From 0ee123d6a8fae58071b514b30c848cd882a33bfd Mon Sep 17 00:00:00 2001
From: Kamyab Nazari <kamyabnazari@icloud.com>
Date: Wed, 5 Jul 2023 21:24:15 +0200
Subject: [PATCH] Changing chunk sizes

---
 backend/ingest.py | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/backend/ingest.py b/backend/ingest.py
index 41938b4..a614a79 100644
--- a/backend/ingest.py
+++ b/backend/ingest.py
@@ -160,9 +160,9 @@ def text_to_docs(text: List[str], metadata: Dict[str, str]) -> List[Document]:
     all_chunks = []
     for page_num, page in text:
         text_splitter = RecursiveCharacterTextSplitter(
-            chunk_size=500,
+            chunk_size=3500,
             separators=["\n\n", "\n", ".", "!", "?", ",", " ", ""],
-            chunk_overlap=100,
+            chunk_overlap=500,
         )
         chunks = text_splitter.split_text(page)
         all_chunks.extend([(page_num, i, chunk) for i, chunk in enumerate(chunks)])