Index body and target

docs/search.md

@@ -109,6 +109,14 @@ The `fts` query accepts the following parameters for each field:
 | operator | Join tokens with `and`, `or` or `not` (default `and`) |
 | prefix   | Treat each token as a prefix (default `True`)         |
 
+!!! info "Full-text search language"
+
+    The dictionary used for full-text searches is defined for each Annotation
+    by the first `language` code found in the Annotation's `body`. If no
+    such language code is found then the server default is used. See the
+    [Configuration](/setup/configuration.md) section for more details.
+
+
 ## fts_phrase
 
 Return Annotations where the specified keys contain a `query` phrase. The
@@ -133,5 +141,7 @@ The `fts_phrase` query accepts the following parameters for each field:
 | prefix   | Treat the query as a prefix (default `True`)          |
 | distance | The distance between tokens (default `1`)             |
 
-Note that all phrase queries will be treated as prefixes; to search for
-exact phrases use `contains` instead.
+!!! note "Exact phrase searches"
+
+    Note that all phrase queries will be treated as prefixes; to search for
+    exact phrases use [`contains`](/search.md#contains) instead.

explicates/core.py

@@ -72,14 +72,20 @@ def setup_search(app):
 
 def setup_db(app):
     """Setup database."""
+    from explicates.model.indexes import indexes
     def create_slave_session(db, bind):
         slave = app.config.get('SQLALCHEMY_BINDS')['slave']
         if slave == app.config.get('SQLALCHEMY_DATABASE_URI'):
             return db.session
         engine = db.get_engine(db.app, bind=bind)
         options = dict(bind=engine, scopefunc=_app_ctx_stack.__ident_func__)
         slave_session = db.create_scoped_session(options=options)
+
+        for idx in indexes:
+            idx.create(bind=engine)
+
         return slave_session
+
     db.app = app
     db.init_app(app)
     db.slave_session = create_slave_session(db, bind='slave')

explicates/model/annotation.py

@@ -3,7 +3,7 @@
 
 from flask import url_for, current_app
 from sqlalchemy.schema import Column, ForeignKey
-from sqlalchemy import Integer, Text
+from sqlalchemy import Integer, String
 from sqlalchemy.ext.hybrid import hybrid_property
 from sqlalchemy.ext.declarative import declarative_base
 
@@ -53,7 +53,7 @@ class Annotation(db.Model, Base):
                             nullable=False)
 
     #: The language used for full-text searches.
-    language = Column(Text, nullable=False, default=get_language)
+    language = Column(String, nullable=False, default=get_language)
 
     @hybrid_property
     def iri(self):

explicates/model/indexes.py

@@ -0,0 +1,15 @@
+# -*- coding: utf8 -*-
+"""Indexes."""
+
+from sqlalchemy.schema import Index
+from sqlalchemy.sql import text
+
+
+indexes = [
+    Index('idx_annotation_body',
+            text("to_tsvector(language::regconfig, _data -> 'body')"),
+            postgresql_using='gin'),
+    Index('idx_annotation_target',
+            text("to_tsvector(language::regconfig, _data -> 'target')"),
+            postgresql_using='gin')
+]

test/test_search.py

@@ -132,6 +132,18 @@ def test_search_by_fts_default(self):
         results = self.search.search(fts=fts_query)
         assert_equal(results, [anno1])
 
+    @with_context
+    def test_search_by_fts_does_not_include_keys(self):
+        """Test search by fts does not include keys."""
+        anno1 = AnnotationFactory(data={'body': {'source': 'foo'}})
+        fts_query = {
+            'body': {
+                'query': 'source'
+            }
+        }
+        results = self.search.search(fts=fts_query)
+        assert_equal(results, [])
+
     @with_context
     def test_search_by_fts_without_prefix(self):
         """Test search by fts without prefix."""