From 2ff207218ef1d2d838202464e109a25bd839d408 Mon Sep 17 00:00:00 2001
From: mattboret <mattboret@gmail.com>
Date: Fri, 3 May 2024 18:04:09 +0200
Subject: [PATCH] Confluence: Add config to index only active pages (#1348)

Co-authored-by: Matthieu Boret <matthieu.boret@fr.clara.net>
---
 backend/danswer/configs/app_configs.py             | 7 ++++++-
 backend/danswer/connectors/confluence/connector.py | 7 +++++++
 2 files changed, 13 insertions(+), 1 deletion(-)

diff --git a/backend/danswer/configs/app_configs.py b/backend/danswer/configs/app_configs.py
index 809e636651..4dbae24952 100644
--- a/backend/danswer/configs/app_configs.py
+++ b/backend/danswer/configs/app_configs.py
@@ -3,7 +3,6 @@ import os
 from danswer.configs.constants import AuthType
 from danswer.configs.constants import DocumentIndexType
 
-
 #####
 # App Configs
 #####
@@ -167,6 +166,12 @@ CONFLUENCE_CONNECTOR_LABELS_TO_SKIP = [
     )
     if ignored_tag
 ]
+
+# Avoid to get archived pages
+CONFLUENCE_CONNECTOR_INDEX_ONLY_ACTIVE_PAGES = (
+    os.environ.get("CONFLUENCE_CONNECTOR_INDEX_ONLY_ACTIVE_PAGES", "").lower() == "true"
+)
+
 JIRA_CONNECTOR_LABELS_TO_SKIP = [
     ignored_tag
     for ignored_tag in os.environ.get("JIRA_CONNECTOR_LABELS_TO_SKIP", "").split(",")
diff --git a/backend/danswer/connectors/confluence/connector.py b/backend/danswer/connectors/confluence/connector.py
index f9f5e7c3bb..b0272d44cb 100644
--- a/backend/danswer/connectors/confluence/connector.py
+++ b/backend/danswer/connectors/confluence/connector.py
@@ -11,6 +11,7 @@ import bs4
 from atlassian import Confluence  # type:ignore
 from requests import HTTPError
 
+from danswer.configs.app_configs import CONFLUENCE_CONNECTOR_INDEX_ONLY_ACTIVE_PAGES
 from danswer.configs.app_configs import CONFLUENCE_CONNECTOR_LABELS_TO_SKIP
 from danswer.configs.app_configs import CONTINUE_ON_CONNECTOR_FAILURE
 from danswer.configs.app_configs import INDEX_BATCH_SIZE
@@ -219,6 +220,9 @@ class ConfluenceConnector(LoadConnector, PollConnector):
                     self.space,
                     start=start_ind,
                     limit=batch_size,
+                    status="current"
+                    if CONFLUENCE_CONNECTOR_INDEX_ONLY_ACTIVE_PAGES
+                    else None,
                     expand="body.storage.value,version",
                 )
             except Exception:
@@ -237,6 +241,9 @@ class ConfluenceConnector(LoadConnector, PollConnector):
                                 self.space,
                                 start=start_ind + i,
                                 limit=1,
+                                status="current"
+                                if CONFLUENCE_CONNECTOR_INDEX_ONLY_ACTIVE_PAGES
+                                else None,
                                 expand="body.storage.value,version",
                             )
                         )