1: #!/usr/bin/perl
2: # The LearningOnline Network
3: # searchcat.pl "Search Catalog" batch script
4: #
5: # $Id: searchcat.pl,v 1.38 2003/08/21 00:19:05 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: ###
30:
31: =pod
32:
33: =head1 NAME
34:
35: B<searchcat.pl> - put authoritative filesystem data into sql database.
36:
37: =head1 SYNOPSIS
38:
39: Ordinarily this script is to be called from a loncapa cron job
40: (CVS source location: F<loncapa/loncom/cron/loncapa>; typical
41: filesystem installation location: F</etc/cron.d/loncapa>).
42:
43: Here is the cron job entry.
44:
45: C<# Repopulate and refresh the metadata database used for the search catalog.>
46: C<10 1 * * 7 www /home/httpd/perl/searchcat.pl>
47:
48: This script only allows itself to be run as the user C<www>.
49:
50: =head1 DESCRIPTION
51:
52: This script goes through a loncapa resource directory and gathers metadata.
53: The metadata is entered into a SQL database.
54:
55: This script also does general database maintenance such as reformatting
56: the C<loncapa:metadata> table if it is deprecated.
57:
58: This script evaluates dynamic metadata from the authors'
59: F<nohist_resevaldata.db> database file in order to store it in MySQL, as
60: well as to compress the filesize (add up all "count"-type metadata).
61:
62: This script is playing an increasingly important role for a loncapa
63: library server. The proper operation of this script is critical for a smooth
64: and correct user experience.
65:
66: =cut
67:
68: use lib '/home/httpd/lib/perl/';
69: use LONCAPA::Configuration;
70:
71: use IO::File;
72: use HTML::TokeParser;
73: use DBI;
74: use GDBM_File;
75: use POSIX qw(strftime mktime);
76:
77: my @metalist;
78:
79: $simplestatus='';
80: my %countext=();
81:
82: sub writesimple {
83: open(SMP,'>/home/httpd/html/lon-status/mysql.txt');
84: print SMP $simplestatus."\n";
85: close(SMP);
86: }
87:
88: sub writecount {
89: open(RSMP,'>/home/httpd/html/lon-status/rescount.txt');
90: foreach (keys %countext) {
91: print RSMP $_.'='.$countext{$_}.'&';
92: }
93: print RSMP 'time='.time."\n";
94: close(RSMP);
95: }
96:
97: sub count {
98: my $file=shift;
99: $file=~/\.(\w+)$/;
100: my $ext=lc($1);
101: if (defined($countext{$ext})) {
102: $countext{$ext}++;
103: } else {
104: $countext{$ext}=1;
105: }
106: }
107: # ----------------------------------------------------- Un-Escape Special Chars
108:
109: sub unescape {
110: my $str=shift;
111: $str =~ s/%([a-fA-F0-9][a-fA-F0-9])/pack("C",hex($1))/eg;
112: return $str;
113: }
114:
115: # -------------------------------------------------------- Escape Special Chars
116:
117: sub escape {
118: my $str=shift;
119: $str =~ s/(\W)/"%".unpack('H2',$1)/eg;
120: return $str;
121: }
122:
123:
124: # ------------------------------------------- Code to evaluate dynamic metadata
125:
126: sub dynamicmeta {
127:
128: my $url=&declutter(shift);
129: $url=~s/\.meta$//;
130: my %returnhash=();
131: my ($adomain,$aauthor)=($url=~/^(\w+)\/(\w+)\//);
132: my $prodir=&propath($adomain,$aauthor);
133: if ((tie(%evaldata,'GDBM_File',
134: $prodir.'/nohist_resevaldata.db',&GDBM_READER(),0640)) &&
135: (tie(%newevaldata,'GDBM_File',
136: $prodir.'/nohist_new_resevaldata.db',&GDBM_WRCREAT(),0640))) {
137: my %sum=();
138: my %cnt=();
139: my %listitems=('count' => 'add',
140: 'course' => 'add',
141: 'avetries' => 'avg',
142: 'stdno' => 'add',
143: 'difficulty' => 'avg',
144: 'clear' => 'avg',
145: 'technical' => 'avg',
146: 'helpful' => 'avg',
147: 'correct' => 'avg',
148: 'depth' => 'avg',
149: 'comments' => 'app',
150: 'usage' => 'cnt'
151: );
152: my $regexp=$url;
153: $regexp=~s/(\W)/\\$1/g;
154: $regexp='___'.$regexp.'___([a-z]+)$';
155: foreach (keys %evaldata) {
156: my $key=&unescape($_);
157: if ($key=~/$regexp/) {
158: my $ctype=$1;
159: if (defined($cnt{$ctype})) {
160: $cnt{$ctype}++;
161: } else {
162: $cnt{$ctype}=1;
163: }
164: unless ($listitems{$ctype} eq 'app') {
165: if (defined($sum{$ctype})) {
166: $sum{$ctype}+=$evaldata{$_};
167: } else {
168: $sum{$ctype}=$evaldata{$_};
169: }
170: } else {
171: if (defined($sum{$ctype})) {
172: if ($evaldata{$_}) {
173: $sum{$ctype}.='<hr>'.$evaldata{$_};
174: }
175: } else {
176: $sum{$ctype}=''.$evaldata{$_};
177: }
178: }
179: if ($ctype ne 'count') {
180: $newevaldata{$_}=$evaldata{$_};
181: }
182: }
183: }
184: foreach (keys %cnt) {
185: if ($listitems{$_} eq 'avg') {
186: $returnhash{$_}=int(($sum{$_}/$cnt{$_})*100.0+0.5)/100.0;
187: } elsif ($listitems{$_} eq 'cnt') {
188: $returnhash{$_}=$cnt{$_};
189: } else {
190: $returnhash{$_}=$sum{$_};
191: }
192: }
193: if ($returnhash{'count'}) {
194: my $newkey=$$.'_'.time.'_searchcat___'.&escape($url).'___count';
195: $newevaldata{$newkey}=$returnhash{'count'};
196: }
197: untie(%evaldata);
198: untie(%newevaldata);
199: }
200: return %returnhash;
201: }
202:
203: # ----------------- Code to enable 'find' subroutine listing of the .meta files
204: require "find.pl";
205: sub wanted {
206: (($dev,$ino,$mode,$nlink,$uid,$gid) = lstat($_)) &&
207: -f _ &&
208: /^.*\.meta$/ && !/^.+\.\d+\.[^\.]+\.meta$/ &&
209: push(@metalist,"$dir/$_");
210: }
211:
212: # --------------- Read loncapa_apache.conf and loncapa.conf and get variables
213: my $perlvarref=LONCAPA::Configuration::read_conf('loncapa.conf');
214: my %perlvar=%{$perlvarref};
215: undef $perlvarref; # remove since sensitive and not needed
216: delete $perlvar{'lonReceipt'}; # remove since sensitive and not needed
217:
218: # ------------------------------------- Only run if machine is a library server
219: exit unless $perlvar{'lonRole'} eq 'library';
220:
221: # ----------------------------- Make sure this process is running from user=www
222:
223: my $wwwid=getpwnam('www');
224: if ($wwwid!=$<) {
225: $emailto="$perlvar{'lonAdmEMail'},$perlvar{'lonSysEMail'}";
226: $subj="LON: $perlvar{'lonHostID'} User ID mismatch";
227: system("echo 'User ID mismatch. searchcat.pl must be run as user www.' |\
228: mailto $emailto -s '$subj' > /dev/null");
229: exit 1;
230: }
231:
232:
233: # ---------------------------------------------------------- We are in business
234:
235: open(LOG,'>'.$perlvar{'lonDaemons'}.'/logs/searchcat.log');
236: print LOG '==== Searchcat Run '.localtime()."====\n\n";
237: $simplestatus='time='.time.'&';
238: my $dbh;
239: # ------------------------------------- Make sure that database can be accessed
240: {
241: unless (
242: $dbh = DBI->connect("DBI:mysql:loncapa","www",$perlvar{'lonSqlAccess'},{ RaiseError =>0,PrintError=>0})
243: ) {
244: print LOG "Cannot connect to database!\n";
245: $simplestatus.='mysql=defunct';
246: &writesimple();
247: exit;
248: }
249:
250: my $make_metadata_table = "CREATE TABLE IF NOT EXISTS metadata (".
251: "title TEXT, author TEXT, subject TEXT, url TEXT, keywords TEXT, ".
252: "version TEXT, notes TEXT, abstract TEXT, mime TEXT, language TEXT, ".
253: "creationdate DATETIME, lastrevisiondate DATETIME, owner TEXT, ".
254: "copyright TEXT, FULLTEXT idx_title (title), ".
255: "FULLTEXT idx_author (author), FULLTEXT idx_subject (subject), ".
256: "FULLTEXT idx_url (url), FULLTEXT idx_keywords (keywords), ".
257: "FULLTEXT idx_version (version), FULLTEXT idx_notes (notes), ".
258: "FULLTEXT idx_abstract (abstract), FULLTEXT idx_mime (mime), ".
259: "FULLTEXT idx_language (language), FULLTEXT idx_owner (owner), ".
260: "FULLTEXT idx_copyright (copyright)) TYPE=MYISAM";
261: # It would sure be nice to have some logging mechanism.
262: $dbh->do($make_metadata_table);
263: }
264:
265: # ------------------------------------------------------------- get .meta files
266: opendir(RESOURCES,"$perlvar{'lonDocRoot'}/res/$perlvar{'lonDefDomain'}");
267: my @homeusers = grep {
268: &ishome("$perlvar{'lonDocRoot'}/res/$perlvar{'lonDefDomain'}/$_")
269: } grep {!/^\.\.?$/} readdir(RESOURCES);
270: closedir RESOURCES;
271:
272: #
273: # Create the statement handlers we need
274: my $delete_sth = $dbh->prepare
275: ("DELETE FROM metadata WHERE url LIKE BINARY ?");
276:
277: my $insert_sth = $dbh->prepare
278: ("INSERT INTO metadata VALUES (".
279: "?,". # title
280: "?,". # author
281: "?,". # subject
282: "?,". # m2???
283: "?,". # version
284: "?,". # current
285: "?,". # notes
286: "?,". # abstract
287: "?,". # mime
288: "?,". # language
289: "?,". # creationdate
290: "?,". # revisiondate
291: "?,". # owner
292: "?)" # copyright
293: );
294:
295: foreach my $user (@homeusers) {
296: print LOG "\n=== User: ".$user."\n\n";
297: # Remove left-over db-files from potentially crashed searchcat run
298: my $prodir=&propath($perlvar{'lonDefDomain'},$user);
299: unlink($prodir.'/nohist_new_resevaldata.db');
300: # Use find.pl
301: undef @metalist;
302: @metalist=();
303: &find("$perlvar{'lonDocRoot'}/res/$perlvar{'lonDefDomain'}/$user");
304: # -- process each file to get metadata and put into search catalog SQL
305: # database. Also, check to see if already there.
306: # I could just delete (without searching first), but this works for now.
307: foreach my $m (@metalist) {
308: print LOG "- ".$m."\n";
309: my $ref=&metadata($m);
310: my $m2='/res/'.&declutter($m);
311: $m2=~s/\.meta$//;
312: &dynamicmeta($m2);
313: &count($m2);
314: $delete_sth->execute($m2);
315: $insert_sth->execute($ref->{'title'},
316: $ref->{'author'},
317: $ref->{'subject'},
318: $m2,
319: $ref->{'keywords'},
320: 'current',
321: $ref->{'notes'},
322: $ref->{'abstract'},
323: $ref->{'mime'},
324: $ref->{'language'},
325: sqltime($ref->{'creationdate'}),
326: sqltime($ref->{'lastrevisiondate'}),
327: $ref->{'owner'},
328: $ref->{'copyright'});
329: # if ($dbh->err()) {
330: # print STDERR "Error:".$dbh->errstr()."\n";
331: # }
332: $ref = undef;
333: }
334:
335: # --------------------------------------------------- Clean up database
336: # Need to, perhaps, remove stale SQL database records.
337: # ... not yet implemented
338:
339: # ------------------------------------------- Copy over the new db-files
340: #
341: # Check the size of nohist_new_resevaldata.db compared to
342: # nohist_resevaldata.db
343: my @stat_result = stat($prodir.'/nohist_new_resevaldata.db');
344: my $new_size = $stat_result[7];
345: @stat_result = stat($prodir.'/nohist_resevaldata.db');
346: my $old_size = $stat_result[7];
347: if ($old_size) {
348: if ($new_size/$old_size > 0.15 ) {
349: system('mv '.$prodir.'/nohist_new_resevaldata.db '.
350: $prodir.'/nohist_resevaldata.db');
351: } else {
352: print LOG "Size of '$user' old nohist_reseval: $old_size ".
353: "Size of new: $new_size. Not overwriting.\n";
354: my $emailto="$perlvar{'lonAdmEMail'},$perlvar{'lonSysEMail'}";
355: my $subj="LON: $perlvar{'lonHostID'} searchcat.pl $user reseval ".
356: "modification error.";
357: system("echo ".
358: "'See /home/httpd/perl/logs/searchcat.txt for information.' ".
359: "| mailto $emailto -s '$subj' > /dev/null");
360: }
361: }
362: }
363: # --------------------------------------------------- Close database connection
364: $dbh->disconnect;
365: print LOG "\n==== Searchcat completed ".localtime()." ====\n";
366: close(LOG);
367: &writesimple();
368: &writecount();
369: exit 0;
370:
371:
372:
373: # =============================================================================
374:
375: # ---------------------------------------------------------------- Get metadata
376: # significantly altered from subroutine present in lonnet
377: sub metadata {
378: my ($uri,$what)=@_;
379: my %metacache;
380: $uri=&declutter($uri);
381: my $filename=$uri;
382: $uri=~s/\.meta$//;
383: $uri='';
384: unless ($metacache{$uri.'keys'}) {
385: unless ($filename=~/\.meta$/) { $filename.='.meta'; }
386: my $metastring=&getfile($perlvar{'lonDocRoot'}.'/res/'.$filename);
387: my $parser=HTML::TokeParser->new(\$metastring);
388: my $token;
389: while ($token=$parser->get_token) {
390: if ($token->[0] eq 'S') {
391: my $entry=$token->[1];
392: my $unikey=$entry;
393: if (defined($token->[2]->{'part'})) {
394: $unikey.='_'.$token->[2]->{'part'};
395: }
396: if (defined($token->[2]->{'name'})) {
397: $unikey.='_'.$token->[2]->{'name'};
398: }
399: if ($metacache{$uri.'keys'}) {
400: $metacache{$uri.'keys'}.=','.$unikey;
401: } else {
402: $metacache{$uri.'keys'}=$unikey;
403: }
404: map {
405: $metacache{$uri.''.$unikey.'.'.$_}=$token->[2]->{$_};
406: } @{$token->[3]};
407: unless (
408: $metacache{$uri.''.$unikey}=$parser->get_text('/'.$entry)
409: ) { $metacache{$uri.''.$unikey}=
410: $metacache{$uri.''.$unikey.'.default'};
411: }
412: }
413: }
414: }
415: return \%metacache;
416: }
417:
418: # ------------------------------------------------------------ Serves up a file
419: # returns either the contents of the file or a -1
420: sub getfile {
421: my $file=shift;
422: if (! -e $file ) { return -1; };
423: my $fh=IO::File->new($file);
424: my $a='';
425: while (<$fh>) { $a .=$_; }
426: return $a;
427: }
428:
429: # ------------------------------------------------------------- Declutters URLs
430: sub declutter {
431: my $thisfn=shift;
432: $thisfn=~s/^$perlvar{'lonDocRoot'}//;
433: $thisfn=~s/^\///;
434: $thisfn=~s/^res\///;
435: return $thisfn;
436: }
437:
438: # --------------------------------------- Is this the home server of an author?
439: # (copied from lond, modification of the return value)
440: sub ishome {
441: my $author=shift;
442: $author=~s/\/home\/httpd\/html\/res\/([^\/]*)\/([^\/]*).*/$1\/$2/;
443: my ($udom,$uname)=split(/\//,$author);
444: my $proname=propath($udom,$uname);
445: if (-e $proname) {
446: return 1;
447: } else {
448: return 0;
449: }
450: }
451:
452: # -------------------------------------------- Return path to profile directory
453: # (copied from lond)
454: sub propath {
455: my ($udom,$uname)=@_;
456: $udom=~s/\W//g;
457: $uname=~s/\W//g;
458: my $subdir=$uname.'__';
459: $subdir =~ s/(.)(.)(.).*/$1\/$2\/$3/;
460: my $proname="$perlvar{'lonUsersDir'}/$udom/$subdir/$uname";
461: return $proname;
462: }
463:
464: # ---------------------------- convert 'time' format into a datetime sql format
465: sub sqltime {
466: my ($sec,$min,$hour,$mday,$mon,$year,$wday,$yday,$isdst) =
467: localtime(&unsqltime(@_[0]));
468: $mon++; $year+=1900;
469: return "$year-$mon-$mday $hour:$min:$sec";
470: }
471:
472: sub maketime {
473: my %th=@_;
474: return POSIX::mktime(($th{'seconds'},$th{'minutes'},$th{'hours'},
475: $th{'day'},$th{'month'}-1,
476: $th{'year'}-1900,0,0,$th{'dlsav'}));
477: }
478:
479:
480: #########################################
481: #
482: # Retro-fixing of un-backward-compatible time format
483:
484: sub unsqltime {
485: my $timestamp=shift;
486: if ($timestamp=~/^(\d+)\-(\d+)\-(\d+)\s+(\d+)\:(\d+)\:(\d+)$/) {
487: $timestamp=&maketime('year'=>$1,'month'=>$2,'day'=>$3,
488: 'hours'=>$4,'minutes'=>$5,'seconds'=>$6);
489: }
490: return $timestamp;
491: }
492:
FreeBSD-CVSweb <freebsd-cvsweb@FreeBSD.org>