Update chatbot.py

chatbot.py

@@ -167,36 +167,6 @@ glaucoma_db_path = 'glaucoma_results.db'
 patient_data = fetch_patient_data(cataract_db_path, glaucoma_db_path)
 readable_patient_data = transform_patient_data(patient_data)
 
-# Function to extract details from the input prompt
-def extract_details_from_prompt(prompt):
-    pattern = re.compile(r"(Glaucoma|Cataract) (\d+)", re.IGNORECASE)
-    matches = pattern.findall(prompt)
-    return [(match[0].capitalize(), int(match[1])) for match in matches]
-
-# Function to fetch specific patient data based on the condition and ID
-def get_specific_patient_data(patient_data, condition, patient_id):
-    specific_data = ""
-    if condition == "Cataract":
-        specific_data = "Cataract Results:\n"
-        for row in patient_data.get('cataract_results', []):
-            if isinstance(row, tuple) and row[0] == patient_id:
-                specific_data += f"Patient ID: {row[0]}, Red Quantity: {row[2]}, Green Quantity: {row[3]}, Blue Quantity: {row[4]}, Stage: {row[5]}\n"
-                break
-    elif condition == "Glaucoma":
-        specific_data = "Glaucoma Results:\n"
-        for row in patient_data.get('results', []):
-            if isinstance(row, tuple) and row[0] == patient_id:
-                specific_data += f"Patient ID: {row[0]}, Cup Area: {row[2]}, Disk Area: {row[3]}, Rim Area: {row[4]}, Rim to Disc Line Ratio: {row[5]}, DDLS Stage: {row[6]}\n"
-                break
-    return specific_data
-
-# Function to aggregate patient history for all mentioned IDs in the question
-def get_aggregated_patient_history(patient_data, details):
-    history = ""
-    for condition, patient_id in details:
-        history += get_specific_patient_data(patient_data, condition, patient_id) + "\n"
-    return history.strip()
-
 # Toggle visibility of input elements based on input type
 def toggle_visibility(input_type):
     if input_type == "Voice":
@@ -211,6 +181,7 @@ def cleanup_response(response):
         response = response[answer_start + len("Answer:"):].strip()
     return response
 
+# Gradio interface for the chatbot
 def chatbot(audio, input_type, text):
     if input_type == "Voice":
         transcription = query_whisper(audio.name)
@@ -220,20 +191,13 @@ def chatbot(audio, input_type, text):
     else:
         query = text
 
-    # Extract details from the prompt
-    details = extract_details_from_prompt(query)
-    
-    # Get aggregated patient history based on the extracted details
-    patient_history = get_aggregated_patient_history(patient_data, details)
-
-    # Create the payload with the patient history and the user's query
+    # Directly use the transformed patient data as context input
     payload = {
-        "inputs": f"role: ophthalmologist assistant patient history: {patient_history} question: {query}"
+        "inputs": f"role: ophthalmologist assistant patient history: {readable_patient_data} question: {query}"
     }
 
     logging.debug(f"Raw input to the LLM: {payload['inputs']}")
 
-    # Query the Hugging Face model with the payload
     response = query_huggingface(payload)
     if isinstance(response, list):
         raw_response = response[0].get("generated_text", "Sorry, I couldn't generate a response.")