File:
[LON-CAPA] /
loncom /
metadata_database /
searchcat.pl
Revision
1.34:
download - view:
text,
annotated -
select for diffs
Thu Jun 19 20:24:57 2003 UTC (21 years, 3 months ago) by
matthew
Branches:
MAIN
CVS tags:
HEAD
&dynamic_metadata
while(each) instead of keys, so we only read the db file once.
study the regular expression before we do a bunch of regex matches
metadata database update code:
Use statement handlers to do quoting
No longer bother to check if something is in the database before
telling MySQL to delete it - we just tell it to delete it and let
it do what must be done.
1: #!/usr/bin/perl
2: # The LearningOnline Network
3: # searchcat.pl "Search Catalog" batch script
4: #
5: # $Id: searchcat.pl,v 1.34 2003/06/19 20:24:57 matthew 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:
80: # ----------------------------------------------------- Un-Escape Special Chars
81:
82: sub unescape {
83: my $str=shift;
84: $str =~ s/%([a-fA-F0-9][a-fA-F0-9])/pack("C",hex($1))/eg;
85: return $str;
86: }
87:
88: # -------------------------------------------------------- Escape Special Chars
89:
90: sub escape {
91: my $str=shift;
92: $str =~ s/(\W)/"%".unpack('H2',$1)/eg;
93: return $str;
94: }
95:
96:
97: # ------------------------------------------- Code to evaluate dynamic metadata
98:
99: sub dynamicmeta {
100:
101: my $url=&declutter(shift);
102: $url=~s/\.meta$//;
103: my %returnhash=();
104: my ($adomain,$aauthor)=($url=~/^(\w+)\/(\w+)\//);
105: my $prodir=&propath($adomain,$aauthor);
106: if ((tie(%evaldata,'GDBM_File',
107: $prodir.'/nohist_resevaldata.db',&GDBM_READER(),0640)) &&
108: (tie(%newevaldata,'GDBM_File',
109: $prodir.'/nohist_new_resevaldata.db',&GDBM_WRCREAT(),0640))) {
110: my %sum=();
111: my %cnt=();
112: my %listitems=('count' => 'add',
113: 'course' => 'add',
114: 'avetries' => 'avg',
115: 'stdno' => 'add',
116: 'difficulty' => 'avg',
117: 'clear' => 'avg',
118: 'technical' => 'avg',
119: 'helpful' => 'avg',
120: 'correct' => 'avg',
121: 'depth' => 'avg',
122: 'comments' => 'app',
123: 'usage' => 'cnt'
124: );
125: my $regexp=$url;
126: $regexp=~s/(\W)/\\$1/g;
127: $regexp='___'.$regexp.'___([a-z]+)$';
128: study($regexp);
129: while (my ($key,$value) = each(%evaldata)) {
130: $key=&unescape($key);
131: next if ($key !~ /$regexp/);
132: my $ctype=$1;
133: if (defined($cnt{$ctype})) {
134: $cnt{$ctype}++;
135: } else {
136: $cnt{$ctype}=1;
137: }
138: unless ($listitems{$ctype} eq 'app') {
139: if (defined($sum{$ctype})) {
140: $sum{$ctype}+=$value;
141: } else {
142: $sum{$ctype}=$value;
143: }
144: } else {
145: if (defined($sum{$ctype})) {
146: if ($value) {
147: $sum{$ctype}.='<hr>'.$value;
148: }
149: } else {
150: $sum{$ctype}=''.$value;
151: }
152: }
153: if ($ctype ne 'count') {
154: $newevaldata{$_}=$value;
155: }
156: }
157: while (my($key,$value) = each(%cnt)) {
158: if ($listitems{$key} eq 'avg') {
159: $returnhash{$key}=int(($sum{$key}/$value)*100.0+0.5)/100.0;
160: } elsif ($listitems{$key} eq 'cnt') {
161: $returnhash{$key}=$value;
162: } else {
163: $returnhash{$key}=$sum{$key};
164: }
165: }
166: if ($returnhash{'count'}) {
167: my $newkey=$$.'_'.time.'_searchcat___'.&escape($url).'___count';
168: $newevaldata{$newkey}=$returnhash{'count'};
169: }
170: untie(%evaldata);
171: untie(%newevaldata);
172: }
173: return %returnhash;
174: }
175:
176: # ----------------- Code to enable 'find' subroutine listing of the .meta files
177: require "find.pl";
178: sub wanted {
179: (($dev,$ino,$mode,$nlink,$uid,$gid) = lstat($_)) &&
180: -f _ &&
181: /^.*\.meta$/ && !/^.+\.\d+\.[^\.]+\.meta$/ &&
182: push(@metalist,"$dir/$_");
183: }
184:
185: # --------------- Read loncapa_apache.conf and loncapa.conf and get variables
186: my $perlvarref=LONCAPA::Configuration::read_conf('loncapa.conf');
187: my %perlvar=%{$perlvarref};
188: undef $perlvarref; # remove since sensitive and not needed
189: delete $perlvar{'lonReceipt'}; # remove since sensitive and not needed
190:
191: # ------------------------------------- Only run if machine is a library server
192: exit unless $perlvar{'lonRole'} eq 'library';
193:
194: # ----------------------------- Make sure this process is running from user=www
195:
196: my $wwwid=getpwnam('www');
197: if ($wwwid!=$<) {
198: $emailto="$perlvar{'lonAdmEMail'},$perlvar{'lonSysEMail'}";
199: $subj="LON: $perlvar{'lonHostID'} User ID mismatch";
200: system("echo 'User ID mismatch. searchcat.pl must be run as user www.' |\
201: mailto $emailto -s '$subj' > /dev/null");
202: exit 1;
203: }
204:
205:
206: # ---------------------------------------------------------- We are in business
207:
208: open(LOG,'>'.$perlvar{'lonDaemons'}.'/logs/searchcat.log');
209: print LOG '==== Searchcat Run '.localtime()."====\n\n";
210: my $dbh;
211: # ------------------------------------- Make sure that database can be accessed
212: {
213: unless (
214: $dbh = DBI->connect("DBI:mysql:loncapa","www",$perlvar{'lonSqlAccess'},{ RaiseError =>0,PrintError=>0})
215: ) {
216: print LOG "Cannot connect to database!\n";
217: exit;
218: }
219: my $make_metadata_table = "CREATE TABLE IF NOT EXISTS metadata (".
220: "title TEXT, author TEXT, subject TEXT, url TEXT, keywords TEXT, ".
221: "version TEXT, notes TEXT, abstract TEXT, mime TEXT, language TEXT, ".
222: "creationdate DATETIME, lastrevisiondate DATETIME, owner TEXT, ".
223: "copyright TEXT, FULLTEXT idx_title (title), ".
224: "FULLTEXT idx_author (author), FULLTEXT idx_subject (subject), ".
225: "FULLTEXT idx_url (url), FULLTEXT idx_keywords (keywords), ".
226: "FULLTEXT idx_version (version), FULLTEXT idx_notes (notes), ".
227: "FULLTEXT idx_abstract (abstract), FULLTEXT idx_mime (mime), ".
228: "FULLTEXT idx_language (language), FULLTEXT idx_owner (owner), ".
229: "FULLTEXT idx_copyright (copyright)) TYPE=MYISAM";
230: # It would sure be nice to have some logging mechanism.
231: $dbh->do($make_metadata_table);
232: }
233:
234: # ------------------------------------------------------------- get .meta files
235: opendir(RESOURCES,"$perlvar{'lonDocRoot'}/res/$perlvar{'lonDefDomain'}");
236: my @homeusers = grep {
237: &ishome("$perlvar{'lonDocRoot'}/res/$perlvar{'lonDefDomain'}/$_")
238: } grep {!/^\.\.?$/} readdir(RESOURCES);
239: closedir RESOURCES;
240:
241: #
242: # Create the statement handlers we need
243: my $delete_sth = $dbh->prepare
244: ("DELETE FROM metadata WHERE url LIKE BINARY ?");
245:
246: my $insert_sth = $dbh->prepare
247: ("INSERT INTO metadata VALUES (".
248: "?,". # title
249: "?,". # author
250: "?,". # subject
251: "?,". # m2???
252: "?,". # version
253: "?,". # current
254: "?,". # notes
255: "?,". # abstract
256: "?,". # mime
257: "?,". # language
258: "?,". # creationdate
259: "?,". # revisiondate
260: "?,". # owner
261: "?)" # copyright
262: );
263:
264: foreach my $user (@homeusers) {
265: print LOG "\n=== User: ".$user."\n\n";
266: # Remove left-over db-files from potentially crashed searchcat run
267: my $prodir=&propath($perlvar{'lonDefDomain'},$user);
268: unlink($prodir.'/nohist_new_resevaldata.db');
269: # Use find.pl
270: undef @metalist;
271: @metalist=();
272: &find("$perlvar{'lonDocRoot'}/res/$perlvar{'lonDefDomain'}/$user");
273: # -- process each file to get metadata and put into search catalog SQL
274: # database. Also, check to see if already there.
275: # I could just delete (without searching first), but this works for now.
276: foreach my $m (@metalist) {
277: print LOG "- ".$m."\n";
278: my $ref=&metadata($m);
279: my $m2='/res/'.&declutter($m);
280: $m2=~s/\.meta$//;
281: &dynamicmeta($m2);
282: $delete_sth->execute($m2);
283: $insert_sth->execute($ref->{'title'},
284: $ref->{'author'},
285: $ref->{'subject'},
286: $m2,
287: $ref->{'keywords'},
288: 'current',
289: $ref->{'notes'},
290: $ref->{'abstract'},
291: $ref->{'mime'},
292: $ref->{'language'},
293: sqltime($ref->{'creationdate'}),
294: sqltime($ref->{'lastrevisiondate'}),
295: $ref->{'owner'},
296: $ref->{'copyright'});
297: # if ($dbh->err()) {
298: # print STDERR "Error:".$dbh->errstr()."\n";
299: # }
300: $ref = undef;
301: }
302:
303: # --------------------------------------------------- Clean up database
304: # Need to, perhaps, remove stale SQL database records.
305: # ... not yet implemented
306:
307: # ------------------------------------------- Copy over the new db-files
308: system('mv '.$prodir.'/nohist_new_resevaldata.db '.
309: $prodir.'/nohist_resevaldata.db');
310: }
311: # --------------------------------------------------- Close database connection
312: $dbh->disconnect;
313: print LOG "\n==== Searchcat completed ".localtime()." ====\n";
314: close(LOG);
315: exit 0;
316:
317:
318:
319: # =============================================================================
320:
321: # ---------------------------------------------------------------- Get metadata
322: # significantly altered from subroutine present in lonnet
323: sub metadata {
324: my ($uri,$what)=@_;
325: my %metacache;
326: $uri=&declutter($uri);
327: my $filename=$uri;
328: $uri=~s/\.meta$//;
329: $uri='';
330: unless ($metacache{$uri.'keys'}) {
331: unless ($filename=~/\.meta$/) { $filename.='.meta'; }
332: my $metastring=&getfile($perlvar{'lonDocRoot'}.'/res/'.$filename);
333: my $parser=HTML::TokeParser->new(\$metastring);
334: my $token;
335: while ($token=$parser->get_token) {
336: if ($token->[0] eq 'S') {
337: my $entry=$token->[1];
338: my $unikey=$entry;
339: if (defined($token->[2]->{'part'})) {
340: $unikey.='_'.$token->[2]->{'part'};
341: }
342: if (defined($token->[2]->{'name'})) {
343: $unikey.='_'.$token->[2]->{'name'};
344: }
345: if ($metacache{$uri.'keys'}) {
346: $metacache{$uri.'keys'}.=','.$unikey;
347: } else {
348: $metacache{$uri.'keys'}=$unikey;
349: }
350: map {
351: $metacache{$uri.''.$unikey.'.'.$_}=$token->[2]->{$_};
352: } @{$token->[3]};
353: unless (
354: $metacache{$uri.''.$unikey}=$parser->get_text('/'.$entry)
355: ) { $metacache{$uri.''.$unikey}=
356: $metacache{$uri.''.$unikey.'.default'};
357: }
358: }
359: }
360: }
361: return \%metacache;
362: }
363:
364: # ------------------------------------------------------------ Serves up a file
365: # returns either the contents of the file or a -1
366: sub getfile {
367: my $file=shift;
368: if (! -e $file ) { return -1; };
369: my $fh=IO::File->new($file);
370: my $a='';
371: while (<$fh>) { $a .=$_; }
372: return $a;
373: }
374:
375: # ------------------------------------------------------------- Declutters URLs
376: sub declutter {
377: my $thisfn=shift;
378: $thisfn=~s/^$perlvar{'lonDocRoot'}//;
379: $thisfn=~s/^\///;
380: $thisfn=~s/^res\///;
381: return $thisfn;
382: }
383:
384: # --------------------------------------- Is this the home server of an author?
385: # (copied from lond, modification of the return value)
386: sub ishome {
387: my $author=shift;
388: $author=~s/\/home\/httpd\/html\/res\/([^\/]*)\/([^\/]*).*/$1\/$2/;
389: my ($udom,$uname)=split(/\//,$author);
390: my $proname=propath($udom,$uname);
391: if (-e $proname) {
392: return 1;
393: } else {
394: return 0;
395: }
396: }
397:
398: # -------------------------------------------- Return path to profile directory
399: # (copied from lond)
400: sub propath {
401: my ($udom,$uname)=@_;
402: $udom=~s/\W//g;
403: $uname=~s/\W//g;
404: my $subdir=$uname.'__';
405: $subdir =~ s/(.)(.)(.).*/$1\/$2\/$3/;
406: my $proname="$perlvar{'lonUsersDir'}/$udom/$subdir/$uname";
407: return $proname;
408: }
409:
410: # ---------------------------- convert 'time' format into a datetime sql format
411: sub sqltime {
412: my ($sec,$min,$hour,$mday,$mon,$year,$wday,$yday,$isdst) =
413: localtime(&unsqltime(@_[0]));
414: $mon++; $year+=1900;
415: return "$year-$mon-$mday $hour:$min:$sec";
416: }
417:
418: sub maketime {
419: my %th=@_;
420: return POSIX::mktime(($th{'seconds'},$th{'minutes'},$th{'hours'},
421: $th{'day'},$th{'month'}-1,
422: $th{'year'}-1900,0,0,$th{'dlsav'}));
423: }
424:
425:
426: #########################################
427: #
428: # Retro-fixing of un-backward-compatible time format
429:
430: sub unsqltime {
431: my $timestamp=shift;
432: if ($timestamp=~/^(\d+)\-(\d+)\-(\d+)\s+(\d+)\:(\d+)\:(\d+)$/) {
433: $timestamp=&maketime('year'=>$1,'month'=>$2,'day'=>$3,
434: 'hours'=>$4,'minutes'=>$5,'seconds'=>$6);
435: }
436: return $timestamp;
437: }
438:
FreeBSD-CVSweb <freebsd-cvsweb@FreeBSD.org>