1 #!/usr/bin/perl -w
   2 #
   3 # $RCSfile: ModifyTextFilesFormat.pl,v $
   4 # $Date: 2008/01/30 21:44:49 $
   5 # $Revision: 1.20 $
   6 #
   7 # Author: Manish Sud <msud@san.rr.com>
   8 #
   9 # Copyright (C) 2004-2008 Manish Sud. All rights reserved.
  10 #
  11 # This file is part of MayaChemTools.
  12 #
  13 # MayaChemTools is free software; you can redistribute it and/or modify it under
  14 # the terms of the GNU Lesser General Public License as published by the Free
  15 # Software Foundation; either version 3 of the License, or (at your option) any
  16 # later version.
  17 #
  18 # MayaChemTools is distributed in the hope that it will be useful, but without
  19 # any warranty; without even the implied warranty of merchantability of fitness
  20 # for a particular purpose.  See the GNU Lesser General Public License for more
  21 # details.
  22 #
  23 # You should have received a copy of the GNU Lesser General Public License
  24 # along with MayaChemTools; if not, see <http://www.gnu.org/licenses/> or
  25 # write to the Free Software Foundation Inc., 59 Temple Place, Suite 330,
  26 # Boston, MA, 02111-1307, USA.
  27 #
  28 
  29 use 5.006;
  30 use strict;
  31 use FindBin; use lib "$FindBin::Bin/../lib";
  32 use Getopt::Long;
  33 use File::Basename;
  34 use Text::ParseWords;
  35 use Benchmark;
  36 use FileUtil;
  37 use TextUtil;
  38 
  39 my($ScriptName, %Options, $StartTime, $EndTime, $TotalTime);
  40 my($TextFile, @TextFilesList, $Index, $OutQuote, $InDelim, $OutDelim, $FileDir, $FileName, $FileExt, $NewTextFileName, $Line, @Words);
  41 
  42 # Autoflush STDOUT
  43 $| = 1;
  44 
  45 # Starting message...
  46 $ScriptName = basename $0;
  47 print "\n$ScriptName:Starting...\n\n";
  48 $StartTime = new Benchmark;
  49 
  50 # Get the options and setup script...
  51 SetupScriptUsage();
  52 if ($Options{help} || @ARGV < 1) {
  53   die GetUsageFromPod("$FindBin::Bin/$ScriptName");
  54 }
  55 
  56 @TextFilesList = ExpandFileNames(\@ARGV, "csv tsv");
  57 
  58 if (@TextFilesList > 1) {
  59   print "Processing Text files...\n";
  60 }
  61 FILELIST: for $Index (0 .. $#TextFilesList) {
  62   $TextFile = $TextFilesList[$Index];
  63   if (@TextFilesList > 1) {
  64     print "\nProcessing file $TextFile...\n";
  65   }
  66   else {
  67     print "Processing file $TextFile...\n"
  68   }
  69   if (!(-e $TextFile)) {
  70     warn "Warning: Ignoring file $TextFile: It doesn't exist\n";
  71     next FILELIST;
  72   }
  73   if (!CheckFileType($TextFile, "csv tsv")) {
  74     warn "Warning: Ignoring file $TextFile: It's not a text file\n";
  75     next FILELIST;
  76   }
  77   ($FileDir, $FileName, $FileExt) = ParseFileName($TextFile);
  78   if ($FileExt =~ /^tsv$/i) {
  79     $InDelim = "\t";
  80   }
  81   else {
  82     $InDelim = "\,";
  83     if ($Options{indelim} !~ /^(comma|semicolon)$/i) {
  84       warn "Warning: Ignoring file $TextFile: The value specified, $Options{indelim}, for option \"--indelim\" is not valid for csv files\n";
  85       next FILELIST;
  86     }
  87     if ($Options{indelim} =~ /^semicolon$/i) {
  88       $InDelim = "\;";
  89     }
  90   }
  91   if (lc($InDelim) eq lc($OutDelim)) {
  92     warn "Warning: Ignoring file $TextFile: The value specified, $Options{outdelim}, for option \"--outdelim\" is same as input delimiter\n";
  93     next FILELIST;
  94   }
  95   if (!open TEXTFILE, "$TextFile") {
  96     warn "Warning: Ignoring file $TextFile: Couldn't open it: $! \n";
  97     next FILELIST;
  98   }
  99   close TEXTFILE;
 100   $FileDir = ""; $FileName = ""; $FileExt = "";
 101   ($FileDir, $FileName, $FileExt) = ParseFileName($TextFile);
 102   $FileExt = "csv";
 103   if ($Options{outdelim} =~ /^tab$/i) {
 104     $FileExt = "tsv";
 105   }
 106   $NewTextFileName = $FileName;
 107   if ($Options{root} && (@TextFilesList == 1)) {
 108     my ($RootFileDir, $RootFileName, $RootFileExt) = ParseFileName($Options{root});
 109     if ($RootFileName && $RootFileExt) {
 110       $NewTextFileName = $RootFileName;
 111     }
 112     else {
 113       $NewTextFileName = $Options{root};
 114     }
 115     $NewTextFileName .= ".$FileExt";
 116   }
 117   else {
 118     $NewTextFileName .= "FormatModified" . ".$FileExt";
 119   }
 120   if (!$Options{overwrite}) {
 121     if (-e $NewTextFileName) {
 122       warn "Warning: Ignoring file $TextFile: New Text file, $NewTextFileName, already exists\n";
 123       next FILELIST;
 124     }
 125   }
 126   open NEWTEXTFILE, ">$NewTextFileName" or die "Error: Can't open $NewTextFileName !$ \n";
 127   print "Generating $NewTextFileName file\n";
 128   open TEXTFILE, "$TextFile" or die "Error: Can't open $TextFile: $! \n";
 129   while ($Line = GetTextLine(\*TEXTFILE)) {
 130     @Words = quotewords($InDelim, 0, $Line);
 131     $Line = JoinWords(\@Words, $OutDelim, $OutQuote);
 132     print NEWTEXTFILE "$Line\n";
 133   }
 134   close NEWTEXTFILE;
 135   close TEXTFILE;
 136 }
 137 
 138 print "$ScriptName:Done...\n\n";
 139 
 140 $EndTime = new Benchmark;
 141 $TotalTime = timediff ($EndTime, $StartTime);
 142 print "Total time: ", timestr($TotalTime), "\n";
 143 
 144 ###############################################################################
 145 
 146 # Setup script usage  and retrieve command line arguments specified using various options...
 147 sub SetupScriptUsage {
 148 
 149   # Retrieve all the options...
 150   %Options = ();
 151   $Options{indelim} = "comma";
 152   $Options{outdelim} = "tab";
 153   $Options{quote} = "yes";
 154   if (!GetOptions(\%Options, "help|h", "indelim=s", "outdelim=s", "overwrite|o", "quote|q=s", "root|r=s", "workingdir|w=s")) {
 155     die "\nTo get a list of valid options and their values, use \"$ScriptName -h\" or\n\"perl -S $ScriptName -h\" command and try again...\n";
 156   }
 157   if ($Options{workingdir}) {
 158     if (! -d $Options{workingdir}) {
 159       die "Error: The value specified, $Options{workingdir},  for option \"-w --workingdir\" is not a directory name.\n";
 160     }
 161     chdir $Options{workingdir} or die "Error: Couldn't chdir $Options{workingdir}: $! \n";
 162   }
 163   if ($Options{indelim} !~ /^(comma|semicolon)$/i) {
 164     die "Error: The value specified, $Options{indelim}, for option \"--indelim\" is not valid. Allowed values: comma or semicolon\n";
 165   }
 166   if ($Options{outdelim} !~ /^(comma|semicolon|tab)$/i) {
 167     die "Error: The value specified, $Options{outdelim}, for option \"--outdelim\" is not valid. Allowed values: comma, tab, or semicolon\n";
 168   }
 169   if ($Options{quote} !~ /^(yes|no)$/i) {
 170     die "Error: The value specified, $Options{quote}, for option \"-q --quote\" is not valid. Allowed values: yes or no\n";
 171   }
 172   $OutDelim = ($Options{outdelim} =~ /^tab$/i ) ? "\t" : (($Options{outdelim} =~ /^semicolon$/i) ? "\;" : "\,");
 173   $OutQuote = ($Options{quote} =~ /^yes$/i) ? 1 : 0;
 174 }
 175