refactor, add more tests #2516

AUTHORS

@@ -29,7 +29,7 @@ Other contributors, listed alphabetically, are:
 * Horst Gutmann -- internationalization support
 * Martin Hans -- autodoc improvements
 * Doug Hellmann -- graphviz improvements
-* Timotheus Kampik - JS enhancements, stop words language fix
+* Timotheus Kampik - JS theme & search enhancements
 * Takeshi Komiya -- numref feature
 * Dave Kuhlman -- original LaTeX writer
 * Blaise Laflamme -- pyramid theme

sphinx/search/__init__.py

@@ -182,11 +182,10 @@ class WordCollector(NodeVisitor):
 
     def is_meta_keywords(self, node, nodetype):
         is_meta = str(nodetype) == '<class \'sphinx.addnodes.meta\'>'
-        if is_meta:
-            language_match = re.search(r'lang\=\"(.*?)\"', str(node))
-            is_correct_language =\
-                language_match == None \
-                or self.lang.lang == language_match.group(1)
+        if is_meta and node.get('name', None) == 'keywords':
+            node_lang = node.get('lang', None)
+            is_correct_language = node_lang == None \
+                or node_lang == self.lang.lang
             return is_meta and is_correct_language
         else:
             return False
@@ -209,8 +208,9 @@ 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 = re.search(r'content\=\"(.*?)\"', str(node)).group(1)
-                self.found_words.extend(self.lang.split(keywords))
+                keywords = node['content']
+                keywords = [keyword.strip() for keyword in keywords.split(',')]
+                self.found_words.extend(keywords)
 
 
 class IndexBuilder(object):

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

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

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

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

tests/roots/test-search/conf.py

@@ -1,11 +1,3 @@
-# -*- coding: utf-8 -*-
-
-import sys
-import os
-
-sys.path.insert(0, os.path.abspath('.'))
-extensions = ['sphinx.ext.autodoc', 'sphinx.ext.viewcode']
 master_doc = 'index'
 exclude_patterns = ['_build']
-
 html_search_language = 'en'

tests/roots/test-search/index.rst

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

tests/test_search.py

@@ -53,10 +53,22 @@ def test_objects_are_escaped(app, status, warning):
     index = jsdump.loads(searchindex[16:-2])
     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(app, status, warning):
+def test_meta_keys_are_handled_for_language_en(app, status, warning):
     app.builder.build_all()
     searchindex = (app.outdir / 'searchindex.js').text()
-    assert 'findnotthiskey' not in searchindex
+    assert_lang_agnostic_key_words(searchindex)
     assert 'findthiskei' in searchindex
-    assert 'thisonetoo' in searchindex
+
+@with_app(testroot='search-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