Upload 4 files

generate_audio.py

@@ -0,0 +1,50 @@
+import os
+from dotenv import load_dotenv
+from elevenlabs.client import ElevenLabs
+from logger_setup import logger
+
+# Load environment variables
+load_dotenv()
+
+# Use absolute path for output
+AUDIO_DIR = os.path.join(os.path.dirname(__file__), "audio_outputs")
+
+# Verify API key
+api_key = os.getenv("ELEVENLABS_API_KEY")
+if not api_key:
+    logger.error("❌ ELEVENLABS_API_KEY is missing or not loaded from .env")
+    raise RuntimeError("ELEVENLABS_API_KEY missing")
+
+client = ElevenLabs(api_key=api_key)
+
+def generate_audio(text: str, voice_id: str, audio_key: str):
+    try:
+        logger.info("🎯 Starting ElevenLabs audio generation")
+        os.makedirs(AUDIO_DIR, exist_ok=True)
+
+        try:
+            audio_stream = client.text_to_speech.convert_as_stream(
+                text=text,
+                voice_id=voice_id,
+                model_id="eleven_multilingual_v2"
+            )
+            logger.info("✅ Audio stream received from ElevenLabs")
+        except Exception as stream_err:
+            logger.error(f"❌ Failed to get audio stream: {stream_err}")
+            raise
+
+        output_path = os.path.join(AUDIO_DIR, f"{audio_key}.mp3")
+
+        try:
+            with open(output_path, "wb") as f:
+                for chunk in audio_stream:
+                    if isinstance(chunk, bytes):
+                        f.write(chunk)
+            logger.info(f"✅ Audio saved to {output_path}")
+        except Exception as write_err:
+            logger.error(f"❌ Failed to save audio to file: {write_err}")
+            raise
+
+    except Exception as e:
+        logger.exception("🔥 Exception in generate_audio")
+        raise

logger_setup.py

@@ -0,0 +1,16 @@
+# logger_setup.py
+import logging
+import os
+
+LOG_FILE = os.path.join(os.path.dirname(__file__), "logfile.log")
+
+logging.basicConfig(
+    level=logging.INFO,
+    format="%(asctime)s [%(levelname)s] %(name)s - %(message)s",
+    handlers=[
+        logging.FileHandler(LOG_FILE, mode='a', encoding='utf-8'),
+        logging.StreamHandler()
+    ]
+)
+
+logger = logging.getLogger("voice-agent")

requirements.txt

@@ -0,0 +1,11 @@
+streamlit
+requests
+openai
+python-dotenv
+PyMuPDF
+python-docx
+elevenlabs
+
+qdrant-client
+fastembed
+firecrawl

utils.py

@@ -0,0 +1,59 @@
+import os
+import re
+from urllib.parse import urlparse
+from bs4 import BeautifulSoup
+
+AUDIO_DIR = "audio_outputs"
+
+voice_map = {'grandma GG': 'rKVm0Cb9J2wrzmZupJea', 'tech wizard': 'ocn9CucaUfmmP6Two6Ik', 'perky sidekick': 'DWR3ijzKmphlRUhbBI7t', 'bill the newscaster': 'R1vZMopVRO75M5xBKX52', 'spunky charlie': 'q3yXDjF0aq4JCEo9u2g4', 'sassy teen': 'mBj2IDD9aXruPJHLGCAv'}
+
+def sanitize_url(url):
+    if not url.startswith(("http://", "https://")):
+        return "https://" + url
+    return url
+
+def extract_internal_links(html_content, base_url):
+    soup = BeautifulSoup(html_content, "html.parser")
+    parsed_base = urlparse(base_url)
+    base_domain = parsed_base.netloc
+
+    links = set()
+    for tag in soup.find_all("a", href=True):
+        href = tag["href"]
+        parsed_href = urlparse(href)
+
+        if parsed_href.netloc == "" or parsed_href.netloc == base_domain:
+            full_url = parsed_href.geturl()
+            if not full_url.startswith("http"):
+                full_url = f"{parsed_base.scheme}://{base_domain}{href}"
+            links.add(full_url)
+
+    return list(links)
+
+def crawl_documentation(url):
+    import requests
+    try:
+        response = requests.get(url, timeout=10)
+        response.raise_for_status()
+        return response.text
+    except Exception as e:
+        return f"Error fetching page: {e}"
+
+def get_voice_prompt_style(voice):
+    tone = {'grandma GG': 'dry, witty, and brutally honest — will roast you if you mess up.', 'tech wizard': 'cryptic, snarky, and a prodigy with code — speaks in digital spells.', 'perky sidekick': 'energetic, cheerful, and endlessly supportive — like a high-five machine.', 'bill the newscaster': 'polished, confident, and composed — delivers everything like breaking news.', 'spunky charlie': 'wildly curious, playful, and full of devil-may-care energy.', 'sassy teen': 'sarcastic, sharp-tongued, and too cool to care — flexes brainpower with attitude.'}
+    return tone.get(voice.lower(), "neutral")
+
+def save_audio_file(audio_path, content):
+    os.makedirs(AUDIO_DIR, exist_ok=True)
+    with open(audio_path, "wb") as f:
+        f.write(content)
+
+__all__ = [
+    "sanitize_url",
+    "extract_internal_links",
+    "crawl_documentation",
+    "get_voice_prompt_style",
+    "save_audio_file",
+    "voice_map",
+    "AUDIO_DIR",
+]