update files

app.py

@@ -589,13 +589,13 @@ def get_text_from_link(url_link):
 
 def get_google_search_results(api_key, search_engine_id, google_search, sort_date, search_string, page=0):
     search_results = []
-    for i in range(3):
+    for i in range(1):
         try:
             search_results += google_search(
                 search_string,
                 api_key,
                 search_engine_id,
-                num=10,
+                num=3,     # num=10,
                 start=0 + 10 * page,
                 sort="date:r:19000101:" + sort_date,
                 dateRestrict=None,
@@ -603,7 +603,23 @@ def get_google_search_results(api_key, search_engine_id, google_search, sort_dat
             )
             break
         except:
-            sleep(3)
+            sleep(1)
+
+    # for i in range(3):
+    #     try:
+    #         search_results += google_search(
+    #             search_string,
+    #             api_key,
+    #             search_engine_id,
+    #             num=10,
+    #             start=0 + 10 * page,
+    #             sort="date:r:19000101:" + sort_date,
+    #             dateRestrict=None,
+    #             gl="US"
+    #         )
+    #         break
+    #     except:
+    #         sleep(3)
 
     return search_results
 
@@ -750,7 +766,6 @@ def averitec_search(claim, generate_question, speaker="they", check_date="2024-0
         for page_num in range(n_pages):
             search_results = get_google_search_results(api_key, search_engine_id, google_search, sort_date,
                                                        this_search_string, page=page_num)
-            search_results = search_results[:5]
 
             for result in search_results:
                 link = str(result["link"])
@@ -829,7 +844,7 @@ def generate_step2_reference_corpus(reference_file):
     return tokenized_corpus, prompt_corpus
 
 
-def decorate_with_questions(claim, retrieve_evidence, top_k=10):  # top_k=100
+def decorate_with_questions(claim, retrieve_evidence, top_k=5):  # top_k=10, 100
     #
     reference_file = "averitec/data/train.json"
     tokenized_corpus, prompt_corpus = generate_step2_reference_corpus(reference_file)