Add author search to search API (bug 1032828)

docs/api/topics/search.rst

@@ -49,6 +49,9 @@ Search
         codes should be provided in ISO 639-1 format, using a comma-separated
         list if supplying multiple languages.
     :type languages: string
+    :param optional author: Filters by author. Requires a case-insensitive
+        exact match of the author field.
+    :type author: string
     :param optional region: Filters apps by a supported region. A region
         code should be provided in ISO 3166 format (e.g., `pl`). If not
         provided, the region is automatically detected via requesting IP

mkt/search/forms.py

@@ -106,6 +106,7 @@ class ApiSearchForm(forms.Form):
                                      label=_lazy('Works offline'))
     languages = forms.CharField(required=False,
                                 label=_lazy('Supported languages'))
+    author = forms.CharField(required=False, label=_lazy('Author name'))
 
     sort = forms.MultipleChoiceField(required=False,
                                      choices=LISTING_SORT_CHOICES)
@@ -170,6 +171,9 @@ def clean_device_and_dev(self):
         elif device:
             raise forms.ValidationError('Invalid device or device type.')
 
+    def clean_author(self):
+        return self.cleaned_data.get('author', '').lower()
+
     def clean(self):
         self.clean_device_and_dev()
         return self.cleaned_data

mkt/search/indexers.py

@@ -182,6 +182,13 @@ def get_analysis(cls):
                        'icu_normalizer'],
         }
 
+        # An analyzer that can do case-insensitive exact matching.
+        analyzers['exact_lowercase'] = {
+            'type': 'custom',
+            'tokenizer': 'keyword',
+            'filter': ['lowercase'],
+        }
+
         for lang, stemmer in amo.STEMMER_MAP.items():
             filters['%s_stem_filter' % lang] = {
                 'type': 'stemmer',

mkt/search/tests/test_filters.py

@@ -15,7 +15,6 @@
 from mkt.search.views import _sort_search, DEFAULT_SORTING
 from mkt.site.fixtures import fixture
 from mkt.webapps.indexers import WebappIndexer
-from mkt.webapps.models import Webapp
 
 
 class TestSearchFilters(BaseOAuth):
@@ -41,6 +40,7 @@ def _filter(self, req, filters, **kwargs):
             form_data = form.cleaned_data
             sq = WebappIndexer.get_app_filter(self.req, {
                 'app_type': form_data['app_type'],
+                'author.raw': form_data['author'],
                 'category': form_data['cat'],
                 'device': form_data['device'],
                 'is_offline': form_data['offline'],
@@ -188,6 +188,11 @@ def test_languages(self):
         ok_({'terms': {'supported_locales': ['ar', 'en-US']}}
             in qs['query']['filtered']['filter']['bool']['must'])
 
+    def test_author(self):
+        qs = self._filter(self.req, {'author': 'Mozilla LABS'})
+        ok_({'term': {'author.raw': u'mozilla labs'}}
+            in qs['query']['filtered']['filter']['bool']['must'])
+
     def test_region_exclusions(self):
         self.req.REGION = regions.CO
         qs = self._filter(self.req, {'q': 'search terms'})

mkt/search/tests/test_views.py

@@ -500,12 +500,26 @@ def test_name_localized_to_default_locale(self):
 
         # Make a request in another language that we know will fail.
         res = self.anon.get(self.url,
-                              data={'q': 'something', 'lang': 'de'})
+                            data={'q': 'something', 'lang': 'de'})
         eq_(res.status_code, 200)
         obj = res.json['objects'][0]
         eq_(obj['slug'], self.webapp.app_slug)
         eq_(obj['name'], u'Algo Algo Steamcube!')
 
+    def test_author(self):
+        res = self.anon.get(self.url,
+                            data={'author': self.webapp.developer_name})
+        eq_(res.status_code, 200)
+        obj = res.json['objects'][0]
+        eq_(obj['slug'], self.webapp.app_slug)
+
+    def test_author_case(self):
+        res = self.anon.get(
+            self.url, data={'author': self.webapp.developer_name.upper()})
+        eq_(res.status_code, 200)
+        obj = res.json['objects'][0]
+        eq_(obj['slug'], self.webapp.app_slug)
+
     def test_device(self):
         AddonDeviceType.objects.create(
             addon=self.webapp, device_type=DEVICE_CHOICES_IDS['desktop'])

mkt/search/views.py

@@ -143,11 +143,12 @@ def search_form_to_es_fields(form_data):
     return {
         'app_type': form_data['app_type'],
         'category': form_data['cat'],
+        'author.raw': form_data['author'],
         'device': form_data['device'],
         'is_offline': form_data['offline'],
         'manifest_url': form_data['manifest_url'],
-        'q': form_data['q'],
         'premium_type': form_data['premium_types'],
+        'q': form_data['q'],
         'supported_locales': form_data['languages'],
         'tags': form_data['tag'],
     }

mkt/webapps/indexers.py

@@ -76,7 +76,16 @@ def _locale_field_mapping(field, analyzer):
                     'id': {'type': 'long'},
                     'app_slug': {'type': 'string'},
                     'app_type': {'type': 'byte'},
-                    'author': {'type': 'string', 'analyzer': 'default_icu'},
+                    'author': {
+                        'type': 'string',
+                        'analyzer': 'default_icu',
+                        'fields': {
+                            # For exact matches. The simple analyzer allows
+                            # for case-insensitive matching.
+                            'raw': {'type': 'string',
+                                    'analyzer': 'exact_lowercase'},
+                        },
+                    },
                     'banner_regions': cls.string_not_indexed(),
                     'bayesian_rating': {'type': 'float', 'doc_values': True},
                     'category': cls.string_not_analyzed(),
@@ -454,6 +463,7 @@ def get_app_filter(cls, request, additional_data=None, sq=None,
 
         data = {
             'app_type': [],
+            'author.raw': None,
             'category': None,  # Slug.
             'device': None,  # ID.
             'gaia': getattr(request, 'GAIA', False),
@@ -466,13 +476,14 @@ def get_app_filter(cls, request, additional_data=None, sq=None,
             'region': getattr(get_region_from_request(request), 'id', None),
             'status': None,
             'supported_locales': [],
-            'tags': '',
             'tablet': getattr(request, 'TABLET', False),
+            'tags': '',
         }
         data.update(additional_data)
 
         # Fields that will be filtered with a term query.
-        term_fields = ('device', 'manifest_url', 'status', 'tags')
+        term_fields = ('author.raw', 'device', 'manifest_url', 'status',
+                       'tags')
         # Fields that will be filtered with a terms query.
         terms_fields = ('category', 'premium_type', 'app_type',
                         'supported_locales')