Add user emoji augmentation for message queries (#937)

* add disposition to text labeling tasks

* add emoji stats to ConversationMessage

* add user emoji augmentation for message queries

* add auth_method,username to message queries (query emoji status)

* add auth_method+username for single message

* fix param name typo

* only join rows when message.emojis != JSON.NULL

* formatting

* make sure emojis and user_emojis default to {}, []

* remove init_user(), use fresh empty default collections

backend/oasst_backend/api/v1/frontend_users.py

@@ -77,7 +77,7 @@ def query_frontend_user_messages(
     """
     Query frontend user messages.
     """
-    pr = PromptRepository(db, api_client)
+    pr = PromptRepository(db, api_client, auth_method=auth_method, username=username)
     messages = pr.query_messages_ordered_by_created_date(
         auth_method=auth_method,
         username=username,

backend/oasst_backend/api/v1/messages.py

@@ -34,7 +34,7 @@ def query_messages(
     """
     Query messages.
     """
-    pr = PromptRepository(db, api_client)
+    pr = PromptRepository(db, api_client, auth_method=auth_method, username=username)
     messages = pr.query_messages_ordered_by_created_date(
         auth_method=auth_method,
         username=username,
@@ -93,7 +93,7 @@ def get_messages_cursor(
 
     qry_max_count = max_count + 1 if before is None or after is None else max_count
 
-    pr = PromptRepository(db, api_client)
+    pr = PromptRepository(db, api_client, auth_method=auth_method, username=username, user_id=user_id)
     items = pr.query_messages_ordered_by_created_date(
         user_id=user_id,
         auth_method=auth_method,
@@ -137,37 +137,49 @@ def get_messages_cursor(
 
 @router.get("/{message_id}", response_model=protocol.Message)
 def get_message(
-    message_id: UUID, api_client: ApiClient = Depends(deps.get_api_client), db: Session = Depends(deps.get_db)
+    message_id: UUID,
+    auth_method: Optional[str] = None,
+    username: Optional[str] = None,
+    api_client: ApiClient = Depends(deps.get_api_client),
+    db: Session = Depends(deps.get_db),
 ):
     """
     Get a message by its internal ID.
     """
-    pr = PromptRepository(db, api_client)
+    pr = PromptRepository(db, api_client, auth_method=auth_method, username=username)
     message = pr.fetch_message(message_id)
     return utils.prepare_message(message)
 
 
 @router.get("/{message_id}/conversation", response_model=protocol.Conversation)
 def get_conv(
-    message_id: UUID, api_client: ApiClient = Depends(deps.get_api_client), db: Session = Depends(deps.get_db)
+    message_id: UUID,
+    auth_method: Optional[str] = None,
+    username: Optional[str] = None,
+    api_client: ApiClient = Depends(deps.get_api_client),
+    db: Session = Depends(deps.get_db),
 ):
     """
     Get a conversation from the tree root and up to the message with given internal ID.
     """
 
-    pr = PromptRepository(db, api_client)
+    pr = PromptRepository(db, api_client, auth_method=auth_method, username=username)
     messages = pr.fetch_message_conversation(message_id)
     return utils.prepare_conversation(messages)
 
 
 @router.get("/{message_id}/tree", response_model=protocol.MessageTree)
 def get_tree(
-    message_id: UUID, api_client: ApiClient = Depends(deps.get_api_client), db: Session = Depends(deps.get_db)
+    message_id: UUID,
+    auth_method: Optional[str] = None,
+    username: Optional[str] = None,
+    api_client: ApiClient = Depends(deps.get_api_client),
+    db: Session = Depends(deps.get_db),
 ):
     """
     Get all messages belonging to the same message tree.
     """
-    pr = PromptRepository(db, api_client)
+    pr = PromptRepository(db, api_client, auth_method=auth_method, username=username)
     message = pr.fetch_message(message_id)
     tree = pr.fetch_message_tree(message.message_tree_id, reviewed=False)
     return utils.prepare_tree(tree, message.message_tree_id)
@@ -175,24 +187,32 @@ def get_tree(
 
 @router.get("/{message_id}/children", response_model=list[protocol.Message])
 def get_children(
-    message_id: UUID, api_client: ApiClient = Depends(deps.get_api_client), db: Session = Depends(deps.get_db)
+    message_id: UUID,
+    auth_method: Optional[str] = None,
+    username: Optional[str] = None,
+    api_client: ApiClient = Depends(deps.get_api_client),
+    db: Session = Depends(deps.get_db),
 ):
     """
     Get all messages belonging to the same message tree.
     """
-    pr = PromptRepository(db, api_client)
+    pr = PromptRepository(db, api_client, auth_method=auth_method, username=username)
     messages = pr.fetch_message_children(message_id)
     return utils.prepare_message_list(messages)
 
 
 @router.get("/{message_id}/descendants", response_model=protocol.MessageTree)
 def get_descendants(
-    message_id: UUID, api_client: ApiClient = Depends(deps.get_api_client), db: Session = Depends(deps.get_db)
+    message_id: UUID,
+    auth_method: Optional[str] = None,
+    username: Optional[str] = None,
+    api_client: ApiClient = Depends(deps.get_api_client),
+    db: Session = Depends(deps.get_db),
 ):
     """
     Get a subtree which starts with this message.
     """
-    pr = PromptRepository(db, api_client)
+    pr = PromptRepository(db, api_client, auth_method=auth_method, username=username)
     message = pr.fetch_message(message_id)
     descendants = pr.fetch_message_descendants(message)
     return utils.prepare_tree(descendants, message.id)
@@ -200,12 +220,16 @@ def get_descendants(
 
 @router.get("/{message_id}/longest_conversation_in_tree", response_model=protocol.Conversation)
 def get_longest_conv(
-    message_id: UUID, api_client: ApiClient = Depends(deps.get_api_client), db: Session = Depends(deps.get_db)
+    message_id: UUID,
+    auth_method: Optional[str] = None,
+    username: Optional[str] = None,
+    api_client: ApiClient = Depends(deps.get_api_client),
+    db: Session = Depends(deps.get_db),
 ):
     """
     Get the longest conversation from the tree of the message.
     """
-    pr = PromptRepository(db, api_client)
+    pr = PromptRepository(db, api_client, auth_method=auth_method, username=username)
     message = pr.fetch_message(message_id)
     conv = pr.fetch_longest_conversation(message.message_tree_id)
     return utils.prepare_conversation(conv)
@@ -213,12 +237,16 @@ def get_longest_conv(
 
 @router.get("/{message_id}/max_children_in_tree", response_model=protocol.MessageTree)
 def get_max_children(
-    message_id: UUID, api_client: ApiClient = Depends(deps.get_api_client), db: Session = Depends(deps.get_db)
+    message_id: UUID,
+    auth_method: Optional[str] = None,
+    username: Optional[str] = None,
+    api_client: ApiClient = Depends(deps.get_api_client),
+    db: Session = Depends(deps.get_db),
 ):
     """
     Get message with the most children from the tree of the provided message.
     """
-    pr = PromptRepository(db, api_client)
+    pr = PromptRepository(db, api_client, auth_method=auth_method, username=username)
     message = pr.fetch_message(message_id)
     message, children = pr.fetch_message_with_max_children(message.message_tree_id)
     return utils.prepare_tree([message, *children], message.id)

backend/oasst_backend/api/v1/users.py

@@ -230,7 +230,7 @@ def query_user_messages(
     """
     Query user messages.
     """
-    pr = PromptRepository(db, api_client)
+    pr = PromptRepository(db, api_client, user_id=user_id)
     messages = pr.query_messages_ordered_by_created_date(
         user_id=user_id,
         api_client_id=api_client_id,

backend/oasst_backend/api/v1/utils.py

@@ -14,7 +14,8 @@ def prepare_message(m: Message) -> protocol.Message:
         lang=m.lang,
         is_assistant=(m.role == "assistant"),
         created_date=m.created_date,
-        emojis=m.emojis,
+        emojis=m.emojis or {},
+        user_emojis=m.user_emojis or [],
     )
 
 
@@ -30,6 +31,8 @@ def prepare_conversation_message_list(messages: list[Message]) -> list[protocol.
             text=message.text,
             lang=message.lang,
             is_assistant=(message.role == "assistant"),
+            emojis=message.emojis or {},
+            user_emojis=message.user_emojis or [],
         )
         for message in messages
     ]