app.py

#to-do: generate graph
import streamlit as st
from crewai import Agent, Task, Crew
import os
from langchain_groq import ChatGroq
from fpdf import FPDF
import pandas as pd
import plotly.express as px
import time

# Title and Sidebar
st.title("🤖 Multi-Agent Business Consultant")

st.sidebar.write(
    "This Business Consultant is built using Multi-Agent system. "
    "Use this application to generate actionable business insights and data-driven analysis!"
)

# User Inputs
business = st.text_input('Enter The Required Business Search Area', value="Artificial Intelligence")
stakeholder = st.text_input('Enter The Stakeholder Team', value="Executives")

# Optional Customization
st.sidebar.subheader("Agent Customization")

# Display customization section in a collapsible expander
with st.sidebar.expander("Customize Agent Goals", expanded=False):
    enable_customization = st.checkbox("Enable Custom Goals")
    if enable_customization:
        planner_goal = st.text_area(
            "Planner Goal",
            value="Develop a comprehensive plan focusing on market trends and strategies."
        )
        writer_goal = st.text_area(
            "Writer Goal",
            value="Craft engaging and actionable content based on analysis."
        )
        analyst_goal = st.text_area(
            "Analyst Goal",
            value="Perform advanced statistical analysis and generate key insights."
        )
    else:
        planner_goal = "Develop a comprehensive plan focusing on market trends and strategies."
        writer_goal = "Craft engaging and actionable content based on analysis."
        analyst_goal = "Perform advanced statistical analysis and generate key insights."


#=================
# LLM Object
#=================
llm = ChatGroq(groq_api_key=os.getenv("GROQ_API_KEY"), model="groq/llama-3.3-70b-versatile")

#=================
# Crew Agents
#=================

planner = Agent(
    role="Business Consultant",
    goal=planner_goal,
    backstory=(
        "You're tasked with providing insights about {topic} to the stakeholder: {stakeholder}. "
        "Your work will form the foundation for the Business Writer and Data Analyst."
    ),
    allow_delegation=False,
    verbose=True,
    llm=llm
)

writer = Agent(
    role="Business Writer",
    goal=writer_goal,
    backstory=(
        "You will write a professional insights document about {topic}, "
        "based on the Business Consultant's plan and the Data Analyst's results."
    ),
    allow_delegation=False,
    verbose=True,
    llm=llm
)

analyst = Agent(
    role="Data Analyst",
    goal=analyst_goal,
    backstory=(
        "You will perform statistical analysis on {topic}, based on the Business Consultant's plan. "
        "Your analysis will support the Business Writer's final document for {stakeholder}."
    ),
    allow_delegation=False,
    verbose=True,
    llm=llm
)

#=================
# Crew Tasks
#=================

plan = Task(
    description=(
        "1. Research trends, key players, and noteworthy news for {topic}.\n"
        "2. Provide structured insights and actionable recommendations.\n"
        "3. Suggest strategies for dealing with international operators.\n"
        "4. Limit content to 500 words."
    ),
    expected_output="A comprehensive consultancy document with insights and recommendations.",
    agent=planner
)

write = Task(
    description=(
        "1. Use the Business Consultant's plan to write a professional document for {topic}.\n"
        "2. Structure the content with engaging sections and visuals.\n"
        "3. Ensure alignment with the stakeholder's goals.\n"
        "4. Limit the document to 400 words."
    ),
    expected_output="A professional document tailored for {stakeholder}.",
    agent=writer
)

analyse = Task(
    description=(
        "1. Perform statistical analysis to provide actionable insights for {topic}.\n"
        "2. Collaborate with the Business Consultant and Writer to align on key metrics.\n"
        "3. Present findings in a format suitable for inclusion in the final document."
    ),
    expected_output="A data-driven analysis tailored for {stakeholder}.",
    agent=analyst
)

#=================
# Execution
#=================

crew = Crew(
    agents=[planner, analyst, writer],
    tasks=[plan, analyse, write],
    verbose=True
)

def generate_pdf_report(result):
    """Generate a professional PDF report from the Crew output."""
    pdf = FPDF()
    pdf.add_page()
    pdf.set_font("Arial", size=12)
    pdf.set_auto_page_break(auto=True, margin=15)

    # Title
    pdf.set_font("Arial", size=16, style="B")
    pdf.cell(200, 10, txt="AI Business Consultant Report", ln=True, align="C")
    pdf.ln(10)

    # Content
    pdf.set_font("Arial", size=12)
    pdf.multi_cell(0, 10, txt=result)

    # Save PDF
    report_path = "Business_Insights_Report.pdf"
    pdf.output(report_path)
    return report_path


if st.button("Generate Insights"):
    with st.spinner('Processing...'):
        try:
            start_time = time.time()
            results = crew.kickoff(inputs={"topic": business, "stakeholder": stakeholder})
            elapsed_time = time.time() - start_time

            # Display Final Report (Writer's Output)
            st.markdown("### Final Report:")
            writer_output = getattr(results.tasks_output[2], "raw", "No details available.")
            if writer_output:
                st.write(writer_output)
            else:
                st.warning("No final report available.")

            # Option for Detailed Insights
            with st.expander("Explore Detailed Insights"):
                tab1, tab2 = st.tabs(["Planner's Insights", "Analyst's Analysis"])

                # Planner's Output
                with tab1:
                    st.markdown("### Planner's Insights")
                    planner_output = getattr(results.tasks_output[0], "raw", "No details available.")
                    st.write(planner_output)

                # Analyst's Output
                with tab2:
                    st.markdown("### Analyst's Analysis")
                    analyst_output = getattr(results.tasks_output[1], "raw", "No details available.")
                    st.write(analyst_output)

            # Display Token Usage and Execution Time
            st.success(f"Analysis completed in {elapsed_time:.2f} seconds.")
            token_usage = getattr(results, "token_usage", None)
            if token_usage:
                st.markdown("#### Token Usage")
                st.json(token_usage)

            # Generate PDF Report
            if writer_output:
                report_path = generate_pdf_report(writer_output)
                with open(report_path, "rb") as report_file:
                    st.download_button("Download Report", data=report_file, file_name="Business_Report.pdf")

        except Exception as e:
            st.error(f"An error occurred during execution: {e}")


# Add reference and credits in the sidebar
st.sidebar.markdown("---")
st.sidebar.markdown("### Reference:")
st.sidebar.markdown("[Multi-Agent Business Consultant - Hanan Tabak](https://medium.com/@hanan.tabak/user-friendly-open-source-multi-agent-ai-business-consultant-on-crewai-and-streamlit-0f972feb1b74)")