Spaces:
Build error
Build error
File size: 44,048 Bytes
d9fb99c |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 |
# azure_unified_assistant.py
"""
Azure OpenAI Unified Assistant with GPT-4o Realtime
Supports: Voice I/O, Streaming Text, Document Q&A, Vision
"""
import os
import json
import asyncio
import base64
import queue
import threading
import time
from datetime import datetime
from typing import Optional, Dict, Any, List, Tuple, AsyncGenerator
from dataclasses import dataclass, field
from enum import Enum
import gradio as gr
import numpy as np
import websockets
from openai import AzureOpenAI, AsyncAzureOpenAI
from dotenv import load_dotenv
import logging
# Load environment variables
load_dotenv()
# Configure logging
logging.basicConfig(
level=logging.INFO,
format='%(asctime)s - %(name)s - %(levelname)s - %(message)s'
)
logger = logging.getLogger(__name__)
# ========================= CONFIGURATION =========================
@dataclass
class UnifiedAssistantConfig:
"""Configuration for GPT-4o Realtime Unified Assistant"""
# Azure Configuration
azure_endpoint: str
api_key: str
api_version: str = "2024-10-01-preview"
# Model Configuration
model_name: str = "gpt-4o-realtime-preview-2024-10-01"
deployment_name: str = "gpt-4o-realtime-preview"
# Vector Storage
vector_store_id: Optional[str] = None
# Audio Configuration
voice: str = "alloy" # alloy, echo, fable, onyx, nova, shimmer
audio_format: str = "pcm16"
sample_rate: int = 24000
# Model Parameters
temperature: float = 0.7
max_tokens: int = 4096
top_p: float = 0.95
# System Configuration
enable_streaming: bool = True
enable_voice: bool = True
enable_vision: bool = True
enable_functions: bool = True
# WebSocket Configuration
ws_timeout: int = 60
ws_max_size: int = 10 * 1024 * 1024 # 10MB
class ResponseType(Enum):
"""Response types from the unified model"""
TEXT = "text"
AUDIO = "audio"
TRANSCRIPT = "transcript"
FUNCTION = "function"
ERROR = "error"
DONE = "done"
# ========================= UNIFIED ASSISTANT =========================
class GPT4oRealtimeAssistant:
"""
Unified Assistant using GPT-4o Realtime Preview
Single model for all capabilities
"""
def __init__(self, config: UnifiedAssistantConfig):
self.config = config
self.websocket = None
self.session_active = False
self.conversation_id = None
# Initialize clients
self._initialize_clients()
# Audio handling
self.audio_queue = queue.Queue()
self.is_recording = False
logger.info(f"✅ Initialized GPT-4o Realtime Assistant with model: {config.model_name}")
def _initialize_clients(self):
"""Initialize both REST and WebSocket clients"""
try:
# REST client for standard operations
self.rest_client = AzureOpenAI(
azure_endpoint=self.config.azure_endpoint,
api_key=self.config.api_key,
api_version=self.config.api_version
)
# Async REST client for streaming
self.async_client = AsyncAzureOpenAI(
azure_endpoint=self.config.azure_endpoint,
api_key=self.config.api_key,
api_version=self.config.api_version
)
logger.info("✅ REST clients initialized successfully")
except Exception as e:
logger.error(f"❌ Failed to initialize clients: {str(e)}")
raise
# ==================== REALTIME WEBSOCKET CONNECTION ====================
async def connect_realtime(self):
"""Establish WebSocket connection for realtime interaction"""
if self.session_active:
logger.info("Session already active")
return
try:
# Construct WebSocket URL
ws_endpoint = self.config.azure_endpoint.replace("https://", "wss://")
ws_url = f"{ws_endpoint}/openai/realtime?api-version={self.config.api_version}&deployment={self.config.deployment_name}"
headers = {
"api-key": self.config.api_key,
"OpenAI-Beta": "realtime=v1"
}
# Connect with timeout
self.websocket = await asyncio.wait_for(
websockets.connect(
ws_url,
extra_headers=headers,
max_size=self.config.ws_max_size
),
timeout=self.config.ws_timeout
)
# Configure session
await self._configure_session()
self.session_active = True
self.conversation_id = f"conv_{datetime.now().strftime('%Y%m%d_%H%M%S')}"
logger.info(f"✅ Realtime session connected: {self.conversation_id}")
except asyncio.TimeoutError:
logger.error("❌ WebSocket connection timeout")
raise
except Exception as e:
logger.error(f"❌ Failed to connect: {str(e)}")
raise
async def _configure_session(self):
"""Configure the realtime session with all capabilities"""
session_config = {
"type": "session.update",
"session": {
"model": self.config.model_name,
"modalities": ["text", "audio"],
"voice": self.config.voice,
"instructions": """You are a helpful AI assistant with access to:
1. Realtime voice conversations
2. Document search and analysis
3. Streaming text responses
4. Image understanding
5. Function calling capabilities
Provide clear, accurate, and helpful responses.
For document questions, cite sources when available.
Keep voice responses concise and natural.""",
# Audio Configuration
"input_audio_format": self.config.audio_format,
"output_audio_format": self.config.audio_format,
"input_audio_transcription": {
"model": "whisper-1"
},
# Voice Activity Detection
"turn_detection": {
"type": "server_vad",
"threshold": 0.5,
"prefix_padding_ms": 300,
"silence_duration_ms": 500
},
# Tools Configuration
"tools": self._get_tools_config(),
"tool_choice": "auto",
# Model Parameters
"temperature": self.config.temperature,
"max_response_output_tokens": self.config.max_tokens
}
}
await self.websocket.send(json.dumps(session_config))
# Wait for session confirmation
response = await self.websocket.recv()
data = json.loads(response)
if data.get("type") == "session.created":
logger.info("✅ Session configured successfully")
else:
logger.warning(f"⚠️ Unexpected session response: {data.get('type')}")
def _get_tools_config(self) -> List[Dict]:
"""Get tools configuration based on settings"""
tools = []
if self.config.vector_store_id:
tools.append({
"type": "file_search",
"file_search": {
"vector_store_ids": [self.config.vector_store_id]
}
})
if self.config.enable_functions:
tools.append({
"type": "code_interpreter"
})
return tools
# ==================== AUDIO HANDLING ====================
async def send_audio(self, audio_data: bytes):
"""Send audio input to the model"""
if not self.session_active:
await self.connect_realtime()
try:
# Convert to base64
audio_base64 = base64.b64encode(audio_data).decode('utf-8')
message = {
"type": "input_audio_buffer.append",
"audio": audio_base64
}
await self.websocket.send(json.dumps(message))
# Commit audio buffer to trigger processing
await self.websocket.send(json.dumps({
"type": "input_audio_buffer.commit"
}))
except Exception as e:
logger.error(f"❌ Error sending audio: {str(e)}")
raise
async def process_audio_stream(self, audio_stream: AsyncGenerator[bytes, None]):
"""Process continuous audio stream"""
if not self.session_active:
await self.connect_realtime()
try:
async for audio_chunk in audio_stream:
await self.send_audio(audio_chunk)
except Exception as e:
logger.error(f"❌ Error processing audio stream: {str(e)}")
raise
# ==================== TEXT HANDLING ====================
async def send_text(self, text: str, use_realtime: bool = False):
"""Send text input - can use either realtime or REST API"""
if use_realtime:
# Use realtime WebSocket
if not self.session_active:
await self.connect_realtime()
message = {
"type": "conversation.item.create",
"item": {
"type": "message",
"role": "user",
"content": [
{
"type": "input_text",
"text": text
}
]
}
}
await self.websocket.send(json.dumps(message))
# Trigger response
await self.websocket.send(json.dumps({
"type": "response.create",
"response": {
"modalities": ["text", "audio"] if self.config.enable_voice else ["text"]
}
}))
else:
# Use REST API for text-only
return self.rest_client.chat.completions.create(
model=self.config.deployment_name,
messages=[
{"role": "user", "content": text}
],
stream=self.config.enable_streaming,
temperature=self.config.temperature,
max_tokens=self.config.max_tokens
)
# ==================== STREAMING RESPONSES ====================
async def stream_response(self) -> AsyncGenerator[Dict[str, Any], None]:
"""Stream responses from the realtime connection"""
if not self.session_active:
logger.error("❌ No active session")
return
try:
while True:
message = await asyncio.wait_for(
self.websocket.recv(),
timeout=self.config.ws_timeout
)
data = json.loads(message)
response_type = data.get("type", "")
# Handle different response types
if response_type == "response.audio.delta":
yield {
"type": ResponseType.AUDIO,
"data": base64.b64decode(data.get("delta", ""))
}
elif response_type == "response.text.delta":
yield {
"type": ResponseType.TEXT,
"data": data.get("delta", "")
}
elif response_type == "response.audio_transcript.delta":
yield {
"type": ResponseType.TRANSCRIPT,
"data": data.get("delta", "")
}
elif response_type == "response.function_call_arguments.delta":
yield {
"type": ResponseType.FUNCTION,
"data": json.loads(data.get("delta", "{}"))
}
elif response_type == "response.done":
yield {
"type": ResponseType.DONE,
"data": None
}
break
elif response_type == "error":
yield {
"type": ResponseType.ERROR,
"data": data.get("error", {})
}
break
except asyncio.TimeoutError:
logger.error("❌ Response timeout")
yield {
"type": ResponseType.ERROR,
"data": {"message": "Response timeout"}
}
except Exception as e:
logger.error(f"❌ Stream error: {str(e)}")
yield {
"type": ResponseType.ERROR,
"data": {"message": str(e)}
}
# ==================== DOCUMENT Q&A ====================
async def query_documents(self, query: str, stream: bool = True):
"""Query documents using vector search"""
if not self.config.vector_store_id:
return "No vector store configured. Please set up a vector store ID."
try:
# Create assistant with vector search
assistant = self.rest_client.beta.assistants.create(
model=self.config.deployment_name,
name="Document Q&A Assistant",
instructions="""Answer questions based on the provided documents.
Cite sources when possible.
If information is not in the documents, say so clearly.""",
tools=[{"type": "file_search"}],
tool_resources={
"file_search": {
"vector_store_ids": [self.config.vector_store_id]
}
},
temperature=self.config.temperature
)
# Create thread
thread = self.rest_client.beta.threads.create()
# Add message
self.rest_client.beta.threads.messages.create(
thread_id=thread.id,
role="user",
content=query
)
# Run with streaming
if stream:
with self.rest_client.beta.threads.runs.create_and_stream(
thread_id=thread.id,
assistant_id=assistant.id,
) as stream:
for event in stream:
if event.event == 'thread.message.delta':
if hasattr(event.data, 'delta') and hasattr(event.data.delta, 'content'):
for content in event.data.delta.content:
if content.type == 'text' and hasattr(content.text, 'value'):
yield content.text.value
else:
# Non-streaming response
run = self.rest_client.beta.threads.runs.create(
thread_id=thread.id,
assistant_id=assistant.id
)
# Wait for completion
while run.status in ['queued', 'in_progress']:
time.sleep(1)
run = self.rest_client.beta.threads.runs.retrieve(
thread_id=thread.id,
run_id=run.id
)
if run.status == 'completed':
messages = self.rest_client.beta.threads.messages.list(
thread_id=thread.id
)
if messages.data:
return messages.data[0].content[0].text.value
else:
return f"Query failed: {run.status}"
except Exception as e:
logger.error(f"❌ Document query error: {str(e)}")
return f"Error querying documents: {str(e)}"
# ==================== MULTIMODAL HANDLING ====================
async def analyze_image(self, image_path: str, question: str):
"""Analyze image with text/voice response"""
try:
# Read and encode image
with open(image_path, "rb") as image_file:
image_data = base64.b64encode(image_file.read()).decode('utf-8')
if self.session_active:
# Use realtime connection for voice response
message = {
"type": "conversation.item.create",
"item": {
"type": "message",
"role": "user",
"content": [
{
"type": "input_text",
"text": question
},
{
"type": "image_url",
"image_url": {
"url": f"data:image/jpeg;base64,{image_data}",
"detail": "high"
}
}
]
}
}
await self.websocket.send(json.dumps(message))
# Trigger response
await self.websocket.send(json.dumps({
"type": "response.create"
}))
# Stream response
async for response in self.stream_response():
yield response
else:
# Use REST API
response = self.rest_client.chat.completions.create(
model=self.config.deployment_name,
messages=[
{
"role": "user",
"content": [
{"type": "text", "text": question},
{
"type": "image_url",
"image_url": {
"url": f"data:image/jpeg;base64,{image_data}",
"detail": "high"
}
}
]
}
],
max_tokens=self.config.max_tokens,
temperature=self.config.temperature
)
yield {
"type": ResponseType.TEXT,
"data": response.choices[0].message.content
}
except Exception as e:
logger.error(f"❌ Image analysis error: {str(e)}")
yield {
"type": ResponseType.ERROR,
"data": {"message": str(e)}
}
# ==================== CLEANUP ====================
async def disconnect(self):
"""Disconnect realtime session"""
if self.websocket:
await self.websocket.close()
self.session_active = False
logger.info("✅ Disconnected from realtime session")
# ========================= GRADIO INTERFACE =========================
class UnifiedGradioInterface:
"""Enhanced Gradio interface for the unified assistant"""
def __init__(self, assistant: GPT4oRealtimeAssistant):
self.assistant = assistant
self.conversation_history = []
self.audio_processor = AudioProcessor()
def create_interface(self) -> gr.Blocks:
"""Create the complete Gradio interface"""
with gr.Blocks(
theme=gr.themes.Soft(),
title="Azure GPT-4o Realtime Assistant",
css=self._get_custom_css()
) as interface:
# Header
gr.Markdown(
"""
# 🌟 Azure GPT-4o Realtime Assistant
**Unified Model with All Capabilities:**
- 🎙️ **Voice:** Realtime voice conversations with ~300ms latency
- ⚡ **Streaming:** Live text streaming responses
- 📚 **Documents:** Vector search and Q&A on your documents
- 🖼️ **Vision:** Image analysis and understanding
- 🔧 **Functions:** Tool calling and code interpretation
**Model:** `gpt-4o-realtime-preview-2024-10-01` | **Context:** 128K tokens
"""
)
# State management
state = gr.State({
"session_active": False,
"conversation_id": None
})
with gr.Tabs() as tabs:
# ============ TAB 1: STREAMING CHAT ============
with gr.Tab("💬 Streaming Chat", elem_id="chat-tab"):
with gr.Row():
with gr.Column(scale=4):
chatbot = gr.Chatbot(
height=500,
show_label=False,
elem_id="chatbot",
bubble_full_width=False
)
with gr.Row():
msg = gr.Textbox(
placeholder="Type your message here... (Press Enter to send)",
label="Message",
scale=4,
lines=2
)
with gr.Column(scale=1):
send_btn = gr.Button("Send", variant="primary", size="lg")
clear_btn = gr.Button("Clear", variant="secondary")
with gr.Column(scale=1):
gr.Markdown("### Settings")
streaming_mode = gr.Radio(
choices=["Stream", "Complete"],
value="Stream",
label="Response Mode"
)
temperature = gr.Slider(
minimum=0,
maximum=1,
value=0.7,
step=0.1,
label="Temperature"
)
gr.Markdown("### Status")
status_display = gr.Markdown("🟢 Ready")
# Chat handlers
async def handle_chat(message, history, mode):
if not message:
return history
history = history or []
history.append([message, ""])
try:
if mode == "Stream":
response = ""
# Stream text response
await self.assistant.send_text(message, use_realtime=True)
async for chunk in self.assistant.stream_response():
if chunk["type"] == ResponseType.TEXT:
response += chunk["data"]
history[-1][1] = response
yield history
elif chunk["type"] == ResponseType.DONE:
break
else:
# Complete response
result = await self.assistant.send_text(message, use_realtime=False)
history[-1][1] = result.choices[0].message.content
yield history
except Exception as e:
history[-1][1] = f"❌ Error: {str(e)}"
yield history
msg.submit(
handle_chat,
[msg, chatbot, streaming_mode],
chatbot
).then(
lambda: "",
outputs=msg
)
send_btn.click(
handle_chat,
[msg, chatbot, streaming_mode],
chatbot
).then(
lambda: "",
outputs=msg
)
clear_btn.click(
lambda: ([], ""),
outputs=[chatbot, msg]
)
# ============ TAB 2: VOICE INTERACTION ============
with gr.Tab("🎙️ Voice Chat", elem_id="voice-tab"):
gr.Markdown(
"""
### Realtime Voice Interaction
Talk naturally with the assistant using your microphone.
"""
)
with gr.Row():
with gr.Column():
# Connection control
with gr.Row():
connect_btn = gr.Button("🔌 Connect Voice", variant="primary")
disconnect_btn = gr.Button("🔌 Disconnect", variant="stop")
connection_status = gr.Markdown("⚪ Not connected")
# Audio interface
audio_input = gr.Audio(
sources=["microphone"],
type="numpy",
streaming=True,
label="🎤 Speak to Assistant"
)
audio_output = gr.Audio(
label="🔊 Assistant Response",
type="numpy",
autoplay=True
)
with gr.Column():
# Conversation display
transcript = gr.Textbox(
label="Conversation Transcript",
lines=20,
max_lines=30,
interactive=False
)
# Voice settings
voice_select = gr.Dropdown(
choices=["alloy", "echo", "fable", "onyx", "nova", "shimmer"],
value="alloy",
label="Voice Selection"
)
# Voice handlers
async def connect_voice():
try:
await self.assistant.connect_realtime()
return "🟢 Connected to realtime voice"
except Exception as e:
return f"🔴 Connection failed: {str(e)}"
async def disconnect_voice():
try:
await self.assistant.disconnect()
return "⚪ Disconnected"
except Exception as e:
return f"🔴 Error: {str(e)}"
async def process_voice(audio_data):
if audio_data is None:
return None, ""
try:
# Convert audio format
sample_rate, audio = audio_data
audio_bytes = (audio * 32767).astype(np.int16).tobytes()
# Send audio
await self.assistant.send_audio(audio_bytes)
# Collect responses
transcript_text = ""
audio_chunks = []
async for response in self.assistant.stream_response():
if response["type"] == ResponseType.AUDIO:
audio_chunks.append(response["data"])
elif response["type"] == ResponseType.TRANSCRIPT:
transcript_text += response["data"]
elif response["type"] == ResponseType.DONE:
break
# Process audio output
if audio_chunks:
combined = b''.join(audio_chunks)
audio_array = np.frombuffer(combined, dtype=np.int16) / 32767.0
return (24000, audio_array), transcript_text
return None, transcript_text
except Exception as e:
return None, f"Error: {str(e)}"
connect_btn.click(
connect_voice,
outputs=connection_status
)
disconnect_btn.click(
disconnect_voice,
outputs=connection_status
)
audio_input.stream(
process_voice,
audio_input,
[audio_output, transcript]
)
# ============ TAB 3: DOCUMENT Q&A ============
with gr.Tab("📚 Document Q&A", elem_id="doc-tab"):
gr.Markdown(
"""
### Query Your Documents
Ask questions about documents in your vector storage.
"""
)
with gr.Row():
with gr.Column():
doc_query = gr.Textbox(
placeholder="Ask a question about your documents...",
label="Query",
lines=3
)
with gr.Row():
search_btn = gr.Button("🔍 Search", variant="primary")
stream_toggle = gr.Checkbox(label="Stream Response", value=True)
doc_response = gr.Textbox(
label="Response",
lines=15,
max_lines=25
)
with gr.Column(scale=1):
gr.Markdown("### Vector Store Info")
vector_status = gr.Markdown(
f"**Store ID:** `{self.assistant.config.vector_store_id or 'Not configured'}`"
)
gr.Markdown(
"""
### Tips
- Be specific in your questions
- Reference document sections if known
- Ask for citations when needed
"""
)
# Document Q&A handler
async def search_documents(query, stream):
if not query:
return "Please enter a query."
try:
if stream:
response = ""
async for chunk in self.assistant.query_documents(query, stream=True):
response += chunk
yield response
else:
response = await self.assistant.query_documents(query, stream=False)
yield response
except Exception as e:
yield f"❌ Error: {str(e)}"
search_btn.click(
search_documents,
[doc_query, stream_toggle],
doc_response
)
doc_query.submit(
search_documents,
[doc_query, stream_toggle],
doc_response
)
# ============ TAB 4: IMAGE ANALYSIS ============
with gr.Tab("🖼️ Vision", elem_id="vision-tab"):
gr.Markdown(
"""
### Image Analysis with Voice
Upload an image and ask questions about it.
"""
)
with gr.Row():
with gr.Column():
image_input = gr.Image(
type="filepath",
label="Upload Image"
)
image_query = gr.Textbox(
placeholder="What would you like to know about this image?",
label="Question",
lines=2
)
analyze_btn = gr.Button("🔍 Analyze", variant="primary")
with gr.Column():
image_response = gr.Textbox(
label="Analysis",
lines=10
)
image_audio = gr.Audio(
label="Voice Response",
type="numpy"
)
# Image analysis handler
async def analyze_image(image, query):
if not image or not query:
return "Please provide both an image and a question.", None
try:
text_response = ""
audio_chunks = []
async for response in self.assistant.analyze_image(image, query):
if response["type"] == ResponseType.TEXT:
text_response += response["data"]
elif response["type"] == ResponseType.AUDIO:
audio_chunks.append(response["data"])
# Process audio if available
audio_output = None
if audio_chunks:
combined = b''.join(audio_chunks)
audio_array = np.frombuffer(combined, dtype=np.int16) / 32767.0
audio_output = (24000, audio_array)
return text_response, audio_output
except Exception as e:
return f"❌ Error: {str(e)}", None
analyze_btn.click(
analyze_image,
[image_input, image_query],
[image_response, image_audio]
)
# ============ TAB 5: SETTINGS ============
with gr.Tab("⚙️ Settings", elem_id="settings-tab"):
gr.Markdown("### Model Configuration")
with gr.Row():
with gr.Column():
gr.Markdown("#### Model Information")
model_info = gr.JSON(
value={
"model": self.assistant.config.model_name,
"deployment": self.assistant.config.deployment_name,
"context_window": 128000,
"capabilities": [
"realtime_audio",
"streaming_text",
"document_qa",
"vision",
"function_calling"
],
"audio_format": self.assistant.config.audio_format,
"sample_rate": self.assistant.config.sample_rate
},
label="Current Configuration"
)
with gr.Column():
gr.Markdown("#### Performance Metrics")
metrics = gr.JSON(
value={
"voice_latency": "~300ms",
"text_streaming": "~50ms/token",
"document_search": "1-2 seconds",
"image_analysis": "2-3 seconds"
},
label="Expected Performance"
)
# Footer
gr.Markdown(
"""
---
<center>
**Azure GPT-4o Realtime** | Single Model for All Tasks |
[Documentation](https://docs.microsoft.com/azure/cognitive-services/openai/)
</center>
"""
)
return interface
def _get_custom_css(self) -> str:
"""Get custom CSS for the interface"""
return """
#chatbot {
height: 500px !important;
}
.gradio-container {
max-width: 1400px !important;
margin: auto !important;
}
#voice-tab .gradio-audio {
height: 120px !important;
}
.markdown-text {
font-size: 14px !important;
}
"""
# ========================= AUDIO PROCESSOR =========================
class AudioProcessor:
"""Handle audio format conversions and processing"""
@staticmethod
def numpy_to_pcm16(audio_array: np.ndarray) -> bytes:
"""Convert numpy array to PCM16 bytes"""
return (audio_array * 32767).astype(np.int16).tobytes()
@staticmethod
def pcm16_to_numpy(audio_bytes: bytes) -> np.ndarray:
"""Convert PCM16 bytes to numpy array"""
return np.frombuffer(audio_bytes, dtype=np.int16) / 32767.0
# ========================= MAIN APPLICATION =========================
async def main():
"""Main application entry point"""
# Load configuration from environment
config = UnifiedAssistantConfig(
azure_endpoint=os.getenv("AZURE_OPENAI_ENDPOINT"),
api_key=os.getenv("AZURE_OPENAI_API_KEY"),
api_version=os.getenv("AZURE_API_VERSION", "2024-10-01-preview"),
deployment_name=os.getenv("AZURE_DEPLOYMENT_NAME", "gpt-4o-realtime-preview"),
vector_store_id=os.getenv("AZURE_VECTOR_STORE_ID"),
voice=os.getenv("AZURE_VOICE", "alloy"),
temperature=float(os.getenv("TEMPERATURE", "0.7")),
max_tokens=int(os.getenv("MAX_TOKENS", "4096"))
)
# Validate configuration
if not config.azure_endpoint or not config.api_key:
logger.error("❌ Missing required environment variables")
raise ValueError(
"Please set AZURE_OPENAI_ENDPOINT and AZURE_OPENAI_API_KEY environment variables"
)
if not config.vector_store_id:
logger.warning("⚠️ No vector store ID configured - document Q&A will be limited")
try:
# Initialize assistant
logger.info("🚀 Initializing GPT-4o Realtime Unified Assistant...")
assistant = GPT4oRealtimeAssistant(config)
# Create Gradio interface
logger.info("🎨 Creating Gradio interface...")
interface_manager = UnifiedGradioInterface(assistant)
interface = interface_manager.create_interface()
# Launch application
logger.info("✅ Launching application...")
interface.queue(max_size=100) # Enable queuing for concurrent users
interface.launch(
server_name="0.0.0.0",
server_port=7860,
share=True, # Create public URL
show_error=True,
quiet=False
)
except KeyboardInterrupt:
logger.info("👋 Shutting down gracefully...")
if assistant.session_active:
await assistant.disconnect()
except Exception as e:
logger.error(f"❌ Fatal error: {str(e)}")
raise
def run():
"""Synchronous entry point"""
asyncio.run(main())
if __name__ == "__main__":
run() |