testing multiprocessing with 10 contexts

ai_ta_backend/filtering_contexts.py

@@ -17,7 +17,7 @@
 
 from concurrent.futures import ThreadPoolExecutor, ProcessPoolExecutor
 from functools import partial
-from multiprocessing import Manager
+from multiprocessing import Pool, Manager
 
 
 from ai_ta_backend.utils_tokenization import count_tokens_and_cost
@@ -41,10 +41,10 @@ def run_context_filtering(contexts, user_query, max_time_before_return=45, max_c
     partial_func2 = partial(select_context, result=filtered_contexts)
 
     with ProcessPoolExecutor(max_workers=200) as executor1:
-      results1 = list(executor1.map(partial_func1, contexts))
+      results1 = list(executor1.map(partial_func1, contexts[:10]))
 
     print(f"⏰ ThreadPool runtime: {(time.monotonic() - start_time):.2f} seconds")
-
+  
     with ProcessPoolExecutor(max_workers=200) as executor:
       executor.map(partial_func2, results1)