Spaces:
Sleeping
Sleeping
File size: 14,750 Bytes
e7afd1c |
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 |
# _virtual_chat.py
import queue
emotion_queue = queue.Queue()
from fer.fer import FER
# GPU optional โ FER runs on CPU by default
emotion_detector = FER(mtcnn=True)
import streamlit as st
import cv2, time, os, tempfile, threading, datetime, glob, base64
import speech_recognition as sr
# from deepface import DeepFace
from db import get_db
import numpy as np
DEEPFACE_AVAILABLE = True
try:
import deepface
except:
DEEPFACE_AVAILABLE = False
def smooth_emotion_ema(emotion: str, confidence: float, alpha: float = 0.3):
ema = st.session_state.emotion_ema
if emotion not in ema:
ema[emotion] = confidence
else:
ema[emotion] = alpha * confidence + (1 - alpha) * ema[emotion]
# Decay other emotions
for e in list(ema.keys()):
if e != emotion:
ema[e] *= (1 - alpha)
# Pick strongest
smoothed_emotion = max(ema, key=ema.get)
return smoothed_emotion, ema[smoothed_emotion]
# ==================== Virtual Chat Mode ====================
def virtual_chat_mode(username=None, detect_text_emotion_func=None, retrieve_answer_func=None):
if username is None:
username = "Guest"
if detect_text_emotion_func is None:
def detect_text_emotion_func(text):
return "neutral", 0.5
if retrieve_answer_func is None:
def retrieve_answer_func(query, emotion):
return "I'm here to help. Please connect the RAG system."
st.title("๐ฅ Virtual Chat - Live Face Emotion Detection")
if not DEEPFACE_AVAILABLE:
st.error("โ ๏ธ DeepFace library not installed. Please install it: `pip install deepface`")
return
st.info("๐ธ Camera stays open! Chat freely with text or voice - bot speaks back!")
EMOJI_MAP = {
"happy": "๐", "sad": "๐ข", "angry": "๐ ", "fear": "๐จ",
"neutral": "๐", "surprise": "๐ฒ", "disgust": "๐คข"
}
# ==================== SESSION STATE ====================
for key, val in {
"live_emotion": "neutral",
"live_confidence": 0.0,
"camera_active": False,
"frame_counter": 0,
"cap": None,
"emotion_ema": {},
"virtual_chat_history": [],
"emotion_timeline": [],
"last_frame": None,
"detecting_emotion": False, # Flag to prevent multiple detections
}.items():
if key not in st.session_state:
st.session_state[key] = val
# ==================== TTS ====================
try:
import pyttsx3
tts_engine = pyttsx3.init()
tts_engine.setProperty('rate', 150)
tts_engine.available = True
except:
tts_engine = type('', (), {"available": False})()
def speak_async(text):
if not tts_engine.available:
return
def speak():
try:
tts_engine.say(text)
tts_engine.runAndWait()
except:
pass
threading.Thread(target=speak, daemon=True).start()
# ==================== ASYNC EMOTION DETECTION ====================
def detect_emotion_async(frame_copy):
if st.session_state.detecting_emotion:
return
st.session_state.detecting_emotion = True
def _detect():
try:
# FER expects RGB
rgb_frame = cv2.cvtColor(frame_copy, cv2.COLOR_BGR2RGB)
result = emotion_detector.detect_emotions(rgb_frame)
if not result:
return
# pick largest face
face = max(result, key=lambda r: r["box"][2] * r["box"][3])
emotions = face["emotions"]
emotion = max(emotions, key=emotions.get)
confidence = emotions[emotion]
emotion_queue.put((emotion, confidence))
except Exception:
pass
finally:
st.session_state.detecting_emotion = False
threading.Thread(target=_detect, daemon=True).start()
# ==================== CRISIS KEYWORDS ====================
CRISIS_KEYWORDS = ["suicide", "kill myself", "end my life", "i want to die", "harm myself"]
# ==================== LAYOUT ====================
col1, col2 = st.columns([1, 1])
# ==================== CAMERA ====================
with col1:
st.subheader("๐น Live Camera Feed")
st.metric(
"Your Current Emotion",
f"{EMOJI_MAP.get(st.session_state.live_emotion, '๐')} {st.session_state.live_emotion.title()}",
f"{st.session_state.live_confidence:.0%}"
)
btn_col1, btn_col2, btn_col3 = st.columns(3)
with btn_col1:
if not st.session_state.camera_active:
if st.button("๐ท Start Camera", use_container_width=True):
try:
# Release any existing camera
if st.session_state.cap is not None:
st.session_state.cap.release()
# Open new camera
st.session_state.cap = cv2.VideoCapture(0)
st.session_state.cap.set(cv2.CAP_PROP_FRAME_WIDTH, 640)
st.session_state.cap.set(cv2.CAP_PROP_FRAME_HEIGHT, 480)
st.session_state.cap.set(cv2.CAP_PROP_FPS, 30)
st.session_state.cap.set(cv2.CAP_PROP_BUFFERSIZE, 1)
if st.session_state.cap.isOpened():
st.session_state.camera_active = True
st.session_state.frame_counter = 0
st.success("โ
Camera started!")
else:
st.error("โ Failed to open camera")
except Exception as e:
st.error(f"Camera error: {str(e)}")
st.rerun()
else:
if st.button("โน Stop Camera", use_container_width=True):
st.session_state.camera_active = False
if st.session_state.cap:
st.session_state.cap.release()
st.session_state.cap = None
save_session_to_mongo(username)
st.success("Camera stopped and session saved!")
st.rerun()
with btn_col2:
if st.button("๐ธ Snapshot", use_container_width=True, disabled=not st.session_state.camera_active):
if st.session_state.last_frame is not None:
os.makedirs("snapshots", exist_ok=True)
fname = f"snapshots/snap_{int(time.time())}.jpg"
cv2.imwrite(fname, st.session_state.last_frame)
st.success(f"๐ท Saved!")
with btn_col3:
if st.button("๐ Detect Now", use_container_width=True, disabled=not st.session_state.camera_active):
if st.session_state.last_frame is not None and not st.session_state.detecting_emotion:
detect_emotion_async(st.session_state.last_frame.copy())
st.info("๐ Detecting...")
video_placeholder = st.empty()
# ==================== CAMERA LOOP ====================
if st.session_state.camera_active:
# โ
Process emotion results from background thread
while not emotion_queue.empty():
emotion, confidence = emotion_queue.get()
smooth_e, smooth_c = smooth_emotion_ema(emotion, confidence)
st.session_state.live_emotion = smooth_e
st.session_state.live_confidence = smooth_c
st.session_state.emotion_timeline.append({
"timestamp": datetime.datetime.utcnow().isoformat(),
"raw_emotion": emotion,
"raw_confidence": confidence,
"smoothed_emotion": smooth_e,
"smoothed_confidence": smooth_c
})
st.session_state.detecting_emotion = False
# Ensure camera is open
if st.session_state.cap is None or not st.session_state.cap.isOpened():
st.session_state.cap = cv2.VideoCapture(0)
st.session_state.cap.set(cv2.CAP_PROP_FRAME_WIDTH, 640)
st.session_state.cap.set(cv2.CAP_PROP_FRAME_HEIGHT, 480)
ret, frame = st.session_state.cap.read()
if ret:
frame = cv2.flip(frame, 1)
st.session_state.last_frame = frame.copy()
st.session_state.frame_counter += 1
# ๐ Emotion detection every ~5 seconds (stable)
if (
st.session_state.frame_counter % 150 == 0
and not st.session_state.detecting_emotion
):
detect_emotion_async(frame.copy())
# Overlay emotion
emotion_text = st.session_state.live_emotion.upper()
confidence_text = f"{st.session_state.live_confidence:.0%}"
cv2.rectangle(frame, (5, 5), (300, 70), (0, 0, 0), -1)
cv2.rectangle(frame, (5, 5), (300, 70), (0, 255, 0), 2)
cv2.putText(
frame, emotion_text, (15, 35),
cv2.FONT_HERSHEY_SIMPLEX, 0.8, (0, 255, 0), 2
)
cv2.putText(
frame, confidence_text, (15, 60),
cv2.FONT_HERSHEY_SIMPLEX, 0.6, (255, 255, 255), 2
)
video_placeholder.image(frame.copy(), channels="BGR")
else:
video_placeholder.error("โ Cannot read from camera")
# ๐ Controlled rerun (ONLY when camera is active)
time.sleep(0.05) # ~20 FPS (CPU-safe)
st.rerun()
# ==================== CHAT ====================
with col2:
st.subheader("๐ฌ Chat Interface")
# Display chat history in a scrollable container
chat_container = st.container()
with chat_container:
for m in st.session_state.virtual_chat_history:
with st.chat_message(m["role"]):
st.markdown(m["content"])
# Input mode selection (stays inside column)
input_mode = st.radio("Choose Input Mode:", ["๐ฌ Type", "๐ค Speak"], horizontal=True)
# ==================== INPUT SECTION (OUTSIDE COLUMNS) ====================
user_input = ""
if input_mode == "๐ฌ Type":
user_input = st.chat_input("Type your message here...")
else:
if st.button("๐ค Press & Speak", use_container_width=True):
recognizer = sr.Recognizer()
with st.spinner("๐ค Listening..."):
try:
with sr.Microphone() as source:
st.info("Speak now...")
recognizer.adjust_for_ambient_noise(source, duration=0.5)
audio = recognizer.listen(source, timeout=5, phrase_time_limit=10)
user_input = recognizer.recognize_google(audio)
st.success(f"โ
You said: {user_input}")
except sr.WaitTimeoutError:
st.error("โฑ๏ธ No speech detected")
except sr.UnknownValueError:
st.error("โ Could not understand audio")
except Exception as e:
st.error(f"โ Error: {str(e)}")
# ==================== PROCESS USER INPUT ====================
if user_input:
# Get emotions
face_emotion = st.session_state.live_emotion
face_confidence = st.session_state.live_confidence
text_emotion, text_confidence = detect_text_emotion_func(user_input)
# Choose emotion based on confidence
final_emotion = face_emotion if face_confidence > 0.5 else text_emotion
# Check for crisis keywords
is_crisis = any(keyword in user_input.lower() for keyword in CRISIS_KEYWORDS)
if is_crisis:
bot_reply = (
"โ ๏ธ I'm very concerned about what you're sharing. Please reach out for help immediately:\n\n"
"๐ฎ๐ณ India Helplines:\n"
"โข AASRA: 91-22-27546669\n"
"โข Vandrevala Foundation: 1860-2662-345\n\n"
"You are not alone. Please talk to someone who can help."
)
else:
# Retrieve answer from RAG system
bot_reply = retrieve_answer_func(user_input, final_emotion)
# Add to chat history
st.session_state.virtual_chat_history.append({
"role": "user",
"content": user_input
})
st.session_state.virtual_chat_history.append({
"role": "assistant",
"content": bot_reply
})
# Speak the reply
speak_async(bot_reply)
st.rerun()
# ==================== SAVE TO MONGODB ====================
def save_session_to_mongo(username: str):
try:
db = get_db()
col = db["virtual_chat_sessions"]
# Read all snapshots
snapshot_data = []
if os.path.exists("snapshots"):
for f in glob.glob("snapshots/*.jpg"):
try:
with open(f, "rb") as img_file:
snapshot_data.append(
base64.b64encode(img_file.read()).decode("utf-8")
)
except Exception:
pass
doc = {
"username": username,
"timestamp": datetime.datetime.utcnow(),
"chat_history": st.session_state.virtual_chat_history,
"emotion_timeline": st.session_state.emotion_timeline,
"snapshots": snapshot_data,
"total_messages": len(st.session_state.virtual_chat_history),
"session_duration_emotions": len(st.session_state.emotion_timeline),
}
col.insert_one(doc)
st.success("โ
Virtual Chat Session saved to MongoDB")
except Exception as e:
st.warning("โ ๏ธ Could not save session to database.")
|