Update app.py

app.py

@@ -5,11 +5,11 @@ from langchain.document_loaders import WebBaseLoader
 from langchain.text_splitter import RecursiveCharacterTextSplitter
 from langchain.llms import HuggingFacePipeline
 from transformers import pipeline
-from gtts import gTTS
 import tempfile
 import os
 from bs4 import BeautifulSoup
 import requests
+import pyttsx3
 
 # CPU-friendly summarization LLM
 summary_pipe = pipeline("text2text-generation", model="google/flan-t5-base", device=-1)
@@ -17,7 +17,7 @@ llm = HuggingFacePipeline(pipeline=summary_pipe)
 
 # Summarization prompt
 summary_prompt = PromptTemplate.from_template("""
-Summarize the following article content in a clear, concise way:
+Summarize the following article content in a clear, concise, and emotionally engaging manner as if you're speaking to a curious listener:
 
 {text}
 
@@ -31,17 +31,36 @@ def extract_main_content(url):
         response = requests.get(url, timeout=10)
         soup = BeautifulSoup(response.content, "html.parser")
 
-        # Remove navigation, header, footer, sidebars, and scripts
         for tag in soup(["nav", "header", "footer", "aside", "script", "style", "noscript"]):
             tag.decompose()
 
-        # Extract main content using tags with significant paragraph text
         paragraphs = soup.find_all("p")
         content = "\n".join([p.get_text() for p in paragraphs if len(p.get_text()) > 60])
         return content.strip()
     except Exception as e:
         return f"Error extracting article content: {str(e)}"
 
+def generate_human_like_audio(text):
+    try:
+        engine = pyttsx3.init()
+        engine.setProperty('rate', 150)  # slower pace
+        engine.setProperty('volume', 1.0)
+        voices = engine.getProperty('voices')
+
+        # Choose a more natural voice if available (optional: pick female)
+        for voice in voices:
+            if 'female' in voice.name.lower():
+                engine.setProperty('voice', voice.id)
+                break
+
+        temp_path = tempfile.NamedTemporaryFile(delete=False, suffix=".mp3")
+        engine.save_to_file(text, temp_path.name)
+        engine.runAndWait()
+
+        return temp_path.name
+    except Exception as e:
+        return None
+
 def url_to_audio_summary(url):
     try:
         article_text = extract_main_content(url)
@@ -50,12 +69,11 @@ def url_to_audio_summary(url):
 
         summary = summary_chain.run(text=article_text)
 
-        # Use gTTS for TTS
-        tts = gTTS(text=summary)
-        temp_path = tempfile.NamedTemporaryFile(delete=False, suffix=".mp3")
-        tts.save(temp_path.name)
+        audio_path = generate_human_like_audio(summary)
+        if not audio_path:
+            return summary, None
 
-        return summary, temp_path.name
+        return summary, audio_path
 
     except Exception as e:
         return f"Error: {str(e)}", None
@@ -68,7 +86,7 @@ iface = gr.Interface(
         gr.Audio(label="Audio Summary")
     ],
     title="URL to Audio Summary Agent",
-    description="Summarizes only the article content from a URL and gives an audio summary. CPU-only using gTTS."
+    description="Summarizes only the article content from a URL and gives a more human-like audio summary using pyttsx3. CPU-only."
 )
 
 if __name__ == "__main__":