Included Business interaction too

main.py

@@ -3,87 +3,43 @@ from pydantic import BaseModel
 from my_agent.agent import build_graph
 import pandas as pd
 from typing import Optional
+from my_agent.utils.initial_interaction import BusinessInteractionChatbot
 
 app = FastAPI()
+interaction_chatbot = BusinessInteractionChatbot()
 graph = build_graph()
 
+
 class RequestInput(BaseModel):
     query: list 
     preferred_topics: Optional[list] = [] 
 
 
-@app.post("/run")
-def run_graph(input_data: RequestInput):
-    result = graph.invoke({'topic' : input_data.query})
-    return {'final_story': result['final_story']}
-    
-
-# *********************INFERENCING PART****************************
-# import asyncio
-# from fastapi import FastAPI, Request, Form
-# from fastapi.responses import HTMLResponse
-# from fastapi.staticfiles import StaticFiles
-# from fastapi.templating import Jinja2Templates
-# from my_agent.agent import build_graph
-# from starlette.concurrency import run_in_threadpool
-
-# app = FastAPI()
-# graph = build_graph()
-
-# app.mount("/static", StaticFiles(directory="static"), name="static")
-# templates = Jinja2Templates(directory="templates")
 
-# # Store session state in memory (simple approach)
-# session_data = {
-#     "topic": [],
-#     "preferred_topics": [],
-# }
 
-# @app.get("/", response_class=HTMLResponse)
-# def get_home(request: Request):
-#     return templates.TemplateResponse("index.html", {"request": request})
+class UserMessage(BaseModel):
+    message: str
 
-# @app.post("/submit", response_class=HTMLResponse)
-# async def handle_query(request: Request, user_query: str = Form(...)):
-#     session_data["topic"] = [user_query]
-#     result = graph.invoke({"topic": session_data["topic"]})
-#     brainstorming = result.get("brainstroming_topics", [{}])[-1].values()
-#     return templates.TemplateResponse("index.html", {
-#         "request": request,
-#         "topics": brainstorming,
-#         "story": None,
-#         "show_buttons": True
-#     })
+details_for_brainstrom = {}
+@app.post("/business-interaction")
+def business_chat(msg: UserMessage):
+    global details_for_brainstrom
+    response = interaction_chatbot.chat(msg.message)
+    if interaction_chatbot.is_complete(response):
+        details = interaction_chatbot.extract_details()
+        details_for_brainstrom = details
+        return {"response": response, "business_details": details, "complete": True}
+    return {"response": response, "complete": False}
 
-# @app.post("/select", response_class=HTMLResponse)
-# async def select_topics(request: Request, selected_topics: list[str] = Form(...), action: str = Form(...)):
-#     session_data["preferred_topics"].append(selected_topics)
 
-#     if action == "go_deeper":
-#         result = await run_in_threadpool(
-#             graph.invoke,
-#             {"topic": session_data["topic"], "preferred_topics": session_data["preferred_topics"]}
-#         )
-#         new_brainstorm = result.get("brainstroming_topics", [{}])[-1].values()
-#         print()
-#         print('*****RESPONSE IS GENERATED********')
-#         print()
-#         return templates.TemplateResponse("index.html", {
-#             "request": request,
-#             "topics": new_brainstorm,
-#             "story": None,
-#             "show_buttons": True
-#         })
+@app.post("/brainstrom")
+def run_graph(input_data: RequestInput):
+    # business_details = details_for_brainstrom
+    result = graph.invoke({'topic' : input_data.query , 'business_details': details_for_brainstrom})
+    # RequestInput.preferred_topics=result['preferred_topics']
+    return {'final_story': result['final_story'],
+            'business_details':result['business_details'],
+            }
+    
 
-#     elif action == "generate_final_story":
-#         result = await run_in_threadpool(
-#             graph.invoke,
-#             {"topic": session_data["topic"], "preferred_topics": session_data["preferred_topics"]}
-#         )
-#         return templates.TemplateResponse("index.html", {
-#             "request": request,
-#             "topics": [],
-#             "story": result.get("final_story"),
-#             "show_buttons": False
-#         })
 

my_agent/utils/initial_interaction.py

@@ -0,0 +1,97 @@
+import os
+from langchain_groq import ChatGroq
+from langgraph.graph import StateGraph, MessagesState, START, END
+from langgraph.checkpoint.memory import MemorySaver
+from langchain_core.messages import SystemMessage
+from pydantic import BaseModel, ConfigDict, Field
+from typing import Optional, List
+from .models_loader import llm
+
+# Pydantic model for extracted business info
+class DetailsFormatter(BaseModel):
+    business_type: str = Field(description="The type of the business")
+    platform: str = Field(description="The platform used for the business")
+    target_audience: str = Field(description="The target audience of the business")
+    business_goals: str = Field(description="The business goals of the business")
+    offerings: str = Field(description="The offerings of the business")
+    Challenges_faced: str = Field(description="The challenges faced by the business")
+
+# State model
+class State(BaseModel):
+    interactions: Optional[list] = []
+    model_config = ConfigDict(arbitrary_types_allowed=True)
+
+# Global business state (shared)
+business_state = State()
+
+class BusinessInteractionChatbot:
+    def __init__(self):
+        self.memory = MemorySaver()
+        # self.llm = ChatGroq(model_name="Gemma2-9b-It")
+        self.llm = llm
+        self.workflow = self._initialize_workflow()
+        self.interact_agent = self.workflow.compile(checkpointer=self.memory)
+        self.messages = []
+
+    def _initialize_workflow(self):
+        workflow = StateGraph(MessagesState)
+        workflow.add_node("chatbot", self._call_model)
+        workflow.add_edge(START, "chatbot")
+        workflow.add_edge("chatbot", END)
+        return workflow
+
+    def _call_model(self, state):
+        template = self._get_prompt_template()
+        messages = [SystemMessage(content=template)] + state["messages"]
+        response = self.llm.invoke(messages)
+        return {"messages": [response]}
+
+    def _get_prompt_template(self):
+        return (
+            '''You are a business assistant who collects only valid and relevant data.
+Your job is to gather details from business owners in a friendly and conversational manner to understand their business better. Ask in very easy and short way.
+
+We need these details:
+1. Business Type (e.g., e-commerce, SaaS, consulting),
+2. Platform(s) used (e.g., website, app, Instagram),
+3. Target Audience (who are their customers or clients),
+4. Business Goals (short-term or long-term objectives),
+5. Offerings (products or services they provide),
+6. Challenges faced (any current business problems or limitations).
+
+Keep interacting until all valid details are collected.
+
+VERY IMPORTANT: Once all valid details are received, say: '**Thanks for providing all your required business details.**'
+'''
+        )
+
+    def chat(self, user_input: str):
+        self.messages.append({"role": "user", "content": user_input})
+        config = {"configurable": {"thread_id": "1"}}
+        response = self.interact_agent.invoke({"messages": [user_input]}, config)['messages'][-1].content
+        self.messages.append({"role": "assistant", "content": response})
+        business_state.interactions.append({'user': user_input, 'agent_response': response})
+        return response
+
+    def is_complete(self, latest_response: str) -> bool:
+        return "Thanks for providing all your required business details" in latest_response
+
+    def extract_details(self):
+        template = f'''Extract the following details of the business from the conversation.
+        1. Business Type (e.g., e-commerce, SaaS, consulting),
+        2. Platform(s) used (e.g., website, app, Instagram),
+        3. Target Audience (who are their customers or clients),
+        4. Business Goals (short-term or long-term objectives),
+        5. Offerings (products or services they provide),
+        6. Challenges faced (any current business problems or limitations).
+        The conversation is:\n{business_state.interactions}'''
+
+        messages = [SystemMessage(content=template)]
+        response = self.llm.bind_tools([DetailsFormatter]).invoke(messages)
+
+        if hasattr(response, 'tool_calls') and response.tool_calls:
+            return response.tool_calls[0]['args']
+        elif hasattr(response, 'content'):
+            return response.content
+        else:
+            return "No response"

my_agent/utils/nodes.py

@@ -10,23 +10,35 @@ from .data_loader import load_influencer_data
 def retrieve(state: State) -> State:
   print('Moving to retrieval process')
   retrievals=[]
-  for topic in state.topic:  # Loop through each topic
-      embedded_query = ST.encode(topic)  # Embed each topic
-      data = load_influencer_data()
-      scores, retrieved_examples = data.get_nearest_examples("embeddings", embedded_query, k=1)
-
-      # Construct a list of dictionaries for this topic
-      result = [{user: story} for user, story in zip(retrieved_examples['username'], retrieved_examples['agentic_story'])]
-      retrievals.append(result)
-      print('Retrieval process completed......')
-  state.retrievals.append(retrievals)
+  if len(state.preferred_topics)==0:
+      for topic in state.topic:  # Loop through each topic
+          embedded_query = ST.encode(topic)  # Embed each topic
+          data = load_influencer_data()
+          scores, retrieved_examples = data.get_nearest_examples("embeddings", embedded_query, k=1)
+
+          # Construct a list of dictionaries for this topic
+          result = [{user: story} for user, story in zip(retrieved_examples['username'], retrieved_examples['agentic_story'])]
+          retrievals.append(result)
+          print('Retrieval process completed......')
+      state.retrievals.append(retrievals)
+
+  if len (state.preferred_topics)>0:
+      for topic in state.preferred_topics[-1]:  # Loop through each topic
+          embedded_query = ST.encode(topic)  # Embed each topic
+          data = load_influencer_data()
+          scores, retrieved_examples = data.get_nearest_examples("embeddings", embedded_query, k=1)
+
+          # Construct a list of dictionaries for this topic
+          result = [{user: story} for user, story in zip(retrieved_examples['username'], retrieved_examples['agentic_story'])]
+          retrievals.append(result)
+          print('Retrieval process completed......')
+      state.retrievals.append(retrievals)
 
   print('The retrieval is:\n',state.retrievals )
   # return State(messages="Retrieved",topic=state.topic,retrievals=state.retrievals)
   return state
 
 def generate_story(state:State)-> State:
-  topic=state.topic
   print('The state retrieval is:',state.retrievals)
   retrieval_list= state.retrievals[-1]
   agentic_stories = []
@@ -116,6 +128,7 @@ def select_preferred_topics(state: State)-> State:
     try:
         preferred_indices = [int(i.strip()) for i in raw_input_str.split(",")]
         preferred_topics = [topic_values[i - 1] for i in preferred_indices if 0 < i <= len(topic_values)]
+        # preferred_topics = user_input
         state.preferred_topics.append(preferred_topics)
     except Exception:
         state.carry_on=False
@@ -133,52 +146,31 @@ def select_preferred_topics(state: State)-> State:
     return state
 
 
-# def select_preferred_topics(state: State) -> State:
-#     print("---API_feedback_mode---")
-
-#     if not state.brainstroming_topics:
-#         print("No brainstormed topics found.")
-#         state.carry_on = False
-#         return state
-
-#     # Get the latest set of brainstormed topics
-#     topic_values = list(state.brainstroming_topics[-1].values())
-#     print(f"Available topics: {topic_values}")
-
-#     # Ensure preferred_topics is well-formed
-#     if state.preferred_topics and isinstance(state.preferred_topics[-1], list):
-#         latest_selection = state.preferred_topics[-1]
-#         if latest_selection:
-#             print("User selected topics:")
-#             print(latest_selection)
-#             state.carry_on = True
-#             return state
-
-#     print("No preferred topics selected via API. Ending feedback loop.")
-#     state.carry_on = False
-#     return state
-
 
 
 
 def generate_final_story(state:State)-> State:
-    template = f'''I want to create a detailed storyline for a video in the given topic. You have to provide me that storyline what to include in the video.
-  Now, i am giving you the topic of the video. But the need is to generate the story focusing on the format that i'll provide to you.
-  You can use this format for the reference purpose, not for the exact similar generation. The format is:\n{state.retrievals[-1]}.
-  \n\n Now let's start creating the storyline for my topic. The topic of the video is: \n\n{state.topic}\n\n
-
-  **Final Reminder** You have to strongly focus on these topics while creating the storyline: {[item for sublist in state.preferred_topics for item in sublist]}'''
-    messages = [SystemMessage(content=template)]
-    response = llm.bind_tools([StoryFormatter]).invoke(messages)
-    print('The final response is:',response)
-    if hasattr(response, 'tool_calls') and response.tool_calls:
-        response = response.tool_calls[0]['args']
-    elif hasattr(response, 'content'):
-        response = response.content
-    else:
-      response = "No response"
-    state.final_story=response
-    state.stories.append(response)
+    if len(state.preferred_topics)>0:
+        template = f'''I want to create a detailed storyline for a video in the given topic. You have to provide me that storyline what to include in the video.
+      Now, i am giving you the topic of the video. But the need is to generate the story focusing on the format that i'll provide to you.
+      You can use this format for the reference purpose, not for the exact similar generation. The format is:\n{state.retrievals[-1]}.
+      \n\n Now let's start creating the storyline for my topic. The topic of the video is: \n\n{state.topic}\n\n
+
+      **Final Reminder** You have to strongly focus on these topics while creating the storyline: {[item for sublist in state.preferred_topics for item in sublist]}'''
+        messages = [SystemMessage(content=template)]
+        response = llm.bind_tools([StoryFormatter]).invoke(messages)
+        print('The final response is:',response)
+        if hasattr(response, 'tool_calls') and response.tool_calls:
+            response = response.tool_calls[0]['args']
+        elif hasattr(response, 'content'):
+            response = response.content
+        else:
+          response = "No response"
+        state.final_story=response
+        state.stories.append(response)
+        return state
+    
+    state.final_story=state.stories[-1]
     return state
 
 

my_agent/utils/state.py

@@ -11,4 +11,6 @@ class State(BaseModel):
     stories : Optional[list]=[]
     final_story: Optional[str]=None
     retrievals : Optional[list]=[]
+    business_details : Optional[dict]={}
+    latest_preferred_topics: Optional[list] = []
     model_config = ConfigDict(arbitrary_types_allowed=True)