better collation support

data/src/entity_manager/system.py

@@ -4418,6 +4418,18 @@ def _lock_query_f(self, entity_class, options, query_buffer):
         if allow_for_update: query_buffer.write(" for update")
         else: self.lock(entity_class)
 
+    def _process_collate(self, query_buffer):
+        # verifies if a valid provider of (case) insensitive collation
+        # exists for the current data source engine if that's not the
+        # case returns the control flow immediately (no collate)
+        if not hasattr(self.engine, "get_insensitive_collate"): return
+        collation = self.engine.get_insensitive_collate()
+        if not collation: return
+
+        # creates the extra collate operation for the current buffer
+        # taking into account the collation for the current engine
+        query_buffer.write(" collate %s" % collation)
+
     def _process_filter(self, entity_class, table_name, filter, query_buffer, is_first = True):
         # in case the is first flag
         # is set, need to write the initial
@@ -4829,9 +4841,9 @@ def _process_filter_like(
                 # query buffer
                 query_buffer.write("'")
 
-            # in case the collate flat is enable the case insensitive collation
+            # in case the collate flat is set, the case insensitive collation
             # mode is enabled so that a more broad search is enabled
-            if collate: query_buffer.write(" collate utf8_general_ci")
+            if collate: self._process_collate(query_buffer)
 
     def _process_filter_rlike(self, entity_class, table_name, filter, query_buffer):
         self._process_filter_like(entity_class, table_name, filter, query_buffer, left = False)

data/src/entity_mysql/system.py

@@ -131,6 +131,9 @@ def get_database_encoding(self):
         encoding = _connection.get_database_encoding()
         return encoding
 
+    def get_insensitive_collate(self):
+        return "utf8_general_ci"
+
     def apply_types(self, types_map):
         types_map["text"] = "longtext"
         types_map["data"] = "longtext"
@@ -516,6 +519,9 @@ def _database_size_result(self, cursor):
         # size from the data source
         return datbase_size
 
+    def _collate_query(self):
+        return "collate utf8_general_ci"
+
     def _index_query(self, entity_class, attribute_name, index_type = "hash"):
         # retrieves the associated table name
         # as the "name" of the entity class

data/src/entity_sqlite/system.py

@@ -109,6 +109,9 @@ def get_database_size(self):
     def get_database_encoding(self):
         return "utf-8"
 
+    def get_insensitive_collate(self):
+        return "nocase"
+
     def connect(self, connection, parameters = {}):
         file_path = parameters.get("file_path", None)
         cache_size = parameters.get("cache_size", 200000)