Updated requirements. More generous spaces GPU timeouts

requirements.txt

@@ -1,8 +1,8 @@
-pandas==2.3.1
+pandas==2.3.2
 gradio==5.44.1
-transformers==4.55.2
+transformers==4.56.0
 spaces==0.40.1
-boto3==1.40.11
+boto3==1.40.22
 pyarrow==21.0.0
 openpyxl==3.1.5
 markdown==3.7

requirements_aws.txt

@@ -1,8 +1,8 @@
-pandas==2.3.1
+pandas==2.3.2
 gradio==5.44.1
-transformers==4.55.2
+transformers==4.56.0
 spaces==0.40.1
-boto3==1.40.11
+boto3==1.40.22
 pyarrow==21.0.0
 openpyxl==3.1.5
 markdown==3.7

requirements_gpu.txt

@@ -1,8 +1,8 @@
-pandas==2.3.1
+pandas==2.3.2
 gradio==5.44.1
-transformers==4.55.2
+transformers==4.56.0
 spaces==0.40.1
-boto3==1.40.11
+boto3==1.40.22
 pyarrow==21.0.0
 openpyxl==3.1.5
 markdown==3.7

tools/dedup_summaries.py

@@ -140,9 +140,6 @@ def deduplicate_topics(reference_df:pd.DataFrame,
     file_data = pd.DataFrame()
     deduplicated_unique_table_markdown = ""
 
-    print('len(reference_df["Response References"].unique()):', len(reference_df["Response References"].unique()))
-    print('len(topic_summary_df["Topic_number"].unique()):', len(topic_summary_df["Topic_number"].unique()))
-
     if (len(reference_df["Response References"].unique()) == 1) | (len(topic_summary_df["Topic_number"].unique()) == 1):
         print("Data file outputs are too short for deduplicating. Returning original data.")
 
@@ -449,9 +446,7 @@ def summarise_output_topics_query(model_choice:str, in_api_key:str, temperature:
 
     return response_text, conversation_history, whole_conversation_metadata
 
-
-
-@spaces.GPU
+@spaces.GPU(duration=300)
 def summarise_output_topics(sampled_reference_table_df:pd.DataFrame,
                             topic_summary_df:pd.DataFrame,
                             reference_table_df:pd.DataFrame,
@@ -670,6 +665,7 @@ def summarise_output_topics(sampled_reference_table_df:pd.DataFrame,
 
         return sampled_reference_table_df, topic_summary_df_revised, reference_table_df_revised, output_files, summarised_outputs, latest_summary_completed, out_metadata_str, summarised_output_markdown, log_output_files, output_files, acc_input_tokens, acc_output_tokens, acc_number_of_calls, time_taken
 
+@spaces.GPU(duration=120)
 def overall_summary(topic_summary_df:pd.DataFrame,
                     model_choice:str,
                     in_api_key:str,

tools/llm_api_call.py

@@ -648,7 +648,7 @@ def generate_zero_shot_topics_df(zero_shot_topics:pd.DataFrame,
         
         return zero_shot_topics_df
 
-@spaces.GPU(duration=60)
+@spaces.GPU(duration=300)
 def extract_topics(in_data_file: GradioFileData,
               file_data:pd.DataFrame,
               existing_topics_table:pd.DataFrame,

tools/llm_funcs.py

@@ -533,7 +533,7 @@ def send_request(prompt: str, conversation_history: List[dict], google_client: a
     progress_bar = range(0,number_of_api_retry_attempts)
 
     # Generate the model's response
-    if "gemini" in model_choice:
+    if "Gemini" in model_source:
 
         for i in progress_bar:
             try:
@@ -541,7 +541,7 @@ def send_request(prompt: str, conversation_history: List[dict], google_client: a
 
                 response = google_client.models.generate_content(model=model_choice, contents=full_prompt, config=config)
 
-                print("Successful call to Gemini model.")
+                #print("Successful call to Gemini model.")
                 break
             except Exception as e:
                 # If fails, try again after X seconds in case there is a throttle limit
@@ -551,13 +551,13 @@ def send_request(prompt: str, conversation_history: List[dict], google_client: a
             
             if i == number_of_api_retry_attempts:
                 return ResponseObject(text="", usage_metadata={'RequestId':"FAILED"}), conversation_history
-    elif "anthropic.claude" in model_choice:
+    elif "AWS" in model_source:
         for i in progress_bar:
             try:
                 print("Calling AWS Claude model, attempt", i + 1)
                 response = call_aws_claude(prompt, system_prompt, temperature, max_tokens, model_choice, bedrock_runtime=bedrock_runtime, assistant_prefill=assistant_prefill)
 
-                print("Successful call to Claude model.")
+                #print("Successful call to Claude model.")
                 break
             except Exception as e:
                 # If fails, try again after X seconds in case there is a throttle limit
@@ -566,7 +566,7 @@ def send_request(prompt: str, conversation_history: List[dict], google_client: a
 
             if i == number_of_api_retry_attempts:
                 return ResponseObject(text="", usage_metadata={'RequestId':"FAILED"}), conversation_history
-    else:
+    elif "Local" in model_source:
         # This is the local model
         for i in progress_bar:
             try:
@@ -577,7 +577,7 @@ def send_request(prompt: str, conversation_history: List[dict], google_client: a
 
                 response = call_llama_cpp_chatmodel(prompt, system_prompt, gen_config, model=local_model)
 
-                print("Successful call to local model.")
+                #print("Successful call to local model.")
                 break
             except Exception as e:
                 # If fails, try again after X seconds in case there is a throttle limit
@@ -586,7 +586,10 @@ def send_request(prompt: str, conversation_history: List[dict], google_client: a
                 time.sleep(timeout_wait)
 
             if i == number_of_api_retry_attempts:
-                return ResponseObject(text="", usage_metadata={'RequestId':"FAILED"}), conversation_history       
+                return ResponseObject(text="", usage_metadata={'RequestId':"FAILED"}), conversation_history
+    else:
+        print("Model source not recognised")
+        return ResponseObject(text="", usage_metadata={'RequestId':"FAILED"}), conversation_history
 
     # Update the conversation history with the new prompt and response
     conversation_history.append({'role': 'user', 'parts': [prompt]})