Update app.py

app.py

@@ -3,34 +3,30 @@ import openai
 import gradio as gr
 from bs4 import BeautifulSoup
 import requests
-import nest_asyncio
-import asyncio
 
-from playwright.sync_api import sync_playwright
-
-nest_asyncio.apply()
 openai.api_key = os.getenv("OPENAI_API_KEY")
 
-# Synchronous version for HF Spaces compatibility
 def extract_text_from_url(url):
-    with sync_playwright() as pw:
-        browser = pw.chromium.launch(headless=True)
-        page = browser.new_page()
-        page.goto(url, timeout=60000)
-        page.wait_for_load_state('networkidle')
-        content = page.content()
-        browser.close()
-
-    soup = BeautifulSoup(content, "html.parser")
-    text = ' '.join(p.get_text(strip=True) for p in soup.find_all(['p', 'span', 'h1', 'h2', 'li']))
-    return text[:4000]  # limit to 4000 characters
+    """Extracts text from HTML for static pages. Warns if content is very thin."""
+    try:
+        resp = requests.get(url, timeout=30, headers={
+            "User-Agent": "Mozilla/5.0 (compatible; Bot/1.0)"
+        })
+        soup = BeautifulSoup(resp.content, "html.parser")
+        # Try to get rich descriptive content
+        candidates = soup.find_all(['h1','h2','h3','h4','p','span','li'])
+        text = ' '.join([c.get_text(strip=True) for c in candidates])
+        text = text[:4000]
+        if len(text) < 100:
+            raise ValueError("Could not extract enough content (site may require JavaScript). Please enter keywords manually.")
+        return text
+    except Exception as e:
+        raise ValueError(f"URL extraction error: {e}")
 
 def extract_keywords(text):
     prompt = f"""
     Extract up to 10 concise, relevant SEO keywords suitable for an automotive advertisement from the following content:
-    
     {text}
-
     Keywords:
     """
     response = openai.ChatCompletion.create(
@@ -39,8 +35,13 @@ def extract_keywords(text):
         temperature=0.6,
         max_tokens=100
     )
-    keywords = response.choices[0].message.content.strip().split(',')
-    return [kw.strip() for kw in keywords]
+    # Handles both comma or newline separation
+    output = response.choices[0].message.content.strip()
+    if ',' in output:
+        keywords = output.split(',')
+    else:
+        keywords = output.split('\n')
+    return [kw.strip() for kw in keywords if kw.strip()]
 
 def generate_ad_copy(platform, keywords):
     prompt = f"""
@@ -72,37 +73,35 @@ def generate_ad_image(keywords):
 
 def main_workflow(input_mode, url_or_keywords):
     error = None
-    # Step 1: Get keywords
+    keywords = []
+    ad_copies = {}
+    image_path = None
+
     if input_mode == "URL":
         try:
             text = extract_text_from_url(url_or_keywords)
             keywords = extract_keywords(text)
         except Exception as e:
-            return None, None, None, f"URL extraction error: {e}"
+            return None, None, None, f"{e}"
     else:
         keywords = [kw.strip() for kw in url_or_keywords.split(",") if kw.strip()]
         if not keywords:
             return None, None, None, "Please provide at least one keyword."
-
-    # Step 2: Generate ad copies
+    # Generate ad copies
     platforms = ["Facebook", "Instagram", "X (Twitter)", "Google Search"]
-    ad_copies = {}
     for platform in platforms:
         ad_copies[platform] = generate_ad_copy(platform, keywords)
-    
-    # Step 3: Generate ad image
+    # Generate image
     try:
         image_path = generate_ad_image(keywords)
     except Exception as e:
-        image_path = None
         error = f"Image generation error: {e}"
 
-    # Step 4: Save ad copies to txt
+    # Save ads to txt
     output_txt = "generated_ads.txt"
     with open(output_txt, "w", encoding="utf-8") as f:
         for platform, content in ad_copies.items():
             f.write(f"--- {platform} Ad Copy ---\n{content}\n\n")
-    
     return keywords, ad_copies, image_path, error
 
 def run_space(input_mode, url, keywords):