Dpuiu Perl

Links

• [1]
• SysAdmin Perl
• Subroutines
• Programming Perl Book
• Functions
• Perl Module Mechanics
• [2]

Style

• Style
• perltidy

POD

• POD; pod2html pod2latex pod2man pod2text pod2usage
• make sure to leave 1 empty line before/after cut

Example:

 $ cat test.pl
 =pod

 =head1 sub max(number,number)
 Returns max of 2 numbers

 =cut

 sub max
 {
       my ($a,$b)=@_;
       return ($a>$b)?$a:$b;
 }

 =head1 sub min(number,number)
 Returns min of 2 numbers

 =cut

 sub min
 {
       my ($a,$b)=@_;
       return ($a<$b)?$a:$b;
 }

 print max(1,2),"\n";
 print min(1,2),"\n";

 $ perdoc test.pl
 sub max(number,number) Returns max of 2 numbers
 sub min(number,number) Returns min of 2 numbers

Functions

Map

 sub addString { return @_[0].$_[1]; }
 @myNames = ('jacob', 'alexander', 'ethan', 'andrew');
 @ucNames = map(addString($_,"."), @myNames);

 @a = (1,3,5);

In place editing

• Edit a file in place: -i

 perl -i.bck script.pl xxx
 => xxx(new), xxx.bck(old)

• dos2unix

 perl -i -pe 's/\r//g' file

• unix2dos

 perl -i -pe 's/\n/\r\n/' file

Cmd line options

• BEGIN/END blocks

 perl -ane 'BEGIN { print "sum\n"}  END { print $sum,"\n" } $sum += $F[0]; '

• Check all files

 ls *pl | perl -pe 'print "perl -c "; '

Regular Expressions

• Maximal match (default)

 echo $PERLLIB | perl -ane '/^(.+):/; print $1,"\n"'

• Minimal match

 echo $PERLLIB | perl -ane '/^(.+?):/; print $1,"\n"'
 /fs/sz-user-supported/Linux-x86_64/lib

• Translation:

 $_ =~ tr/abc/ABC/

• All matches

 while($text=~m/($pattern)/g) { print "$` $& $'\n"; }

Search directories

  perl -Idir_name  file.pl
  <=>
 cat file.pl
   use lib dir_name ;

 default: /usr/include & /usr/lib/perl
 add dirs to PERLLIB env variable <=>  @INC 
 perl -e '$,="\n"; print  @INC ; '

• Find installed modules

 perl -e 'foreach(@INC) { if(-d $_) {system "find $_ -type d -name Bio\n";}} '
 perl -e 'foreach(@INC) { if(-f $_) {system "find $_ -type f -name SeqIO.pm\n";}} '
 perl -e 'foreach(@INC) { if(-d $_) {system "find $_ -type f -path Bio/SeqIO.pm\n";}} ' # not working though it should!

 ~/bin/findPerlModule.pl Bio::SeqIO

Command line arguments

• PERL5OPT : -[DIMUdmtw] switches

Debug

• caller function : shows the context

 print STDERR "DEBUG: context: ",(caller(0))[3],"\n";

 sub myDie { print STDERR "DEBUG: context: ",(caller(1))[3],"\n"; }
 sub b { myDie(); }

• Example

 use Data::Dumper;
 ...
 Dumper(\@list);
 Dumper(\%hash);

• perldebug

!    : re-exectutes prev command
!<n> : re-exectutes command <n>
H    : show command history

save   debug.history : saves debug history
source debug.history : runs debug history

Serialization

• Writes object to a file
• Example

 $ cat dump.pl

 use Data::Dumper;
 require "delta.pm";

 my $result=delta2result();
 print Data::Dumper->Dump([$result],['result']);

 $cat eval.pl

 use Data::Dumper;
 require "delta.pm";
 
 undef $/;
 eval <>;
  
 print Data::Dumper->Dump([$result],['result']);

Redefine commands

• Example:

 use subs 'cmp';

Libraries

• prefix.pm

 ...
 1;

• prefix.pl

 require "prefix.pm":
 ...

Modules

• -M moldule: same as "use module_name"

 perl -MLWP::Simple -e' print head "http://www.example.com"'
 perl -e 'use LWP::Simple; print head "http://www.example.com"'

• Installation(Makefile)

 $ perl Makefile.PL PREFIX=~/szdevel/CPAN
 $ make
 $ make test
 $ make install
   
 $ find ~/szdevel/CPAN -type f
 bin/xmltidy
 lib/perl5/site_perl/5.8.5/XML/Tidy.pm
 lib64/perl5/site_perl/5.8.5/x86_64-linux-thread-multi/auto/XML/Tidy/.packlist
 lib64/perl5/5.8.5/x86_64-linux-thread-multi/perllocal.pod
 share/man/man3/XML::Tidy.3pm

• Installation(Build)

 $ perl Build.PL --prefix ~/szdevel/CPAN
 $ ./Build test
 $ ./Build install

Useful modules

• Getopt::Long
• File::Find
• Algorithm::Cluster
• Bioperl

 git clone git://github.com/bioperl/bioperl-live.git
 cd bioperl-live/
 ./Build.PL --prefix /fs/szdevel/core-cbcb-software/Linux-x86_64
 ./Build test
 ./Build install
 cd ..

 git clone git://github.com/bioperl/bioperl-run.git
 cd bioperl-run/
 ./Build.PL --prefix /fs/szdevel/core-cbcb-software/Linux-x86_64
 ./Build test
 ./Build install
 cd ..

 git clone git://github.com/bioperl/bioperl-ext.git
 perl Makefile.PL PREFIX=/fs/szdevel/core-cbcb-software/Linux-x86_64  "DIR=Bio/Ext Bio/SeqIO/staden"
 make
 make test
 make install

 mv Makefile Makefile.old
 mv ./Bio/Ext/Align/Makefile ./Bio/Ext/Align/Makefile.old
 mv ./Bio/SeqIO/staden/Makefile ./Bio/SeqIO/staden/Makefile.old

 #if error ...
 cat Makefile.old                    | p 'if(/^(\w+FLAGS\s+=)(.+)/) { print $1," -fPIC ",$2,"\n" } else { print $_ }' > Makefile
 cat ./Bio/Ext/Align/Makefile.old    | p 'if(/^(\w+FLAGS\s+=)(.+)/) { print $1," -fPIC ",$2,"\n" } else { print $_ }' > ./Bio/Ext/Align/Makefile
 cat ./Bio/SeqIO/staden/Makefile.old | p 'if(/^(\w+FLAGS\s+=)(.+)/) { print $1," -fPIC ",$2,"\n" } else { print $_ }' > ./Bio/SeqIO/staden/Makefile

TO_INST_PM

Variables

• spec_variables
• $0 : program name
• $. : input line number
• $, : output field separator (default="")

 $ perl -e '$, = "\n";print (1,2,3,"");'
 1
 2
 3

 $ perl -e '$, = " "; @l=("a","b","c"); print map {uc($_)} @l; '
 A B C

• $/ : the input record separators (default="\n")
• $\ : output record separators (default="")

Argument passing

Example:

 sub test
 {
       my ($options)=@_;
       my %options=%{$options};
       print $options{a},"\n";
 }

 my %options;
 $options{a}="a";
 test(\%options);

Files

• open(IN, '-') # Open standard input
• open(IN, '>-') # Open standard output
• @lines = <IN> # read all lines in the file
• $line = <IN> # read one line

Hashes

• foreach $person (keys %ages) {}
• foreach $age (values %ages) { }
• while (($person, $age) = each(%ages)) {}
• cmd hash

 my %cmd_hash = (scarf2std=>\&scarf2std, fqint2std=>\&fqint2std);  
 &{$cmd_hash{$cmd}};  
 sub scarf2std {...} 
 sub fqint2std {...}

Security

File names

• use open(IN,"<" ,filename) instead of open(IN,filename)

 filename could be "rm -fr * |" 

• check file names for illegal characters

 $filename =~ /^([\w.-]+)$/)

Taint mode

 #!/usr/bin/perl -wT
 use strict;
 delete @ENV{qw(IFS CDPATH ENV BASH_ENV)};
 $ENV{PATH} = "/usr/bin/:/usr/local/bin";

Profiling

 perl -d:DProf test.pl        => tmon.out
 dprofpp

CPAN

• FAQ
• Check if a module is installed

 perl -MModule -e 1
 perl -MXML::Merge -e 1

• Edit CPAN configuration

 ~/.cpan/CPAN/MyConfig.pm
 'makepl_arg' => q[LIB=~/szdevel/CPAN/lib \
                   INSTALLBIN=~/szdevel/CPAN/bin \
                   INSTALLSCRIPT=~/szdevel/CPAN/bin \
                   INSTALLMAN1DIR=~/szdevel/CPAN/share/man/man1/ 
                   INSTALLMAN3DIR=~/szdevel/CPAN/share/man/man3/],

Bioperl

• Main Page