From bf60a9c88cd223e53acd1d8e58508eb6d1febbb5 Mon Sep 17 00:00:00 2001
From: Fridolin Somers <fridolin.somers@biblibre.com>
Date: Mon, 19 Nov 2018 14:46:05 +0100
Subject: [PATCH] Bug 21865: improve remove_unused_authorities.pl script

remove_unused_authorities.pl script can be improved.

This patch changes changes verbosity so than test mode can be used
to know the autorities that are used and those that can be deleted.
It also writes a line in output if limited authority type(s).

This patch also removes the unused vars $thresholdmin and $thresholdmax.

It also changes the query to use SQL with parameters for authority types.

Test plan :
1) On a catalog create a new authority
2) Be sure catalog is well indexed
3) Run misc/migration_tools/remove_unused_authorities.pl -t
4) You will see the line :
   *** Testing only, authorities will not be deleted. ***
5) You will see lines of :
   authid=x type=y : used X time(s)
6) You will see the line for the authority created in 1) :
   authid=x type=y : can be deleted
7) You will see at the end :
   x authorities parsed
   y can be deleted because unused
   z unchanged because used
8) Run misc/migration_tools/remove_unused_authorities.pl
9) You don't see the line :
   *** Testing only, authorities will not be deleted. ***
10) You will see lines of :
   authid=x type=y : used X time(s)
11) You will see the line for the authority created in 1) :
   authid=x type=y : deleted
12) You will see at the end :
   x authorities parsed
   y deleted because unused
   z unchanged because used
13) Run misc/migration_tools/remove_unused_authorities.pl --auth NP --auth CO
14) You see the line :
    Restricted to authority type(s) : NP,CO.

Signed-off-by: Bernardo Gonzalez Kriegel <bgkriegel@gmail.com>
Signed-off-by: Jonathan Druart <jonathan.druart@bugs.koha-community.org>
Signed-off-by: Martin Renvoize <martin.renvoize@ptfs-europe.com>
---
 .../remove_unused_authorities.pl              | 53 +++++++++++--------
 1 file changed, 32 insertions(+), 21 deletions(-)

diff --git a/misc/migration_tools/remove_unused_authorities.pl b/misc/migration_tools/remove_unused_authorities.pl
index bef6cbe4c7..70f2037f6a 100755
--- a/misc/migration_tools/remove_unused_authorities.pl
+++ b/misc/migration_tools/remove_unused_authorities.pl
@@ -20,8 +20,7 @@
 # You should have received a copy of the GNU General Public License
 # along with Koha; if not, see <http://www.gnu.org/licenses>.
 
-use strict;
-use warnings;
+use Modern::Perl;
 
 use Koha::Script;
 use C4::Context;
@@ -43,55 +42,67 @@ if ($want_help) {
     print_usage();
     exit 0;
 }
-
 if ($test) {
-    print "testing only, authorities will not be deleted.\n";
+    print "*** Testing only, authorities will not be deleted. ***\n";
+}
+if (@authtypes) {
+    print "Restricted to authority type(s) : ".join(',', @authtypes).".\n";
 }
 
 my $errZebraConnection = C4::Context->Zconn("biblioserver",0)->errcode();
 if ( $errZebraConnection == 10000 ) {
-    die "Zebra server seems not to be available. This script needs Zebra runs."
+    die "Zebra server seems not to be available. This script needs Zebra runs.";
 } elsif ( $errZebraConnection ) {
     die "Error from Zebra: $errZebraConnection";
 }
 
 my $dbh=C4::Context->dbh;
-my $thresholdmin=0;
-my $thresholdmax=0;
 my @results;
 # prepare the request to retrieve all authorities of the requested types
-my $rqsql = "SELECT * from auth_header where 1";
-$rqsql .= " AND authtypecode IN (".join(",",map{$dbh->quote($_)}@authtypes).")" if @authtypes;
+my $rqsql = q{ SELECT authid,authtypecode FROM auth_header };
+$rqsql .= q{ WHERE authtypecode IN (}.join(',',map{ '?' }@authtypes).')' if @authtypes;
 my $rqselect = $dbh->prepare($rqsql);
 $|=1;
 
-$rqselect->execute;
+$rqselect->execute(@authtypes);
 my $counter=0;
 my $totdeleted=0;
 my $totundeleted=0;
 my $searcher = Koha::SearchEngine::Search->new({index => 'biblios'});
 while (my $data=$rqselect->fetchrow_hashref){
-    my $query;
-    $query= "an=".$data->{'authid'};
+    $counter++;
+    print 'authid='.$data->{'authid'};
+    print ' type='.$data->{'authtypecode'};
+    my $bibliosearch = 'an:'.$data->{'authid'};
     # search for biblios mapped
-    my ($err,$res,$used) = $searcher->simple_search_compat($query,0,10);
+    my ($err,$res,$used) = $searcher->simple_search_compat($bibliosearch,0,10);
     if (defined $err) {
-        warn "error: $err on search $query\n";
+        print "\n";
+        warn "Error: $err on search for biblios $bibliosearch\n";
         next;
     }
-    print ".";
-    print "$counter\n" unless $counter++ % 100;
-    # if found, delete, otherwise, just count
-    if ($used>=$thresholdmin and $used<=$thresholdmax){
-        DelAuthority({ authid => $data->{'authid'} }) unless $test;
+    unless ($used > 0){
+        unless ($test) {
+            DelAuthority({ authid => $data->{'authid'} });
+            print " : deleted";
+        } else {
+            print " : can be deleted";
+        }
         $totdeleted++;
     } else {
         $totundeleted++;
+        print " : used $used time(s)";
     }
+    print "\n";
 }
 
-print "$counter authorities parsed, $totdeleted deleted and $totundeleted unchanged because used\n";
-
+print "$counter authorities parsed\n";
+unless ($test) {
+    print "$totdeleted deleted because unused\n";
+} else {
+    print "$totdeleted can be deleted because unused\n";
+}
+print "$totundeleted unchanged because used\n";
 
 sub print_usage {
     print <<_USAGE_;
-- 
2.39.5