1: #!/usr/bin/perl
2: # The LearningOnline Network
3: # searchcat.pl "Search Catalog" batch script
4: #
5: # $Id: searchcat.pl,v 1.24 2002/10/18 13:54:31 www Exp $
6: #
7: # Copyright Michigan State University Board of Trustees
8: #
9: # This file is part of the LearningOnline Network with CAPA (LON-CAPA).
10: #
11: # LON-CAPA is free software; you can redistribute it and/or modify
12: # it under the terms of the GNU General Public License as published by
13: # the Free Software Foundation; either version 2 of the License, or
14: # (at your option) any later version.
15: #
16: # LON-CAPA is distributed in the hope that it will be useful,
17: # but WITHOUT ANY WARRANTY; without even the implied warranty of
18: # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
19: # GNU General Public License for more details.
20: #
21: # You should have received a copy of the GNU General Public License
22: # along with LON-CAPA; if not, write to the Free Software
23: # Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
24: #
25: # /home/httpd/html/adm/gpl.txt
26: #
27: # http://www.lon-capa.org/
28: #
29: # YEAR=2001
30: # 04/14/2001, 04/16/2001 Scott Harrison
31: #
32: # YEAR=2002
33: # 05/11/2002 Scott Harrison
34: #
35: ###
36:
37: # This script goes through a LON-CAPA resource
38: # directory and gathers metadata.
39: # The metadata is entered into a SQL database.
40:
41: use lib '/home/httpd/lib/perl/';
42: use LONCAPA::Configuration;
43:
44: use IO::File;
45: use HTML::TokeParser;
46: use DBI;
47: use GDBM_File;
48: use POSIX qw(strftime mktime);
49:
50: my @metalist;
51:
52:
53: # ----------------------------------------------------- Un-Escape Special Chars
54:
55: sub unescape {
56: my $str=shift;
57: $str =~ s/%([a-fA-F0-9][a-fA-F0-9])/pack("C",hex($1))/eg;
58: return $str;
59: }
60:
61: # -------------------------------------------------------- Escape Special Chars
62:
63: sub escape {
64: my $str=shift;
65: $str =~ s/(\W)/"%".unpack('H2',$1)/eg;
66: return $str;
67: }
68:
69:
70: # ------------------------------------------- Code to evaluate dynamic metadata
71:
72: sub dynamicmeta {
73: #
74: #
75: # Do nothing for now ...
76: #
77: #
78: return;
79: #
80: # ..., but stuff below already works
81: #
82: my $url=&declutter(shift);
83: $url=~s/\.meta$//;
84: my %returnhash=();
85: my ($adomain,$aauthor)=($url=~/^(\w+)\/(\w+)\//);
86: my $prodir=&propath($adomain,$aauthor);
87: if (tie(%evaldata,'GDBM_File',
88: $prodir.'/nohist_resevaldata.db',&GDBM_WRCREAT(),0640)) {
89: my %sum=();
90: my %cnt=();
91: my %listitems=('count' => 'add',
92: 'course' => 'add',
93: 'avetries' => 'avg',
94: 'stdno' => 'add',
95: 'difficulty' => 'avg',
96: 'clear' => 'avg',
97: 'technical' => 'avg',
98: 'helpful' => 'avg',
99: 'correct' => 'avg',
100: 'depth' => 'avg',
101: 'comments' => 'app',
102: 'usage' => 'cnt'
103: );
104: my $regexp=$url;
105: $regexp=~s/(\W)/\\$1/g;
106: $regexp='___'.$regexp.'___([a-z]+)$';
107: foreach (keys %evaldata) {
108: my $key=&unescape($_);
109: if ($key=~/$regexp/) {
110: my $ctype=$1;
111: if (defined($cnt{$ctype})) {
112: $cnt{$ctype}++;
113: } else {
114: $cnt{$ctype}=1;
115: }
116: unless ($listitems{$ctype} eq 'app') {
117: if (defined($sum{$ctype})) {
118: $sum{$ctype}+=$evaldata{$_};
119: } else {
120: $sum{$ctype}=$evaldata{$_};
121: }
122: } else {
123: if (defined($sum{$ctype})) {
124: if ($evaldata{$_}) {
125: $sum{$ctype}.='<hr>'.$evaldata{$_};
126: }
127: } else {
128: $sum{$ctype}=''.$evaldata{$_};
129: }
130: }
131: if ($ctype eq 'count') {
132: delete($evaldata{$_});
133: }
134: }
135: }
136: foreach (keys %cnt) {
137: if ($listitems{$_} eq 'avg') {
138: $returnhash{$_}=int(($sum{$_}/$cnt{$_})*100.0+0.5)/100.0;
139: } elsif ($listitems{$_} eq 'cnt') {
140: $returnhash{$_}=$cnt{$_};
141: } else {
142: $returnhash{$_}=$sum{$_};
143: }
144: }
145: if ($returnhash{'count'}) {
146: my $newkey=$$.'_'.time.'_searchcat___'.&escape($url).'___count';
147: $evaldata{$newkey}=$returnhash{'count'};
148: }
149: untie(%evaldata);
150: }
151: return %returnhash;
152: }
153:
154: # ----------------- Code to enable 'find' subroutine listing of the .meta files
155: require "find.pl";
156: sub wanted {
157: (($dev,$ino,$mode,$nlink,$uid,$gid) = lstat($_)) &&
158: -f _ &&
159: /^.*\.meta$/ && !/^.+\.\d+\.[^\.]+\.meta$/ &&
160: push(@metalist,"$dir/$_");
161: }
162:
163: # --------------- Read loncapa_apache.conf and loncapa.conf and get variables
164: my $perlvarref=LONCAPA::Configuration::read_conf('loncapa.conf');
165: my %perlvar=%{$perlvarref};
166: undef $perlvarref; # remove since sensitive and not needed
167: delete $perlvar{'lonReceipt'}; # remove since sensitive and not needed
168:
169: # ------------------------------------- Only run if machine is a library server
170: exit unless $perlvar{'lonRole'} eq 'library';
171:
172: my $dbh;
173: # ------------------------------------- Make sure that database can be accessed
174: {
175: unless (
176: $dbh = DBI->connect("DBI:mysql:loncapa","www",$perlvar{'lonSqlAccess'},{ RaiseError =>0,PrintError=>0})
177: ) {
178: print "Cannot connect to database!\n";
179: exit;
180: }
181: my $make_metadata_table = "CREATE TABLE IF NOT EXISTS metadata (".
182: "title TEXT, author TEXT, subject TEXT, url TEXT, keywords TEXT, ".
183: "version TEXT, notes TEXT, abstract TEXT, mime TEXT, language TEXT, ".
184: "creationdate DATETIME, lastrevisiondate DATETIME, owner TEXT, ".
185: "copyright TEXT, FULLTEXT idx_title (title), ".
186: "FULLTEXT idx_author (author), FULLTEXT idx_subject (subject), ".
187: "FULLTEXT idx_url (url), FULLTEXT idx_keywords (keywords), ".
188: "FULLTEXT idx_version (version), FULLTEXT idx_notes (notes), ".
189: "FULLTEXT idx_abstract (abstract), FULLTEXT idx_mime (mime), ".
190: "FULLTEXT idx_language (language), FULLTEXT idx_owner (owner), ".
191: "FULLTEXT idx_copyright (copyright)) TYPE=MYISAM";
192: # It would sure be nice to have some logging mechanism.
193: $dbh->do($make_metadata_table);
194: }
195:
196: # ------------------------------------------------------------- get .meta files
197: opendir(RESOURCES,"$perlvar{'lonDocRoot'}/res/$perlvar{'lonDefDomain'}");
198: my @homeusers=grep
199: {&ishome("$perlvar{'lonDocRoot'}/res/$perlvar{'lonDefDomain'}/$_")}
200: grep {!/^\.\.?$/} readdir(RESOURCES);
201: closedir RESOURCES;
202: foreach my $user (@homeusers) {
203: &find("$perlvar{'lonDocRoot'}/res/$perlvar{'lonDefDomain'}/$user");
204: }
205:
206: # -- process each file to get metadata and put into search catalog SQL database
207: # Also, check to see if already there.
208: # I could just delete (without searching first), but this works for now.
209: foreach my $m (@metalist) {
210: my $ref=&metadata($m);
211: my $m2='/res/'.&declutter($m);
212: $m2=~s/\.meta$//;
213: &dynamicmeta($m2);
214: my $q2="select * from metadata where url like binary '$m2'";
215: my $sth = $dbh->prepare($q2);
216: $sth->execute();
217: my $r1=$sth->fetchall_arrayref;
218: if (@$r1) {
219: $sth=$dbh->prepare("delete from metadata where url like binary '$m2'");
220: $sth->execute();
221: }
222: $sth=$dbh->prepare('insert into metadata values ('.
223: '"'.delete($ref->{'title'}).'"'.','.
224: '"'.delete($ref->{'author'}).'"'.','.
225: '"'.delete($ref->{'subject'}).'"'.','.
226: '"'.$m2.'"'.','.
227: '"'.delete($ref->{'keywords'}).'"'.','.
228: '"'.'current'.'"'.','.
229: '"'.delete($ref->{'notes'}).'"'.','.
230: '"'.delete($ref->{'abstract'}).'"'.','.
231: '"'.delete($ref->{'mime'}).'"'.','.
232: '"'.delete($ref->{'language'}).'"'.','.
233: '"'.sqltime(delete($ref->{'creationdate'})).'"'.','.
234: '"'.sqltime(delete($ref->{'lastrevisiondate'})).'"'.','.
235: '"'.delete($ref->{'owner'}).'"'.','.
236: '"'.delete($ref->{'copyright'}).'"'.')');
237: $sth->execute();
238: }
239:
240: # ----------------------------------------------------------- Clean up database
241: # Need to, perhaps, remove stale SQL database records.
242: # ... not yet implemented
243:
244: # --------------------------------------------------- Close database connection
245: $dbh->disconnect;
246:
247: # ---------------------------------------------------------------- Get metadata
248: # significantly altered from subroutine present in lonnet
249: sub metadata {
250: my ($uri,$what)=@_;
251: my %metacache;
252: $uri=&declutter($uri);
253: my $filename=$uri;
254: $uri=~s/\.meta$//;
255: $uri='';
256: unless ($metacache{$uri.'keys'}) {
257: unless ($filename=~/\.meta$/) { $filename.='.meta'; }
258: my $metastring=&getfile($perlvar{'lonDocRoot'}.'/res/'.$filename);
259: my $parser=HTML::TokeParser->new(\$metastring);
260: my $token;
261: while ($token=$parser->get_token) {
262: if ($token->[0] eq 'S') {
263: my $entry=$token->[1];
264: my $unikey=$entry;
265: if (defined($token->[2]->{'part'})) {
266: $unikey.='_'.$token->[2]->{'part'};
267: }
268: if (defined($token->[2]->{'name'})) {
269: $unikey.='_'.$token->[2]->{'name'};
270: }
271: if ($metacache{$uri.'keys'}) {
272: $metacache{$uri.'keys'}.=','.$unikey;
273: } else {
274: $metacache{$uri.'keys'}=$unikey;
275: }
276: map {
277: $metacache{$uri.''.$unikey.'.'.$_}=$token->[2]->{$_};
278: } @{$token->[3]};
279: unless (
280: $metacache{$uri.''.$unikey}=$parser->get_text('/'.$entry)
281: ) { $metacache{$uri.''.$unikey}=
282: $metacache{$uri.''.$unikey.'.default'};
283: }
284: }
285: }
286: }
287: return \%metacache;
288: }
289:
290: # ------------------------------------------------------------ Serves up a file
291: # returns either the contents of the file or a -1
292: sub getfile {
293: my $file=shift;
294: if (! -e $file ) { return -1; };
295: my $fh=IO::File->new($file);
296: my $a='';
297: while (<$fh>) { $a .=$_; }
298: return $a
299: }
300:
301: # ------------------------------------------------------------- Declutters URLs
302: sub declutter {
303: my $thisfn=shift;
304: $thisfn=~s/^$perlvar{'lonDocRoot'}//;
305: $thisfn=~s/^\///;
306: $thisfn=~s/^res\///;
307: return $thisfn;
308: }
309:
310: # --------------------------------------- Is this the home server of an author?
311: # (copied from lond, modification of the return value)
312: sub ishome {
313: my $author=shift;
314: $author=~s/\/home\/httpd\/html\/res\/([^\/]*)\/([^\/]*).*/$1\/$2/;
315: my ($udom,$uname)=split(/\//,$author);
316: my $proname=propath($udom,$uname);
317: if (-e $proname) {
318: return 1;
319: } else {
320: return 0;
321: }
322: }
323:
324: # -------------------------------------------- Return path to profile directory
325: # (copied from lond)
326: sub propath {
327: my ($udom,$uname)=@_;
328: $udom=~s/\W//g;
329: $uname=~s/\W//g;
330: my $subdir=$uname.'__';
331: $subdir =~ s/(.)(.)(.).*/$1\/$2\/$3/;
332: my $proname="$perlvar{'lonUsersDir'}/$udom/$subdir/$uname";
333: return $proname;
334: }
335:
336: # ---------------------------- convert 'time' format into a datetime sql format
337: sub sqltime {
338: my ($sec,$min,$hour,$mday,$mon,$year,$wday,$yday,$isdst) =
339: localtime(&unsqltime(@_[0]));
340: $mon++; $year+=1900;
341: return "$year-$mon-$mday $hour:$min:$sec";
342: }
343:
344: sub maketime {
345: my %th=@_;
346: return POSIX::mktime(
347: ($th{'seconds'},$th{'minutes'},$th{'hours'},
348: $th{'day'},$th{'month'}-1,$th{'year'}-1900,0,0,$th{'dlsav'}));
349: }
350:
351:
352: #########################################
353: #
354: # Retro-fixing of un-backward-compatible time format
355:
356: sub unsqltime {
357: my $timestamp=shift;
358: if ($timestamp=~/^(\d+)\-(\d+)\-(\d+)\s+(\d+)\:(\d+)\:(\d+)$/) {
359: $timestamp=&maketime(
360: 'year'=>$1,'month'=>$2,'day'=>$3,
361: 'hours'=>$4,'minutes'=>$5,'seconds'=>$6);
362: }
363: return $timestamp;
364: }
365:
FreeBSD-CVSweb <freebsd-cvsweb@FreeBSD.org>
![[BACK]](/icons/back.gif){kind=icon}
Return to searchcat.pl CVS log ![[TXT]](/icons/text.gif){kind=icon}
![[DIR]](/icons/dir.gif){kind=icon}
Up to [LON-CAPA] / loncom / metadata_database