Corrected prompt. Now runs Haiku correctly

Dockerfile

@@ -14,14 +14,14 @@ RUN apt-get update && apt-get install -y \
 
 WORKDIR /src
 
-COPY requirements_cpu.txt .
+COPY requirements_aws.txt .
 
 RUN pip uninstall -y typing_extensions \
 && pip install --no-cache-dir --target=/install typing_extensions==4.12.2 \
 && pip install torch==2.5.1+cpu --target=/install --index-url https://download.pytorch.org/whl/cpu \
-&& pip install --no-cache-dir --target=/install -r requirements_cpu.txt
+&& pip install --no-cache-dir --target=/install -r requirements_aws.txt
 
-RUN rm requirements_cpu.txt
+RUN rm requirements_aws.txt
 
 # Stage 2: Final runtime image
 FROM public.ecr.aws/docker/library/python:3.11.9-slim-bookworm

app.py

@@ -1,7 +1,7 @@
 import os
 import socket
 from tools.helper_functions import ensure_output_folder_exists, add_folder_to_path, put_columns_in_df, get_connection_params, output_folder, get_or_create_env_var, reveal_feedback_buttons, wipe_logs, model_full_names, view_table, empty_output_vars_extract_topics, empty_output_vars_summarise, RUN_LOCAL_MODEL
-from tools.aws_functions import upload_file_to_s3
+from tools.aws_functions import upload_file_to_s3, RUN_AWS_FUNCTIONS
 from tools.llm_api_call import extract_topics, load_in_data_file, load_in_previous_data_files, sample_reference_table_summaries, summarise_output_topics, batch_size_default
 from tools.auth import authenticate_user
 from tools.prompts import initial_table_prompt, prompt2, prompt3, system_prompt, add_existing_topics_system_prompt, add_existing_topics_prompt
@@ -107,6 +107,14 @@ local_model_type = "Gemma 2b"
 if RUN_LOCAL_MODEL == "1":
     load_model(local_model_type, chatf.gpu_layers, chatf.context_length, chatf.gpu_config, chatf.cpu_config, chatf.torch_device)
 
+    default_model_choice = "gemma_2b_it_local"
+
+elif RUN_AWS_FUNCTIONS == "1":
+    default_model_choice = "anthropic.claude-3-haiku-20240307-v1:0"
+
+else:
+    default_model_choice = "gemini-1.5-flash-002"
+
 # Create the gradio interface
 app = gr.Blocks(theme = gr.themes.Base())
 
@@ -167,7 +175,7 @@ with app:
         """
         )
         with gr.Row():
-            model_choice = gr.Dropdown(value = "gemini-1.5-flash-002", choices = model_full_names, label="LLM model to use", multiselect=False)
+            model_choice = gr.Dropdown(value = default_model_choice, choices = model_full_names, label="LLM model to use", multiselect=False)
             in_api_key = gr.Textbox(value = "", label="Enter Gemini API key (only if using Google API models)", lines=1, type="password")
 
         with gr.Accordion("Upload xlsx or csv files with consultation responses", open = True):

requirements_aws.txt

@@ -0,0 +1,16 @@
+pandas==2.2.3
+gradio==5.8.0
+boto3==1.35.71
+pyarrow==18.1.0
+openpyxl==3.1.3
+markdown==3.7
+tabulate==0.9.0
+lxml==5.3.0
+google-generativeai==0.8.3
+html5lib==1.1
+beautifulsoup4==4.12.3
+rapidfuzz==3.10.1
+llama-cpp-python==0.2.90 --extra-index-url https://abetlen.github.io/llama-cpp-python/whl/cpu
+transformers==4.47.0
+numpy==1.26.4
+typing_extensions==4.12.2

tools/llm_api_call.py

@@ -436,7 +436,10 @@ def send_request(prompt: str, conversation_history: List[dict], model: object, c
 # output_str = output['choices'][0]['text']
 
     # Check if is a LLama.cpp model response
-    if 'choices' in response:
+        # Check if the response is a ResponseObject
+    if isinstance(response, ResponseObject):
+        conversation_history.append({'role': 'assistant', 'parts': [response.text]})
+    elif 'choices' in response:
         conversation_history.append({'role': 'assistant', 'parts': [response['choices'][0]['text']]})
     else:
         conversation_history.append({'role': 'assistant', 'parts': [response.text]})
@@ -477,19 +480,17 @@ def process_requests(prompts: List[str], system_prompt: str, conversation_histor
 
         response, conversation_history = send_request(prompt, conversation_history, model=model, config=config, model_choice=model_choice, system_prompt=system_prompt, temperature=temperature)
 
-        if 'choices' in response:            
-            responses.append(response)      
-
+        if isinstance(response, ResponseObject):
+            responses.append(response)
+            whole_conversation.append(prompt)
+            whole_conversation.append(response.text)
+        elif 'choices' in response:
+            responses.append(response)
             # Create conversation txt object
             whole_conversation.append(prompt)
             whole_conversation.append(response['choices'][0]['text'])
-
         else:
             responses.append(response)
-            #print("response.usage_metadata:", response.usage_metadata)
-            #print("Response.text:", response.text)
-            #print("responses:", responses)
-            # Create conversation txt object
             whole_conversation.append(prompt)
             whole_conversation.append(response.text)
 
@@ -714,14 +715,22 @@ def write_llm_output_and_logs(responses: List[ResponseObject],
             return input_string[last_index:]  # Return everything from the last match onward
         return input_string  # Return the original string if the term is not found
 
-    if "choices" in responses[-1]:
-        print("Text response:", responses[-1]["choices"][0]['text'])
+        # Check if the last response is a ResponseObject
+    if isinstance(responses[-1], ResponseObject):
+        #print("Text response:", responses[-1].text)
+        start_of_table_response = remove_before_last_term(responses[-1].text)
+        cleaned_response = clean_markdown_table(start_of_table_response)
+        print("cleaned_response:", cleaned_response)
+    elif "choices" in responses[-1]:
+        #print("Text response:", responses[-1]["choices"][0]['text'])
         start_of_table_response = remove_before_last_term(responses[-1]["choices"][0]['text'])
         cleaned_response = clean_markdown_table(start_of_table_response)
         print("cleaned_response:", cleaned_response)
     else:
+        #print("Text response:", responses[-1].text)
         start_of_table_response = remove_before_last_term(responses[-1].text)
         cleaned_response = clean_markdown_table(start_of_table_response)
+        print("cleaned_response:", cleaned_response)
     
     markdown_table = markdown.markdown(cleaned_response, extensions=['tables'])
 
@@ -1328,15 +1337,20 @@ def extract_topics(in_data_file,
                 try:
                     final_table_output_path = output_folder + batch_file_path_details + "_full_final_response_" + model_choice_clean + "_temp_" + str(temperature) + ".txt"
 
-                    if "choices" in responses[-1]:
+                    if isinstance(responses[-1], ResponseObject):
+                        with open(final_table_output_path, "w", encoding='utf-8', errors='replace') as f:
+                            f.write(responses[-1].text)
+                        display_table = responses[-1].text
+                    elif "choices" in responses[-1]:
                         with open(final_table_output_path, "w", encoding='utf-8', errors='replace') as f:
                             f.write(responses[-1]["choices"][0]['text'])
                         display_table =responses[-1]["choices"][0]['text']
-
                     else:
                         with open(final_table_output_path, "w", encoding='utf-8', errors='replace') as f:
                             f.write(responses[-1].text)
                         display_table = responses[-1].text
+                    
+
 
                     log_files_output_paths.append(final_table_output_path)
 
@@ -1539,12 +1553,15 @@ def summarise_output_topics_query(model_choice:str, in_api_key:str, temperature:
 
     print("Finished summary query")
 
-    # Extract text from the `responses` list
-    if "choices" in responses[-1]:
+    if isinstance(responses[-1], ResponseObject):
+        response_texts = [resp.text for resp in responses]
+    elif "choices" in responses[-1]:
         response_texts = [resp["choices"][0]['text'] for resp in responses]
     else:
         response_texts = [resp.text for resp in responses]
 
+
+
     latest_response_text = response_texts[-1]
 
     #print("latest_response_text:", latest_response_text)

tools/prompts.py

@@ -3,7 +3,7 @@ system_prompt = """You are a researcher analysing responses from an open text da
 initial_table_prompt = """The open text data is shown in the following table that contains two columns, Reference and Response. Response table: 
 {response_table}
 
-Your task is to create one new markdown table with the headings 'General Topic', 'Subtopic', 'Sentiment', 'Summary', and 'Response references'.
+Your task is to create one new markdown table with the headings 'General Topic', 'Subtopic', 'Sentiment', 'Response references', and 'Summary'.
 In the first column identify general topics relevant to responses. Create as many general topics as you can.
 In the second column list subtopics relevant to responses. Make the subtopics as specific as possible and make sure they cover every issue mentioned.
 In the third column write the sentiment of the subtopic: Negative, Neutral, or Positive.
@@ -28,7 +28,7 @@ Topics known to be relevant to this dataset are shown in the following Topics ta
 {topics}
 
 Your task is to create one new markdown table, assigning responses from the Response table to existing topics, or to create new topics if no existing topics are relevant.  
-Create a new markdown table with the headings 'General Topic', 'Subtopic', 'Sentiment', 'Summary', and 'Response references'.
+Create a new markdown table with the headings 'General Topic', 'Subtopic', 'Sentiment', 'Response references', and 'Summary'.
 In the first and second columns, assign General Topics and Subtopics to Responses. Assign topics from the Topics table above if they are very relevant to the text of the Response. Fill in the General Topic and Sentiment for the Subtopic if they do not already exist. If you find a new topic that does not exist in the Topics table, add a new row to the new table. Make the General Topic and Subtopic as specific as possible.
 In the third column, write the sentiment of the Subtopic: Negative, Neutral, or Positive.
 In the fourth column list each specific Response reference number that is relevant to the Subtopic, separated by commas. Do no write any other text in this column.