Fix typos (#1143)

Found via `codespell -S .mypy_cache,yarn.lock,*.json,*.ipynb -L
rouge,nam,vie`

backend/oasst_backend/api/v1/hugging_face.py

@@ -18,7 +18,7 @@ async def get_text_toxicity(
 
     Args:
         msg (str): the message that we want to analyze.
-        api_client (ApiClient, optional): authentification of the user of the request.
+        api_client (ApiClient, optional): authentication of the user of the request.
             Defaults to Depends(deps.get_trusted_api_client).
 
     Returns:

backend/oasst_backend/config.py

@@ -14,7 +14,7 @@ class TreeManagerConfiguration(BaseModel):
     number is reached."""
 
     max_initial_prompt_review: int = 100
-    """Maximum number of initial prompts under review before no more inital prompt tasks will be handed out."""
+    """Maximum number of initial prompts under review before no more initial prompt tasks will be handed out."""
 
     max_tree_depth: int = 3
     """Maximum depth of message tree."""
@@ -75,7 +75,7 @@ class TreeManagerConfiguration(BaseModel):
 
     min_active_rankings_per_lang: int = 0
     """When the number of active ranking tasks is below this value when a tree enters a terminal
-    state an available trees in BACKLOG_RANKING will be actived (i.e. enters the RANKING state)."""
+    state an available trees in BACKLOG_RANKING will be activated (i.e. enters the RANKING state)."""
 
     labels_initial_prompt: list[TextLabel] = [
         TextLabel.spam,

backend/oasst_backend/models/message_tree_state.py

@@ -12,7 +12,7 @@ class State(str, Enum):
     """States of the Open-Assistant message tree state machine."""
 
     INITIAL_PROMPT_REVIEW = "initial_prompt_review"
-    """In this state the message tree consists only of a single inital prompt root node.
+    """In this state the message tree consists only of a single initial prompt root node.
     Initial prompt labeling tasks will determine if the tree goes into `growing` or
     `aborted_low_grade` state."""
 
@@ -33,11 +33,11 @@ class State(str, Enum):
     compute the aggergated ranking scores that will appear in the dataset."""
 
     READY_FOR_EXPORT = "ready_for_export"
-    """The Scoring algorithm computed rankings scores for all childern. The message tree can be
+    """The Scoring algorithm computed rankings scores for all children. The message tree can be
     exported as part of an Open-Assistant message tree dataset."""
 
     SCORING_FAILED = "scoring_failed"
-    """An exception occured in the scoring algorithm."""
+    """An exception occurred in the scoring algorithm."""
 
     ABORTED_LOW_GRADE = "aborted_low_grade"
     """The system received too many bad reviews and stopped handing out tasks for this message tree."""

backend/oasst_backend/prompt_repository.py

@@ -484,7 +484,7 @@ class PromptRepository:
                     OasstErrorCode.TASK_PAYLOAD_TYPE_MISMATCH,
                 )
 
-            logger.debug(f"text_labels relpy: {valid_labels=}, {mandatory_labels=}")
+            logger.debug(f"text_labels reply: {valid_labels=}, {mandatory_labels=}")
 
             if valid_labels:
                 if not all([label in valid_labels for label in text_labels.labels.keys()]):

backend/oasst_backend/task_repository.py

@@ -177,7 +177,7 @@ class TaskRepository:
         if not allow_personal_tasks and not task.collective:
             raise OasstError("This is not a collective task", OasstErrorCode.TASK_NOT_COLLECTIVE)
         if task.done:
-            raise OasstError("Allready closed", OasstErrorCode.TASK_ALREADY_DONE)
+            raise OasstError("Already closed", OasstErrorCode.TASK_ALREADY_DONE)
 
         task.done = True
         self.db.add(task)

backend/oasst_backend/tree_manager.py

@@ -320,7 +320,7 @@ class TreeManager:
             if num_red_flag is not None and num_red_flag >= self.cfg.auto_mod_red_flags:
                 if m.parent_id is None:
                     logger.warning(
-                        f"[AUTO MOD] Halting tree {m.message_tree_id}, inital prompt got too many red flags ({m.emojis})."
+                        f"[AUTO MOD] Halting tree {m.message_tree_id}, initial prompt got too many red flags ({m.emojis})."
                     )
                     self.enter_low_grade_state(m.message_tree_id)
                 else:
@@ -895,7 +895,7 @@ class TreeManager:
                     logger.warning("The intersection of ranking results ID sets has less than two elements. Skipping.")
                     continue
 
-                # keep only elements in commond set
+                # keep only elements in command set
                 ordered_ids_list = [list(filter(lambda x: x in common_set, ids)) for ids in ordered_ids_list]
                 assert all(len(x) == len(common_set) for x in ordered_ids_list)
 
@@ -1069,7 +1069,7 @@ HAVING(COUNT(mr.message_id) FILTER (WHERE mr.user_id = :user_id) = 0)
 """
 
     def query_incomplete_rankings(self, lang: str) -> list[IncompleteRankingsRow]:
-        """Query parents which have childern that need further rankings"""
+        """Query parents which have children that need further rankings"""
 
         user_id = self.pr.user_id if not settings.DEBUG_ALLOW_DUPLICATE_TASKS else None
         r = self.db.execute(
@@ -1256,7 +1256,7 @@ LEFT JOIN message_reaction mr ON mr.task_id = t.id AND mr.payload_type = 'Rankin
 
     @managed_tx_method(CommitMode.COMMIT)
     def ensure_tree_states(self) -> None:
-        """Add message tree state rows for all root nodes (inital prompt messages)."""
+        """Add message tree state rows for all root nodes (initial prompt messages)."""
 
         missing_tree_ids = self.query_misssing_tree_states()
         for id in missing_tree_ids:
@@ -1598,7 +1598,7 @@ DELETE FROM message WHERE message_tree_id = :message_tree_id;
         total_messages = sum(len(x) for x in replies_by_tree.values())
         logger.debug(f"found: {len(replies_by_tree)} trees; {len(prompts)} prompts; {total_messages} messages;")
 
-        # remove all trees based on inital prompts of the user
+        # remove all trees based on initial prompts of the user
         if purge_initial_prompts:
             for p in prompts:
                 self.purge_message_tree(p.message_tree_id)
@@ -1636,7 +1636,7 @@ DELETE FROM message WHERE message_tree_id = :message_tree_id;
                     logger.debug(f"purging message: {m.id}")
                     self._purge_message_internal(m.id)
 
-            # update childern counts
+            # update children counts
             self.pr.update_children_counts(m.message_tree_id)
 
             # reactivate tree

backend/oasst_backend/utils/ranking.py

@@ -66,7 +66,7 @@ def get_winner(pairs):
 
 def get_ranking(pairs):
     """
-    Abuses concordance property to get a (not necessarily unqiue) ranking.
+    Abuses concordance property to get a (not necessarily unique) ranking.
     The lack of uniqueness is due to the potential existence of multiple
     equally ranked winners. We have to pick one, which is where
     the non-uniqueness comes from