sed\
		    -e "s/^\s*/PM \"NLM/" \
		    -e 's/\s*$/"/' \
		    -e '$! s/$/ OR /'
		;;
	    * )
	        # unknown argument
		echo "ERROR: Unknown idtype ${idtype}" >&2
		exit 1
		;;
	esac
	;;

    dimensions )
	# Search in [Dimensions](https://app.dimensions.ai/) using DOI search
	case "$idtype" in
	    doi )
		# Nothing to do here
		;;
	    * )
	        # unknown argument
		echo "ERROR: Unknown idtype ${idtype}" >&2
		exit 1
		;;
	esac
	# Steps: 
	#  * Delete empty lines or lines containig only whitespace.
	#  * Add '"' to beginning of each line of the input.
	#  * Add '"' to end of each line of the input.
	#    Whitespace at beginning or end of line will be deleted.
	sed\
	    -e '/^\s*$/d' - | \
	sed\
	    -e 's/^\s*/"/' \
	    -e 's/\s*$/"/' \
	;;

    dimensions_api )
	# Fields to search for in [Dimensions Analytics API](https://app.dimensions.ai/)
        # using [Dimensions Search Language (DSL)](https://docs.dimensions.ai/dsl/index.html)
	# Beware: Not yet tested in API!
	case "$idtype" in
	    doi )
		DIMENSIONS_FIELD=doi
		;;
	    * )
	        # unknown argument
		echo "ERROR: Unknown idtype ${idtype}" >&2
		exit 1
		;;
	esac
	# Steps: 
	#  * Delete empty lines or lines containig only whitespace.
	#  * Add '"' to beginning of each line of the input.
	#  * Add '"' to end of each line of the input.
	#    Whitespace at beginning or end of line will be deleted.
	#  * Add ', ' to end of each line except the last one.
	#  * Add "search publications where ${DIMENSIONS_FIELD} in [" to beginning of file
	#  * Add ")" and field specification to end of file
	sed\
	    -e '/^\s*$/d' - | \
	sed\
	    -e 's/^\s*/"/' \
	    -e 's/\s*$/"/' \
	    -e '$! s/$/, /' | \
	sed\
	    -e "1 i search publications where ${DIMENSIONS_FIELD} in [" \
	    -e "\$ a ] return publications"
	;;

    * )	
	echo "ERROR: Unknown database syntax ${syntax}" >&2
	exit 1
	;;

esac