refactor code/tests: keyword support for search index #2516

sphinx/search/__init__.py

@@ -15,6 +15,7 @@ from six.moves import cPickle as pickle
 from docutils.nodes import raw, comment, title, Text, NodeVisitor, SkipNode
 from os import path
 
+import sphinx
 from sphinx.util import jsdump, rpartition
 from sphinx.util.pycompat import htmlescape
 
@@ -181,14 +182,14 @@ class WordCollector(NodeVisitor):
         self.lang = lang
 
     def is_meta_keywords(self, node, nodetype):
-        is_meta = str(nodetype) == '<class \'sphinx.addnodes.meta\'>'
+        if isinstance(node, sphinx.addnodes.meta) and node.get('name') == 'keywords':
-        if is_meta and node.get('name', None) == 'keywords':
+            meta_lang = node.get('lang')
-            node_lang = node.get('lang', None)
+            if meta_lang is None:  # lang not specified
-            is_correct_language = node_lang == None \
+                return True
-                or node_lang == self.lang.lang
+            elif meta_lang == self.lang.lang:  # matched to html_search_language
-            return is_meta and is_correct_language
+                return True
-        else:
+
-            return False
+        return False
 
     def dispatch_visit(self, node):
         nodetype = type(node)
@@ -208,9 +209,9 @@ class WordCollector(NodeVisitor):
         elif issubclass(nodetype, title):
             self.found_title_words.extend(self.lang.split(node.astext()))
         elif self.is_meta_keywords(node, nodetype):
-                keywords = node['content']
+            keywords = node['content']
-                keywords = [keyword.strip() for keyword in keywords.split(',')]
+            keywords = [keyword.strip() for keyword in keywords.split(',')]
-                self.found_words.extend(keywords)
+            self.found_words.extend(keywords)
 
 
 class IndexBuilder(object):

tests/roots/test-search-de/conf.py

@@ -1,3 +0,0 @@
-master_doc = 'index'
-exclude_patterns = ['_build']
-html_search_language = 'de'

tests/roots/test-search-de/index.rst

@@ -1,8 +0,0 @@
-meta keywords
-=============
-
-.. meta::
-   :keywords lang=de: findthiskey, thistoo
-   :keywords: thisonetoo
-   :keywords lang=en: findnotthiskey
-   :description: thisnoteither

tests/roots/test-search/index.rst

@@ -2,7 +2,7 @@ meta keywords
 =============
 
 .. meta::
-   :keywords lang=en: findthiskey, thistoo
+   :keywords lang=en: findthiskey, thistoo, notgerman
    :keywords: thisonetoo
-   :keywords lang=de: findnotthiskey
+   :keywords lang=de: onlygerman, onlytoogerman
-   :description: thisnoteither
+   :description: thisnoteither

tests/test_search.py

@@ -8,6 +8,7 @@
     :copyright: Copyright 2007-2016 by the Sphinx team, see AUTHORS.
     :license: BSD, see LICENSE for details.
 """
+import os
 
 from docutils import frontend, utils
 from docutils.parsers import rst
@@ -54,21 +55,24 @@ def test_objects_are_escaped(app, status, warning):
     assert 'n::Array&lt;T, d&gt;' in index.get('objects').get('')  # n::Array<T,d> is escaped
 
 def assert_lang_agnostic_key_words(searchindex):
-    assert 'findnotthiskey' not in searchindex
     assert 'thisnoteith' not in searchindex
-    assert 'thistoo' in searchindex
     assert 'thisonetoo' in searchindex
 
 @with_app(testroot='search')
 def test_meta_keys_are_handled_for_language_en(app, status, warning):
+    os.remove(app.outdir / 'searchindex.js')
     app.builder.build_all()
     searchindex = (app.outdir / 'searchindex.js').text()
     assert_lang_agnostic_key_words(searchindex)
     assert 'findthiskei' in searchindex
+    assert 'onlygerman' not in searchindex
+    assert 'thistoo' in searchindex
 
-@with_app(testroot='search-de')
+@with_app(testroot='search', confoverrides={'html_search_language': 'de'})
 def test_meta_keys_are_handled_for_language_de(app, status, warning):
     app.builder.build_all()
     searchindex = (app.outdir / 'searchindex.js').text()
     assert_lang_agnostic_key_words(searchindex)
-    assert 'findthiskey' in searchindex
+    assert 'onlygerman' in searchindex
+    assert 'notgerman' not in searchindex
+    assert 'onlytoogerman' in searchindex