refactor

app.py

@@ -15,9 +15,6 @@ from auditqa.reader import nvidia_client, dedicated_endpoint
 from auditqa.utils import make_html_source, parse_output_llm_with_sources, save_logs, get_message_template, get_client_location, get_client_ip
 from dotenv import load_dotenv
 from threading import Lock
-# import json
-# from functools import partial
-# import time
 from gradio.routes import Request
 from qdrant_client import QdrantClient
 
@@ -26,7 +23,7 @@ from auditqa.logging_config import setup_logging
 setup_logging()
 import logging
 logger = logging.getLogger(__name__)
-logger.setLevel(logging.DEBUG)  # Ensure debug logging is enabled
+logger.setLevel(logging.DEBUG) 
 
 load_dotenv()
 
@@ -44,12 +41,12 @@ JSON_DATASET_PATH = JSON_DATASET_DIR / f"logs-{uuid4()}.json"
 # the logs are written to dataset repo periodically from local logs
 # https://huggingface.co/spaces/Wauplin/space_to_dataset_saver
 scheduler = CommitScheduler(
-     repo_id="mtyrrell/spaces_log",
+     repo_id=model_config.get('app','repo_id'),
      repo_type="dataset",
      folder_path=JSON_DATASET_DIR,
-     path_in_repo=".",
+     path_in_repo="audit_chatbot",
      token=SPACES_LOG,
-     every=2)
+     every=2) # TESTING: every 2 seconds
 
 #####--------------- VECTOR STORE -------------------------------------------------
 # Configure cloud Qdrant client
@@ -511,7 +508,7 @@ with gr.Blocks(title="Audit Q&A", css= "style.css", theme=theme,elem_id = "main-
                 
                 
 
-    #-------------------- Feedback UI elements + state management -------------------------
+    #-------------------- New UI elements for Feedback -------------------------
     with gr.Row(visible=False) as feedback_row:
         gr.Markdown("Was this response helpful?")
         with gr.Row():
@@ -520,7 +517,6 @@ with gr.Blocks(title="Audit Q&A", css= "style.css", theme=theme,elem_id = "main-
     
     feedback_thanks = gr.Markdown("Thanks for the feedback!", visible=False)
     feedback_state = gr.State()  # Add state to store logs data
-    debug_ip = gr.Markdown(visible=False)  # Add debug display for IP
 
     def show_feedback(logs):
         """Show feedback buttons and store logs in state"""
@@ -528,29 +524,25 @@ with gr.Blocks(title="Audit Q&A", css= "style.css", theme=theme,elem_id = "main-
 
     def submit_feedback_okay(logs_data):
         """Handle 'okay' feedback submission"""
-        ip = logs_data.get('client_ip', 'No IP found')
-        location = logs_data.get('client_location', 'No location found')
-        return submit_feedback("okay", logs_data) + (gr.update(visible=True, value=f"TESTING - Client IP: {ip}, Location: {location}"),)
+        return submit_feedback("okay", logs_data)
 
     def submit_feedback_not_okay(logs_data):
         """Handle 'not okay' feedback submission"""
-        ip = logs_data.get('client_ip', 'No IP found')
-        location = logs_data.get('client_location', 'No location found')
-        return submit_feedback("not_okay", logs_data) + (gr.update(visible=True, value=f"TESTING - Client IP: {ip}, Location: {location}"),)
+        return submit_feedback("not_okay", logs_data)
     
     okay_btn.click(
         submit_feedback_okay,
         [feedback_state],
-        [feedback_row, feedback_thanks, debug_ip]
+        [feedback_row, feedback_thanks]
     )
     
     not_okay_btn.click(
         submit_feedback_not_okay,
         [feedback_state],
-        [feedback_row, feedback_thanks, debug_ip]
+        [feedback_row, feedback_thanks]
     )
 
-    #-------------------- Gradio voodoo continued -------------------------
+   #-------------------- Session Management + Geolocation -------------------------
 
     # Add these state components at the top level of the Blocks
     session_id = gr.State(None)
@@ -561,6 +553,8 @@ with gr.Blocks(title="Audit Q&A", css= "style.css", theme=theme,elem_id = "main-
         """Handler for getting client IP in Gradio context"""
         return get_client_ip(request)
     
+    #-------------------- Gradio voodoo -------------------------
+    
     # Update the event handlers
     (textbox
         .submit(get_client_ip_handler, [textbox], [client_ip], api_name="get_ip_textbox")
@@ -584,5 +578,4 @@ with gr.Blocks(title="Audit Q&A", css= "style.css", theme=theme,elem_id = "main-
 
     demo.queue()
 
-demo.launch()
-logger.info("App launched")
+demo.launch()

auditqa/reader.py

@@ -14,9 +14,8 @@ import logging
 logger = logging.getLogger(__name__)
 
 model_config = getconfig("model_params.cfg")
-# NVIDIA_SERVER = os.environ["NVIDIA_SERVERLESS"]
-# HF_token = os.environ["LLAMA_3_1"]
-# HF_token = os.getenv('LLAMA_3_1') # TESTING
+NVIDIA_SERVER = os.environ["NVIDIA_SERVERLESS"]
+HF_token = os.environ["LLAMA_3_1"]
 
 def nvidia_client():
     logger.info("NVIDIA client activated")
@@ -35,8 +34,8 @@ def nvidia_client():
 def dedicated_endpoint():
     logger.info("Serverless endpoint activated")
     try:
-        hf_api_key = os.environ["LLAMA_3_1"]
-        if not hf_api_key:
+        HF_token = os.environ["LLAMA_3_1"]
+        if not HF_token:
             raise ValueError("LLAMA_3_1 environment variable is empty")
             
         model_id = "meta-llama/Meta-Llama-3-8B-Instruct"
@@ -44,7 +43,7 @@ def dedicated_endpoint():
 
         client = InferenceClient(
             model=model_id,
-            api_key=hf_api_key,
+            api_key=HF_token,
         )
         
         logger.info("Serverless InferenceClient initialization successful")

auditqa/retriever.py

@@ -41,12 +41,11 @@ def create_filter(reports:list = [],sources:str =None,
     return filter
 
 
-
 def get_context(vectorstore,query,reports,sources,subtype,year):
     logger.info("Retriever activated")
     # create metadata filter
     # filter = create_filter(reports=reports,sources=sources,subtype=subtype,year=year)
-    filter = None
+    filter = None #TESTING 
 
     # getting context
     retriever = vectorstore.as_retriever(search_type="similarity_score_threshold", 

auditqa/utils.py

@@ -36,7 +36,6 @@ def save_logs(scheduler, JSON_DATASET_PATH, logs, feedback=None) -> None:
         raise
 
 
-
 def get_message_template(type, SYSTEM_PROMPT, USER_PROMPT):
     if type == 'NVIDIA':
         messages =  [{"role": "system", "content": SYSTEM_PROMPT},

logs/app.log

@@ -1367,3 +1367,4 @@ Make sure your token has the correct permissions.
 2025-02-13 13:39:44,129 - auditqa.reader - INFO - Serverless InferenceClient initialization successful
 2025-02-13 13:40:00,385 - auditqa.utils - INFO - logging done
 2025-02-13 13:40:05,605 - auditqa.utils - INFO - logging done
+2025-02-13 13:45:56,822 - __main__ - INFO - App launched

model_params.cfg

@@ -11,4 +11,6 @@ DEDICATED_MODEL = meta-llama/Llama-3.1-8B-Instruct
 DEDICATED_ENDPOINT = https://qu2d8m6dmsollhly.us-east-1.aws.endpoints.huggingface.cloud
 NVIDIA_MODEL = meta-llama/Llama-3.1-8B-Instruct
 NVIDIA_ENDPOINT = https://huggingface.co/api/integrations/dgx/v1
-MAX_TOKENS = 512
+MAX_TOKENS = 512
+[app]
+repo_id = mtyrrell/audit_chatbot