Update app.py

app.py

@@ -1,45 +1,128 @@
 import streamlit as st
 import cv2
 import numpy as np
+import mediapipe as mp
+import requests
+from gtts import gTTS
+import os
+import tempfile
 
-# -------------------------------
-# Page Configuration
-# -------------------------------
-st.set_page_config(
-    page_title="Hand2Voice",
-    layout="wide"
+# ---------------- CONFIG ----------------
+st.set_page_config(page_title="Hand2Voice", layout="wide")
+
+GESTURE_URL = "https://raw.githubusercontent.com/YOUR_USERNAME/Hand2Voice/main/gestures/gesture_rules.json"
+
+# ---------------- LOAD GESTURES ----------------
+@st.cache_data
+def load_gestures():
+    response = requests.get(GESTURE_URL)
+    return response.json()["gestures"]
+
+gesture_rules = load_gestures()
+
+# ---------------- MEDIAPIPE SETUP ----------------
+mp_hands = mp.solutions.hands
+hands = mp_hands.Hands(
+    static_image_mode=True,
+    max_num_hands=1,
+    min_detection_confidence=0.7
 )
 
-st.title("🤚 Hand2Voice")
-st.subheader("Hand Gestures to Speech Interface")
+# ---------------- FINGER LOGIC ----------------
+def get_finger_states(hand_landmarks):
+    finger_tips = [4, 8, 12, 16, 20]
+    finger_bases = [2, 6, 10, 14, 18]
+
+    states = []
+
+    # Thumb
+    states.append(
+        1 if hand_landmarks.landmark[4].x >
+             hand_landmarks.landmark[3].x else 0
+    )
+
+    # Other fingers
+    for tip, base in zip(finger_tips[1:], finger_bases[1:]):
+        states.append(
+            1 if hand_landmarks.landmark[tip].y <
+                 hand_landmarks.landmark[base].y else 0
+        )
+
+    return states
+
+# ---------------- GESTURE MATCHING ----------------
+def match_gesture(finger_states):
+    for name, info in gesture_rules.items():
+        if finger_states == info["pattern"]:
+            return name
+    return "Unknown Gesture"
+
+# ---------------- GESTURE RECOGNIZER ----------------
+def recognize_gesture(frame):
+    rgb = cv2.cvtColor(frame, cv2.COLOR_BGR2RGB)
+    results = hands.process(rgb)
+
+    if results.multi_hand_landmarks:
+        for hand_landmarks in results.multi_hand_landmarks:
+            states = get_finger_states(hand_landmarks)
+            return match_gesture(states)
+
+    return "No Hand Detected"
+
+# ---------------- TEXT TO SPEECH ----------------
+def speak_text(text):
+    tts = gTTS(text=text, lang="en")
+    temp_file = tempfile.NamedTemporaryFile(delete=False, suffix=".mp3")
+    tts.save(temp_file.name)
+    return temp_file.name
+
+# ---------------- UI ----------------
+st.markdown(
+    "<h1 style='text-align:center;'>🤟 Hand2Voice</h1>"
+    "<p style='text-align:center;'>Hand Gesture to Voice Conversion</p>",
+    unsafe_allow_html=True
+)
+
+st.divider()
 
-# -------------------------------
-# Layout
-# -------------------------------
 col1, col2 = st.columns(2)
 
-# -------------------------------
-# Camera Section (Left)
-# -------------------------------
+# -------- CAMERA COLUMN --------
 with col1:
-    st.markdown("### 📷 Camera Input")
-    image = st.camera_input("Capture hand gesture")
+    st.subheader("📷 Camera Input")
+    image = st.camera_input("Capture your hand gesture")
 
-# -------------------------------
-# Text Section (Right)
-# -------------------------------
+# -------- OUTPUT COLUMN --------
 with col2:
-    st.markdown("### 📝 Recognized Text")
-    st.markdown(
-        "<h2>HELLO <span style='background-color:yellow;'>HOW</span> ARE YOU</h2>",
-        unsafe_allow_html=True
-    )
+    st.subheader("📝 Output")
+
+    if image is not None:
+        bytes_data = image.getvalue()
+        np_img = np.frombuffer(bytes_data, np.uint8)
+        frame = cv2.imdecode(np_img, cv2.IMREAD_COLOR)
+
+        gesture_text = recognize_gesture(frame)
+
+        # Highlight text
+        st.markdown(
+            f"""
+            <div style="
+                padding:20px;
+                background-color:#e0f2fe;
+                border-radius:12px;
+                font-size:24px;
+                font-weight:bold;
+                text-align:center;">
+                🔊 {gesture_text}
+            </div>
+            """,
+            unsafe_allow_html=True
+        )
 
-    st.markdown("### 🔊 Spoken Output")
-    st.info("Currently speaking: HOW")
+        # Speak text
+        if gesture_text not in ["No Hand Detected", "Unknown Gesture"]:
+            audio_path = speak_text(gesture_text)
+            st.audio(audio_path, format="audio/mp3")
 
-# -------------------------------
-# Status
-# -------------------------------
-st.markdown("---")
-st.success("Capture an image of your hand to begin")
+    else:
+        st.info("Show your hand and capture an image")