|
0
|
1 #!/usr/bin/perl -w
|
|
|
2 #
|
|
|
3 # $RCSfile: JoinSDFiles.pl,v $
|
|
|
4 # $Date: 2015/02/28 20:46:20 $
|
|
|
5 # $Revision: 1.35 $
|
|
|
6 #
|
|
|
7 # Author: Manish Sud <msud@san.rr.com>
|
|
|
8 #
|
|
|
9 # Copyright (C) 2015 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 strict;
|
|
|
30 use FindBin; use lib "$FindBin::Bin/../lib";
|
|
|
31 use Getopt::Long;
|
|
|
32 use File::Basename;
|
|
|
33 use Benchmark;
|
|
|
34 use SDFileUtil;
|
|
|
35 use FileUtil;
|
|
|
36
|
|
|
37 my($ScriptName, %Options, $StartTime, $EndTime, $TotalTime);
|
|
|
38
|
|
|
39 # Autoflush STDOUT
|
|
|
40 $| = 1;
|
|
|
41
|
|
|
42 # Starting message...
|
|
|
43 $ScriptName = basename $0;
|
|
|
44 print "\n$ScriptName:Starting...\n\n";
|
|
|
45 $StartTime = new Benchmark;
|
|
|
46
|
|
|
47 # Get the options and setup script...
|
|
|
48 SetupScriptUsage();
|
|
|
49 if ($Options{help} || @ARGV < 1) {
|
|
|
50 die GetUsageFromPod("$FindBin::Bin/$ScriptName");
|
|
|
51 }
|
|
|
52
|
|
|
53 my(@SDFilesList);
|
|
|
54 @SDFilesList = ExpandFileNames(\@ARGV, "sdf sd");
|
|
|
55 if (@SDFilesList == 1) {
|
|
|
56 die "Error: Specify more than one SD file.\n";
|
|
|
57 }
|
|
|
58
|
|
|
59 # Process options...
|
|
|
60 print "Processing options...\n";
|
|
|
61 my(%OptionsInfo);
|
|
|
62 ProcessOptions();
|
|
|
63
|
|
|
64 # Setup information about input files...
|
|
|
65 print "Checking input SD files...\n";
|
|
|
66 my(%SDFilesInfo);
|
|
|
67 RetrieveSDFilesInfo();
|
|
|
68
|
|
|
69 # Join files...
|
|
|
70 print "\nGenerating new SD file $OptionsInfo{NewSDFile}...\n";
|
|
|
71 JoinSDFiles();
|
|
|
72
|
|
|
73 print "\n$ScriptName:Done...\n\n";
|
|
|
74
|
|
|
75 $EndTime = new Benchmark;
|
|
|
76 $TotalTime = timediff ($EndTime, $StartTime);
|
|
|
77 print "Total time: ", timestr($TotalTime), "\n";
|
|
|
78
|
|
|
79 ###############################################################################
|
|
|
80
|
|
|
81 # Join all valid SD files...
|
|
|
82 sub JoinSDFiles {
|
|
|
83 my($FileIndex, $SDFile, $NewSDFile);
|
|
|
84
|
|
|
85 $NewSDFile = $OptionsInfo{NewSDFile};
|
|
|
86
|
|
|
87 open NEWSDFILE, ">$NewSDFile" or die "Error: Couldn't open $NewSDFile: $! \n";
|
|
|
88 FILELIST: for $FileIndex (0 .. $#SDFilesList) {
|
|
|
89 if (!$SDFilesInfo{FileOkay}[$FileIndex]) {
|
|
|
90 next FILELIST;
|
|
|
91 }
|
|
|
92 $SDFile = $SDFilesList[$FileIndex];
|
|
|
93 print "\nProcessing file $SDFile...\n";
|
|
|
94
|
|
|
95 open SDFILE, "$SDFile" or die "Error: Couldn't open $SDFile: $! \n";
|
|
|
96 while (<SDFILE>) {
|
|
|
97 s/(\r\n)|(\r)/\n/g;
|
|
|
98 print NEWSDFILE;
|
|
|
99 }
|
|
|
100 close SDFILE;
|
|
|
101 }
|
|
|
102
|
|
|
103 close NEWSDFILE;
|
|
|
104 }
|
|
|
105
|
|
|
106 # Retrieve information about SD files...
|
|
|
107 sub RetrieveSDFilesInfo {
|
|
|
108 my($Index, $SDFile);
|
|
|
109
|
|
|
110 %SDFilesInfo = ();
|
|
|
111 @{$SDFilesInfo{FileOkay}} = ();
|
|
|
112
|
|
|
113 FILELIST: for $Index (0 .. $#SDFilesList) {
|
|
|
114 $SDFilesInfo{FileOkay}[$Index] = 0;
|
|
|
115
|
|
|
116 $SDFile = $SDFilesList[$Index];
|
|
|
117 if (!(-e $SDFile)) {
|
|
|
118 warn "Warning: Ignoring file $SDFile: It doesn't exist\n";
|
|
|
119 next FILELIST;
|
|
|
120 }
|
|
|
121 if (!CheckFileType($SDFile, "sdf sd")) {
|
|
|
122 warn "Warning: Ignoring file $SDFile: It's not a SD file\n";
|
|
|
123 next FILELIST;
|
|
|
124 }
|
|
|
125 if (! open SDFILE, "$SDFile") {
|
|
|
126 warn "Warning: Ignoring file $SDFile: Couldn't open it: $! \n";
|
|
|
127 next FILELIST;
|
|
|
128 }
|
|
|
129 close SDFILE;
|
|
|
130
|
|
|
131 $SDFilesInfo{FileOkay}[$Index] = 1;
|
|
|
132 }
|
|
|
133 }
|
|
|
134
|
|
|
135 # Process option values...
|
|
|
136 sub ProcessOptions {
|
|
|
137 my($FileDir, $FileName, $FileExt, $NewSDFile);
|
|
|
138
|
|
|
139 %OptionsInfo = ();
|
|
|
140
|
|
|
141 $OptionsInfo{OutFileRoot} = $Options{root} ? $Options{root} : undef;
|
|
|
142 $OptionsInfo{Overwrite} = $Options{overwrite} ? $Options{overwrite} : undef;
|
|
|
143
|
|
|
144 if ($Options{root}) {
|
|
|
145 $FileDir = ""; $FileName = ""; $FileExt = "";
|
|
|
146 ($FileDir, $FileName, $FileExt) = ParseFileName($Options{root});
|
|
|
147 if ($FileName && $FileExt) {
|
|
|
148 $NewSDFile = $FileName . "." . $FileExt;
|
|
|
149 }
|
|
|
150 else {
|
|
|
151 $NewSDFile = $Options{root} . ".sdf";
|
|
|
152 }
|
|
|
153 }
|
|
|
154 else {
|
|
|
155 $FileDir = ""; $FileName = ""; $FileExt = "";
|
|
|
156 ($FileDir, $FileName, $FileExt) = ParseFileName($SDFilesList[0]);
|
|
|
157 $NewSDFile = $FileName . "1To" . @SDFilesList . "Joined.sdf";
|
|
|
158 }
|
|
|
159
|
|
|
160 if (!$Options{overwrite}) {
|
|
|
161 if (-e $NewSDFile) {
|
|
|
162 die "Error: The file $NewSDFile already exists.\n";
|
|
|
163 }
|
|
|
164 }
|
|
|
165 if ($Options{root}) {
|
|
|
166 my($FileIndex);
|
|
|
167 for $FileIndex (0 .. $#SDFilesList) {
|
|
|
168 if (lc($NewSDFile) eq lc($SDFilesList[$FileIndex])) {
|
|
|
169 die "Error: Output filename, $NewSDFile, is similar to a input file name.\nSpecify a different name using \"-r --root\" option or use default name.\n";
|
|
|
170 }
|
|
|
171 }
|
|
|
172 }
|
|
|
173 $OptionsInfo{NewSDFile} = $NewSDFile;
|
|
|
174
|
|
|
175 }
|
|
|
176
|
|
|
177 # Setup script usage and retrieve command line arguments specified using various options...
|
|
|
178 sub SetupScriptUsage {
|
|
|
179
|
|
|
180 # Retrieve all the options...
|
|
|
181 %Options = ();
|
|
|
182 if (!GetOptions(\%Options, "help|h", "overwrite|o", "root|r=s", "workingdir|w=s")) {
|
|
|
183 die "\nTo get a list of valid options and their values, use \"$ScriptName -h\" or\n\"perl -S $ScriptName -h\" command and try again...\n";
|
|
|
184 }
|
|
|
185 if ($Options{workingdir}) {
|
|
|
186 if (! -d $Options{workingdir}) {
|
|
|
187 die "Error: The value specified, $Options{workingdir}, for option \"-w --workingdir\" is not a directory name.\n";
|
|
|
188 }
|
|
|
189 chdir $Options{workingdir} or die "Error: Couldn't chdir $Options{workingdir}: $! \n";
|
|
|
190 }
|
|
|
191 }
|
|
|
192
|
|
|
193 __END__
|
|
|
194
|
|
|
195 =head1 NAME
|
|
|
196
|
|
|
197 JoinSDFiles.pl - Join multiple SDFiles into a single SDFile
|
|
|
198
|
|
|
199 =head1 SYNOPSIS
|
|
|
200
|
|
|
201 JoinSDFiles.pl SDFiles...
|
|
|
202
|
|
|
203 JoinSDFiles.pl [B<-h, --help>] [B<-o, --overwrite>] [B<-r, --root> rootname]
|
|
|
204 [B<-w, --workingdir> dirname] SDFiles...
|
|
|
205
|
|
|
206 =head1 DESCRIPTION
|
|
|
207
|
|
|
208 Multiple I<SDFiles> are joined to generate a single SDFile. The file names
|
|
|
209 are separated by spaces. The valid file extensions are I<.sdf> and I<.sd>.
|
|
|
210 All other file names are ignored. All the SD files in a current directory can be
|
|
|
211 specified either by I<*.sdf> or the current directory name.
|
|
|
212
|
|
|
213 =head1 OPTIONS
|
|
|
214
|
|
|
215 =over 4
|
|
|
216
|
|
|
217 =item B<-h, --help>
|
|
|
218
|
|
|
219 Print this help message.
|
|
|
220
|
|
|
221 =item B<-o, --overwrite>
|
|
|
222
|
|
|
223 Overwrite existing files.
|
|
|
224
|
|
|
225 =item B<-r, --root> I<rootname>
|
|
|
226
|
|
|
227 New SD file name is generated using the root: <Root>.sdf. Default file
|
|
|
228 name:<FirstSDFileName>1To<Count>Joined.sdf.
|
|
|
229
|
|
|
230 =item B<-w, --workingdir> I<dirname>
|
|
|
231
|
|
|
232 Location of working directory. Default: current directory.
|
|
|
233
|
|
|
234 =back
|
|
|
235
|
|
|
236 =head1 EXAMPLES
|
|
|
237
|
|
|
238 To join SD files, type:
|
|
|
239
|
|
|
240 % JoinSDFiles.pl -o Sample1.sdf Sample2.sdf
|
|
|
241 % JoinSDFiles.pl -o *.sdf
|
|
|
242
|
|
|
243 To join all Sample*.sdf files in a directory, SomeDir, and generate a new file NewSample.sdf, type:
|
|
|
244
|
|
|
245 % JoinSDFiles.pl -r NewSample -w SomeDir -o *.sdf
|
|
|
246
|
|
|
247 =head1 AUTHOR
|
|
|
248
|
|
|
249 Manish Sud <msud@san.rr.com>
|
|
|
250
|
|
|
251 =head1 SEE ALSO
|
|
|
252
|
|
|
253 InfoSDFiles.pl, MolFilesToSD.pl, SDToMolFiles.pl, SplitSDFiles.pl
|
|
|
254
|
|
|
255 =head1 COPYRIGHT
|
|
|
256
|
|
|
257 Copyright (C) 2015 Manish Sud. All rights reserved.
|
|
|
258
|
|
|
259 This file is part of MayaChemTools.
|
|
|
260
|
|
|
261 MayaChemTools is free software; you can redistribute it and/or modify it under
|
|
|
262 the terms of the GNU Lesser General Public License as published by the Free
|
|
|
263 Software Foundation; either version 3 of the License, or (at your option)
|
|
|
264 any later version.
|
|
|
265
|
|
|
266 =cut
|
