1 package Maasha::Biopieces;
4 # Copyright (C) 2007-2009 Martin A. Hansen.
6 # This program is free software; you can redistribute it and/or
7 # modify it under the terms of the GNU General Public License
8 # as published by the Free Software Foundation; either version 2
9 # of the License, or (at your option) any later version.
11 # This program is distributed in the hope that it will be useful,
12 # but WITHOUT ANY WARRANTY; without even the implied warranty of
13 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
14 # GNU General Public License for more details.
16 # You should have received a copy of the GNU General Public License
17 # along with this program; if not, write to the Free Software
18 # Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
20 # http://www.gnu.org/copyleft/gpl.html
23 # >>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> DESCRIPTION <<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<
26 # Routines for manipulation, parsing and emitting of human/machine readable biopieces records.
29 # >>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>><<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<
32 use Getopt::Long qw( :config bundling );
37 use vars qw( @ISA @EXPORT_OK );
41 @ISA = qw( Exporter );
51 # >>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> SIGNAL HANDLER <<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<
54 $SIG{ '__DIE__' } = \&sig_handler;
55 $SIG{ 'INT' } = \&sig_handler;
56 $SIG{ 'TERM' } = \&sig_handler;
61 my ( $time_stamp, $script, $user, $pid, $file, $fh );
63 $time_stamp = Maasha::Common::time_stamp();
64 $user = Maasha::Common::get_user();
65 $script = Maasha::Common::get_scriptname();
66 $pid = Maasha::Common::get_processid();
68 $file = "$ENV{ 'BP_TMP' }/" . join( ".", $user, $script, $pid ) . ".status";
69 $fh = Maasha::Filesys::file_write_open( $file );
71 print $fh join( ";", $time_stamp, join( " ", @ARGV ) ) . "\n";
79 # Martin A. Hansen, June 2009.
81 # Retrieves initial status information written with status_set and uses this
82 # to write a status entry to the log file.
84 my ( $status, # status - OPTIONAL
89 my ( $time0, $time1, $script, $user, $pid, $file, $fh, $elap, $fh_global, $fh_local, $line, $args, $tmp_dir );
93 $time1 = Maasha::Common::time_stamp();
94 $user = Maasha::Common::get_user();
95 $script = Maasha::Common::get_scriptname();
96 $pid = Maasha::Common::get_processid();
98 $file = "$ENV{ 'BP_TMP' }/" . join( ".", $user, $script, $pid ) . ".status";
100 return if not -f $file;
102 $fh = Maasha::Filesys::file_read_open( $file );
109 ( $time0, $args, $tmp_dir ) = split /;/, $line;
111 Maasha::Filesys::dir_remove( $tmp_dir ) if defined $tmp_dir;
113 $elap = Maasha::Common::time_stamp_diff( $time0, $time1 );
115 $fh_global = Maasha::Filesys::file_append_open( "$ENV{ 'BP_LOG' }/biopieces.log" );
116 $fh_local = Maasha::Filesys::file_append_open( "$ENV{ 'HOME' }/.biopieces.log" );
118 print $fh_global join( "\t", $time0, $time1, $elap, $user, $status, "$script $args" ) . "\n";
119 print $fh_local join( "\t", $time0, $time1, $elap, $user, $status, "$script $args" ) . "\n";
121 $fh_global->autoflush( 1 );
122 $fh_local->autoflush( 1 );
131 # Martin A. Hansen, January 2008.
133 # Log messages to logfile.
137 my ( $time_stamp, $user, $script, $fh_global, $fh_local );
139 $time_stamp = Maasha::Common::time_stamp();
140 $user = Maasha::Common::get_user();
141 $script = Maasha::Common::get_scriptname();
143 $fh_global = Maasha::Filesys::file_append_open( "$ENV{ 'BP_LOG' }/biopieces.log" );
144 $fh_local = Maasha::Filesys::file_append_open( "$ENV{ 'HOME' }/.biopieces.log" );
146 print $fh_global "$time_stamp\t$user\t$script ", join( " ", @ARGV ), "\n";
147 print $fh_local "$time_stamp\t$user\t$script ", join( " ", @ARGV ), "\n";
149 $fh_global->autoflush( 1 );
150 $fh_local->autoflush( 1 );
159 # Martin A. Hansen, July 2007.
161 # Opens a stream to STDIN or a file,
163 my ( $file, # file - OPTIONAL
166 # Returns filehandle.
170 if ( not -t STDIN ) {
171 $fh = Maasha::Filesys::stdin_read();
172 } elsif ( not $file ) {
173 # Maasha::Common::error( qq(no data stream) );
175 $fh = Maasha::Filesys::file_read_open( $file );
184 # Martin A. Hansen, August 2007.
186 # Opens a stream to STDOUT or a file.
188 my ( $path, # path - OPTIONAL
189 $gzip, # compress data - OPTIONAL
192 # Returns filehandle.
197 $fh = Maasha::Filesys::file_write_open( $path, $gzip );
199 $fh = Maasha::Filesys::stdout_write();
208 # Martin A. Hansen, May 2009.
210 # Close stream if open.
212 my ( $fh, # filehandle
217 close $fh if defined $fh;
223 # Martin A. Hansen, July 2007.
225 # Reads one record at a time and converts that record
226 # to a Perl data structure (a hash) which is returned.
228 my ( $fh, # handle to stream
233 my ( $block, @lines, $line, $key, $value, %record );
235 return if not defined $fh;
237 local $/ = "\n---\n";
241 return if not defined $block;
245 @lines = split "\n", $block;
247 foreach $line ( @lines )
249 ( $key, $value ) = split ": ", $line, 2;
251 $record{ $key } = $value;
254 return wantarray ? %record : \%record;
260 # Martin A. Hansen, July 2007.
262 # Given a Perl datastructure (a hash ref) emits this to STDOUT or a filehandle.
264 my ( $data, # data structure
265 $fh, # file handle - OPTIONAL
270 if ( scalar keys %{ $data } )
274 map { print $fh "$_: $data->{ $_ }\n" } keys %{ $data };
279 map { print "$_: $data->{ $_ }\n" } keys %{ $data };
290 # Martin A. Hansen, May 2009
294 my ( $arg_list, # data structure with argument description
299 my ( $arg, @list, $options );
301 # ---- Adding the mandatory arguments to the arg_list ----
303 push @{ $arg_list }, (
305 { long => 'help', short => '?', type => 'flag', mandatory => 'no', default => undef, allowed => undef, disallowed => undef },
306 { long => 'stream_in', short => 'I', type => 'file!', mandatory => 'no', default => undef, allowed => undef, disallowed => undef },
307 { long => 'stream_out', short => 'O', type => 'file', mandatory => 'no', default => undef, allowed => undef, disallowed => undef },
308 { long => 'verbose', short => 'v', type => 'flag', mandatory => 'no', default => undef, allowed => undef, disallowed => undef },
311 check_duplicates_args( $arg_list );
313 # ---- Compiling options list ----
315 foreach $arg ( @{ $arg_list } )
317 if ( $arg->{ 'type' } eq 'flag' ) {
318 push @list, "$arg->{ 'long' }|$arg->{ 'short' }";
320 push @list, "$arg->{ 'long' }|$arg->{ 'short' }=s";
324 # ---- Parsing options from @ARGV ----
328 Getopt::Long::GetOptions( $options, @list );
330 # print Dumper( $options );
332 check_print_usage( $options );
334 # ---- Expanding and checking options ----
336 foreach $arg ( @{ $arg_list } )
338 check_mandatory( $arg, $options );
339 set_default( $arg, $options );
340 check_uint( $arg, $options );
341 check_int( $arg, $options );
342 set_list( $arg, $options );
343 check_dir( $arg, $options );
344 check_file( $arg, $options );
345 set_files( $arg, $options );
346 check_files( $arg, $options );
347 check_allowed( $arg, $options );
348 check_disallowed( $arg, $options );
351 # print Dumper( $options );
353 # return wantarray ? $options : %{ $options }; # WTF! Someone changed the behaviour of wantarray???
359 sub check_duplicates_args
361 # Martin A. Hansen, May 2009
363 # Check if there are duplicate long or short arguments,
364 # and raise an error if so.
366 my ( $arg_list, # List of argument hashrefs,
371 my ( $arg, %check_hash );
373 foreach $arg ( @{ $arg_list } )
375 Maasha::Common::error( qq(Duplicate long argument: $arg->{ 'long' }) ) if exists $check_hash{ $arg->{ 'long' } };
376 Maasha::Common::error( qq(Duplicate short argument: $arg->{ 'short' }) ) if exists $check_hash{ $arg->{ 'short' } };
378 $check_hash{ $arg->{ 'long' } } = 1;
379 $check_hash{ $arg->{ 'short' } } = 1;
384 sub check_print_usage
386 # Martin A. Hansen, May 2009.
388 # Check if we need to print usage and print usage
389 # and exit if that is the case.
391 my ( $options, # option hash
396 my ( $script, $wiki );
398 $script = Maasha::Common::get_scriptname();
400 if ( $script ne 'print_wiki' )
402 if ( exists $options->{ 'help' } or -t STDIN )
404 if ( not ( exists $options->{ 'stream_in' } or $options->{ 'data_in' } ) )
406 if ( scalar keys %{ $options } <= 1 )
408 $wiki = $ENV{ 'BP_DIR' } . "/bp_usage/$script.wiki";
410 if ( exists $options->{ 'help' } ) {
411 `print_wiki --data_in=$wiki --help`;
412 } elsif ( $script =~ /^(list_biopieces|list_genomes)$/ ) {
415 `print_wiki --data_in=$wiki`;
428 # Martin A. Hansen, May 2009.
430 # Check if mandatory arguments are set and raises an error if not.
433 $options, # options hash
438 if ( $arg->{ 'mandatory' } eq 'yes' and not defined $options->{ $arg->{ 'long' } } ) {
439 Maasha::Common::error( qq(Argument --$arg->{ 'long' } is mandatory) );
446 # Martin A. Hansen, May 2009.
448 # Set default values in option hash.
451 $options, # options hash
456 if ( not defined $options->{ $arg->{ 'long' } } ) {
457 $options->{ $arg->{ 'long' } } = $arg->{ 'default' }
464 # Martin A. Hansen, May 2009.
466 # Check if value to argument is an unsigned integer and
467 # raises an error if not.
470 $options, # options hash
475 if ( $arg->{ 'type' } eq 'uint' and defined $options->{ $arg->{ 'long' } } )
477 if ( $options->{ $arg->{ 'long' } } !~ /^\d+$/ ) {
478 Maasha::Common::error( qq(Argument --$arg->{ 'long' } must be an unsigned integer - not $options->{ $arg->{ 'long' } }) );
486 # Martin A. Hansen, May 2009.
488 # Check if value to argument is an integer and
489 # raises an error if not.
492 $options, # options hash
497 if ( $arg->{ 'type' } eq 'int' and defined $options{ $arg->{ 'long' } } )
499 if ( $options->{ $arg->{ 'long' } } !~ /^-?\d+$/ ) {
500 Maasha::Common::error( qq(Argument --$arg->{ 'long' } must be an integer - not $options->{ $arg->{ 'long' } }) );
508 # Martin A. Hansen, May 2009.
510 # Splits an argument of type 'list' into a list that is put
511 # in the options hash.
514 $options, # options hash
519 if ( $arg->{ 'type' } eq 'list' and defined $options->{ $arg->{ 'long' } } ) {
520 $options->{ $arg->{ 'long' } } = [ split /,/, $options->{ $arg->{ 'long' } } ];
527 # Martin A. Hansen, May 2009.
529 # Check if an argument of type 'dir!' truly is a directory and
530 # raises an error if not.
533 $options, # options hash
538 if ( $arg->{ 'type' } eq 'dir!' and defined $options->{ $arg->{ 'long' } } )
540 if ( not -d $options->{ $arg->{ 'long' } } ) {
541 Maasha::Common::error( qq(No such directory: "$options->{ $arg->{ 'long' } }") );
549 # Martin A. Hansen, May 2009.
551 # Check if an argument of type 'file!' truly is a file and
552 # raises an error if not.
555 $options, # options hash
560 if ( $arg->{ 'type' } eq 'file!' and defined $options->{ $arg->{ 'long' } } )
562 if ( not -f $options->{ $arg->{ 'long' } } ) {
563 Maasha::Common::error( qq(No such file: "$options->{ $arg->{ 'long' } }") );
571 # Martin A. Hansen, May 2009.
573 # Split the argument to 'files' into a list that is put on the options hash.
576 $options, # options hash
581 if ( $arg->{ 'type' } eq 'files' and defined $options->{ $arg->{ 'long' } } ) {
582 $options->{ $arg->{ 'long' } } = [ split /,/, $options->{ $arg->{ 'long' } } ];
589 # Martin A. Hansen, May 2009.
591 # Split the argument to 'files!' and check if each file do exists before adding
592 # the file list to the options hash.
595 $options, # options hash
600 my ( $elem, @files );
602 if ( $arg->{ 'type' } eq 'files!' and defined $options->{ $arg->{ 'long' } } )
604 foreach $elem ( split /,/, $options->{ $arg->{ 'long' } } )
608 } elsif ( $elem =~ /\*/ ) {
609 push @files, glob( $elem );
613 if ( scalar @files == 0 ) {
614 Maasha::Common::error( qq(Argument to --$arg->{ 'long' } must be a valid file or fileglob expression - not $options->{ $arg->{ 'long' } }) );
617 $options->{ $arg->{ 'long' } } = [ @files ];
624 # Martin A. Hansen, May 2009.
626 # Check if all values to all arguement are allowed and raise an
630 $options, # options hash
637 if ( defined $arg->{ 'allowed' } and defined $options->{ $arg->{ 'long' } } )
639 map { $val_hash{ $_ } = 1 } split /,/, $arg->{ 'allowed' };
641 if ( $arg->{ 'type' } =~ /^(list|files|files!)$/ )
643 foreach $elem ( @{ $options->{ $arg->{ 'long' } } } )
645 if ( not exists $val_hash{ $elem } ) {
646 Maasha::Common::error( qq(Argument to --$arg->{ 'long' } $elem is not allowed) );
652 if ( not exists $val_hash{ $options->{ $arg->{ 'long' } } } ) {
653 Maasha::Common::error( qq(Argument to --$arg->{ 'long' } $options->{ $arg->{ 'long' } } is not allowed) );
662 # Martin A. Hansen, May 2009.
664 # Check if any values to all arguemnts are disallowed and raise an error if so.
667 $options, # options hash
672 my ( $val, %val_hash );
674 if ( defined $arg->{ 'disallowed' } and defined $options->{ $arg->{ 'long' } } )
676 foreach $val ( split /,/, $arg->{ 'disallowed' } )
678 if ( $options->{ $arg->{ 'long' } } eq $val ) {
679 Maasha::Common::error( qq(Argument to --$arg->{ 'long' } $val is disallowed) );
688 # Martin A. Hansen, November 2007.
690 # Extracts files from an explicit GetOpt::Long argument
691 # allowing for the use of glob. E.g.
693 # --data_in=test.fna,test2.fna
695 # --data_in=test.fna,/dir/*.fna
697 my ( $option, # option from GetOpt::Long
702 my ( $elem, @files );
704 foreach $elem ( split ",", $option )
708 } elsif ( $elem =~ /\*/ ) {
709 push @files, glob( $elem );
713 return wantarray ? @files : \@files;
719 # Martin A. Hansen, April 2008.
721 # Removes temporary directory and exits gracefully.
722 # This subroutine is meant to be run always as the last
723 # thing even if a script is dies or is interrupted
726 my ( $sig, # signal from the %SIG
729 # print STDERR "signal->$sig<-\n";
731 my $script = Maasha::Common::get_scriptname();
737 if ( $sig =~ /MAASHA_ERROR/ )
739 print STDERR "\nProgram '$script' had an error" . " - Please wait for temporary data to be removed\n";
740 status_log( "ERROR" );
742 elsif ( $sig eq "INT" )
744 print STDERR "\nProgram '$script' interrupted (ctrl-c was pressed)" . " - Please wait for temporary data to be removed\n";
745 status_log( "INTERUPTED" );
747 elsif ( $sig eq "TERM" )
749 print STDERR "\nProgram '$script' terminated (someone used kill?)" . " - Please wait for temporary data to be removed\n";
750 status_log( "TERMINATED" );
754 print STDERR "\nProgram '$script' died->$sig" . " - Please wait for temporary data to be removed\n";
755 status_log( "DIED" );
766 # Martin A. Hansen, July 2008.
768 # Cleans out any unused temporary files and directories in BP_TMP.
772 my ( $tmpdir, @dirs, $curr_pid, $dir, $user, $sid, $pid );
774 $tmpdir = $ENV{ 'BP_TMP' } || Maasha::Common::error( 'No BP_TMP variable in environment.' );
776 $curr_pid = Maasha::Common::get_processid();
778 @dirs = Maasha::Filesys::ls_dirs( $tmpdir );
780 foreach $dir ( @dirs )
782 if ( $dir =~ /^$tmpdir\/(.+)_(\d+)_(\d+)_bp_tmp$/ )
788 # next if $user eq "maasha"; # DEBUG
790 if ( $user eq Maasha::Common::get_user() )
792 if ( not Maasha::Common::process_running( $pid ) )
794 # print STDERR "Removing stale dir: $dir\n";
795 Maasha::Filesys::dir_remove( $dir );
797 elsif ( $pid == $curr_pid )
799 # print STDERR "Removing current dir: $dir\n";
800 Maasha::Filesys::dir_remove( $dir );
810 # Martin A. Hansen, May 2009.
812 # Returns a precision timestamp for calculating
815 return Maasha::Common::get_time_hires();
821 # Martin A. Hansen, May 2009
823 # Print the run time to STDERR for the current script if
824 # the verbose switch is set in the option hash.
826 my ( $t0, # run time begin
828 $options, # options hash
833 my $script = Maasha::Common::get_scriptname();
835 print STDERR "Program: $script" . ( " " x ( 25 - length( $script ) ) ) . sprintf( "Run time: %.4f\n", ( $t1 - $t0 ) ) if $options->{ 'verbose' };
842 # Martin A. Hansen, April 2008.
844 # Create a temporary directory based on
845 # $ENV{ 'BP_TMP' } and sessionid. The directory
846 # name is written to the status file.
850 my ( $user, $sid, $pid, $script, $path, $file, $fh, $line );
852 Maasha::Common::error( qq(no BP_TMP set in %ENV) ) if not -d $ENV{ 'BP_TMP' };
854 $user = Maasha::Common::get_user();
855 $sid = Maasha::Common::get_sessionid();
856 $pid = Maasha::Common::get_processid();
857 $script = Maasha::Common::get_scriptname();
859 $path = "$ENV{ 'BP_TMP' }/" . join( "_", $user, $sid, $pid, "bp_tmp" );
860 $file = "$ENV{ 'BP_TMP' }/" . join( ".", $user, $script, $pid ) . ".status";
862 $fh = Maasha::Filesys::file_read_open( $file );
867 $fh = Maasha::Filesys::file_write_open( $file );
868 print $fh "$line;$path\n";
871 Maasha::Filesys::dir_create( $path );
883 # >>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>><<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<