Update app.py

app.py

@@ -1,5 +1,5 @@
 """
-B2B Content Strategy Planner (Scoring Model Edition)
+B2B Content Strategy Planner (GPT-4o-mini Integrated Edition)
 Deployed for Hugging Face Spaces.
 """
 
@@ -8,6 +8,8 @@ import requests
 import re
 import time
 import random
+import os
+import json
 import pandas as pd
 from urllib.parse import urlparse, urljoin
 from collections import defaultdict, Counter
@@ -15,6 +17,7 @@ from bs4 import BeautifulSoup
 import plotly.graph_objects as go
 import plotly.express as px
 from typing import List, Dict, Set, Tuple, Optional
+from openai import OpenAI
 
 # ============================================================================
 # 1. CONFIGURATION
@@ -26,17 +29,18 @@ HEADERS = {
     'Connection': 'keep-alive'
 }
 
+# Extensive Industry Packs (Preserved)
 TOPIC_PACKS = {
-    "cloud": ["cloud", "saas", "paas", "iaas", "migration", "serverless", "aws", "azure", "gcp"],
-    "devops": ["devops", "kubernetes", "docker", "ci-cd", "terraform", "ansible"],
-    "data": ["analytics", "bi", "data-science", "etl", "big-data", "visualization", "tableau", "power-bi"],
-    "security": ["cybersecurity", "zero-trust", "threat-intelligence", "compliance", "gdpr", "soc2"],
-    "ai": ["ai", "machine-learning", "generative-ai", "llm", "nlp", "computer-vision"],
-    "martech": ["martech", "crm", "seo", "content-marketing", "lead-gen", "hubspot", "salesforce"],
-    "fintech": ["fintech", "payments", "banking", "blockchain", "crypto", "defi"],
-    "healthtech": ["healthtech", "telemedicine", "ehr", "hipaa", "medtech"],
-    "ecommerce": ["ecommerce", "shopify", "magento", "dtc", "omnichannel"],
-    "web": ["web-dev", "frontend", "backend", "react", "node", "javascript"]
+    "cloud": ["cloud", "saas", "paas", "iaas", "migration", "serverless", "aws", "azure", "gcp", "multicloud"],
+    "devops": ["devops", "kubernetes", "docker", "ci-cd", "terraform", "ansible", "jenkins", "gitops"],
+    "data": ["analytics", "bi", "data-science", "etl", "big-data", "visualization", "tableau", "power-bi", "warehouse"],
+    "security": ["cybersecurity", "zero-trust", "threat-intelligence", "soc", "ciso", "ransomware", "compliance"],
+    "ai": ["ai", "machine-learning", "generative-ai", "llm", "nlp", "computer-vision", "neural-networks"],
+    "martech": ["martech", "crm", "seo", "content-marketing", "lead-gen", "hubspot", "salesforce", "abm"],
+    "fintech": ["fintech", "payments", "banking", "blockchain", "crypto", "defi", "wealth-management"],
+    "healthtech": ["healthtech", "telemedicine", "ehr", "hipaa", "medtech", "digital-health"],
+    "ecommerce": ["ecommerce", "shopify", "magento", "dtc", "omnichannel", "retail-tech"],
+    "web": ["web-dev", "frontend", "backend", "react", "node", "javascript", "full-stack"]
 }
 
 CATEGORY_HINTS = {
@@ -47,7 +51,13 @@ CATEGORY_HINTS = {
     "Product": ["product", "platform", "tool", "software"]
 }
 
-AI_MODELS = ["GPT-4o", "Claude 3.5 Sonnet", "Gemini 1.5 Pro", "GPT-4 Turbo", "Claude 3 Opus"]
+AI_MODELS = [
+    "GPT-4o", 
+    "GPT-4 Turbo", 
+    "Claude 3.5 Sonnet", 
+    "Claude 3 Opus", 
+    "Gemini 1.5 Pro"
+]
 
 # ============================================================================
 # 2. CRAWLER & UTILS
@@ -185,69 +195,106 @@ def categorize_links(urls):
     return dict(cats)
 
 # ============================================================================
-# 3. AI IDEA FACTORY (Detailed + 7-Point Scoring)
+# 3. GPT-4o-mini INTEGRATION (Multi-Model Simulation)
 # ============================================================================
 
 SCORING_CRITERIA = """
-### ℹ️ Scoring Model Criteria
-1. **Topic Relevance (20 pts)**: Frequency and natural integration of the core topic.
-2. **Stage Alignment (20 pts)**: Fit for the selected funnel stage (e.g., Guide for TOFU).
-3. **Idea Count (15 pts)**: Optimal range is 4-6 ideas. Penalties for <4 or >6.
-4. **Structure Quality (15 pts)**: Completeness (Title, Desc, Type, Topics, Diff).
-5. **Specificity (10 pts)**: Presence of metrics, tool names, and concrete details.
-6. **Creativity (10 pts)**: Uniqueness of angles and hooks.
-7. **Actionability (10 pts)**: Practicality and clear outcomes.
+### ℹ️ Scoring Criteria
+1. **Topic Relevance (20)**: Focus on core keywords.
+2. **Stage Alignment (20)**: Fit for TOFU/MOFU/BOFU.
+3. **Idea Count (15)**: Optimal range 4-6.
+4. **Structure (15)**: Title, Desc, Format, Topics, Diff.
+5. **Specificity (10)**: Metrics and tools.
+6. **Creativity (10)**: Unique hooks.
+7. **Actionability (10)**: Practical steps.
 """
 
-def generate_detailed_idea(i, topic, stage, model_style):
-    titles = [
-        f"Top 5 Hurdles in {topic}: How to Surmount Them",
-        f"Is Your {topic} Strategy Ready? 5 Challenges to Prepare For",
-        f"The Survival Guide: Overcoming {topic} Challenges",
-        f"Why {topic} Projects Fail: 5 Challenges You Must Address",
-        f"Transforming Chaos to Order: Navigating {topic}",
-        f"The Hidden Costs of Bad {topic} Management",
-        f"{topic} in 2026: What Leaders Need to Know"
-    ]
-    formats = ["Blog Post", "Checklist Guide", "E-book", "Video Explainer", "Webinar", "Case Study", "Infographic"]
+def call_openai_api(system_prompt, user_prompt):
+    """Real call to GPT-4o-mini"""
+    api_key = os.environ.get("OPENAI_API_KEY")
+    if not api_key:
+        return None  # Fallback to mock if no key
     
-    return f"""
-    <div style="margin-bottom: 20px; padding: 15px; border-left: 3px solid #e5e7eb;">
-        <div style="font-weight: 600; font-size: 1.05em; color: #1f2937;">{i+1}. Title: "{titles[i]}"</div>
-        <div style="margin-top: 5px; font-size: 0.95em;"><b>Description:</b> Discover the common challenges organizations face with {topic} and learn {model_style} strategies to overcome them.</div>
-        <div style="margin-top: 5px; font-size: 0.95em;"><b>Content Type:</b> <span style="background:#f3f4f6; padding:2px 6px; border-radius:4px; font-size:0.9em;">{formats[i]}</span></div>
-        <div style="margin-top: 5px; font-size: 0.95em;"><b>Key Topics:</b> Adoption, integration, budget, training, {stage} metrics.</div>
-        <div style="margin-top: 5px; color: #4b5563; font-size: 0.95em;"><b>Differentiation:</b> Focus on practical, real-world scenarios rather than theory. Uses {model_style} storytelling.</div>
-    </div>
+    client = OpenAI(api_key=api_key)
+    
+    try:
+        response = client.chat.completions.create(
+            model="gpt-4o-mini",
+            messages=[
+                {"role": "system", "content": system_prompt},
+                {"role": "user", "content": user_prompt}
+            ],
+            response_format={ "type": "json_object" }
+        )
+        return json.loads(response.choices[0].message.content)
+    except Exception as e:
+        print(f"OpenAI Error: {e}")
+        return None
+
+def generate_model_response_real(model_name, topic, stage):
+    """
+    Uses GPT-4o-mini to SIMULATE the other models.
+    This provides distinct writing styles using a single API key.
     """
-
-def generate_model_response(model_name, topic, stage):
-    style = "data-driven" if "GPT" in model_name else "human-centric" if "Claude" in model_name else "action-oriented"
     
-    # Simulate variable idea counts to test the Scoring Model
-    # Optimal is 4-6. Some models will generate 3 or 7 to show penalty.
-    num_ideas = random.choice([3, 4, 5, 5, 5, 6, 7]) 
+    # Define Personas for the Simulation
+    personas = {
+        "GPT-4o": "You are a data-driven strategist. Focus on metrics, ROI, and comprehensive guides.",
+        "GPT-4 Turbo": "You are a technical expert. Focus on implementation details, architecture, and how-to guides.",
+        "Claude 3.5 Sonnet": "You are a creative storyteller. Focus on narrative, brand voice, and human-centric angles.",
+        "Claude 3 Opus": "You are a visionary thought leader. Focus on future trends, high-level strategy, and executive insights.",
+        "Gemini 1.5 Pro": "You are an actionable coach. Focus on checklists, templates, and step-by-step execution."
+    }
     
-    html = ""
-    for i in range(num_ideas):
-        html += generate_detailed_idea(i, topic, stage, style)
+    system_prompt = f"""
+    {personas.get(model_name, "You are a content strategist.")}
+    Generate exactly 5 content ideas for the topic '{topic}' at the '{stage}' funnel stage.
+    Output MUST be a JSON object with a key 'ideas' containing a list of objects.
+    Each object must have: 'title', 'description', 'content_type', 'key_topics', 'differentiation'.
+    Ensure the tone matches your persona.
+    """
+    
+    user_prompt = f"Generate 5 ideas for {topic} ({stage})."
     
-    return html, num_ideas
+    # Try API Call
+    data = call_openai_api(system_prompt, user_prompt)
+    
+    # Fallback if API fails or no key
+    if not data:
+        return generate_model_response_mock(model_name, topic, stage), 5
+        
+    # Format HTML from JSON
+    html = ""
+    for i, idea in enumerate(data.get('ideas', [])):
+        html += f"""
+        <div style="margin-bottom: 20px; padding: 15px; border-left: 3px solid #e5e7eb;">
+            <div style="font-weight: 600; font-size: 1.05em; color: #1f2937;">{i+1}. Title: "{idea['title']}"</div>
+            <div style="margin-top: 5px; font-size: 0.95em;"><b>Description:</b> {idea['description']}</div>
+            <div style="margin-top: 5px; font-size: 0.95em;"><b>Content Type:</b> <span style="background:#f3f4f6; padding:2px 6px; border-radius:4px; font-size:0.9em;">{idea['content_type']}</span></div>
+            <div style="margin-top: 5px; font-size: 0.95em;"><b>Key Topics:</b> {idea['key_topics']}</div>
+            <div style="margin-top: 5px; color: #4b5563; font-size: 0.95em;"><b>Differentiation:</b> {idea['differentiation']}</div>
+        </div>
+        """
+    return html, len(data.get('ideas', []))
+
+def generate_model_response_mock(model_name, topic, stage):
+    # Mock fallback if API key is missing
+    html = ""
+    for i in range(5):
+        html += f"""
+        <div style="margin-bottom: 20px; padding: 15px; border-left: 3px solid #e5e7eb;">
+            <div style="font-weight: 600; font-size: 1.05em; color: #1f2937;">{i+1}. Title: "Mock Idea for {topic}"</div>
+            <div style="margin-top: 5px; font-size: 0.95em;"><b>Description:</b> This is a simulated response because OpenAI API Key is missing.</div>
+        </div>
+        """
+    return html
 
 def score_model_output(num_ideas):
-    # 1. Idea Count Score (15 pts) - Penalize if not 4-6
-    if 4 <= num_ideas <= 6:
-        score_count = 15
-    elif num_ideas == 3 or num_ideas == 7:
-        score_count = 10
-    else:
-        score_count = 5
-
-    # 2. Other Criteria (Randomized within logic)
+    # Auto-scoring logic based on criteria
     metrics = {
         "Topic Relevance": random.randint(16, 20),
         "Stage Alignment": random.randint(16, 20),
-        "Idea Count": score_count,
+        "Idea Count": 15 if 4 <= num_ideas <= 6 else 10,
         "Structure Quality": random.randint(12, 15),
         "Specificity": random.randint(7, 10),
         "Creativity": random.randint(7, 10),
@@ -261,22 +308,21 @@ def ui_generate_multimodel(topic, stage):
     
     results = []
     
-    # Run all 5 models
+    # Run models
     for model in AI_MODELS:
-        resp_html, count = generate_model_response(model, topic, stage)
+        resp_html, count = generate_model_response_real(model, topic, stage)
         score, metrics = score_model_output(count)
         results.append({
             "model": model,
             "score": score,
             "metrics": metrics,
-            "resp": resp_html,
-            "count": count
+            "resp": resp_html
         })
     
     # Rank by Score
     results.sort(key=lambda x: x['score'], reverse=True)
     
-    # 1. Full Comparison Table
+    # 1. Comparison Table
     table_html = """
     <table style="width:100%; border-collapse: collapse; margin-bottom: 20px; font-family:sans-serif; font-size:0.85em;">
         <tr style="background: #f8fafc; border-bottom: 2px solid #e2e8f0; text-align:center;">
@@ -334,7 +380,7 @@ def ui_generate_multimodel(topic, stage):
     return table_html, cards_html
 
 # ============================================================================
-# 4. COMPETITOR INTELLIGENCE (Fixed Links)
+# 4. COMPETITOR ANALYSIS (Fixed Links & Multi-Domain)
 # ============================================================================
 
 def get_tokens(urls: List[str]) -> Set[str]:
@@ -356,7 +402,6 @@ def generate_pill_html(items: List[str], style: str, empty_msg: str):
     
     html = '<div style="display:flex; flex-wrap:wrap; gap:10px; margin-top:15px;">'
     for item in sorted(list(items))[:40]:
-        # Google Search Link
         link = f"https://www.google.com/search?q={item.replace(' ', '+')}"
         html += f"""
         <a href="{link}" target="_blank" style="text-decoration:none;">
@@ -401,49 +446,16 @@ def generate_radar(user_urls, competitors):
     tot = len(stages) or 1
     user_vals = [(stages.count(s)/tot)*100 for s in cats]
     
-    fig.add_trace(go.Scatterpolar(
-        r=user_vals, theta=cats, fill='toself', 
-        name='Your Company', line_color='#3b82f6', opacity=0.7
-    ))
+    fig.add_trace(go.Scatterpolar(r=user_vals, theta=cats, fill='toself', name='Your Company', line_color='#3b82f6', opacity=0.7))
     
     colors = ['#ef4444', '#10b981', '#f59e0b', '#8b5cf6', '#ec4899']
     for i, c in enumerate(competitors):
         s = c['stats']
-        fig.add_trace(go.Scatterpolar(
-            r=[s['TOFU'], s['MOFU'], s['BOFU']], 
-            theta=cats, fill='none',
-            name=c['domain'], line_color=colors[i % len(colors)],
-            line_width=2
-        ))
+        fig.add_trace(go.Scatterpolar(r=[s['TOFU'], s['MOFU'], s['BOFU']], theta=cats, fill='none', name=c['domain'], line_color=colors[i % len(colors)], line_width=2))
         
-    fig.update_layout(
-        polar=dict(radialaxis=dict(visible=True, range=[0, 100])), 
-        title="Funnel Distribution", 
-        height=400,
-        margin=dict(l=40, r=40, t=40, b=40),
-        legend=dict(orientation="h", y=-0.1)
-    )
+    fig.update_layout(polar=dict(radialaxis=dict(visible=True, range=[0, 100])), title="Funnel Distribution", height=400)
     return fig
 
-def generate_radar_interpretation(user_urls, competitors):
-    stages = [classify_stage(u) for u in user_urls]
-    tot = len(stages) or 1
-    u_tofu = (stages.count('TOFU')/tot)*100
-    
-    comp_tofu_avg = 0
-    if competitors:
-        comp_tofu_sum = sum([c['stats']['TOFU'] for c in competitors])
-        comp_tofu_avg = comp_tofu_sum / len(competitors)
-    
-    diff = u_tofu - comp_tofu_avg
-    
-    if diff > 10:
-        return "✅ **Awareness Leader:** Your content is heavily focused on TOFU compared to competitors. Consider adding more MOFU/BOFU to convert this traffic."
-    elif diff < -10:
-        return "⚠️ **Awareness Gap:** Competitors have significantly more TOFU content. You may struggle to attract new leads compared to them."
-    else:
-        return "⚖️ **Balanced Strategy:** Your funnel allocation aligns closely with the market average."
-
 # ============================================================================
 # 5. UI EVENTS
 # ============================================================================
@@ -455,10 +467,7 @@ def ui_fetch(url, force):
     cats = categorize_links(urls)
     state['urls'] = urls
     cat_opts = ["All"] + sorted(list(cats.keys()))
-    return (urls, cats, log, 
-            gr.update(choices=cat_opts, value="All"), 
-            gr.update(choices=[], value=None),
-            gr.update(choices=[], value=None))
+    return (urls, cats, log, gr.update(choices=cat_opts, value="All"), gr.update(choices=[], value=None), gr.update(choices=[], value=None))
 
 def ui_update_topics(category, cat_data):
     if not cat_data: return gr.update(choices=[])
@@ -488,11 +497,10 @@ def ui_add_comp(url, current_comps):
     return current_comps, txt, ""
 
 def ui_run_full_comp_analysis(user_urls, competitors):
-    if not user_urls: return None, "", "", ""
+    if not user_urls: return None, "", ""
     radar = generate_radar(user_urls, competitors)
-    interp = generate_radar_interpretation(user_urls, competitors)
     g_html, s_html = analyze_gaps_strengths(user_urls, competitors, None)
-    return radar, interp, g_html, s_html
+    return radar, g_html, s_html
 
 def ui_analyze_gaps_wrapper(user_urls, competitors, topic):
     return analyze_gaps_strengths(user_urls, competitors, topic)[0]
@@ -512,7 +520,7 @@ with gr.Blocks(theme=gr.themes.Soft(), title="Content Strategy Master") as demo:
     gr.Markdown("# 🚀 B2B Content Strategy Master")
     
     with gr.Tabs():
-        # TAB 1
+        # TAB 1: PLANNER
         with gr.Tab("📊 Analytics"):
             with gr.Row():
                 url_in = gr.Textbox(label="Your Website URL")
@@ -524,31 +532,27 @@ with gr.Blocks(theme=gr.themes.Soft(), title="Content Strategy Master") as demo:
                 analyze_btn = gr.Button("Analyze")
             plan_urls = gr.Textbox(label="Matched Pages", lines=5)
 
-        # TAB 2
+        # TAB 2: AI IDEAS
         with gr.Tab("✨ AI Ideas"):
             with gr.Row():
                 idea_topic = gr.Dropdown(label="Topic", choices=[], allow_custom_value=True)
                 idea_stage = gr.Dropdown(["TOFU", "MOFU", "BOFU"], label="Stage", value="TOFU")
                 gen_btn = gr.Button("Generate Multi-Model Ideas", variant="primary")
-            
-            with gr.Accordion("ℹ️ How Scoring Works (7-Point Model)", open=False):
+            with gr.Accordion("ℹ️ Scoring Criteria", open=False):
                 gr.Markdown(SCORING_CRITERIA)
-            
             comp_table = gr.HTML(label="Model Ranking")
             idea_results = gr.HTML(label="Detailed Ideas")
 
-        # TAB 3
+        # TAB 3: COMPETITORS
         with gr.Tab("⚔️ Competitors"):
             with gr.Row():
                 comp_url = gr.Textbox(label="Competitor URL")
                 add_comp = gr.Button("Add Competitor")
             comp_list = gr.Markdown("No competitors added.")
-            
             run_full = gr.Button("📊 Run Full Market Analysis", variant="primary")
             
             with gr.Row():
                 radar_plot = gr.Plot(label="Market Radar")
-                radar_txt = gr.Markdown(label="Insight")
             
             gr.Markdown("---")
             with gr.Group():
@@ -575,7 +579,7 @@ with gr.Blocks(theme=gr.themes.Soft(), title="Content Strategy Master") as demo:
     gen_btn.click(ui_generate_multimodel, [idea_topic, idea_stage], [comp_table, idea_results])
     
     add_comp.click(ui_add_comp, [comp_url, state_comps], [state_comps, comp_list, comp_url])
-    run_full.click(ui_run_full_comp_analysis, [state_urls, state_comps], [radar_plot, radar_txt, gap_html, str_html])
+    run_full.click(ui_run_full_comp_analysis, [state_urls, state_comps], [radar_plot, gap_html, str_html])
     
     gap_btn.click(ui_analyze_gaps_wrapper, [state_urls, state_comps, gap_search], [gap_html])
     str_btn.click(ui_analyze_strengths_wrapper, [state_urls, state_comps, str_search], [str_html])