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

if ($#argv == 2) then
  set probe = "$2"
  if ("$1" != '-r') then
     echo "two arguments given but first is not -r, (it is $1)"
     exit
  endif
  set justreference = true
else
  set justreference = false
  set probe = "$1"
  if ("$1" == '-r') then
     echo "missing key for paper"
     exit
  endif
endif

if ($justreference == false) then
echo version = 1.25 of pullpaper 2010 Aug 10
endif
# 2011 Aug 10, 1.25: don't report tmp2 at the end for no reason ...
# 2011 Apr 13, 1.24: locate Freier1986
# 2011 Apr 13, 1.23: use regular grep
# 2011 Mar 11, 1.22: shut up the Klaxon!
# 2011 Mar 11, 1.21: upgrade for Mac OS X
# 2008 Jun 18, 1.20: location: paper -> papers
# 2007 Oct 24, 1.19: use yvp to get to the paper in pubmed if klaxon goes off
# 2006 Jul 05, 1.18: ignore capitalization in key by using 'grep -i'
# 2005 Nov 01, 1.17: give paper reference at end
# 2005 Sep 22, 1.16: could not find Krawczak.Reiss1992 ... missing from all.bib
# 2005 Sep 10, 1.15: handle arguments better when -r flag only.
# 2005 Sep  6, 1.14: no blanks when -r flag
# 2005 Sep  6, 1.13: -r flag
# 2005 Aug 31, 1.12: cleanup output
# 2005 Apr  1, 1.11: pullpaper SantaLucia1998 missing }
# 2004 Dec 10, 1.10: pullpaper Kornberg2000 crash on echo
# 2004 Sep 10, 1.09: pullpaper Harbison.Young2004 missed pmid?
# 2004 May 11, 1.08: pullpaper Zhang.Zhang1998 : C T should be removed
# 2004 Apr  9, 1.07: use first and last author
# 2004 Apr  9, 1.06: remove '-{}\' from author list
# 2004 Apr  9, 1.05: use first author if pmid is not there
# 2004 Apr  9, 1.04: set term = +++ form
# 2004 Apr  9, 1.03: fix usage message
# 2004 Apr  9, 1.02: sound
# 2004 Apr  8, 1.01: functional!!! extract numbers if fail to find pmid
# 2004 Apr  8, 1.00: origin 

set source = ~/papers/references/all.bib
if (-f $source) then
#   echo using $source
else
   set source = ~toms/papers/references/all.bib
endif

if ($justreference == false) then
  echo pull the paper whose key name is given
  echo by finding it in $source
endif

if ($#argv == 0) then
   echo 'usage: pullpaper [-r] [key]'
   echo "where 'key' is a key for a reference in"
   echo $source
   echo 'The key is used to find the Pubmed ID (pmid).'
   echo 'The pubmed id is then used to direct the browser'
   echo 'to the pubmed entry.'
   echo '-r flag: just give the bibtex reference and stop'
   echo 'Capitilaztion in the key is ignored to simplify typing.'
   exit
else
   if ($justreference == false) then
      echo using $source
   endif
endif

set key = $1

set from = 0
# to = 20 failed for Harbison.Young200
# and gave a Klaxon
set to = 20

set from = 1
set to = 60
set target = $source
set tmp1 = /tmp/`whoami`1.pullpaper
set tmp2 = /tmp/`whoami`2.pullpaper

grep -i -A $to -B $from $probe $target > $tmp1
set bot = `sed 's/^$/@@/' $tmp1|number|grep @@|head -2|tail -1|tr -d '@ '`
#  echo $bot

if ("$bot" == '') then
  echo $probe not found in $target
  exit
endif

head -$bot $tmp1 > $tmp2

if ($justreference == true) then

  cat $tmp2 |\
  grep -v '^$' |\
  cat

  exit
endif

echo Looking for entry $probe

set l="`cat $tmp2|fmti 700|grep pmid`"

echo -----
echo The reference is "$l":
echo "$l"
echo -----

if ("$l" == '') then
  echo "no pmid in $key"
  # extract pages, volume and year for another attempt
  grep -i -A $to -B $from $probe $target > $tmp1
echo -------- "$tmp1"
cat $tmp1
echo --------
#  fmti 800 $tmp1 | head -2 | tail -1 > $tmp2
  fmti 800 $tmp1 | head -2 | tail -1 | tr -d '{}' > $tmp2
echo -------- "$tmp2"
cat $tmp2
echo --------

  set aline = `cat "$tmp2"|sed -e "s/, /,@/g"|tr "@" "\012"|grep author`
echo aline is "$aline"
# extract the author list:
  set a1 = `echo "$aline"|tr '"' "\012"|head -2|tail -1`
                      # remove the periods:
  set a2 = `echo "$a1"|sed -e "s/\.//g"`
                      # change and to AND for pubmed to search:
  set a3 = `echo "$a2"|sed -e "s/ and / AND /g"`

  # make sure that dashes in author name are zapped:
  set a31 = `echo "$a3"|sed -e "s/-/ /g"`

                      # remove single characters:
  set a4 = `echo "$a31"|sed -e "s/ . / /g"`
                      # remove single characters from the first one:
  set a5 = `echo "$a4"|sed -e "s/^. / /g"`
                      # remove single characters AGAIN (not sure why):
                      # but it is needed for Schneider1986
  set a6 = `echo "$a5"|sed -e "s/ . / /g"`

                      # remove '-{}\' from author list
#  set a7 = `echo "$a6"|tr -d '{-}\'`
  set a7 = `echo "$a6"|sed -e 's/-/ /g'|sed -e 's/{/ /g'|sed -e 's/}/ /g'`
  set a8 = `echo "$a7"|sed -e 's,/, ,g'`

  set alist = "$a8"

while (0)
echo a6: $a6
echo a7: $a7
echo a8: $a8
echo alist: $alist
exit
end

# author = "T. D. Schneider and G. D. Stormo and L. Gold and A. Ehrenfeucht",
# Schneider AND Stormo AND Gold AND Ehrenfeucht
#  set author = "$a6"
# pubmed refuses to search with this.  Use only first and last authors.

echo alist: $alist

  set b1 = `echo "$alist"|sed -e "s/ AND /@/g"`
  set b2 = `echo "$b1"|tr '@' '\012'|head -1`
  set b3 = `echo "$b1"|tr '@' '\012'|tail -1`
# remove dashes in names as Zhang.Zhang1998 which contains:
# aline is author = "C.-T. Zhang and Z.-S. Lin and M. Yan and R. Zhang",

# echo b1: $b1
# echo b2: $b2
# echo b3: $b3

  set author = "$b2 AND $b3"
echo "$author"

  set vline = `cat $tmp2|sed -e "s/, /,@/g"|tr "@" "\012"|grep volume`
echo vline is "$vline"
  set volume = `echo "$vline"|tr '"' "\012"|head -2|tail -1`

  set pline = `cat $tmp2|sed -e "s/, /,@/g"|tr "@" "\012"|grep pages`
echo pline is "$pline"
  set pages = `echo "$pline"|tr '"' "\012"|head -2|tail -1`

  set yline = `cat $tmp2|sed -e "s/, /,@/g"|tr "@" "\012"|grep year`
  set year = `echo "$yline"|tr '"' "\012"|head -2|tail -1`

echo author: $author
echo volume: $volume
echo pages: $pages
echo year: $year

# the last page number is not in PubMed, remove it
set firstpage = `echo $pages|tr '-' '\012'|head -1`
echo firstpage: $firstpage
#  set s = "$volume&term=$firstpage&term=$year"
#  set s = "$author+AND+$volume+AND+$firstpage+AND+$year"
#  # sound klaxon &
#  echo "*********************************************"
#  echo "**** The Klaxon WARNING means:          *****"
#  echo "**** There is no Pubmed ID found for    *****"
#  echo "**** this reference.                    *****"
#  echo "**** (NOTE: is $to too small?)          *****"
#  echo "**** Using volume, first page and year: *****"
#  echo "*********************************************"
#  echo "search string s is $s"
#  echo "volume is $volume"
#  echo "year is $year"
#  echo "firstpage is $firstpage"
#exit # for now
#  pubmedsearch "$s"
  yvp $year $volume $firstpage
  exit
endif

# at this point the $l is a string for the ENTIRE entry
# Break it up and isolate the pmid

# echo "$l" |\
# sed -e "s/, /,@/g" |\
# tr "@" "\012" |\
# cat

set pmidline = `echo "$l"|sed -e "s/, /,@/g"|tr "@" "\012"|grep pmid`

echo "$pmidline"
echo -----

# convert the pmidline to pmid

set pmid = `echo "$pmidline"|tr '"' "\012"|head -2|tail -1`

# safety check
if ($pmid == '') then
  echo 'Failed to find a pmid\!'
  exit

else
  echo the pmid is $pmid for $key
  pmid $pmid $1
endif

echo ----------------------------------------------------------
exit
# the following might be useful for debugging, but not for general use!
date
echo -----
# echo "$l"
# echo '-----'
cat $tmp2
echo -----