Redis-dataprep: Make Redis connection consistent (#359)

There are both 'from langchain_community.vectorstores import Redis' and
'import redis' in prepare_doc_redis.py. Redis.from_texts_return_keys()
is with REDIS_URL as connection, while redis.Redis() takes the
combination of REDIS_HOST and REDIS_PORT as connection. When the former
and the latter are not the same, it will fail to get data from vector
database. Change to use REDIS_URL as the unified way.

Signed-off-by: Cathy Zhang <[email protected]>
Co-authored-by: Sihan Chen <[email protected]>

comps/dataprep/redis/langchain/prepare_doc_redis.py

@@ -10,7 +10,7 @@
 
 # from pyspark import SparkConf, SparkContext
 import redis
-from config import EMBED_MODEL, INDEX_NAME, KEY_INDEX_NAME, REDIS_HOST, REDIS_PORT, REDIS_URL
+from config import EMBED_MODEL, INDEX_NAME, KEY_INDEX_NAME, REDIS_URL
 from fastapi import Body, File, Form, HTTPException, UploadFile
 from langchain.text_splitter import RecursiveCharacterTextSplitter
 from langchain_community.embeddings import HuggingFaceBgeEmbeddings, HuggingFaceHubEmbeddings
@@ -35,7 +35,7 @@
 
 tei_embedding_endpoint = os.getenv("TEI_ENDPOINT")
 upload_folder = "./uploaded_files/"
-redis_pool = redis.ConnectionPool(host=REDIS_HOST, port=REDIS_PORT)
+redis_pool = redis.ConnectionPool.from_url(REDIS_URL)
 
 
 def check_index_existance(client):
@@ -318,7 +318,7 @@ async def delete_single_file(file_path: str = Body(..., embed=True)):
 
     # partially delete files/folders
     if delete_path.exists():
-        r = redis.Redis(host=REDIS_HOST, port=REDIS_PORT)
+        r = redis.Redis(connection_pool=redis_pool)
         client = r.ft(KEY_INDEX_NAME)
         client2 = r.ft(INDEX_NAME)
         doc_id = "file:" + encode_filename(file_path)