EntrezDirect scripting - lmmx/devnotes GitHub Wiki

function cutf		(){ cut -d $'\t' -f "$@"; }

function striptoalpha	(){ for thisword in $(echo "$@" | tr -dc "[A-Z][a-z]\n" | tr [A-Z] [a-z]); do echo $thisword; done; }

function pubmed		(){ esearch -db pubmed -query "$@" | efetch -format docsum | xtract -pattern DocumentSummary -present Author -and Title -element Id -first "Author/Name" -element Title; }

function pubmeddocsum	(){ esearch -db pubmed -query "$@" | efetch -format docsum; } 

function pubmedextractdoi (){ pubmeddocsum "$@" | xtract -pattern DocumentSummary -element Id -first "Author/Name" -element Title SortPubDate -block ArticleId -match "IdType:doi" -element Value | awk '{split($0,a,"\t"); split(a[4],b,"/"); print a[1]"\t"a[2]"\t"a[3]"\t"a[5]"\t"b[1]}'; }

function pubmeddoi	(){ pubmedextractdoi "$@" | cutf 4; }

function pubmeddoimulti (){ 
	xtracted=$(pubmedextractdoi "$@")
	if [[ $(echo "$xtracted" | cutf 4) == '' ]]
	then
		xtractedpmid=$(echo "$xtracted" | cutf 1)
		pmid2doirestful "$xtractedpmid"
	else
		echo "$xtracted" | cutf 4
	fi
}

function pmid2doi	(){ curl -s www.pmid2doi.org/rest/json/doi/"$@" | awk '{split($0,a,",\"doi\":\"|\"}"); print a[2]}'; }
function pmid2doimulti 	(){
	curleddoi=$(pmid2doi "$@")
	if [[ $curleddoi == '' ]]
	then
		pmid2doincbi "$@"
	else
		echo "$curleddoi"
	fi
}
function pmid2doincbi 	(){
	xtracteddoi=$(pubmedextractdoi "$@")
	if [[ $xtracteddoi == '' ]]
	then
		echo "DOI NA"
	else
		echo "$xtracteddoi"
	fi
}

function AddPubTableDOIsSimple   () {
        old_IFS=$IFS
        IFS=$'\n'
        for line in $(cat "$@"); do
                AddPubDOI "$line"
        done
        IFS=$old_IFS
}

# Came across NCBI rate throttling while trying to call AddPubDOI in parallel, so added a second attempt for "DOI NA"
# and also writing STDOUT output to STDERR as this function will be used on a file (meaning STDOUT will get silenced)
# so you can see progress through the lines, as in:
#     AddPubTableDOIs table.tsv > outputfile.tsv
# I'd recommend it's not wise to overwrite unless you're using version control.

function AddPubTableDOIs   () {
	old_IFS=$IFS
	IFS=$'\n'
	for line in $(cat "$@"); do
		DOIresp=$(AddPubDOI "$line" 2>/dev/null)
		if [[ $DOIresp =~ 'DOI NA' ]]; then
			# try again in case it's just NCBI rate throttling, but just the once
			DOIresp2=$(AddPubDOI "$line" 2>/dev/null)
			if [[ $(echo "$DOIresp2" | awk 'BEGIN{FS="\t"};{print NF}' | uniq | wc -l) == '1' ]]; then
				echo "$DOIresp2"
				>&2 echo "$DOIresp"
			else
				DOIinput=$(echo "$line" | cutf 1-3)
				echo -e "$DOIinput\tDOI NA: Parse error"
				>&2 echo "$DOIinput\tDOI NA: Parse error"
			fi
		else
			if [[ $(echo "$DOIresp" | awk 'BEGIN{FS="\t"};{print NF}' | uniq | wc -l) == '1' ]]; then
				echo "$DOIresp"
				>&2 echo "$DOIresp"
			else
				DOIinput=$(echo "$line" | cutf 1-3)
				echo -e "$DOIinput\tDOI NA: Parse error"
				>&2 echo "$DOIinput\tDOI NA: Parse error"
			fi
		fi
	done
	IFS=$old_IFS
}

function AddPubDOI	(){
	if [[ $(echo "$@" | cutf 4) != '' ]]; then
		echo "$@"
		continue
	fi
	printf "$(echo "$@" | cutf 1-3)\t"
	thistitle=$(echo "$@" | cutf 3)
	if [[ $thistitle != 'Title' ]]; then
		thisauthor=$(echo "$@" | cutf 1)
		thisyear=$(echo "$@" | cutf 2)
		round1=$(pubmeddoimulti "$thistitle AND $thisauthor [AUTHOR]")
		round1hits=$(echo "$round1" | wc -l)
		if [[ "$round1hits" -gt '1' ]]; then
			round2=$(pubmeddoimulti "$thistitle AND $thisauthor [AUTHOR] AND ("$thisyear"[Date - Publication] : "$thisyear"[Date - Publication])")
			round2hits=$(echo "$round2" | wc -l)
			if [[ "$round2hits" -gt '1' ]]; then
				round3=$(
					xtracted=$(pubmedextractdoi "$@")
					xtractedtitles=$(echo "$xtracted" | cutf 3 | tr -dc "[A-Z][a-z]\n")
					alphatitles=$(striptoalpha "$xtractedtitles")
					thistitlealpha=$(striptoalpha "$thistitle")
					presearchIFS=$IFS
					IFS=$'\n'
					titlecounter="1"
					for searchtitle in $(echo "$alphatitles"); do
						(( titlecounter++ ))
						if [[ "$searchtitle" == *"$thistitlealpha"* ]]; then
							echo "$xtracted" | sed $titlecounter'q;d' | cutf 4
						fi
					done
					IFS=$presearchIFS
				)
				round3hits=$(echo "$round3" | wc -l)
				if [[ "$round3hits" -gt '1' ]]; then
					echo "ERROR multiple DOIs after 3 attempts to reduce - "$round3
				else
					echo $round3
				fi
			else
				echo $round2
			fi
		else
			echo $round1
		fi
	fi
}

function pmid2doirestful (){
	curleddoi=$(pmid2doi "$@")
	if [[ $curleddoi == '' ]]
	then
		echo "DOI NA"
	else
		echo "$curleddoi"
	fi
}

function mmrlit { cat ~/Dropbox/Y3/MMR/Essay/literature_table.tsv; }
function mmrlitedit { vim ~/Dropbox/Y3/MMR/Essay/literature_table.tsv; }
function mmrlitgrep     (){ grep -i "$@" ~/Dropbox/Y3/MMR/Essay/literature_table_with_DOIs.tsv; }
function mmrlitdoi      (){ mmrlitgrep "$@" | cut -d $'\t' -f 4 | tr -d '\n' | xclip -sel p; clipconfirm;  }
function mmrlitdoicite  (){ mmrlitgrep "$@" | cut -d $'\t' -f 4 | awk '{print "`r citet(\""$0"\")`"}' | tr -d '\n' | xclip -sel p; clipconfirm; }