From a9b7b2ffc1bc19f7770f3a1aadbaca1d95c91d45 Mon Sep 17 00:00:00 2001
From: Graham Knop <haarg@haarg.org>
Date: Mon, 4 Nov 2024 22:40:09 +0100
Subject: [PATCH] contributors: fetch all contrib pauseids in one query

Rather than trying to fill in each pause id on a separate query, run one
query for all contributors. Significantly speeds up fetches for large
contributor lists.
---
 lib/MetaCPAN/Query/Release.pm | 34 +++++++++++++++++++++-------------
 1 file changed, 21 insertions(+), 13 deletions(-)

diff --git a/lib/MetaCPAN/Query/Release.pm b/lib/MetaCPAN/Query/Release.pm
index 8a3a49802..d114f1cdf 100644
--- a/lib/MetaCPAN/Query/Release.pm
+++ b/lib/MetaCPAN/Query/Release.pm
@@ -153,6 +153,7 @@ sub get_contributors {
         $dupe ? () : $info;
     } ( @$authors, @$contribs );
 
+    my %want_email;
     for my $contrib (@contribs) {
 
         # heuristic to autofill pause accounts
@@ -165,20 +166,27 @@ sub get_contributors {
 
         }
 
-        # check if contributor's email points to a registered author
-        if ( !$contrib->{pauseid} ) {
-            for my $email ( @{ $contrib->{email} } ) {
-                my $check_author = $self->es->search(
-                    es_doc_path('author'),
-                    body => {
-                        query => { term => { email => $email } },
-                        size  => 10,
-                    }
-                );
+        push @{ $want_email{$_} }, $contrib for @{ $contrib->{email} };
+    }
+
+    if (%want_email) {
+        my $check_author = $self->es->search(
+            es_doc_path('author'),
+            body => {
+                query => { terms => { email => [ sort keys %want_email ] } },
+                _source => [ 'email', 'pauseid' ],
+                size    => 100,
+            },
+        );
 
-                if ( hit_total($check_author) ) {
-                    $contrib->{pauseid}
-                        = uc $check_author->{hits}{hits}[0]{_source}{pauseid};
+        for my $author ( @{ $check_author->{hits}{hits} } ) {
+            my $emails = $author->{_source}{email};
+            $emails = [$emails]
+                if !ref $emails;
+            my $pauseid = uc $author->{_source}{pauseid};
+            for my $email (@$emails) {
+                for my $contrib ( @{ $want_email{$email} } ) {
+                    $contrib->{pauseid} = $pauseid;
                 }
             }
         }