new: highlight() accepts custom values on Solr and ES

This allows the default values to be overriden and arbitrary
backend-specific parameters may be provided to Solr or ElasticSearch.

Thanks to @tymofij for the patch

Closes django-haystack#1334

Author: acdha

haystack/backends/__init__.py

@@ -854,9 +854,9 @@ def add_stats_query(self,stats_field,stats_facets):
         """Adds stats and stats_facets queries for the Solr backend."""
         self.stats[stats_field] = stats_facets
 
-    def add_highlight(self):
+    def add_highlight(self, **kwargs):
         """Adds highlighting to the search results."""
-        self.highlight = True
+        self.highlight = kwargs or True
 
     def add_within(self, field, point_1, point_2):
         """Adds bounding box parameters to search query."""

haystack/backends/elasticsearch_backend.py

@@ -322,13 +322,19 @@ def build_search_kwargs(self, query_string, sort_by=None, start_offset=0, end_of
         # if end_offset is not None:
         #     kwargs['size'] = end_offset - start_offset
 
-        if highlight is True:
+        if highlight:
+            # `highlight` can either be True or a dictionary containing custom parameters
+            # which will be passed to the backend and may override our default settings:
+
             kwargs['highlight'] = {
                 'fields': {
                     content_field: {'store': 'yes'},
                 }
             }
 
+            if isinstance(highlight, dict):
+                kwargs['highlight'].update(highlight)
+
         if self.include_spelling:
             kwargs['suggest'] = {
                 'suggest': {

haystack/backends/solr_backend.py

@@ -180,10 +180,16 @@ def build_search_kwargs(self, query_string, sort_by=None, start_offset=0, end_of
         if end_offset is not None:
             kwargs['rows'] = end_offset - start_offset
 
-        if highlight is True:
+        if highlight:
+            # `highlight` can either be True or a dictionary containing custom parameters
+            # which will be passed to the backend and may override our default settings:
+
             kwargs['hl'] = 'true'
             kwargs['hl.fragsize'] = '200'
 
+            if isinstance(highlight, dict):
+                kwargs.update(highlight)
+
         if self.include_spelling is True:
             kwargs['spellcheck'] = 'true'
             kwargs['spellcheck.collate'] = 'true'

haystack/query.py

@@ -9,7 +9,7 @@
 
 from haystack import connection_router, connections
 from haystack.backends import SQ
-from haystack.constants import DEFAULT_OPERATOR, ITERATOR_LOAD_PER_QUERY, REPR_OUTPUT_SIZE
+from haystack.constants import DEFAULT_OPERATOR, ITERATOR_LOAD_PER_QUERY
 from haystack.exceptions import NotHandled
 from haystack.inputs import AutoQuery, Raw
 from haystack.utils import log as logging
@@ -313,10 +313,10 @@ def order_by(self, *args):
 
         return clone
 
-    def highlight(self):
+    def highlight(self, **kwargs):
         """Adds highlighting to the results."""
         clone = self._clone()
-        clone.query.add_highlight()
+        clone.query.add_highlight(**kwargs)
         return clone
 
     def models(self, *models):

test_haystack/elasticsearch_tests/test_elasticsearch_backend.py

@@ -424,6 +424,9 @@ def test_search(self):
         self.assertEqual(self.sb.search('Index', highlight=True)['hits'], 3)
         self.assertEqual(sorted([result.highlighted[0] for result in self.sb.search('Index', highlight=True)['results']]),
                          [u'<em>Indexed</em>!\n1', u'<em>Indexed</em>!\n2', u'<em>Indexed</em>!\n3'])
+        self.assertEqual(sorted([result.highlighted[0] for result in self.sb.search('Index', highlight={'pre_tags': ['<start>'],'post_tags': ['</end>']})['results']]),
+                         [u'<start>Indexed</end>!\n1', u'<start>Indexed</end>!\n2', u'<start>Indexed</end>!\n3'])
+
 
         self.assertEqual(self.sb.search('Indx')['hits'], 0)
         self.assertEqual(self.sb.search('indaxed')['spelling_suggestion'], 'indexed')
@@ -764,6 +767,17 @@ def test_count(self):
         # Should only execute one query to count the length of the result set.
         self.assertEqual(len(connections['elasticsearch'].queries), 1)
 
+    def test_highlight(self):
+        reset_search_queries()
+        results = self.sqs.filter(content='index').highlight()
+        self.assertEqual(results[0].highlighted, [u'<em>Indexed</em>!\n1'])
+
+    def test_highlight_options(self):
+        reset_search_queries()
+        results = self.sqs.filter(content='index')
+        results = results.highlight(pre_tags=['<i>'], post_tags=['</i>'])
+        self.assertEqual(results[0].highlighted, [u'<i>Indexed</i>!\n1'])
+
     def test_manual_iter(self):
         results = self.sqs.all()