- Fixing query to reduce the amount of processing on the code side

shared/selene/device/repository/device.py

@@ -20,16 +20,7 @@ def get_device_by_id(db, device_id: str) -> Device:
         singleton=True
     )
     sql_results = fetch(db, query)
-    device = {k: v for k, v in sql_results.items() if k in ('id', 'name', 'platform', 'enclosure_version', 'core_version')}
-
-    wake_word = {k: v for k, v in sql_results.items() if k in ('wake_word_id', 'wake_word', 'engine')}
-    wake_word['id'] = wake_word.pop('wake_word_id')
-    device['wake_word'] = wake_word
-
-    text_to_speech = {k: v for k, v in sql_results.items() if k in ('text_to_speech_id', 'setting_name', 'display_name', 'engine')}
-    text_to_speech['id'] = text_to_speech.pop('text_to_speech_id')
-    device['text_to_speech'] = text_to_speech
-    return Device(**device)
+    return Device(**sql_results)
 
 
 def get_devices_by_account_id(db, account_id: str) -> List[Device]:

shared/selene/device/repository/sql/get_device_by_id.sql

@@ -5,13 +5,8 @@ SELECT
   dev.enclosure_version,
   dev.core_version,
   dev.placement,
-  wk_word.id as wake_word_id,
-  wk_word.wake_word,
-  wk_word.engine,
-  tts.id as text_to_speech_id,
-  tts.setting_name,
-  tts.display_name,
-  tts.engine
+  json_build_object('id', wk_word.id, 'wake_word', wk_word.wake_word, 'engine', wk_word.engine) as wake_word,
+  json_build_object('id', tts.id, 'setting_name', tts.setting_name, 'display_name', tts.display_name, 'engine', tts.engine) as text_to_speech
 FROM device.device dev
 INNER JOIN
   device.wake_word wk_word ON dev.wake_word_id = wk_word.id