add caching

data_access.py

@@ -5,6 +5,7 @@ from typing import Optional
 
 import asyncpg
 import psycopg2
+from cachetools import TTLCache, cached
 from dotenv import load_dotenv
 import pandas as pd
 
@@ -36,6 +37,7 @@ async def get_questions(conn: asyncpg.Connection):
     questions = await conn.fetch("SELECT id, question_text FROM questions ORDER BY id")
     return [{"id": q["id"], "text": q["question_text"]} for q in questions]
 
+@cached(cache=TTLCache(ttl=1800, maxsize=1024))
 async def get_metadata(conn: asyncpg.Connection, question_id: int, source_finder_id_run_id: int):
     metadata = await conn.fetchrow('''
         SELECT metadata 
@@ -54,6 +56,7 @@ async def get_source_finders(conn: asyncpg.Connection):
 
 
 # Get distinct run IDs for a question
+@cached(cache=TTLCache(ttl=1800, maxsize=1024))
 async def get_run_ids(conn: asyncpg.Connection, source_finder_id: int, question_id: int = None):
     query = """
     select distinct sfr.description, srs.source_finder_run_id as run_id
@@ -243,6 +246,7 @@ async def get_unified_sources(conn: asyncpg.Connection, question_id: int, source
     return unified_results, stats_df
 
 
+@cached(cache=TTLCache(ttl=1800, maxsize=1024))
 async def get_source_text(conn: asyncpg.Connection, tractate_chunk_id: int):
     """
     Retrieves the text content for a given tractate chunk ID.

requirements.txt

@@ -1,4 +1,5 @@
 asyncpg
 gradio
 dotenv
-psycopg2
+psycopg2
+cachetools