From e5a73815f580d296572e11b71b5f3ed320bbe912 Mon Sep 17 00:00:00 2001
From: Gareth Rees <gareth@mysociety.org>
Date: Fri, 30 May 2014 12:42:48 +0100
Subject: Add helper to highlight and excerpt by regex

Backport of https://github.com/rails/rails/pull/11793/

Contains integration tests to check that it works
as expected with ActsAsXapian.
---
 .../integration/xapian_search_highlighting_spec.rb | 29 ++++++++++++++++++++++
 1 file changed, 29 insertions(+)
 create mode 100644 spec/integration/xapian_search_highlighting_spec.rb

(limited to 'spec/integration')

diff --git a/spec/integration/xapian_search_highlighting_spec.rb b/spec/integration/xapian_search_highlighting_spec.rb
new file mode 100644
index 000000000..7bd64c995
--- /dev/null
+++ b/spec/integration/xapian_search_highlighting_spec.rb
@@ -0,0 +1,29 @@
+require File.expand_path(File.dirname(__FILE__) + '/../spec_helper')
+
+describe 'highlighting search results' do
+    include HighlightHelper
+
+    it 'ignores stopwords' do
+        phrase = 'department of humpadinking'
+        search = ActsAsXapian::Search.new([PublicBody], phrase, :limit => 1)
+        matches = search.words_to_highlight(:regex => true)
+        highlight_matches(phrase, matches).should == '<mark>department</mark> of <mark>humpadinking</mark>'
+    end
+
+    it 'ignores case' do
+        search_phrase = 'department of humpadinking'
+        search = ActsAsXapian::Search.new([PublicBody], search_phrase, :limit => 1)
+        matches = search.words_to_highlight(:regex => true)
+        highlight_matches('Department of Humpadinking', matches).should == '<mark>Department</mark> of <mark>Humpadinking</mark>'
+    end
+
+    it 'highlights stemmed words' do
+        phrase = 'department'
+        search = ActsAsXapian::Search.new([PublicBody], phrase, :limit => 1)
+        matches = search.words_to_highlight(:regex => true)
+
+        search.words_to_highlight(:regex => false).should == ['depart']
+        highlight_matches(phrase, matches).should == '<mark>department</mark>'
+    end
+
+end
-- 
cgit v1.2.3


From f23b89f3474847cdd14ba892c5a7259964e18148 Mon Sep 17 00:00:00 2001
From: Gareth Rees <gareth@mysociety.org>
Date: Tue, 24 Jun 2014 16:08:32 +0100
Subject: Handle unhelpful stemming

Stemming returns 'bore' as the word to highlight which can't be matched
in the original phrase.

Also removes duplicates from the results
---
 spec/integration/xapian_search_highlighting_spec.rb | 10 ++++++++++
 1 file changed, 10 insertions(+)

(limited to 'spec/integration')

diff --git a/spec/integration/xapian_search_highlighting_spec.rb b/spec/integration/xapian_search_highlighting_spec.rb
index 7bd64c995..65a34cf91 100644
--- a/spec/integration/xapian_search_highlighting_spec.rb
+++ b/spec/integration/xapian_search_highlighting_spec.rb
@@ -26,4 +26,14 @@ describe 'highlighting search results' do
         highlight_matches(phrase, matches).should == '<mark>department</mark>'
     end
 
+    it 'highlights stemmed words even if the stem is unhelpful' do
+        # Stemming returns 'bore' as the word to highlight which can't be
+        # matched in the original phrase.
+        phrase = 'boring'
+        search = ActsAsXapian::Search.new([PublicBody], phrase, :limit => 1)
+        matches = search.words_to_highlight(:regex => true, :include_original => true)
+
+        highlight_matches(phrase, matches).should == '<mark>boring</mark>'
+    end
+
 end
-- 
cgit v1.2.3