--- trunk/scripts/est-spider 2006/01/16 21:34:14 77
+++ trunk/scripts/est-spider 2006/08/15 16:38:06 178
@@ -3,14 +3,11 @@
use File::Find;
use Getopt::Long;
use File::Which;
-use HyperEstraier;
use Search::Estraier;
use Text::Iconv;
#use File::MMagic;
use File::MMagic::XS qw/:compat/;
-
-# do we use Node API?
-my $node_url;
+use Time::HiRes qw/time/;
my $collection; # name which will be inserted
my $path_add; # add additional info in path
@@ -20,7 +17,7 @@
#$verbose = 1;
my $debug = 0;
my $force = 0;
-my $native = 0;
+my $all = 0;
my $result = GetOptions(
"collection=s" => \$collection,
@@ -28,12 +25,24 @@
"verbose!" => \$verbose,
"debug!" => \$debug,
"exclude=s" => \$exclude,
- "node=s" => \$node_url,
"force!" => \$force,
- "native!" => \$native,
+ "all!" => \$all,
);
-my $dir = shift @ARGV || die "usage: $0 [dir]";
+my ($node_url,$dir) = @ARGV;
+
+die <<"_END_OF_USAGE_" if (! $node_url || ! $dir);
+usage: $0 http://localhost:1978/node/my_dir /path/to/directory
+
+options:
+ --collection="name of collection"
+ --path=/path/to/add/at/end
+ --exclude=regex_to_exclude
+ --verbose
+ --force
+ --debug
+ --all save placeholders for all files
+_END_OF_USAGE_
if (! -e $dir) {
warn "directory $dir doesn't exist, skipping\n";
@@ -44,7 +53,17 @@
#$basedir =~ s,/[^/]+$,/,;
#require "$basedir/filter.pm";
-my $pdftotext = which('pdftotext');
+my $docs = 0;
+my $start_t = time();
+
+my $filter;
+foreach my $f (qw/pdftotext pstotext/) {
+ my $w = which($f);
+ if ($f) {
+ $filter->{$f} = $w;
+ print STDERR "using $f filter at $w\n" if ($verbose);
+ }
+}
#my $mm = new File::MMagic('/usr/share/misc/file/magic');
my $mm = new File::MMagic::XS();
@@ -54,32 +73,12 @@
select(STDERR); $|=1;
select(STDOUT); $|=1;
-print STDERR "using $pdftotext to convert pdf into html\n" if ($pdftotext && $verbose);
-
-my $db;
-if ($node_url) {
- if ($native) {
- $db = HyperEstraier::Node->new($node_url);
- } else {
- $db = new Search::Estraier::Node;
- $db->set_url($node_url);
- }
- $db->set_auth('admin', 'admin');
-} else {
- # open the database
- $db = HyperEstraier::Database->new();
- $db->open('/tmp/casket', $HyperEstraier::Database::DBWRITER | $HyperEstraier::Database::DBCREAT);
-
- sub signal {
- my($sig) = @_;
- print "\nCaught a SIG$sig--syncing database and shutting down\n";
- $db->sync();
- exit(0);
- }
-
- $SIG{'INT'} = \&signal;
- $SIG{'QUIT'} = \&signal;
-}
+my $db = new Search::Estraier::Node(
+ url => $node_url,
+ user => 'admin',
+ passwd => 'admin',
+ croak_on_error => 1,
+);
find({ wanted => \&file,
follow => 1,
@@ -87,19 +86,20 @@
no_chdir => 1,
}, $dir);
-unless ($node_url) {
- print "--- sync\n";
- $db->sync();
+my $dur = (time() - $start_t) || 1;
+printf STDERR "%d documents in %.2fs [%.2f docs/s]\n", $docs, $dur, ($docs / $dur);
+
+$db->master(
+ action => 'sync'
+);
+
- print "--- optimize...\n";
- $db->optimize(0);
-}
exit;
-sub dump_contents($$$$) {
- my ($db,$contents,$mtime,$path) = @_;
+sub dump_contents {
+ my ($db,$contents,$mtime,$path,$size) = @_;
- return unless ($contents); # don't die on empty files
+ return unless (defined($contents)); # don't die on empty files
if ($exclude && $path =~ m/$exclude/i) {
print STDERR "skip: $path\n" if ($verbose);
@@ -107,17 +107,14 @@
}
use bytes;
- my $size = length $contents;
+ if (! $size) {
+ $size = length $contents;
+ }
print STDERR " [$size]" if ($verbose);
# create a document object
- my $doc;
- if ($native) {
- $doc = HyperEstraier::Document->new;
- } else {
- $doc = new Search::Estraier::Document;
- }
+ my $doc = new Search::Estraier::Document;
my $title = $1 if ($contents =~ m#
(.+?)#is);
@@ -132,58 +129,35 @@
$doc->add_attr('@size', $size);
$doc->add_attr('@mtime', $mtime);
- # html2text
- $contents =~ s#<[^>]+/*>##gs;
- $contents =~ s#\s\s+# #gs;
+ if ($contents) {
+ # html2text
+ $contents =~ s#<[^>]+/*>##gs;
+ $contents =~ s#\s\s+# #gs;
+
+ $doc->add_text($iconv->convert($contents));
+ }
+ # store path
+ $doc->add_hidden_text($path);
+ # boost title
+ $doc->add_hidden_text($title);
- $doc->add_text($iconv->convert($contents));
-
-# print $doc->dump_draft if ($verbose);
+ print $doc->dump_draft if ($debug);
# register the document object to the database
- if ($node_url) {
- $db->put_doc($doc);
- } else {
- $db->put_doc($doc, $HyperEstraier::Database::PDCLEAN);
- }
-
-}
-
-sub file {
-
- my $path = $_;
- my $contents;
+ $db->put_doc($doc);
- return if (! $force && -l $path || $path =~ m#/.svn# || $path =~ m/(~|.bak)$/);
+ $docs++;
- my $mtime = (stat($path))[9] || -1;
- my $mtime_db = $db->get_doc_attr_by_uri("file:///$path", '@mtime') || -2;
-
- if ($mtime == $mtime_db) {
- print STDERR "# same: $path $mtime\n" if ($verbose);
- return unless($force);
- } else {
- print STDERR "# changed: $path $mtime != $mtime_db\n" if ($debug);
- }
-
- # skip files on which File::MMagic::XS croaks
- return if ($path =~ m#\.au$#);
-
- my $type = $mm->checktype_filename($path);
- $type =~ s/\s+/ /gs;
-
- print STDERR "# $path $type\n" if ($debug);
+}
- if ($pdftotext && -f $path && $type =~ m/pdf/i) {
+sub filter_to_pages {
+ my ($path, $mtime, $command) = @_;
print STDERR "$path {converting}" if ($verbose);
- open(F,"$pdftotext -htmlmeta \"$path\" - |") || die "can't open $pdftotext with '$path'";
+ open(F,"$command |") || die "can't open $command with '$path': $!";
my $html;
while() {
- # XXX why pdftotext barks if I try to use this is beyond me.
- #$contents .= $_;
-
$html .= $_;
}
close(F);
@@ -216,17 +190,67 @@
$page_nr++;
}
+
+
+}
+
+sub file {
+
+ my $path = $_;
+ my $contents;
+
+ return if (! $force && -l $path || $path =~ m#/.svn# || $path =~ m/(~|.bak|.gif)$/);
+ return if (! $all && -d $path);
+
+ my $mtime = (stat($path))[9] || -1;
+ my $mtime_db = eval { $db->get_doc_attr_by_uri("file:///$path", '@mtime') };
+ $mtime_db ||= -2;
+
+ if ($mtime == $mtime_db) {
+ print STDERR "# same: $path $mtime\n" if ($verbose);
+ return unless($force);
+ } else {
+ print STDERR "# changed: $path $mtime != $mtime_db\n" if ($debug);
+ }
+
+ # skip files on which File::MMagic::XS croaks
+ if ($path =~ m#\.au$#) {
+ warn "skipped '$path' to prevent File::MMagic::XS bug\n" if ($debug);
+ return;
+ }
+
+ my $type = $mm->checktype_filename($path);
+ $type =~ s/\s+/ /gs;
+
+ print STDERR "# $path $type\n" if ($debug);
+
+ if ($type =~ m/pdf/i) {
+ if ($filter->{pdftotext}) {
+ filter_to_pages($path, $mtime, qq( $filter->{pdftotext} -htmlmeta "$path" - ));
+ } else {
+ warn "skipping '$path', no pdftotext filter\n" if ($verbose);
+ return;
+ }
+ } elsif ($type eq 'application/postscript') {
+ if ($filter->{pstotext}) {
+ filter_to_pages($path, $mtime, qq( $filter->{pstotext} "$path" ));
+ } else {
+ warn "skipping '$path', no pstotext filter\n" if ($verbose);
+ return;
+ }
} else {
# return if (! -f $path || ! m/\.(html*|php|pl|txt|info|log|text)$/i);
- if (-f $path && $type =~ m/html/ ||
- ($type !~ m#text# && $path =~ m/\.(php|pl|txt|info|log|text)$/io)
+ if (-f $path &&
+ $type !~ m/html/ &&
+ $path !~ m/\.(php|pl|txt|info|log|text)$/io
) {
- dump_contents($db, "$path $type", $mtime, $path);
+ dump_contents($db, '', $mtime, $path, -s $path) if ($all);
+ return;
}
# skip index files
- return if (m/index_[a-z]\.html*/i || m/index_symbol\.html*/i);
+ return if ($path =~ m/index_(?:[a-z]+|symbol)\.html*/i);
open(F,"$path") || die "can't open file: $path";
print STDERR "$path ($type)" if ($verbose);