From 5d86ad6961795edb71c5cee555b76b307da2b9a6 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Nielson=20Jann=C3=A9?= <nielsonj@gmail.com>
Date: Fri, 28 Mar 2025 18:56:01 +0100
Subject: [PATCH] Add logging warning when no page number is found

---
 generic_rag/graphs/ret_gen.py | 7 +++++++
 1 file changed, 7 insertions(+)

diff --git a/generic_rag/graphs/ret_gen.py b/generic_rag/graphs/ret_gen.py
index b3c97f4..c30d6e2 100644
--- a/generic_rag/graphs/ret_gen.py
+++ b/generic_rag/graphs/ret_gen.py
@@ -1,3 +1,4 @@
+import logging
 from pathlib import Path
 from typing import Any, Union
 
@@ -7,6 +8,9 @@ from langgraph.checkpoint.memory import MemorySaver
 from langgraph.graph import END, START, StateGraph
 from typing_extensions import List, TypedDict
 
+logging.basicConfig(level=logging.INFO)
+logger = logging.getLogger(__name__)
+
 
 class State(TypedDict):
     question: str
@@ -79,6 +83,9 @@ class RetGenLangGraph:
             else:
                 pdf_sources[source].add(page_number)
 
+            if len(pdf_sources[source]) == 0:
+                logging.warning(f"PDF source {source} has no page number. Please check the metadata of the document.")
+
         return pdf_sources
 
     def get_last_web_sources(self) -> set: