--- trunk2/lib/WebPAC.pm 2004/09/10 22:24:42 421 +++ trunk2/lib/WebPAC.pm 2004/09/15 21:21:36 453 @@ -9,6 +9,7 @@ use XML::Simple; use Template; use Log::Log4perl qw(get_logger :levels); +use Time::HiRes qw(time); use Data::Dumper; @@ -33,34 +34,19 @@ my $webpac = new WebPAC( config_file => 'name.conf', - [code_page => 'ISO-8859-2',] - [low_mem => 1,] + code_page => 'ISO-8859-2', + low_mem => 1, ); Default C is C. +Default is not to use C options (see L below). + This method will also read configuration files C (used by indexer and Web font-end) and configuration file specified by C which describes databases to be indexed. -C options is double-edged sword. If enabled, WebPAC -will run on memory constraint machines (which doesn't have enough -physical RAM to create memory structure for whole ISIS database). - -If your machine has 512Mb or more and database is around 10000 records, -memory shouldn't be an issue. If you don't have enough physical RAM, you -might consider using virtual memory (if your operating system is handling it -well, like on FreeBSD or Linux) instead of dropping to L to handle -parsed structure of ISIS database. - -However, when WebPAC is running on desktop machines (or laptops :-), it's -highly undesireable for system to start swapping. Using C option can -reduce WecPAC memory usage to 16Mb for same database with lookup fields and -sorted indexes which stay in RAM. Performance will suffer, but memory usage -will really be minimal. It might be also more confortable to run WebPAC reniced -on those machines. - =cut # mapping between data type and tag which specify @@ -77,6 +63,8 @@ my $self = {@_}; bless($self, $class); + $self->{'start_t'} = time(); + my $log_file = $self->{'log'} || "log.conf"; Log::Log4perl->init($log_file); @@ -126,7 +114,7 @@ # running with low_mem flag? well, use DBM::Deep then. if ($self->{'low_mem'}) { - $log->info("running with low_mem which impacts performance (<64 Mb memory usage)"); + $log->info("running with low_mem which impacts performance (<32 Mb memory usage)"); my $db_file = "data.db"; @@ -135,7 +123,7 @@ $log->debug("removed '$db_file' from last run"); } - use DBM::Deep; + require DBM::Deep; my $db = new DBM::Deep $db_file; @@ -144,7 +132,7 @@ if ($db->error()) { $log->logdie("can't open '$db_file' under low_mem: ",$db->error()); } else { - $log->debug("using file $db_file for DBM::Deep"); + $log->debug("using file '$db_file' for DBM::Deep"); } $self->{'db'} = $db; @@ -160,12 +148,16 @@ $webpac->open_isis( filename => '/data/ISIS/ISIS', code_page => '852', - limit_mfn => '500', + limit_mfn => 500, + start_mfn => 6000, lookup => [ ... ], ); By default, ISIS code page is assumed to be C<852>. +If optional parametar C is set, this will be first MFN to read +from database (so you can skip beginning of your database if you need to). + If optional parametar C is set, it will read just 500 records from database in example above. @@ -193,6 +185,8 @@ $log->logcroak("need filename") if (! $arg->{'filename'}); my $code_page = $arg->{'code_page'} || '852'; + $log->logdie("can't find database ",$arg->{'filename'}) unless (glob($arg->{'filename'}.'.*')); + # store data in object $self->{'isis_filename'} = $arg->{'filename'}; $self->{'isis_code_page'} = $code_page; @@ -210,13 +204,21 @@ my $isis_db = OpenIsis::open($arg->{'filename'}); my $maxmfn = OpenIsis::maxRowid( $isis_db ) || 1; + my $startmfn = 1; + + if (my $s = $self->{'start_mfn'}) { + $log->info("skipping to MFN $s"); + $startmfn = $s; + } else { + $self->{'start_mfn'} = $startmfn; + } - $maxmfn = $self->{limit_mfn} if ($self->{limit_mfn}); + $maxmfn = $startmfn + $self->{limit_mfn} if ($self->{limit_mfn}); - $log->info("processing $maxmfn records..."); + $log->info("processing ",($maxmfn-$startmfn)." records..."); # read database - for (my $mfn = 1; $mfn <= $maxmfn; $mfn++) { + for (my $mfn = $startmfn; $mfn <= $maxmfn; $mfn++) { $log->debug("mfn: $mfn\n"); @@ -264,7 +266,7 @@ } - $self->{'current_mfn'} = 1; + $self->{'current_mfn'} = -1; $self->{'last_pcnt'} = 0; $log->debug("max mfn: $maxmfn"); @@ -287,7 +289,15 @@ my $log = $self->_get_logger(); - my $mfn = $self->{'current_mfn'}++ || $log->logconfess("it seems that you didn't load database!"); + $log->logconfess("it seems that you didn't load database!") unless ($self->{'current_mfn'}); + + if ($self->{'current_mfn'} == -1) { + $self->{'current_mfn'} = $self->{'start_mfn'}; + } else { + $self->{'current_mfn'}++; + } + + my $mfn = $self->{'current_mfn'}; if ($mfn > $self->{'max_mfn'}) { $self->{'current_mfn'} = $self->{'max_mfn'}; @@ -304,6 +314,19 @@ } } +=head2 mfn + +Returns current record number (MFN). + + print $webpac->mfn; + +=cut + +sub mfn { + my $self = shift; + return $self->{'current_mfn'}; +} + =head2 progress_bar Draw progress bar on STDERR. @@ -329,16 +352,52 @@ $self->{'last_pcnt'} ||= 1; - $self->{'last_pcnt'} = $curr if ($curr < $self->{'last_pcnt'}); - my $p = int($curr * 100 / $max); + + # reset on re-run + if ($p < $self->{'last_pcnt'}) { + $self->{'last_pcnt'} = $p; + $self->{'last_t'} = time(); + $self->{'last_curr'} = undef; + } + if ($p != $self->{'last_pcnt'}) { - printf STDERR ("%5d / %5d [%-51s] %-2d %% \r",$curr,$max,"=" x ($p/2).">", $p ); + + my $last_curr = $self->{'last_curr'} || $curr; + my $t = time(); + my $rate = ($curr - $last_curr) / (($t - $self->{'last_t'} || 1)); + my $eta = ($max-$curr) / ($rate || 1); + printf STDERR ("%5d [%-38s] %-5d %0.1f/s %s\r",$curr,"=" x ($p/3)."$p%>", $max, $rate, $self->fmt_time($eta)); $self->{'last_pcnt'} = $p; + $self->{'last_t'} = time(); + $self->{'last_curr'} = $curr; } print STDERR "\n" if ($p == 100); } +=head2 fmt_time + +Format time (in seconds) for display. + + print $webpac->fmt_time(time()); + +This method is called by L to display remaining time. + +=cut + +sub fmt_time { + my $self = shift; + + my $t = shift || 0; + my $out = ""; + + my ($ss,$mm,$hh) = gmtime($t); + $out .= "${hh}h" if ($hh); + $out .= sprintf("%02d:%02d", $mm,$ss); + $out .= " " if ($hh == 0); + return $out; +} + =head2 open_import_xml Read file from C directory and parse it. @@ -718,6 +777,31 @@ return @arr; } +=head2 sort_arr + +Sort array ignoring case and html in data + + my @sorted = $webpac->sort_arr(@unsorted); + +=cut + +sub sort_arr { + my $self = shift; + + my $log = $self->_get_logger(); + + # FIXME add Schwartzian Transformation? + + my @sorted = sort { + $a =~ s#<[^>]+/*>##; + $b =~ s#<[^>]+/*>##; + lc($b) cmp lc($a) + } @_; + $log->debug("sorted values: ",sub { join(", ",@sorted) }); + + return @sorted; +} + =head2 data_structure @@ -774,6 +858,11 @@ } next if (! @v); + if ($tag->{'sort'}) { + @v = $self->sort_arr(@v); + $log->warn("sort within tag is usually not what you want!"); + } + # use format? if ($tag->{'format_name'}) { @v = map { $self->apply_format($tag->{'format_name'},$tag->{'format_delimiter'},$_) } @v; @@ -788,12 +877,35 @@ next; # don't return headline in data_structure! } - # does tag have type? - if ($tag->{'type'}) { - push @{$row->{$tag->{'type'}}}, @v; - } else { - push @{$row->{'display'}}, @v; - push @{$row->{'swish'}}, @v; + # delimiter will join repeatable fields + if ($tag->{'delimiter'}) { + @v = ( join($tag->{'delimiter'}, @v) ); + } + + # default types + my @types = qw(display swish); + # override by type attribute + @types = ( $tag->{'type'} ) if ($tag->{'type'}); + + foreach my $type (@types) { + # append to previous line? + $log->debug("type: $type ",sub { join(" ",@v) }, $row->{'append'} || 'no append'); + if ($tag->{'append'}) { + + # I will delimit appended part with + # delimiter (or ,) + my $d = $tag->{'delimiter'}; + # default delimiter + $d ||= ", "; + + my $last = pop @{$row->{$type}}; + $d = "" if (! $last); + $last .= $d . join($d, @v); + push @{$row->{$type}}, $last; + + } else { + push @{$row->{$type}}, @v; + } } @@ -806,6 +918,11 @@ my $name = $self->{'import_xml'}->{'indexer'}->{$field}->{'name'}; $row->{'name'} = $name ? $self->_x($name) : $field; + # post-sort all values in field + if ($self->{'import_xml'}->{'indexer'}->{$field}->{'sort'}) { + $log->warn("sort at field tag not implemented"); + } + push @ds, $row; $log->debug("row $field: ",sub { Dumper($row) }); @@ -1021,6 +1138,37 @@ also use method names, and not only classes (which are just few) to filter logging. + +=head1 MEMORY USAGE + +C options is double-edged sword. If enabled, WebPAC +will run on memory constraint machines (which doesn't have enough +physical RAM to create memory structure for whole source database). + +If your machine has 512Mb or more of RAM and database is around 10000 records, +memory shouldn't be an issue. If you don't have enough physical RAM, you +might consider using virtual memory (if your operating system is handling it +well, like on FreeBSD or Linux) instead of dropping to L to handle +parsed structure of ISIS database (this is what C option does). + +Hitting swap at end of reading source database is probably o.k. However, +hitting swap before 90% will dramatically decrease performance and you will +be better off with C and using rest of availble memory for +operating system disk cache (Linux is particuallary good about this). +However, every access to database record will require disk access, so +generation phase will be slower 10-100 times. + +Parsed structures are essential - you just have option to trade RAM memory +(which is fast) for disk space (which is slow). Be sure to have planty of +disk space if you are using C and thus L. + +However, when WebPAC is running on desktop machines (or laptops :-), it's +highly undesireable for system to start swapping. Using C option can +reduce WecPAC memory usage to around 64Mb for same database with lookup +fields and sorted indexes which stay in RAM. Performance will suffer, but +memory usage will really be minimal. It might be also more confortable to +run WebPAC reniced on those machines. + =cut 1;