Annotation of /trunk/perl/scripts/est-spider

#!/usr/bin/perl -w
use strict;
use File::Find;
use Getopt::Long;
use File::Which;
use HyperEstraier;
use Text::Iconv;
#use File::MMagic;
use File::MMagic::XS qw/:compat/;

# do we use Node API?
my $node_url;

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 $result = GetOptions(
        "collection=s" => \$collection,
        "path=s" => \$path_add,
        "verbose!" => \$verbose,
        "debug!" => \$debug,
        "exclude=s" => \$exclude,
        "node=s" => \$node_url,
        "force!" => \$force,
);

my $dir = shift @ARGV || die "usage: $0 [dir]";

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

my $pdftotext = which('pdftotext');

#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;

print STDERR "using $pdftotext to convert pdf into html\n" if ($pdftotext && $verbose);

my $db;
if ($node_url) {
        $db = HyperEstraier::Node->new($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;
}

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

unless ($node_url) {
        print "--- sync\n";
        $db->sync();

        print "--- optimize...\n";
        $db->optimize(0);
}
exit;

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

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

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

        use bytes;
        my $size = length $contents;

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

        # create a document object 
        my $doc = HyperEstraier::Document->new;

        my $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);

        # html2text
        $contents =~ s#<[^>]+/*>##gs;
        $contents =~ s#\s\s+# #gs;

        $doc->add_text($iconv->convert($contents));

#       print $doc->dump_draft if ($verbose);

        # 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;

        return if (! $force && -l $path || $path =~ m#/.svn# || $path =~ m/(~|.bak)$/);

        my $mtime = (stat($path))[9];
        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) {

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

                open(F,"$pdftotext -htmlmeta \"$path\" - |") || die "can't open $pdftotext with '$path'";
                my $html;
                while(<F>) {
                        # XXX why pdftotext barks if I try to use this is beyond me.
                        #$contents .= $_;

                        $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;
                }

                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++;
                }

        } else {

#               return if (! -f $path || ! m/\.(html*|php|pl|txt|info|log|text)$/i);
                return unless (-f $path && $type =~ m/html/ ||
                        ($type =~ m#text# && $path =~ m/\.(php|pl|txt|info|log|text)$/io)
                );

                # skip index files
                return if (m/index_[a-z]\.html*/i || m/index_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	5	#!/usr/bin/perl -w
2			use strict;
3			use File::Find;
4			use Getopt::Long;
5			use File::Which;
6			use HyperEstraier;
7			use Text::Iconv;
8	dpavlin	27	#use File::MMagic;
9			use File::MMagic::XS qw/:compat/;
10	dpavlin	5
11	dpavlin	18	# do we use Node API?
12			my $node_url;
13
14	dpavlin	5	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	dpavlin	28	my $debug = 0;
21	dpavlin	29	my $force = 0;
22	dpavlin	5
23			my $result = GetOptions(
24			"collection=s" => \$collection,
25			"path=s" => \$path_add,
26			"verbose!" => \$verbose,
27	dpavlin	28	"debug!" => \$debug,
28	dpavlin	5	"exclude=s" => \$exclude,
29	dpavlin	18	"node=s" => \$node_url,
30	dpavlin	29	"force!" => \$force,
31	dpavlin	5	);
32
33			my $dir = shift @ARGV \|\| die "usage: $0 [dir]";
34
35			#my $basedir = $0;
36			#$basedir =~ s,/[^/]+$,/,;
37			#require "$basedir/filter.pm";
38
39			my $pdftotext = which('pdftotext');
40
41	dpavlin	27	#my $mm = new File::MMagic('/usr/share/misc/file/magic');
42			my $mm = new File::MMagic::XS();
43
44	dpavlin	5	my $iconv = new Text::Iconv('iso-8859-2', 'utf-8');
45
46			select(STDERR); $\|=1;
47			select(STDOUT); $\|=1;
48
49			print STDERR "using $pdftotext to convert pdf into html\n" if ($pdftotext && $verbose);
50
51	dpavlin	18	my $db;
52			if ($node_url) {
53			$db = HyperEstraier::Node->new($node_url);
54			$db->set_auth('admin', 'admin');
55			} else {
56			# open the database
57			$db = HyperEstraier::Database->new();
58			$db->open('/tmp/casket', $HyperEstraier::Database::DBWRITER \| $HyperEstraier::Database::DBCREAT);
59	dpavlin	5
60	dpavlin	18	sub signal {
61			my($sig) = @_;
62			print "\nCaught a SIG$sig--syncing database and shutting down\n";
63			$db->sync();
64			exit(0);
65			}
66	dpavlin	5
67	dpavlin	18	$SIG{'INT'} = \&signal;
68			$SIG{'QUIT'} = \&signal;
69	dpavlin	5	}
70
71			find({ wanted => \&file,
72			follow => 1,
73	dpavlin	26	follow_skip => 2,
74			no_chdir => 1,
75	dpavlin	5	}, $dir);
76
77	dpavlin	18	unless ($node_url) {
78			print "--- sync\n";
79			$db->sync();
80	dpavlin	5
81	dpavlin	18	print "--- optimize...\n";
82			$db->optimize(0);
83			}
84	dpavlin	5	exit;
85
86			sub dump_contents($$$$) {
87			my ($db,$contents,$mtime,$path) = @_;
88
89			return unless ($contents); # don't die on empty files
90
91			if ($exclude && $path =~ m/$exclude/i) {
92			print STDERR "skip: $path\n" if ($verbose);
93			return;
94			}
95
96			use bytes;
97			my $size = length $contents;
98
99			print STDERR " [$size]" if ($verbose);
100
101			# create a document object
102			my $doc = HyperEstraier::Document->new;
103
104	dpavlin	29	my $title = $1 if ($contents =~ m#<title>(.+?)</title>#is);
105	dpavlin	5
106	dpavlin	27	# chop long titles to 100 chars
107			$title = substr($title, 0, 100) . '...' if ($title && length($title) > 100);
108			# use path if no title is found
109			$title \|\|= $path;
110
111	dpavlin	5	# add attributes to the document object
112			$doc->add_attr('@uri', "file:///$path");
113	dpavlin	27	$doc->add_attr('@title', $iconv->convert($title));
114	dpavlin	5	$doc->add_attr('@size', $size);
115			$doc->add_attr('@mtime', $mtime);
116
117			# html2text
118			$contents =~ s#<[^>]+/*>##gs;
119			$contents =~ s#\s\s+# #gs;
120
121			$doc->add_text($iconv->convert($contents));
122
123			# print $doc->dump_draft if ($verbose);
124
125			# register the document object to the database
126	dpavlin	18	if ($node_url) {
127			$db->put_doc($doc);
128			} else {
129			$db->put_doc($doc, $HyperEstraier::Database::PDCLEAN);
130			}
131	dpavlin	5
132			}
133
134			sub file {
135
136			my $path = $_;
137			my $contents;
138
139	dpavlin	29	return if (! $force && -l $path \|\| $path =~ m#/.svn# \|\| $path =~ m/(~\|.bak)$/);
140	dpavlin	5
141	dpavlin	28	my $mtime = (stat($path))[9];
142			my $mtime_db = $db->get_doc_attr_by_uri("file:///$path", '@mtime') \|\| -2;
143
144			if ($mtime == $mtime_db) {
145			print STDERR "# same: $path $mtime\n" if ($verbose);
146	dpavlin	29	return unless($force);
147	dpavlin	28	} else {
148			print STDERR "# changed: $path $mtime != $mtime_db\n" if ($debug);
149			}
150
151			# skip files on which File::MMagic::XS croaks
152			return if ($path =~ m#\.au$#);
153
154	dpavlin	27	my $type = $mm->checktype_filename($path);
155			$type =~ s/\s+/ /gs;
156	dpavlin	5
157	dpavlin	28	print STDERR "# $path $type\n" if ($debug);
158	dpavlin	27
159			if ($pdftotext && -f $path && $type =~ m/pdf/i) {
160
161	dpavlin	5	print STDERR "$path {converting}" if ($verbose);
162
163			open(F,"$pdftotext -htmlmeta \"$path\" - \|") \|\| die "can't open $pdftotext with '$path'";
164			my $html;
165			while(<F>) {
166			# XXX why pdftotext barks if I try to use this is beyond me.
167			#$contents .= $_;
168
169			$html .= $_;
170			}
171			close(F);
172
173			return if (! $html);
174
175			my $file_only = $path;
176			$file_only =~ s/^.*\/([^\/]+)$/$1/g;
177
178			my ($pre_html,$pages,$post_html) = ('<html><head><title>$path :: page ##page_nr##</title></head><body><pre>',$html,'</pre></body></html>');
179
180			($pre_html,$pages,$post_html) = ($1,$2,$3) if ($html =~ m/^(<html>.+?<pre>)(.+)(<\/pre>.+?)$/si);
181
182			if ($collection) {
183			$pre_html =~ s/<title>(.+?)<\/title>/<title>$collection :: page ##page_nr##<\/title>/si;
184			} else {
185			$pre_html =~ s/<title>(.+?)<\/title>/<title>$1 :: page ##page_nr##<\/title>/si \|\|
186			$pre_html =~ s/<title><\/title>/<title>$file_only :: page ##page_nr##<\/title>/si;
187			}
188
189			my $page_nr = 1;
190			foreach my $page (split(/\f/s,$pages)) {
191			print STDERR " $page_nr" if ($verbose);
192			my $pre_tmp = $pre_html;
193			$pre_tmp =~ s/##page_nr##/$page_nr<\/title>/s;
194	dpavlin	30	dump_contents($db, $pre_tmp . $page . $post_html, $mtime, "$path#$page_nr") if ($page !~ m/^\s*$/s);
195	dpavlin	5	$page_nr++;
196			}
197
198			} else {
199
200	dpavlin	27	# return if (! -f $path \|\| ! m/\.(html*\|php\|pl\|txt\|info\|log\|text)$/i);
201	dpavlin	28	return unless (-f $path && $type =~ m/html/ \|\|
202			($type =~ m#text# && $path =~ m/\.(php\|pl\|txt\|info\|log\|text)$/io)
203			);
204	dpavlin	5
205			# skip index files
206			return if (m/index_[a-z]\.html/i \|\| m/index_symbol\.html/i);
207
208			open(F,"$path") \|\| die "can't open file: $path";
209	dpavlin	27	print STDERR "$path ($type)" if ($verbose);
210	dpavlin	5	while(<F>) {
211			$contents .= "$_";
212			}
213			$contents .= "\n\n";
214
215			#$contents = filter($contents,$collection);
216
217			# add optional components to path
218			$path .= " $path_add" if ($path_add);
219
220	dpavlin	28	dump_contents($db, $contents, $mtime, $path);
221	dpavlin	5	}
222
223			print STDERR "\n" if ($verbose);
224			# die "zero size content in '$path'" if (! $contents);
225
226			}
227