Upload 3 files

app.py

@@ -1,137 +1,101 @@
+import os
+import uuid
+import logging
 import streamlit as st
 import requests
-import time
-import os
-
-API_URL = "http://127.0.0.1:8000"
-
-st.set_page_config(page_title="Voice Agent", layout="centered")
-st.title(" Ask Me Anything – With Voice!")
-st.caption("Summarized answers with expressive AI voices.")
-
-# 💚 Pulse animation CSS
-st.markdown("""
-    <style>
-    .pulse-box {
-        animation: pulse 1.5s ease-in-out;
-        border: 2px solid #4CAF50 !important;
-        border-radius: 0.5rem;
-        padding: 0.5rem;
-    }
-    @keyframes pulse {
-        0% { box-shadow: 0 0 0px rgba(0, 255, 0, 0.5); }
-        50% { box-shadow: 0 0 20px rgba(0, 255, 0, 0.9); }
-        100% { box-shadow: 0 0 0px rgba(0, 255, 0, 0.5); }
-    }
-    </style>
-""", unsafe_allow_html=True)
-
-voice_options = {
-    "Grandma GG": "grandma GG",
-    "Tech Wizard": "tech wizard",
-    "Perky Sidekick": "perky sidekick",
-    "Bill the Newscaster": "bill the newscaster",
-    "Spunky Charlie": "spunky charlie",
-    "Sassy Teen": "sassy teen"
-}
-
-st.sidebar.header("🔊 Voice Settings")
-voice_choice = st.sidebar.selectbox("Choose a voice:", list(voice_options.keys()))
-voice = voice_options[voice_choice]
-
-for key in ["query", "url", "file_text"]:
-    if key not in st.session_state:
-        st.session_state[key] = ""
+from dotenv import load_dotenv
+from utils import voice_map, get_voice_prompt_style, AUDIO_DIR
+from generate_audio import generate_audio
+
+# Load secrets
+load_dotenv()
+OPENAI_API_KEY = os.getenv("OPENAI_API_KEY")
+ELEVENLABS_API_KEY = os.getenv("ELEVENLABS_API_KEY")
+
+# Setup
+os.makedirs(AUDIO_DIR, exist_ok=True)
+logging.basicConfig(filename="frontend.log", level=logging.INFO)
+
+# Page config
+st.set_page_config(page_title="Voice Agent Pro", page_icon="🎤")
+st.title("🎙️ Voice Agent Pro")
+st.markdown("Summarized answers with expressive AI voices.")
+
+# Sidebar voice selector
+st.sidebar.header("🎚️ Voice Settings")
+voice_label = st.sidebar.selectbox("Choose a voice:", list(voice_map.keys()))
+voice_id = voice_map[voice_label]
+tone_prompt = get_voice_prompt_style(voice_label)
+
+# App state
+if "answer" not in st.session_state: st.session_state.answer = ""
+if "audio_key" not in st.session_state: st.session_state.audio_key = None
+if "file_text" not in st.session_state: st.session_state.file_text = ""
+if "key_points" not in st.session_state: st.session_state.key_points = []
+
+# Query box
+query = st.text_area(
+    "🗨️ Ask or refine something based on the bullets:",
+    value="",
+    placeholder="e.g., What makes you so cool, Grandma?",
+    key="query"
+)
+
+url = st.text_input("🌐 Optional URL to summarize:")
+uploaded_file = st.file_uploader("📎 Or upload a file (PDF, TXT, DOCX)", type=["pdf", "txt", "docx"])
 
+# Clear all
 if st.button("🧹 Clear All"):
     st.session_state.query = ""
-    st.session_state.url = ""
     st.session_state.file_text = ""
-    st.rerun()
-
-# Bullet-to-query safe logic BEFORE rendering input box
-for i in range(10):
-    if st.session_state.get(f"use_bullet_{i}", False):
-        st.session_state["load_bullet_text"] = st.session_state.get(f"bullet_text_{i}", "")
-        st.rerun()
-
-pulse = False
-if "load_bullet_text" in st.session_state:
-    st.session_state.query = st.session_state["load_bullet_text"]
-    del st.session_state["load_bullet_text"]
-    pulse = True
-
-# Query box with optional pulse animation
-if pulse:
-    st.markdown('<div class="pulse-box">', unsafe_allow_html=True)
-    query = st.text_area("💬 Ask or refine something based on the bullets:", key="query")
-    st.markdown("</div>", unsafe_allow_html=True)
-else:
-    query = st.text_area("💬 Ask or refine something based on the bullets:", key="query")
-
-url = st.text_input("🌐 Optional URL to summarize:", placeholder="https://example.com", key="url")
-uploaded_file = st.file_uploader("📎 Or upload a file (PDF, TXT, DOCX)", type=["pdf", "txt", "docx"])
-
-file_text = ""
-if uploaded_file is not None:
-    file_text = uploaded_file.read().decode("utf-8", errors="ignore")
-    st.session_state.file_text = file_text
-    st.success("✅ File uploaded successfully!")
-else:
-    file_text = st.session_state.get("file_text", "")
-
-if st.button("🎯 Summarize"):
-    with st.spinner("Generating response..."):
-        try:
-            payload = {
-                "query": query,
-                "url": url,
-                "voice": voice,
-                "file_text": file_text
-            }
-
+    st.session_state.answer = ""
+    st.session_state.audio_key = None
+    st.session_state.key_points = []
+
+# Summarize
+if st.button("🔁 Summarize"):
+    if not query and not url and not uploaded_file:
+        st.warning("Please enter a question, a URL, or upload a file.")
+    else:
+        with st.spinner("Talking to GPT..."):
             try:
-                response = requests.post(f"{API_URL}/process", json=payload, timeout=15)
-            except requests.exceptions.RequestException as req_err:
-                st.error(f"�� Connection to backend failed:\n{req_err}")
-                st.stop()
-
-            if response.status_code == 200:
-                data = response.json()
-
-                if not query.strip() and data.get("key_points"):
-                    st.markdown("### 📋 Key Points from File")
-                    for i, point in enumerate(data["key_points"]):
-                        st.markdown(f"- {point}")
-                        st.session_state[f"bullet_text_{i}"] = point
-                        st.button("↪️ Use This", key=f"use_bullet_{i}")
-                else:
-                    st.markdown("### 📜 Answer")
-                    st.success(data["answer"])
-
-                # ✅ Audio playback with retries
-                if data.get("audio_key"):
-                    audio_key = data["audio_key"]
-                    audio_url = f"{API_URL}/get-audio/{audio_key}"
-
-                    for attempt in range(10):
-                        try:
-                            audio_check = requests.get(audio_url)
-                            if audio_check.status_code == 200:
-                                audio_bytes = audio_check.content
-                                if audio_bytes and len(audio_bytes) > 2000:
-                                    st.audio(audio_bytes, format="audio/mp3")
-                                    break
-                                else:
-                                    st.warning("⚠️ Audio not ready yet. Waiting...")
-                            else:
-                                st.warning("⚠️ Couldn't reach audio endpoint.")
-                        except Exception as e:
-                            st.warning(f"⚠️ Audio request failed: {e}")
-                        time.sleep(0.5)
-                    else:
-                        st.warning("⚠️ Audio not ready or empty after retries.")
-            else:
-                st.error(f"❌ Backend error: {response.status_code} - {response.text}")
-        except Exception as e:
-            st.error(f"🔥 Unexpected error:\n{e}")
+                if uploaded_file:
+                    st.session_state.file_text = uploaded_file.read().decode("utf-8")
+
+                # Compose prompt
+                context = ""
+                if st.session_state.file_text:
+                    context += st.session_state.file_text + "\n\n"
+                if url:
+                    context += f"Summarize this page: {url}\n\n"
+                context += f"{tone_prompt}\n\nNow answer: {query}"
+
+                # GPT call
+                headers = {"Authorization": f"Bearer {OPENAI_API_KEY}"}
+                response = requests.post(
+                    "https://api.openai.com/v1/chat/completions",
+                    headers=headers,
+                    json={
+                        "model": "gpt-4o",
+                        "messages": [{"role": "user", "content": context}],
+                        "temperature": 0.7
+                    }
+                )
+                answer = response.json()["choices"][0]["message"]["content"]
+                st.session_state.answer = answer
+
+                # Generate audio
+                audio_key = str(uuid.uuid4())
+                generate_audio(answer, voice_id, audio_key)
+                st.session_state.audio_key = audio_key
+
+            except Exception as e:
+                st.error(f"🔥 Error: {e}")
+
+# Display answer
+if st.session_state.answer:
+    st.subheader("📜 Answer")
+    st.success(st.session_state.answer)
+    if st.session_state.audio_key:
+        audio_path = os.path.join(AUDIO_DIR, f"{st.session_state.audio_key}.mp3")
+        st.audio(audio_path)

generate_audio.py

@@ -2,6 +2,9 @@ import os
 import logging
 from elevenlabs import stream
 from elevenlabs.client import ElevenLabs
+from dotenv import load_dotenv
+load_dotenv()
+
 
 AUDIO_DIR = "audio_outputs"
 logger = logging.getLogger(__name__)