#!/usr/bin/perl

# eap_search.pl v1.2
#
# modified version of bibsearch.pl (see below) for the Edinburgh
# Anisotropy Project at the British Geological Survey.
#
# last modification: 25 Jun 2002
#
# author: Peter Hanssen
# email : mail@Peter.Hanssen.name
# web   : http://Peter.Hanssen.name
#
# Use of environment variable DOCUMENT_ROOT to determine path to files
#
# Additional HTML-field: bibtex [0/1] - switch BibTeX output on/off
#     
# Additional HTML-output of the 'offline' location for:
#  - @inproceedings
#  - @article
#  - @inbook
#  - @phdthesis
#  - @techreport
#
# Bug removed for last field of item not delimited by comma.
#
# Additional hard-wired HTML-formating.
#
# Using DOCUMENT_ROOT to build directory path.
#
#==============================================================================
#
# See following for basic setup and instructions...

# ****
#
# bibsearch.pl v1.1b
#
# (c) Andy Wood ... 1994
#
# Using ReadParse from cgi-lib.pl - Copyright 1993 Steven E. Brenner  
#
# This script will respond to a call from a HTML form with the following
# fields:
#
#   'header' - url filename of the header template file.
#   'footer' - url filename of the footer template file.
#   'term'   - the search term(s).
#   'field'  - the field(s) to search in.
#   'type'   - one of 'exact', 'substr', or 'regexp'.
#   'files'  - list of url filenames to be searched seperated by `\0'
#
# It copies the specified header template file to stdout. This file should
# contain valid html and can be used to put a title on the search results.
# Any instances of "$term" will be replaced with the search term text, and
# any instances of "$type" with the search type text.
#
# Then it searches the list of files, using the search term, in the manner
# specified by the search type. The list of files should be a valid BibTeX
# database or one with HTML markup (such as those created with bibmarkup).
# Any entries in the database that match the search will be outputted with
# a HTML title.
#
# When it has scanned all the files it will then copy the specified footer
# template to stdout, again replacing "$term" and "$type".
#
# A url filename is of the form "~user/path/file" or "ftp:/path/file" and
# these only work on the local filesystem - see GetPath for more details.
#
# For an example of this script in use, examine:
#
#                  http://www.cs.bham.ac.uk/~amw/agents/bibtex/search.html
#
# Andy...
# __
#
# Andy Wood : amw@cs.bham.ac.uk               The University of Birmingham
#             tel: +44 (0)21 414 3736           School of Computer Science
#             fax: +44 (0)21 414 4281                Edgbaston, Birmingham
#             http://www.cs.bham.ac.uk/~amw               B15 2TT  England
#
# ****

# ****
#
# Modifications
#
# v1.1 - updated ReadParse to version 1.6
#      - separated out hard coded paths from the main code
#
# v1.1a - fixed bug in ReadParse version 1.6
#
# v1.1b - unfixed bug in ReadParse and fixed corresponding bug in own code ;0)
#
# ****

# ****
#
# Constants
#
# ****

# For use in GetPath
# USERHTMLDIR - name of directory for user supplied pages
# LOCALFTPPATH - local path for public ftp site
#
# Use of DOCUMENT_ROOT to switch between zweb-server and test-server path
  
   $USERHTMLDIR = "$ENV{'DOCUMENT_ROOT'}/PUBLICATIONS/SEARCH/";
   $LOCALFTPPATH = "$ENV{'DOCUMENT_ROOT'}/PUBLICATIONS/SEARCH/";

# ****
#
# ReadParse
#
# Reads in GET or POST data, converts it to unescaped text, and puts
# one key=value in each member of the list "@in"
# Also creates key/value pairs in %in, using '\0' to separate multiple
# selections
#
# If a variable-glob parameter (e.g., *cgi_input) is passed to ReadParse,
# information is stored there, rather than in $in, @in, and %in.
#
# From cgi-lib.pl - Copyright 1993/1994 Steven E. Brenner  
# http://www.bio.cam.ac.uk/web
#
# ****

sub ReadParse {
  local (*in) = @_ if @_;

  local ($i, $loc, $key, $val);

  # Read in text
  if ($ENV{'REQUEST_METHOD'} eq "GET") {
    $in = $ENV{'QUERY_STRING'};
  } elsif ($ENV{'REQUEST_METHOD'} eq "POST") {
    read(STDIN,$in,$ENV{'CONTENT_LENGTH'});
  }

  @in = split(/&/,$in);

  foreach $i (0 .. $#in) {
    # Convert plus's to spaces
    $in[$i] =~ s/\+/ /g;

    # Split into key and value.  
    ($key, $val) = split(/=/,$in[$i],2); # splits on the first =.

    # Convert %XX from hex numbers to alphanumeric
    $key =~ s/%(..)/pack("c",hex($1))/ge;
    $val =~ s/%(..)/pack("c",hex($1))/ge;

    # Associate key and value
    $in{$key} .= "\0" if (defined($in{$key})); # \0 is the multiple separator
    $in{$key} .= $val;

  }

  return 1; # just for fun
}


# ****
#
# GetPath
#
# Converts it's argument from a partial path ("~amw/file" or "ftp:/path/file") into
# it's full equivalent ("/home/pg/amw/public_html/file" or "/scratch/ftp/path/file")
# ensuring that we don't inadvertently allow external users full access to the file
# system. We also remove any instances of ".." in the path.
#
# ****

sub GetPath
  {
  local( $filename ) = $_[ 0 ];

  if ( $filename =~ /^~/ )
    {
    local( $name ) = $filename;
    $name =~ s/^~([^\/]*)\/(.*)/$1/;
    local( $file ) = $2;
    local( @entry ) = getpwnam( $name );
    
    $filename = $entry[ $#entry - 1 ].$USERHTMLDIR.$file;
    }
  elsif ( $filename =~ /^ftp:(.*)/ )
    {
    $filename = $LOCALFTPPATH.$1;
    }
  else
    {
#   $filename = ""; 
    $filename = $USERHTMLDIR.$filename;
    }

  $filename =~ s/\.\.//g;     # Make sure we don't allow any ..'ing
  return $filename;
  }


# ****
#
# PrintHeader
#
# Prints the line that tells WWW that we're an HTML document (honest!)
#
# ****

sub PrintHeader
  {
  print "Content-type: text/html\n\n";
  }


# ****
# 
# Search
#
# See if entry ($_) matches required search term. First argument is
# the search term (in this case $in{ 'term' }), the second is the
# search type ($in{ 'type' }), the third is the field ($in{ 'field' }).
#
# ****

sub Search
  {
  local( $found ) = 0;
  local( $searchin );

  if ( $_[2] ne "" && $_[2] ne "all" )
    {
    local( @fields ) = split( /[\s]+/, $_[2] );

    foreach $field ( 0..$#fields )
      {
      $searchin .= &GetField( $fields[$field], $_ );
      }
    }
  else
    {
    $searchin = $_;
    }

  if ( $_[1] eq 'regexp' )
    {
    $found = 1 if $searchin =~ /($_[0])/i;
    }
  else
    {
    local( @keyword ) = split( /[\s]+/, $_[0] );

    foreach $word ( 0..$#keyword )
      {
      if ( $_[1] eq 'exact' )
        {
        $found = 1 if $searchin =~ /\b($keyword[$word])\b/i;
        }
      elsif( $_[1] eq 'substr' )
        {
        $found = 1 if $searchin =~ /($keyword[$word])/i;
        }
      }
    }

  return $found;
  }


# ****
#
# MarkupEntry
#
# Create a line of HTML for each entry in the file. This pulls the
# title and the author (or editor) from the BibTeX entry in $_, and
# prints a HTML heading, followed by an availablity list of urls if
# there are any, followed by the full entry in <PRE>formatted form.
#
# ****

sub MarkupEntry
  {
  local( $junk, $bibtex ) = split( '@', $_, 2 );
  $bibtex = '@'.$bibtex;
  local( $key, $rest ) = split( ',', $bibtex, 2 );
  $key =~ s/^@.*[{(]\s*(.*)\s*$/$1/;                     # Retrieve the key

  local( $author, $title );
  $title = &GetField( "title", $bibtex );

  if ( /author\s*=/i )
    {
    $author = &GetField( "author", $bibtex );
    }
  elsif ( /editor\s*=/i )
    {
    $author = &GetField( "editor", $bibtex );
    if ( $author =~ /\band\b/ )
      {
      $author .= " (Eds)";
      }
    else
      {
      $author .= " (Ed)";
      }
    }
  else
    {
    $author = "";
    }
               
  print "\n<TABLE ALIGN=CENTER WIDTH=80%>\n";
  print "<TR BGCOLOR=\"#33CC33\"><TD COLSPAN=2><FONT COLOR=\"#FFFFFF\"><B>", $title, "</B></FONT><BR>\n";
  print "<I>", $author, "</I><BR>\n";

# Printing reference location for different media

  local( $bt, $jl, $uv, $yr, $ps, $vl, $pb, $nu );
  if ( /\@inproceedings/i )
    {
    $bt = &GetField( "booktitle", $bibtex );
    $yr = &GetField( "year", $bibtex );
    $ps = &GetField( "pages", $bibtex );
    print $bt, ", ", $yr, "</TD></TR>\n";
    }
  elsif ( /\@article/i )
    {
    $jl = &GetField( "journal", $bibtex );
    $yr = &GetField( "year", $bibtex );
    $vl = &GetField( "volume", $bibtex );
    $nu = &GetField( "number", $bibtex );
    $ps = &GetField( "pages", $bibtex );
    print $jl, ", ", $yr, ", <B>", $vl, "</B>, ", $nu, ", ", $ps, "</TD></TR>\n";
    }
  elsif ( /\@inbook/i )
    {
    $bt = &GetField( "booktitle", $bibtex );
    $yr = &GetField( "year", $bibtex );
    $ed = &GetField( "editor", $bibtex );
    $pb = &GetField( "publisher", $bibtex );
    $ps = &GetField( "pages", $bibtex );
    print $bt, ", ", $yr, ", ", $ed, "(Ed.), ", $pb, ", ", $ps, "</TD></TR>\n";
    }
  elsif ( /\@book/i )
    {
    $bt = &GetField( "booktitle", $bibtex );
    $yr = &GetField( "year", $bibtex );
    $pb = &GetField( "publisher", $bibtex );
    $ps = &GetField( "pages", $bibtex );
    print $bt, ", ", $yr, ", ", $pb, ", ", $ps, "</TD></TR>\n";
    }
  elsif ( /\@phdthesis/i )
    {
    $uv = &GetField( "university", $bibtex );
    $yr = &GetField( "year", $bibtex );
    print $uv, ", ", $yr, "</TD></TR>\n";
    }
  elsif ( /\@techreport/i )
    {
    $jl = &GetField( "journal", $bibtex );
    $yr = &GetField( "year", $bibtex );
    $vl = &GetField( "volume", $bibtex );
    $ps = &GetField( "pages", $bibtex );
    print $jl, ", ", $yr, ", ", $vl, ", ", $ps, "</TD></TR>\n";
    }
  else
    {
    print "<FONT COLOR=\"#FF0000\">There\'s something wrong with the database format!</FONT></TD></TR>\n";
    }

  if ( /url\s*=/i )
    {
    print "<TR><TD COLSPAN=2>Available as", &MarkupURL( &GetField( "url", $bibtex ) ), ".</TD></TR>\n";
    }

  s/<PRE>//i;
  s/<\/PRE>//i;
  if ( $in{'bibtex'} > 0 )                # ADDED FOR NON BIBTEX USERS
    {
    print "<TR><TD VALIGN=TOP><FONT SIZE=\"-1\">BibTeX:</FONT></TD>\n";
    print "<TD><FONT SIZE=\"-2\"><PRE>\n", $bibtex, "</PRE></FONT></TD></TR>\n";
    print "</TABLE>\n";
    }
  else
    {
    print "<TR><TD>&nbsp;</TD></TR></TABLE>\n";
    }
  
  }


# ****
#
# MarkupURL
#
# Takes a list of URLs seperated by commas and expands them into a html list
# that you can click on.
#
# ****

sub MarkupURL
  {
  local( $url, $format, $html ) = "";
  local( @urls ) = split( ',', $_[0] );

  foreach $url ( 0..$#urls )
    {
    $html .= "\nand" if ( $url == $#urls && $#urls >= 1 );
    $html .= "," if ( $url >= 1 && $url < $#urls );

    $format = "";

    if ( $urls[ $url ] =~ /\.Z/ ||  $urls[ $url ] =~ /\.gz/ )
      {
      $format = " compressed file";
      }

    if ( $urls[ $url ] =~ /\.ps/i )
      {
      $format .= " postscript file";
      }
    elsif ( $urls[ $url ] =~ /\.pdf/i )
      {
      $format .= " PDF";
      }
    elsif ( $urls[ $url ] =~ /\.txt/i )
      {
      $format .= " text only";
      }
    elsif ( $urls[ $url ] =~ /\.gif/i )
      {
      $format .= " gif image";
      }
    elsif ( $urls[ $url ] =~ /\.html/i )
      {
      $format .= " HTML file";
      }
    else
      {
      $format = $urls[ $url ];
      $format =~ s/.*\/([^\/]*)$/$1/;
      }
    
    $urls[ $url ] =~ s/^\s+//;

    if ( $urls[ $url ] =~ /\SPONSORS/i )
      {
      $html .= "<A CLASS=SearchLink HREF=\"".$urls[ $url ]."\">".$format."</A> for <FONT COLOR=\"#FF0000\">sponsors only<\/FONT>";
      }
    else
      {
      $html .= "<A CLASS=SearchLink HREF=\"".$urls[ $url ]."\">".$format."</A>";
      }

    }

  return $html;
  }
   

# ****
#
# GetField
#
# Gets the field specified in the first argument and strips it of quotes and/or
# squiggly brackets, removes excess spaces and returns it.
#
# ****

sub GetField
  {
  local( $field, $contents ) = @_;                   # Arguments: field name, bibtex entry

  $contents =~ s/\n/ /g;                             # Remove all \n's

  if ( $contents =~ /.*\b($field)\s*=\s*"([^"]*)"\s*,/i )
    {
    $contents = $2;
    }
# elsif ( $contents =~ /.*\b($field)\s*=\s*{(.*)}\s*,/i )
  elsif ( $contents =~ /.*\b($field)\s*=\s*{(.*)}\s*/i )   # CHANGED FOR LAST ITEM WITHOUT COMMA ENDING
    {
    $contents = $2;                                  # Contains remaining fields too
    $contents =~ s/}\s*,.*//g;                       # So remove everything after },
    }
  elsif ( $contents =~ /.*\b($field)\s*=\s*(\d*)\s*,/i )
    {
    $contents = $2;
    }
  else
    {
    $contents = "";
    }

  $contents =~ s/"|{|}//g;                           # Remove ""`s and {}'s
  $contents =~ s/\s+/ /g;                            # Make lots of spaces into 1.

  return $contents;
  }


# ****
#
# ProcessFile
#
# Process the file (filename specified by the first argument). This involves opening
# it, printing the last part of the filename, stripping off the html header (if it
# is not a .bib file), printing all the entries that match the search term (or "none
# found" if there aren't any) and closing the file again.
#
# ****

sub ProcessFile
  {
  local( $filename ) = &GetPath( $_[0] );

  open( FILE, $filename ) || print "<P><H2>Couldn't Open Input file - ".$filename."</H2>\n";

  if ( $filename =~ /_/ )  # Works with filenames like bibsearch_papers.bib otherwise old version
    {
    local( $filepart ) = substr( $filename, rindex( $filename, '_' ) + 1);
    $filepart =~ s/\.bib/$1/;
    print "<P><H3>Results for ".$filepart."</H3>\n";  # Prints just e.g. papers
    }
  else
    {
    local( $filepart ) = substr( $filename, rindex( $filename, '/' ) + 1);
    print "<P><H2>Results of ".$filepart."</H2>\n";  # Prints whole filename without path
    }

  local( $found ) = 0;

  if ( $filename =~ /\.bib$/ )
    {
    $stage='body';
    $/ = "";
    }
  else
    {
    $stage='header';
    }

  while ( <FILE> )
    {
    if ( $stage eq 'body' )
      {
      if ( /@.*{/ && &Search( $in{'term'}, $in{'type'}, $in{'field'} ) )
                                                              # If bibtex entry and matches search.
        {
        &MarkupEntry();
        $found = 1;
        }

      if ( /<!--BibTeX Ends-->/i )                            # If end of entries
        {
        $stage='footer';
        $/ = "\n";
        }
      }

    if ( ( $stage eq 'header' ) && /<!--BibTeX Starts-->/i )  # If end of header
      {
      $stage='body';
      $/ = "";
      }
    }

  print "<I>No matches found.</I><P>\n" if $found == 0;
  close( FILE );
  }


# ****
#
# PrintTemplate
#
# Copy a Template file substituting variables where necessary.
#
# ****

sub PrintTemplate
  {
  local( $filename ) = &GetPath( $_[0] );

  open( TEMPLATE, $filename ) || print "<P><H2>Incorrect or can't find Template File!\n", $_[0] , "</H2>\n";

  while ( <TEMPLATE> )
    {
    s/\$term/$in{'term'}/;
    s/\$type/$in{'type'}/;
    s/\$field/$in{'field'}/;
    print;
    }

  close( TEMPLATE );
  }


# ****
#
# Main part of the code
#
# ****

$* = 1;

&ReadParse();
&PrintHeader();
&PrintTemplate( $in{'header'} );
print "<!--BibTex Starts-->\n";

if ( $in{'files'} eq "" )
  {
  print "<P><H2>No Search Files Specified!</H2>\n";
  }
elsif ( $in{'term'} eq "" )
  {
  print "<P><H2>No Search Term Given!</H2>\n";
  }
else
  {
  local( @files ) = split( /\0/, $in{'files'} );

  foreach $file ( 0..$#files )
    {
    &ProcessFile( $files[$file] );
    }
  }

print "\n<!--BibTeX Ends-->\n";
&PrintTemplate( $in{'footer'} );

exit;