Annotation of /trunk/scripts/est-spider

#!/usr/bin/perl -w
use strict;
use File::Find;
use Getopt::Long;
use File::Which;
use Search::Estraier;
use Text::Iconv;
#use File::MMagic;
use File::MMagic::XS qw/:compat/;
use Time::HiRes qw/time/;
use HTML::TreeBuilder;
use Data::Dump qw/dump/;

my $collection;         # name which will be inserted
my $path_add;           # add additional info in path
my $verbose;
my $exclude;

#$verbose = 1;
my $debug = 0;
my $force = 0;
my $all = 0;
my $skip_images = 0;

my $result = GetOptions(
        "collection=s" => \$collection,
        "path=s" => \$path_add,
        "verbose!" => \$verbose,
        "debug!" => \$debug,
        "exclude=s" => \$exclude,
        "force!" => \$force,
        "all!" => \$all,
        "skip-images!" => \$skip_images,
);

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
        --skip-images
        --verbose
        --force
        --debug
        --all   save placeholders for all files
_END_OF_USAGE_

if (! -e $dir) {
        warn "directory $dir doesn't exist, skipping\n";
        exit 1;
}

#my $basedir = $0;
#$basedir =~ s,/[^/]+$,/,;
#require "$basedir/filter.pm";

my $docs = 0;
my $start_t = time();

my $filter;
foreach my $f (qw/pdftotext pstotext/) {
        my $w = which($f);
        if ($w) {
                $filter->{$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();

my $iconv = new Text::Iconv('iso-8859-2', 'utf-8');

select(STDERR); $|=1;
select(STDOUT); $|=1;

my $db = new Search::Estraier::Node(
        url => $node_url,
        user => 'admin',
        passwd => 'admin',
        croak_on_error => 1,
        create => 1,
);

#
# check if hhc file exists, and if it does, extract information from it
#

my $hhc_file;
# try to find hhc
find({ wanted => sub {
                return unless( m!\.hhc$!i );
                $hhc_file = $_;
                warn "using $hhc_file for tree structure\n";
        },
        follow => 1,
        follow_skip => 2,
        no_chdir => 1,
}, $dir);

my $meta;

if ($hhc_file) {

        sub param {
                my ($el) = @_;

                my $n;
                foreach my $p ( $el->find('param') ) {
                        $n->{ $p->attr('name') } = $p->attr('value');
                }

                if ( ! defined($n->{Local}) ) {
                        warn "### skipped = ",dump($n),$/;
                        return;
                }

                my $path = $dir . '/' . $n->{Local};
                $path =~ s!//!/!g;
                $path = lc($path);

                $n->{path} = $path;

                my $nr = $n->{ImageNumber} || next;

                if ($nr == 27) {
                        $meta->{title} = $n->{Name};
                        $meta->{index_path} = $path;
                } elsif ($nr == 21) {
                        $meta->{toc_path} = $path;
                } elsif ($nr == 1) {
                        $meta->{foreword_path} = $path;
                } elsif ($nr == 11) {
                        # nop
                } else {
                        warn "unknown ImageNumber: $nr\n";
                }

                return $n;
        }

        my $tree = HTML::TreeBuilder->new;
        $tree->parse_file($hhc_file);

        my $prefix = $collection ? ( $collection . ' :: ' ) : '';

        my @prefix;
        my $depth = 0;

        foreach my $e ( $tree->look_down( sub { $_[0]->tag =~ m/(object)/ } ) ) {

#               printf("%05s %s\n", $e->parent->address(), $e->as_HTML() );

                my $l = ($e->depth() / 2) - 1;

                $prefix[ 0 ] = $meta->{title} || '';

                my $n = param($e);
                $prefix[ $l ] = $n->{Name};

                next unless ($n->{path});

                my $t = '';
                my @p;
                foreach my $i ( 0 .. $l ) {
                        push @p, $prefix[ $i ] if ($prefix[ $i ]);
                }
                $t = join(' :: ', @p ) if (@p);

                $meta->{path2title}->{ $n->{path} } = $t;

        }

        $tree->delete;

}


find({ wanted => \&file, 
        follow => 1,
        follow_skip => 2,
        no_chdir => 1,
}, $dir);

my $dur = (time() - $start_t) || 1;
printf STDERR "%d documents in %.2fs [%.2f docs/s]\n", $docs, $dur, ($docs / $dur);

$db->master(
        action => 'sync'
);


exit;

sub dump_contents {
        my ($db,$contents,$mtime,$path,$size) = @_;

        return unless (defined($contents));     # don't die on empty files

        if ($exclude && $path =~ m/$exclude/i) {
                print STDERR "skip: $path\n" if ($verbose);
                return;
        }

        use bytes;
        if (! $size) {
                $size = length $contents;
        }

        print STDERR " [$size]" if ($verbose);

        # create a document object 
        my $doc = new Search::Estraier::Document;

        my $title;

        if ( defined($meta->{path2title}->{lc($path)}) ) {
                $title = $meta->{path2title}->{lc($path)};
                warn " $title\n";
        } else {

                $title = $1 if ($contents =~ m#<title>(.+?)</title>#is);

                # chop long titles to 100 chars
                $title = substr($title, 0, 100) . '...' if ($title && length($title) > 100);
                # use path if no title is found
                $title ||= $path;
        
        }

        # add attributes to the document object 
        $doc->add_attr('@uri', "file:///$path");
        $doc->add_attr('@title', $iconv->convert($title));
        $doc->add_attr('@size', $size);
        $doc->add_attr('@mtime', $mtime);

        if ($contents) {
                # html2text
                $contents =~ s#<script.*?</script>##gis;
                $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);

        print $doc->dump_draft if ($debug);

        # register the document object to the database
        $db->put_doc($doc);

        $docs++;

}

sub filter_to_pages {
                my ($path, $mtime, $command) = @_;

                print STDERR "$path {converting}" if ($verbose);

                open(F,"$command |") || die "can't open $command with '$path': $!";
                my $html;
                while(<F>) {
                        $html .= $_;
                }
                close(F);

                return if (! $html);

                my $file_only = $path;
                $file_only =~ s/^.*\/([^\/]+)$/$1/g;

                my ($pre_html,$pages,$post_html) = ('<html><head><title>$path :: page ##page_nr##</title></head><body><pre>',$html,'</pre></body></html>');

                ($pre_html,$pages,$post_html) = ($1,$2,$3) if ($html =~ m/^(<html>.+?<pre>)(.+)(<\/pre>.+?)$/si);

                if ($collection) {
                        $pre_html =~ s/<title>(.+?)<\/title>/<title>$collection :: page ##page_nr##<\/title>/si;
                } else {
                        $pre_html =~ s/<title>(.+?)<\/title>/<title>$1 :: page ##page_nr##<\/title>/si ||
                        $pre_html =~ s/<title><\/title>/<title>$file_only :: page ##page_nr##<\/title>/si;
                }

                # save empty entry as a placeholder
                dump_contents($db, ' ', $mtime, "$path");

                my $page_nr = 1;
                foreach my $page (split(/\f/s,$pages)) {
                        print STDERR " $page_nr" if ($verbose);
                        my $pre_tmp = $pre_html;
                        $pre_tmp =~ s/##page_nr##/$page_nr<\/title>/s;
                        dump_contents($db, $pre_tmp . $page . $post_html, $mtime, "$path#$page_nr") if ($page !~ m/^\s*$/s);
                        $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') } unless ($force);
        $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;
                }
        } elsif ($type =~ m!^image/! && $skip_images) {
                warn "skipping image '$path'\n" if ($verbose);
                return; # don't index images
        } else {

#               return if (! -f $path || ! m/\.(html*|php|pl|txt|info|log|text)$/i);
                if (-f $path &&
                        $type !~ m/html/ &&
                        $path !~ m/\.(php|pl|txt|info|log|text)$/io
                ) {
                        dump_contents($db, '', $mtime, $path, -s $path) if ($all);
                        return;
                }

                # skip index files
                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);
                while(<F>) {
                        $contents .= "$_";
                }
                $contents .= "\n\n";

                #$contents = filter($contents,$collection);

                # add optional components to path
                $path .= " $path_add" if ($path_add);

                dump_contents($db, $contents, $mtime, $path);
        }

        print STDERR "\n" if ($verbose);
#       die "zero size content in '$path'" if (! $contents);

}

1	dpavlin	64	#!/usr/bin/perl -w
2			use strict;
3			use File::Find;
4			use Getopt::Long;
5			use File::Which;
6			use Search::Estraier;
7			use Text::Iconv;
8			#use File::MMagic;
9			use File::MMagic::XS qw/:compat/;
10	dpavlin	118	use Time::HiRes qw/time/;
11	dpavlin	179	use HTML::TreeBuilder;
12			use Data::Dump qw/dump/;
13	dpavlin	64
14			my $collection; # name which will be inserted
15			my $path_add; # add additional info in path
16			my $verbose;
17			my $exclude;
18
19			#$verbose = 1;
20			my $debug = 0;
21			my $force = 0;
22	dpavlin	119	my $all = 0;
23	dpavlin	196	my $skip_images = 0;
24	dpavlin	64
25			my $result = GetOptions(
26			"collection=s" => \$collection,
27			"path=s" => \$path_add,
28			"verbose!" => \$verbose,
29			"debug!" => \$debug,
30			"exclude=s" => \$exclude,
31			"force!" => \$force,
32	dpavlin	119	"all!" => \$all,
33	dpavlin	196	"skip-images!" => \$skip_images,
34	dpavlin	64	);
35
36	dpavlin	89	my ($node_url,$dir) = @ARGV;
37	dpavlin	64
38	dpavlin	89	die <<"_END_OF_USAGE_" if (! $node_url \|\| ! $dir);
39			usage: $0 http://localhost:1978/node/my_dir /path/to/directory
40
41			options:
42			--collection="name of collection"
43			--path=/path/to/add/at/end
44			--exclude=regex_to_exclude
45	dpavlin	196	--skip-images
46	dpavlin	89	--verbose
47			--force
48			--debug
49	dpavlin	119	--all save placeholders for all files
50	dpavlin	89	_END_OF_USAGE_
51
52	dpavlin	64	if (! -e $dir) {
53			warn "directory $dir doesn't exist, skipping\n";
54			exit 1;
55			}
56
57			#my $basedir = $0;
58			#$basedir =~ s,/[^/]+$,/,;
59			#require "$basedir/filter.pm";
60
61	dpavlin	118	my $docs = 0;
62			my $start_t = time();
63
64	dpavlin	90	my $filter;
65			foreach my $f (qw/pdftotext pstotext/) {
66			my $w = which($f);
67	dpavlin	180	if ($w) {
68	dpavlin	90	$filter->{$f} = $w;
69			print STDERR "using $f filter at $w\n" if ($verbose);
70			}
71			}
72	dpavlin	64
73			#my $mm = new File::MMagic('/usr/share/misc/file/magic');
74			my $mm = new File::MMagic::XS();
75
76			my $iconv = new Text::Iconv('iso-8859-2', 'utf-8');
77
78			select(STDERR); $\|=1;
79			select(STDOUT); $\|=1;
80
81	dpavlin	119	my $db = new Search::Estraier::Node(
82			url => $node_url,
83			user => 'admin',
84			passwd => 'admin',
85			croak_on_error => 1,
86	dpavlin	179	create => 1,
87	dpavlin	119	);
88	dpavlin	64
89	dpavlin	179	#
90			# check if hhc file exists, and if it does, extract information from it
91			#
92
93			my $hhc_file;
94			# try to find hhc
95			find({ wanted => sub {
96			return unless( m!\.hhc$!i );
97			$hhc_file = $_;
98			warn "using $hhc_file for tree structure\n";
99			},
100			follow => 1,
101			follow_skip => 2,
102			no_chdir => 1,
103			}, $dir);
104
105			my $meta;
106
107			if ($hhc_file) {
108
109			sub param {
110			my ($el) = @_;
111
112			my $n;
113			foreach my $p ( $el->find('param') ) {
114			$n->{ $p->attr('name') } = $p->attr('value');
115			}
116
117			if ( ! defined($n->{Local}) ) {
118			warn "### skipped = ",dump($n),$/;
119			return;
120			}
121
122			my $path = $dir . '/' . $n->{Local};
123			$path =~ s!//!/!g;
124			$path = lc($path);
125
126			$n->{path} = $path;
127
128			my $nr = $n->{ImageNumber} \|\| next;
129
130			if ($nr == 27) {
131			$meta->{title} = $n->{Name};
132			$meta->{index_path} = $path;
133			} elsif ($nr == 21) {
134			$meta->{toc_path} = $path;
135			} elsif ($nr == 1) {
136			$meta->{foreword_path} = $path;
137			} elsif ($nr == 11) {
138			# nop
139			} else {
140			warn "unknown ImageNumber: $nr\n";
141			}
142
143			return $n;
144			}
145
146			my $tree = HTML::TreeBuilder->new;
147			$tree->parse_file($hhc_file);
148
149			my $prefix = $collection ? ( $collection . ' :: ' ) : '';
150
151			my @prefix;
152			my $depth = 0;
153
154			foreach my $e ( $tree->look_down( sub { $_[0]->tag =~ m/(object)/ } ) ) {
155
156			# printf("%05s %s\n", $e->parent->address(), $e->as_HTML() );
157
158			my $l = ($e->depth() / 2) - 1;
159
160			$prefix[ 0 ] = $meta->{title} \|\| '';
161
162			my $n = param($e);
163			$prefix[ $l ] = $n->{Name};
164
165			next unless ($n->{path});
166
167			my $t = '';
168			my @p;
169			foreach my $i ( 0 .. $l ) {
170			push @p, $prefix[ $i ] if ($prefix[ $i ]);
171			}
172			$t = join(' :: ', @p ) if (@p);
173
174			$meta->{path2title}->{ $n->{path} } = $t;
175
176			}
177
178			$tree->delete;
179
180			}
181
182
183	dpavlin	64	find({ wanted => \&file,
184			follow => 1,
185			follow_skip => 2,
186			no_chdir => 1,
187			}, $dir);
188
189	dpavlin	118	my $dur = (time() - $start_t) \|\| 1;
190			printf STDERR "%d documents in %.2fs [%.2f docs/s]\n", $docs, $dur, ($docs / $dur);
191	dpavlin	64
192	dpavlin	178	$db->master(
193			action => 'sync'
194			);
195
196
197	dpavlin	64	exit;
198
199	dpavlin	90	sub dump_contents {
200			my ($db,$contents,$mtime,$path,$size) = @_;
201	dpavlin	64
202	dpavlin	90	return unless (defined($contents)); # don't die on empty files
203	dpavlin	64
204			if ($exclude && $path =~ m/$exclude/i) {
205			print STDERR "skip: $path\n" if ($verbose);
206			return;
207			}
208
209			use bytes;
210	dpavlin	90	if (! $size) {
211			$size = length $contents;
212			}
213	dpavlin	64
214			print STDERR " [$size]" if ($verbose);
215
216			# create a document object
217	dpavlin	89	my $doc = new Search::Estraier::Document;
218	dpavlin	64
219	dpavlin	179	my $title;
220	dpavlin	64
221	dpavlin	179	if ( defined($meta->{path2title}->{lc($path)}) ) {
222			$title = $meta->{path2title}->{lc($path)};
223			warn " $title\n";
224			} else {
225	dpavlin	64
226	dpavlin	179	$title = $1 if ($contents =~ m#<title>(.+?)</title>#is);
227
228			# chop long titles to 100 chars
229			$title = substr($title, 0, 100) . '...' if ($title && length($title) > 100);
230			# use path if no title is found
231			$title \|\|= $path;
232
233			}
234
235	dpavlin	64	# add attributes to the document object
236			$doc->add_attr('@uri', "file:///$path");
237			$doc->add_attr('@title', $iconv->convert($title));
238			$doc->add_attr('@size', $size);
239			$doc->add_attr('@mtime', $mtime);
240
241	dpavlin	90	if ($contents) {
242			# html2text
243	dpavlin	181	$contents =~ s#<script.*?</script>##gis;
244	dpavlin	90	$contents =~ s#<[^>]+/*>##gs;
245			$contents =~ s#\s\s+# #gs;
246	dpavlin	64
247	dpavlin	90	$doc->add_text($iconv->convert($contents));
248			}
249			# store path
250			$doc->add_hidden_text($path);
251			# boost title
252			$doc->add_hidden_text($title);
253	dpavlin	64
254	dpavlin	120	print $doc->dump_draft if ($debug);
255	dpavlin	64
256			# register the document object to the database
257	dpavlin	89	$db->put_doc($doc);
258	dpavlin	64
259	dpavlin	118	$docs++;
260
261	dpavlin	64	}
262
263	dpavlin	90	sub filter_to_pages {
264			my ($path, $mtime, $command) = @_;
265	dpavlin	64
266			print STDERR "$path {converting}" if ($verbose);
267
268	dpavlin	90	open(F,"$command \|") \|\| die "can't open $command with '$path': $!";
269	dpavlin	64	my $html;
270			while(<F>) {
271			$html .= $_;
272			}
273			close(F);
274
275			return if (! $html);
276
277			my $file_only = $path;
278			$file_only =~ s/^.*\/([^\/]+)$/$1/g;
279
280			my ($pre_html,$pages,$post_html) = ('<html><head><title>$path :: page ##page_nr##</title></head><body><pre>',$html,'</pre></body></html>');
281
282			($pre_html,$pages,$post_html) = ($1,$2,$3) if ($html =~ m/^(<html>.+?<pre>)(.+)(<\/pre>.+?)$/si);
283
284			if ($collection) {
285			$pre_html =~ s/<title>(.+?)<\/title>/<title>$collection :: page ##page_nr##<\/title>/si;
286			} else {
287			$pre_html =~ s/<title>(.+?)<\/title>/<title>$1 :: page ##page_nr##<\/title>/si \|\|
288			$pre_html =~ s/<title><\/title>/<title>$file_only :: page ##page_nr##<\/title>/si;
289			}
290
291			# save empty entry as a placeholder
292			dump_contents($db, ' ', $mtime, "$path");
293
294			my $page_nr = 1;
295			foreach my $page (split(/\f/s,$pages)) {
296			print STDERR " $page_nr" if ($verbose);
297			my $pre_tmp = $pre_html;
298			$pre_tmp =~ s/##page_nr##/$page_nr<\/title>/s;
299			dump_contents($db, $pre_tmp . $page . $post_html, $mtime, "$path#$page_nr") if ($page !~ m/^\s*$/s);
300			$page_nr++;
301			}
302
303	dpavlin	90
304
305			}
306
307			sub file {
308
309			my $path = $_;
310			my $contents;
311
312	dpavlin	119	return if (! $force && -l $path \|\| $path =~ m#/.svn# \|\| $path =~ m/(~\|.bak\|.gif)$/);
313	dpavlin	121	return if (! $all && -d $path);
314	dpavlin	90
315			my $mtime = (stat($path))[9] \|\| -1;
316	dpavlin	182	my $mtime_db;
317			eval { $db->get_doc_attr_by_uri("file:///$path", '@mtime') } unless ($force);
318	dpavlin	119	$mtime_db \|\|= -2;
319	dpavlin	90
320			if ($mtime == $mtime_db) {
321			print STDERR "# same: $path $mtime\n" if ($verbose);
322			return unless($force);
323	dpavlin	64	} else {
324	dpavlin	90	print STDERR "# changed: $path $mtime != $mtime_db\n" if ($debug);
325			}
326	dpavlin	64
327	dpavlin	90	# skip files on which File::MMagic::XS croaks
328			if ($path =~ m#\.au$#) {
329			warn "skipped '$path' to prevent File::MMagic::XS bug\n" if ($debug);
330			return;
331			}
332
333			my $type = $mm->checktype_filename($path);
334			$type =~ s/\s+/ /gs;
335
336			print STDERR "# $path $type\n" if ($debug);
337
338			if ($type =~ m/pdf/i) {
339			if ($filter->{pdftotext}) {
340			filter_to_pages($path, $mtime, qq( $filter->{pdftotext} -htmlmeta "$path" - ));
341			} else {
342			warn "skipping '$path', no pdftotext filter\n" if ($verbose);
343			return;
344			}
345			} elsif ($type eq 'application/postscript') {
346			if ($filter->{pstotext}) {
347			filter_to_pages($path, $mtime, qq( $filter->{pstotext} "$path" ));
348			} else {
349			warn "skipping '$path', no pstotext filter\n" if ($verbose);
350			return;
351			}
352	dpavlin	196	} elsif ($type =~ m!^image/! && $skip_images) {
353			warn "skipping image '$path'\n" if ($verbose);
354			return; # don't index images
355	dpavlin	90	} else {
356
357	dpavlin	64	# return if (! -f $path \|\| ! m/\.(html*\|php\|pl\|txt\|info\|log\|text)$/i);
358	dpavlin	119	if (-f $path &&
359			$type !~ m/html/ &&
360			$path !~ m/\.(php\|pl\|txt\|info\|log\|text)$/io
361	dpavlin	77	) {
362	dpavlin	119	dump_contents($db, '', $mtime, $path, -s $path) if ($all);
363	dpavlin	90	return;
364	dpavlin	77	}
365	dpavlin	64
366			# skip index files
367	dpavlin	119	return if ($path =~ m/index_(?:[a-z]+\|symbol)\.html*/i);
368	dpavlin	64
369			open(F,"$path") \|\| die "can't open file: $path";
370			print STDERR "$path ($type)" if ($verbose);
371			while(<F>) {
372			$contents .= "$_";
373			}
374			$contents .= "\n\n";
375
376			#$contents = filter($contents,$collection);
377
378			# add optional components to path
379			$path .= " $path_add" if ($path_add);
380
381			dump_contents($db, $contents, $mtime, $path);
382			}
383
384			print STDERR "\n" if ($verbose);
385			# die "zero size content in '$path'" if (! $contents);
386
387			}
388