From 17005fb705d346cc0d68a0ff46b3ed41e4c11b9e Mon Sep 17 00:00:00 2001
From: hagen-danswer <hagen@danswer.ai>
Date: Mon, 22 Jul 2024 10:36:26 -0700
Subject: [PATCH] switched default pruning behavior and removed some logging
 (#1898)

---
 backend/danswer/background/celery/celery_utils.py | 6 ++----
 backend/danswer/configs/app_configs.py            | 4 ++--
 2 files changed, 4 insertions(+), 6 deletions(-)

diff --git a/backend/danswer/background/celery/celery_utils.py b/backend/danswer/background/celery/celery_utils.py
index 6b9b5a8960..eb9943b41e 100644
--- a/backend/danswer/background/celery/celery_utils.py
+++ b/backend/danswer/background/celery/celery_utils.py
@@ -6,8 +6,8 @@ from sqlalchemy.orm import Session
 from danswer.background.task_utils import name_cc_cleanup_task
 from danswer.background.task_utils import name_cc_prune_task
 from danswer.background.task_utils import name_document_set_sync_task
+from danswer.configs.app_configs import ALLOW_SIMULTANEOUS_PRUNING
 from danswer.configs.app_configs import MAX_PRUNING_DOCUMENT_RETRIEVAL_PER_MINUTE
-from danswer.configs.app_configs import PREVENT_SIMULTANEOUS_PRUNING
 from danswer.connectors.cross_connector_utils.rate_limit_wrapper import (
     rate_limit_builder,
 )
@@ -80,7 +80,7 @@ def should_prune_cc_pair(
             return True
         return False
 
-    if PREVENT_SIMULTANEOUS_PRUNING:
+    if not ALLOW_SIMULTANEOUS_PRUNING:
         pruning_type_task_name = name_cc_prune_task()
         last_pruning_type_task = get_latest_task_by_type(
             pruning_type_task_name, db_session
@@ -89,11 +89,9 @@ def should_prune_cc_pair(
         if last_pruning_type_task and check_task_is_live_and_not_timed_out(
             last_pruning_type_task, db_session
         ):
-            logger.info("Another Connector is already pruning. Skipping.")
             return False
 
     if check_task_is_live_and_not_timed_out(last_pruning_task, db_session):
-        logger.info(f"Connector '{connector.name}' is already pruning. Skipping.")
         return False
 
     if not last_pruning_task.start_time:
diff --git a/backend/danswer/configs/app_configs.py b/backend/danswer/configs/app_configs.py
index ccebfdb213..e927d76b8b 100644
--- a/backend/danswer/configs/app_configs.py
+++ b/backend/danswer/configs/app_configs.py
@@ -214,8 +214,8 @@ EXPERIMENTAL_CHECKPOINTING_ENABLED = (
 
 DEFAULT_PRUNING_FREQ = 60 * 60 * 24  # Once a day
 
-PREVENT_SIMULTANEOUS_PRUNING = (
-    os.environ.get("PREVENT_SIMULTANEOUS_PRUNING", "").lower() == "true"
+ALLOW_SIMULTANEOUS_PRUNING = (
+    os.environ.get("ALLOW_SIMULTANEOUS_PRUNING", "").lower() == "true"
 )
 
 # This is the maxiumum rate at which documents are queried for a pruning job. 0 disables the limitation.