#!/bin/tcsh -f
#(ie run the tshell on this but don't read the .cshrc or .tcshrc)

set ver     = "version = 1.06 of pmid2pmcid 2009 Dec 19"
#           2009 Dec 19, 1.06: fix varible name to be ver instead of version
#           2008 Jul 25, 1.05: switch to Monica Romiti's method
#           2008 Jul 24, 1.04: set agent
#           2008 Jul 24, 1.03: handle error
#           2008 Mar 25, 1.02: official documentation
#           2008 Mar 25, 1.01: turn on verbosemode flag
#           2008 Mar 25, 1.00: origin 

if ($#argv == 2) then
   set verbosemode = 1
else
   set verbosemode = 0
endif

# swap these lines to force verbose mode:
#set verbosemode = 1
#set verbosemode = 0

if ($#argv == 1) then
   if ($verbosemode) then
      echo "$ver"
   endif
endif

if ($#argv == 0) then
   echo "$ver"
   echo 'usage: pmid2pmcid [PMID] [turn-on-verbose]'
   echo 'Convert the PubMed ID (PMID) number of a paper from PubMed'
   echo 'to a PubMed Central ID (PMCID) number.'
   echo 'The program uses wget to get web pages.'
   echo 'The last line output is the PMCID.'
   echo ''
   echo 'If there is any second argument, the program runs in'
   echo 'verbose mode, showing results.  Use this for testing.'
   echo ''
   echo 'If there is no PubMed Central paper and no PMCID is found,'
   echo 'A blank line is returned.'

   echo ''
   echo 'Pubmed:'
   echo 'http://www.ncbi.nlm.nih.gov/sites/entrez'
   echo ''
   echo 'Pubmed Central:'
   echo 'http://www.pubmedcentral.nih.gov/'
   echo ''
   echo 'wget:'
   echo 'http://www.ccrnp.ncifcrf.gov/~toms/wget.html'

   echo ''
   echo 'PMCID is required by NIH:'
   echo 'Policy No. 118 3/10/08 Public Access Policy'
   echo 'http://web.ncifcrf.gov/campus/administrative/policies/100admin/118.asp'

   echo ''
   echo 'Examples:'
   echo ''
   echo '=== no argument ==================================================='
   echo 'pmid2pmcid'
   echo '=== argument for existing PMID ===================================='
   echo 'pmid2pmcid 15130839'
   echo '=== argument for existing PMID with verbose ======================='
   echo 'pmid2pmcid 15130839 verbose'
   echo '=== argument for non-existing PMID ================================'
   echo 'pmid2pmcid 17377584'
   echo '==================================================================='

   echo ''
   echo 'See also: bibtex2pmcid'

   echo ''
   echo '---'
   echo ''
   echo 'Dr. Thomas D. Schneider'
   echo 'National Institutes of Health'
   echo 'National Cancer Institute'
   echo 'Center for Cancer Research Nanobiology Program'
   echo 'Molecular Information Theory Group'
   echo 'Frederick, Maryland  21702-1201'
   echo 'toms@ncifcrf.gov'
   echo 'permanent email: toms@alum.mit.edu'
   echo 'http://www.ccrnp.ncifcrf.gov/~toms/'
   exit
endif

# 2008 Jul 25 New method from Monica Romiti:
# wget -O - -q "http://eutils.ncbi.nlm.nih.gov/entrez/eutils/efetch.fcgi?db=pubmed&r
# etmode=txt&rettype=medline&id=17921503" | grep "PMC - " 
# # PMC - PMC2189734 
# echo ----
# wget -O - -q "http://eutils.ncbi.nlm.nih.gov/entrez/eutils/efetch.fcgi?db=pubmed&r
# etmode=txt&rettype=medline&id=1792" | grep "PMC - " 

set PMID = $1

set source = "http://eutils.ncbi.nlm.nih.gov/entrez/eutils/efetch.fcgi?db=pubmed&retmode=txt&rettype=medline&id=${PMID}"

wget -O - -q "$source" |\
grep "PMC - "  |\
sed "s/PMC - //"

exit
********************************************************************************

# Old method:

# set agent = ""
# set tmp1 = /tmp/`whoami`1.pmid2pmc
# set tmp2 = /tmp/`whoami`2.pmid2pmc

set source = "http://www.pubmedcentral.nih.gov/articlerender.fcgi?tool=pubmed&pubmedid=${PMID}"

if ($verbosemode) then
   # give the wget message
   echo wget $agent -O "$tmp1" "$source"
   wget $agent -O "$tmp1" "$source"
else
   # silence the wget message
   wget $agent -O "$tmp1" "$source" >& $tmp1
endif

# head $tmp1
# set PMCID = `cat $tmp1 | grep 'PMCID:'`
# echo "$PMCID"

#cat $tmp1 | grep 'PMCID:' > $tmp2
#cat $tmp2

set g = ""

cat $tmp1 |\
grep 'PMCID:' |\
sed "s,<div ,${g},g" |\
sed "s,</div>,${g},g" |\
tr "$g" "\n" |\
grep 'PMCID:' |\
cat > $tmp2

if ($verbosemode) then
   echo ----- isolated code for the PMCID:
   cat   $tmp2
   echo ----- the final PMCID:
endif

set PMCID = `cat $tmp2|tr '>' '\n'|sed 's/PMC//' |tail -1`

# handle error
if ("$PMCID" == "") then
   # report the error
   grep ERROR $tmp1 | head -1
else
   # give the PMCID
   echo $PMCID
endif



exit
********************************************************************************

Example:

http://www.ncbi.nlm.nih.gov/pubmed/15130839?ordinalpos=1&itool=EntrezSystem2.PEntrez.Pubmed.Pubmed_ResultsPanel.Pubmed_RVDocSum

points to

http://www.pubmedcentral.nih.gov/articlerender.fcgi?tool=pubmed&pubmedid=15130839

that page gives

PMCID: PMC1852464

which is the desired number.

The line on the page is:

<div class="fm-citation-pmcid"><span class="fm-citation-ids-label">PMCID: </span>PMC1852464</div><div class="fm-citation-manuscriptid"><span class="fm-citation-ids-label">NIHMSID: </span><span>NIHMS8281</span></div>

strawberry 7% grep PMCID toms.pmid2pmc | wc
       1     338   11028
strawberry 8% wc toms.pmid2pmc
       4    7861  116441 toms.pmid2pmc

So since they are using really long lines (UGH!!) this cut it down
only 4 fold.

# for possible use later with wget:
# --user-agent=netscape
set agent = ""
set agent = "--user-agent=netscape"

# Error message:
# % pmid2pmcid 17921503
# 
# --16:32:55--  http://www.pubmedcentral.nih.gov/articlerender.fcgi?tool=pubmed&pu
# bmedid=17921503
#            => `/tmp/toms1.pmid2pmc'
# Resolving www.pubmedcentral.nih.gov... 130.14.29.110
# Connecting to www.pubmedcentral.nih.gov[130.14.29.110]:80... connected.
# HTTP request sent, awaiting response... 503 Service Unavailable
# 16:32:55 ERROR 503: Service Unavailable.