import os
from phi.agent import Agent
from phi.model.openai import OpenAIChat
import openai
import markdown2
import pdfkit
from datetime import datetime
from bni_agent import get_bni_benefits
from rag_agent import recommend_bni_connections
# Load environment variables (API keys, etc.)
from dotenv import load_dotenv
load_dotenv()
#####################################################################################
# PHASE 3 #
#####################################################################################
##############################
# 1️⃣ Reasoning Agent #
##############################
reasoning_agent = Agent(
name="Reasoning Agent",
model=OpenAIChat(id="gpt-4o"),
description="Processes all collected data and generates structured AI adoption strategies.",
show_tool_calls=True,
markdown=True,
)
def generate_ai_strategy(company_data: str, industry_trends: str) -> str:
"""Generates a structured AI adoption strategy based on company details and industry insights."""
query = f"""
You are an AI business strategist analyzing a company's AI adoption potential. Given:
- **Company Overview:** {company_data}
- **Industry Trends:** {industry_trends}
Generate a structured AI strategy including:
1. Key AI Opportunities
"""
response = reasoning_agent.run(query)
return response.content # Returns the generated AI strategy
##############################
# 2️⃣ AI Integration Advisor #
##############################
# ai_integration_agent = Agent(
# name="AI Integration Advisor",
# model=OpenAIChat(id="gpt-4o"),
# description="Suggests AI implementation strategies based on industry insights and company operations.",
# show_tool_calls=True,
# markdown=True,
# )
# def suggest_ai_integration(company_data: str, ai_strategy: str) -> str:
# """Suggests a structured AI implementation plan."""
# query = f"""
# Based on the AI adoption strategy:
# - **Company Context:** {company_data}
# - **AI Strategy Summary:** {ai_strategy}
# Provide a structured AI implementation plan including:
# 1. Step-by-step AI adoption process
# 2. Required AI Technologies & Infrastructure
# 3. Workforce training & AI skill development
# 4. Risk considerations (data security, compliance, ethical AI)
# 5. Key performance indicators (KPIs) for AI success.
# """
# response = ai_integration_agent.run(query)
# return response.content # Returns AI integration plan
##############################
# 3️⃣ Revenue Growth Agent #
##############################
revenue_growth_agent = Agent(
name="Revenue Growth Agent",
model=OpenAIChat(id="gpt-4o"),
description="Identifies AI-driven opportunities to enhance revenue and efficiency.",
show_tool_calls=True,
markdown=True,
)
def identify_revenue_opportunities(company_data: str, ai_strategy: str) -> str:
"""Identifies AI-driven revenue generation opportunities for a business."""
query = f"""
You are an AI business analyst identifying practical AI-driven revenue growth opportunities for:
Company Overview: {company_data}
AI Strategy: {ai_strategy}
Provide a concise, single-paragraph summary covering:
AI-based Monetization Strategies – How can the company create new revenue streams using AI?
Cost Reduction & Operational Efficiency – How can AI streamline operations and reduce costs?
Market Expansion – How can AI open new markets or customer segments?
Competitive Positioning – How can AI help differentiate the company from competitors?
Keep it short, crisp, and actionable, avoiding complex or unnecessary details.
"""
response = revenue_growth_agent.run(query)
return response.content # Returns revenue opportunities
##############################
# 4️⃣ BNI Agent: Membership Benefits #
##############################
def get_bni_membership_benefits(company_data: str) -> str:
"""Fetches BNI membership benefits tailored to the user's company."""
return get_bni_benefits(company_data)
##############################
# 5️⃣ BNI RAG Agent: Pearl Chapter Connections #
##############################
def get_bni_recommendations(company_data: str) -> str:
"""Recommends relevant BNI Pearl Chapter members based on the user's company data."""
return recommend_bni_connections(company_data)
##############################
# 6️⃣ Report Generation Agent #
##############################
# Ensure your OpenAI API key is set in your environment variables
openai.api_key = os.getenv("OPENAI_API_KEY")
def generate_report(company_name: str, company_data: str,
ai_strategy: str,
revenue_opportunities: str,
bni_benefits: str,
bni_connections: str) -> str:
"""
Generates a structured AI strategy report in PDF format using `pdfkit`.
"""
# 📝 **Step 1: Prepare the Report Content**
report_content = f"""
# AI Strategy Report for {company_name}
## 1️⃣ Company Overview
{company_data}
## 2️⃣ How BNI Can Help Your Business
{bni_benefits}
## 3️⃣ Recommended BNI Pearl Chapter Connections
{bni_connections}
## 4️⃣ AI Adoption Strategy
{ai_strategy}
## 5️⃣ Revenue Growth Opportunities
{revenue_opportunities}
"""
# 🎯 **Step 2: Use OpenAI GPT-4o-mini to Format Markdown**
response = openai.chat.completions.create(
model="gpt-4o-mini",
messages=[
{"role": "system", "content": "You are an expert in formatting business reports in Markdown."},
{"role": "user", "content": f"Format the following report in well-structured Markdown. Use bold headings, bullet points, and clear sections. Ensure the text is concise, professional, and structured for readability. Do not include any emojis or unnecessary text:\n\n{report_content}"}
],
max_tokens=5000,
temperature=0.5,
)
enhanced_markdown = response.choices[0].message.content # Extract the formatted response
# 🖼 **Step 3: Convert Enhanced Markdown to HTML**
html_report = markdown2.markdown(enhanced_markdown)
# 📄 **Step 4: Define the HTML Template with Styling**
html_template = f"""
Giant Analytics
{html_report}
"""
# 📌 **Step 5: Generate the PDF**
current_date = datetime.now().strftime("%Y-%m-%d")
pdf_filename = f"{company_name}_AI_Report_{current_date}.pdf"
# Path to wkhtmltopdf executable (Make sure it's installed)
path_to_wkhtmltopdf = '/usr/bin/wkhtmltopdf' # Update this path if different
config = pdfkit.configuration(wkhtmltopdf=path_to_wkhtmltopdf)
# Convert HTML to PDF
pdfkit.from_string(html_template, pdf_filename, configuration=config)
return pdf_filename # Returns the generated PDF filename