From: Bill Erickson <berickxx@gmail.com>
Date: Fri, 27 Sep 2019 19:44:34 +0000 (-0400)
Subject: Avoid passing non-indexed fields to save disk space
X-Git-Url: https://old-git.evergreen-ils.org/?a=commitdiff_plain;h=1c00e23a8cfccc1fd4c4e0e3d2985ebc784b7ffd;p=working%2FEvergreen.git

Avoid passing non-indexed fields to save disk space

Signed-off-by: Bill Erickson <berickxx@gmail.com>
---

diff --git a/Open-ILS/src/perlmods/lib/OpenILS/Elastic.pm b/Open-ILS/src/perlmods/lib/OpenILS/Elastic.pm
index d049630d4a..be430e5e4b 100644
--- a/Open-ILS/src/perlmods/lib/OpenILS/Elastic.pm
+++ b/Open-ILS/src/perlmods/lib/OpenILS/Elastic.pm
@@ -263,6 +263,10 @@ sub truncate_value {
     return $value;
 }
 
+sub get_index_def {
+    my ($self) = @_;
+    return $self->es->indices->get(index => $self->index_name);
+}
 
 
 
diff --git a/Open-ILS/src/perlmods/lib/OpenILS/Elastic/Bib/Search.pm b/Open-ILS/src/perlmods/lib/OpenILS/Elastic/Bib/Search.pm
index 8e15361abf..d026360861 100644
--- a/Open-ILS/src/perlmods/lib/OpenILS/Elastic/Bib/Search.pm
+++ b/Open-ILS/src/perlmods/lib/OpenILS/Elastic/Bib/Search.pm
@@ -392,6 +392,13 @@ sub populate_bib_index_batch {
 
     my $holdings = $self->load_holdings($bib_ids);
     my $marc = $self->load_marc($bib_ids);
+    my $def = $self->get_index_def;
+
+    # Ask ES what the index properties are so we can avoid passing data
+    # that will not be indexed, since ES will store the data on the source
+    # object even if it's not indexed.  This reduces bulk.
+    my $properties = # nestier than expected, not sure why.
+        $def->{$self->index_name}->{mappings}->{record}->{properties}->{record}->{properties};
 
     for my $bib_id (@$bib_ids) {
 
@@ -429,7 +436,7 @@ sub populate_bib_index_batch {
             } elsif ($fname eq 'identifier|issn') {
                 index_issns($body, $value);
             } else {
-                append_field_value($body, $fname, $value);
+                append_field_value($body, $fname, $value, $properties);
             }
         }
 
@@ -495,7 +502,11 @@ sub index_issns {
 }
 
 sub append_field_value {
-    my ($body, $fname, $value) = @_;
+    my ($body, $fname, $value, $properties) = @_;
+
+    # Confirm the data is wanted in the index before passing to ES to
+    # reduce the overall data footprint.
+    return unless $properties->{$fname};
 
     if ($body->{$fname}) {
         if (ref $body->{$fname}) {