added new chain

.gitignore

@@ -164,3 +164,6 @@ cython_debug/
 *.sqlite3
 *.bin
 *.pickle
+
+
+*.db

app_gui.py

@@ -1,92 +1,107 @@
 # Import Gradio for UI, along with other necessary libraries
 import gradio as gr
+from fastapi import FastAPI
+from rag_app.agents.react_agent import agent_executor, llm
+from rag_app.chains import user_response_sentiment_prompt
+from typing import Dict
+import re
+from rag_app.utils.utils import extract_responses
+from rag_app.loading_data.load_S3_vector_stores import get_chroma_vs
 from rag_app.agents.react_agent import agent_executor
-from config import db
 # need to import the qa!
-db.create_new_session()
 
 
-def add_text(history, text):
-    """Function to add a new input to the chat history
-    
-    Return: return_description
-    """
-    
-    # Append the new text to the history with a placeholder for the response
-    history = history + [(text, None)]
-    return history, ""
-
 
-def bot(history):
-    """Function representing the bot's response mechanism
-    
-    """
-    
-    # Obtain the response from the 'infer' function using the latest input
-    response = infer(history[-1][0], history)
-    #sources = [doc.metadata.get("source") for doc in response['source_documents']]
-    #src_list = '\n'.join(sources)
-    #print_this = response['result'] + "\n\n\n Sources: \n\n\n" + src_list
 
+app = FastAPI()
+get_chroma_vs()
+user_sentiment_chain = user_response_sentiment_prompt | llm
+# data = user_sentiment_chain.invoke({"user_reponse":"thanks for the help"})
+data = user_sentiment_chain.invoke({"user_reponse":"OMG I AM SO LOST!!! HELP!!!"})
+responses = extract_responses(data)
+if responses['AI'] == "1":
+    print("GG")
 
-    #history[-1][1] = print_this #response['answer']
-    # Update the history with the bot's response
-    history[-1][1] = response['output']
-    return history
+if __name__ == "__main__":
 
+    # Function to add a new input to the chat history
+    def add_text(history, text):
+        # Append the new text to the history with a placeholder for the response
+        history = history + [(text, None)]
+        return history, ""
 
-def infer(question, history):
-    """Function to infer the response using the RAG model
-    
-    """
-    
-    # Use the question and history to query the RAG model
-    try:
-        result = agent_executor.invoke(
-            {
-                "input": question,
-                "chat_history": history
-            }
-        )
-        return result
-    except Exception:
-        raise gr.Error("Model is Overloaded, Please retry later!")
-    
+    # Function representing the bot's response mechanism
+    def bot(history):
+        # Obtain the response from the 'infer' function using the latest input
+        response = infer(history[-1][0], history)
+        history[-1][1] = response['output']
+        return history
 
+    # Function to infer the response using the RAG model
+    def infer(question, history):
+        # Use the question and history to query the RAG model
+        #result = qa({"query": question, "history": history, "question": question})
+        # try:
+        #     data = user_sentiment_chain.invoke({"user_reponse":question})
+        #     responses = extract_responses(data)
+        #     if responses['AI'] == "1":
+        #         pass
+        #         # Do important stuff here plox
+        #         # store into database
+        # except Exception as e:
+        #     raise e
+        try:
+            result = agent_executor.invoke(
+                {
+                    "input": question,
+                    "chat_history": history
+                }
+            )
+            return result
+        except Exception as e:
+            # raise gr.Error("Model is Overloaded, Please retry later!")
+            raise e
+        
+    def vote(data: gr.LikeData):
+        if data.liked:
+            print("You upvoted this response: ")
+        else:
+            print("You downvoted this response: ")
 
-# CSS styling for the Gradio interface
-css = """
-#col-container {max-width: 700px; margin-left: auto; margin-right: auto;}
-"""
+    # CSS styling for the Gradio interface
+    css = """
+    #col-container {max-width: 700px; margin-left: auto; margin-right: auto;}
+    """
 
-# HTML content for the Gradio interface title
-title = """
-<div style="text-align:left;">
-    <p>Hello, I BotTina 2.0, your intelligent AI assistant. I can help you explore Wuerttembergische Versicherungs products.<br />
-</div>
-"""
+    # HTML content for the Gradio interface title
+    title = """
+    <div style="text-align:left;">
+        <p>Hello, I BotTina 2.0, your intelligent AI assistant. I can help you explore Wuerttembergische Versicherungs products.<br />
+    </div>
+    """
 
-# Building the Gradio interface
-with gr.Blocks(theme=gr.themes.Soft()) as demo:
-    with gr.Column(elem_id="col-container"):
-        gr.HTML(title)  # Add the HTML title to the interface
-        chatbot = gr.Chatbot([], elem_id="chatbot",
-                                    label="BotTina 2.0",
-                                    bubble_full_width=False,
-                                    avatar_images=(None, "https://dacodi-production.s3.amazonaws.com/store/87bc00b6727589462954f2e3ff6f531c.png"),
-                                    height=680,)  # Initialize the chatbot component
-        clear = gr.Button("Clear")  # Add a button to clear the chat
+    # Building the Gradio interface
+    with gr.Blocks(theme=gr.themes.Soft()) as demo:
+        with gr.Column(elem_id="col-container"):
+            gr.HTML(title)  # Add the HTML title to the interface
+            chatbot = gr.Chatbot([], elem_id="chatbot",
+                                        label="BotTina 2.0",
+                                        bubble_full_width=False,
+                                        avatar_images=(None, "https://dacodi-production.s3.amazonaws.com/store/87bc00b6727589462954f2e3ff6f531c.png"),
+                                        height=680,)  # Initialize the chatbot component
+            chatbot.like(vote, None, None)
+            clear = gr.Button("Clear")  # Add a button to clear the chat
 
-        # Create a row for the question input
-        with gr.Row():
-            question = gr.Textbox(label="Question", placeholder="Type your question and hit Enter ")
+            # Create a row for the question input
+            with gr.Row():
+                question = gr.Textbox(label="Question", placeholder="Type your question and hit Enter ")
 
-    # Define the action when the question is submitted
-    question.submit(add_text, [chatbot, question], [chatbot, question], queue=False).then(
-        bot, chatbot, chatbot
-    )
-    # Define the action for the clear button
-    clear.click(lambda: None, None, chatbot, queue=False)
+        # Define the action when the question is submitted
+        question.submit(add_text, [chatbot, question], [chatbot, question], queue=False).then(
+            bot, chatbot, chatbot
+        )
+        # Define the action for the clear button
+        clear.click(lambda: None, None, chatbot, queue=False)
 
-# Launch the Gradio demo interface
-demo.launch(share=False, debug=True)
+    # Launch the Gradio demo interface
+    demo.queue().launch(share=False, debug=True)

rag_app/chains/__init__.py

@@ -0,0 +1 @@
+# from rag_app.chains.s

rag_app/chains/user_response_sentiment_chain.py

@@ -0,0 +1,42 @@
+from langchain_core.prompts import PromptTemplate
+
+
+user_response_sentiment_template = """ 
+You will be given a user response to an agent.
+=================
+{user_reponse} 
+====================
+You must determine if the user has has their questions answered. 
+If the user seems satisfied respond saying "1" or "0" ONLY.
+
+Examples:
+================
+
+Example 1
+USER: Great Work! 
+YOUR RESPONSE: 1
+=================
+
+USER: I still need help! 
+YOUR RESPONSE: 0
+Example 2
+================================
+
+USER: I don't understand what you mean 
+YOUR RESPONSE: 0
+Example 3
+================================
+
+USER: That makes sense! 
+YOUR RESPONSE: 1
+Example 4
+================================
+
+  
+
+"""
+
+user_response_sentiment_prompt = PromptTemplate.from_template(user_response_sentiment_template)
+
+
+# llm_chain = prompt | llms