Upload 5 files

README_DOCKER.md

@@ -0,0 +1,127 @@
+# 🚗 Safe Driving Assistant — Docker Integration Guide
+
+Welcome to the premium Docker integration guide for your **Safe Driving Assistant**. This guide provides step-by-step instructions on how to package, configure, build, and run the assistant inside a fully isolated, headless-capable Docker container.
+
+The containerized environment includes all C/C++ libraries, graphics backends, sound drivers, and speech synthesizers, pre-configured with a virtual framebuffer (**Xvfb**) to prevent OpenCV graphical crashes.
+
+---
+
+## 🌟 Highlights of the Docker Configuration
+
+*   **Zero-Dependency Setup:** No need to install `dlib`, `cmake`, `opencv`, or `pyaudio` manually on your host machine.
+*   **Virtual Framebuffer (Xvfb):** Headless-safe. If no display is detected, it automatically spawns a virtual X11 server so that `cv2.imshow` calls do not crash.
+*   **Dynamic Network Routing:** Pre-configured to easily bridge to your local **Ollama** SLM instance.
+*   **Unified Sound Backend:** Installs PulseAudio, ALSA, and `espeak` dependencies required for speech synthesis and playback.
+*   **Environment-Driven Configuration:** Highly customizable via environment variables (`FLASK_HOST`, `CAMERA_ID`, `OLLAMA_API_URL`, etc.).
+
+---
+
+## 🛠️ Step 1: Build the Docker Image
+
+Open your terminal in the project directory (`Safe Driving Assistant`) and run the following command to build your custom co-pilot image:
+
+```bash
+docker build -t safe-driving-assistant:latest .
+```
+
+*This compilation will take several minutes during the first run because it builds `dlib` (facial recognition) and compiles native C extensions.*
+
+---
+
+## 🚀 Step 2: Run the Docker Container
+
+Depending on your host Operating System and hardware setup, select one of the premium run configurations below:
+
+### Option A: Standard Headless / Web-HUD Only (Recommended)
+This runs the assistant, serves the live futuristic Web-HUD on port `5000`, and bridges network queries to your host's local Ollama service.
+
+```bash
+docker run -d \
+  --name driving_assistant \
+  --add-host=host.docker.internal:host-gateway \
+  -e OLLAMA_API_URL="http://host.docker.internal:11434/api/generate" \
+  -p 5000:5000 \
+  safe-driving-assistant:latest
+```
+
+*   **Web Dashboard:** Once started, open your browser and navigate to **`http://localhost:5000`** to view the live dashboard!
+*   **How it works:** The container runs headlessly. Face recognition and eye-aspect ratio tracking are active, and the live stream is pushed directly to the dashboard.
+
+---
+
+### Option B: Linux Run with Full Hardware Passthrough (Camera & Audio)
+If you are running native Linux and want the container to access your physical USB webcam and audio hardware directly, run:
+
+```bash
+docker run -it \
+  --name driving_assistant \
+  --device /dev/video0:/dev/video0 \
+  --device /dev/snd:/dev/snd \
+  --add-host=host.docker.internal:host-gateway \
+  -e CAMERA_ID=0 \
+  -e OLLAMA_API_URL="http://host.docker.internal:11434/api/generate" \
+  -p 5000:5000 \
+  safe-driving-assistant:latest
+```
+
+*   **`--device /dev/video0`:** Mounts your primary webcam inside the container.
+*   **`--device /dev/snd`:** Exposes speaker and mic controls.
+
+---
+
+### Option C: Windows Host (WSL2) with Web USB Webcam Passthrough
+If you are using WSL2 on Windows and want to feed your physical webcam into the Docker container:
+
+1.  Bind your USB camera to WSL2 using [usbipd-win](https://github.com/dorssel/usbipd-win):
+    ```powershell
+    usbipd list
+    usbipd bind --busid <BUSID>
+    usbipd attach --wsl --busid <BUSID>
+    ```
+2.  Start the container with device mapping:
+    ```bash
+    docker run -it \
+      --name driving_assistant \
+      --device /dev/video0:/dev/video0 \
+      --add-host=host.docker.internal:host-gateway \
+      -e CAMERA_ID=0 \
+      -e OLLAMA_API_URL="http://host.docker.internal:11434/api/generate" \
+      -p 5000:5000 \
+      safe-driving-assistant:latest
+    ```
+
+---
+
+## ⚙️ Environment Variables Customization
+
+You can dynamically tune your assistant container at runtime by passing `-e KEY=VALUE` parameters to `docker run`:
+
+| Variable Name | Default Value | Description |
+| :--- | :--- | :--- |
+| `FLASK_HOST` | `0.0.0.0` | Binding IP address for Flask web dashboard. |
+| `FLASK_PORT` | `5000` | Port on which the web HUD dashboard will be served. |
+| `CAMERA_ID` | `0` | Camera index. |
+| `OLLAMA_API_URL` | `http://localhost:11434/api/generate` | The API endpoint for the Ollama voice co-pilot. Use `http://host.docker.internal:11434/api/generate` for bridging to host. |
+| `OLLAMA_MODEL` | `drivesafe` | Name of the custom SLM model configured inside Ollama. |
+| `FRAME_WIDTH` | `640` | Video frame capture width. |
+| `FRAME_HEIGHT` | `480` | Video frame capture height. |
+
+---
+
+## 🧼 Housekeeping & Diagnostics
+
+### View Real-Time Logs
+To see the system warnings, detected EAR, and conversational chatbot interactions:
+```bash
+docker logs -f driving_assistant
+```
+
+### Stop & Remove Container
+To stop and clean up the assistant container instance:
+```bash
+docker stop driving_assistant
+docker rm driving_assistant
+```
+
+---
+**Safe Driving Assistant** — *Keep your eyes on the road, your hands on the wheel, and drive safely!* 🚗💨

app.py

@@ -27,6 +27,7 @@ import re
 from dotenv import load_dotenv
 import urllib.parse
 import urllib.request
+import subprocess
 
 load_dotenv()
 
@@ -77,6 +78,25 @@ def index():
 def api_status():
     return jsonify(system_status)
 
+@app.route('/api/music_url', methods=['GET'])
+def api_music_url():
+    video_id = request.args.get('video_id')
+    if not video_id:
+        return jsonify({"error": "No video_id provided"}), 400
+    try:
+        cmd = ["yt-dlp", "-J", "-f", "bestaudio", f"https://www.youtube.com/watch?v={video_id}"]
+        result = subprocess.run(cmd, capture_output=True, text=True, check=True)
+        info = json.loads(result.stdout)
+        return jsonify({
+            "url": info.get("url"),
+            "title": info.get("title"),
+            "artist": info.get("uploader"),
+            "thumbnail": info.get("thumbnail")
+        })
+    except Exception as e:
+        print(f"[yt-dlp error] {repr(e)}")
+        return jsonify({"error": str(e)}), 500
+
 @app.route('/api/frame', methods=['POST'])
 def api_frame():
     global start_closed_time, last_announced_state
@@ -195,13 +215,12 @@ def api_voice():
                 html = urllib.request.urlopen(f'https://www.youtube.com/results?search_query={query}').read().decode()
                 video_ids = re.findall(r'watch\?v=(\S{11})', html)
                 if video_ids:
-                    url = f'https://music.youtube.com/watch?v={video_ids[0]}'
-                    return jsonify({"speak": f"Playing {stripped} on YouTube Music to keep you alert.", "action": "open_url", "url": url})
+                    return jsonify({"speak": f"Playing {stripped} to keep you alert.", "action": "play_native", "video_ids": video_ids[:20], "title": stripped})
                 else:
                     return jsonify({"speak": "I couldn't find that song."})
             except Exception as e:
                 print(f"[YouTube Search Error] {repr(e)}")
-                return jsonify({"speak": "I had trouble searching for the song on YouTube. It might be blocking automated requests."})
+                return jsonify({"speak": "I had trouble searching for the song on YouTube Music. It might be blocking automated requests."})
         return jsonify({})
 
     # Normal wake-word logic
@@ -212,9 +231,9 @@ def api_voice():
             active_listening = False
             return jsonify({"speak": "Goodbye."})
     else:
-        if 'eliot' not in text:
+        if 'lara' not in text:
             return jsonify({})
-        idx = text.find('eliot') + len('eliot')
+        idx = text.find('lara') + len('lara')
         command = re.sub(r'[^\w\s]', '', text[idx:]).strip()
 
     if len(command) < 3:
@@ -223,26 +242,43 @@ def api_voice():
 
     print(f'[Voice API] Command: "{command}"')
 
-    # Media Commands (Cloud servers can't press media keys, so we handle it gracefully)
-    is_media_control = command.lower() in ['pause', 'stop', 'play', 'resume']
-    if is_media_control:
-        return jsonify({"speak": "Media controls are not supported in the web browser."})
+    # Media Commands (Native Player Control)
+    cmd = command.lower()
+    if 'pause' in cmd:
+        return jsonify({"speak": "Pausing.", "action": "pause_native"})
+    if 'resume' in cmd or cmd == 'play':
+        return jsonify({"speak": "Resuming.", "action": "resume_native"})
+    if 'stop' in cmd:
+        return jsonify({"speak": "Stopping.", "action": "stop_native"})
+    if 'next' in cmd or 'skip' in cmd:
+        return jsonify({"speak": "Skipping.", "action": "next_native"})
+    if 'previous' in cmd or 'back' in cmd:
+        return jsonify({"speak": "Going back.", "action": "prev_native"})
+    if 'volume up' in cmd or 'louder' in cmd or 'increase volume' in cmd:
+        return jsonify({"speak": "Volume up.", "action": "vol_up_native"})
+    if 'volume down' in cmd or 'quieter' in cmd or 'decrease volume' in cmd:
+        return jsonify({"speak": "Volume down.", "action": "vol_down_native"})
+    if 'unmute' in cmd:
+        return jsonify({"speak": "Unmuting.", "action": "unmute_native"})
+    elif 'mute' in cmd:
+        return jsonify({"speak": "Muting.", "action": "mute_native"})
 
     if command.lower().startswith('play '):
         song = command[5:].strip()
+        if song.lower() in ['music', 'song', 'the music', 'the song', 'it']:
+            return jsonify({"speak": "Resuming.", "action": "resume_native"})
         if song:
             try:
                 query = urllib.parse.quote_plus(song)
                 html = urllib.request.urlopen(f'https://www.youtube.com/results?search_query={query}').read().decode()
                 video_ids = re.findall(r'watch\?v=(\S{11})', html)
                 if video_ids:
-                    url = f'https://music.youtube.com/watch?v={video_ids[0]}'
-                    return jsonify({"speak": f"Playing {song} on YouTube Music.", "action": "open_url", "url": url})
+                    return jsonify({"speak": f"Playing {song}.", "action": "play_native", "video_ids": video_ids[:20], "title": song})
                 else:
                     return jsonify({"speak": "I couldn't find that song."})
             except Exception as e:
                 print(f"[YouTube Search Error] {repr(e)}")
-                return jsonify({"speak": "I had trouble searching for the song on YouTube. It might be blocking automated requests."})
+                return jsonify({"speak": "I had trouble searching for the song on YouTube Music. It might be blocking automated requests."})
 
     # LLM Interaction
     api_key = os.getenv('API_KEY', '')
@@ -252,9 +288,11 @@ def api_voice():
     try:
         llm = OpenAI(api_key=api_key, base_url='https://integrate.api.nvidia.com/v1')
         SYS_PROMPT = (
-            'You are eliot, a helpful voice assistant. '
-            'Reply in plain text only — no markdown, no bullet points. '
-            'Be very concise: 1-2 sentences maximum.'
+            'You are Lara, a voice assistant and copilot for a driver. '
+            'You MUST ONLY assist with driving, road-related questions, directions, routes, and playing music. '
+            'If the user asks about ANYTHING else, politely refuse to answer and state that you can only help with driving and music. '
+            'Provide helpful, concise, and clear answers so the driver can stay focused on the road. '
+            'Keep your responses brief (1-2 sentences max) and do not use conversational filler.'
         )
         resp = llm.chat.completions.create(
             model='meta/llama-3.1-8b-instruct',
@@ -288,7 +326,13 @@ def api_chat():
         
     try:
         client = OpenAI(api_key=api_key, base_url="https://integrate.api.nvidia.com/v1")
-        sys_prompt = "You are a helpful assistant. Answer clearly and concisely."
+        sys_prompt = (
+            'You are Lara, a voice assistant and copilot for a driver. '
+            'You MUST ONLY assist with driving, road-related questions, directions, routes, and playing music. '
+            'If the user asks about ANYTHING else, politely refuse to answer and state that you can only help with driving and music. '
+            'Provide helpful, concise, and clear answers so the driver can stay focused on the road. '
+            'Keep your responses brief (1-2 sentences max) and do not use conversational filler.'
+        )
         completion = client.chat.completions.create(
             model='meta/llama-3.1-8b-instruct',
             max_tokens=300,

templates/index.html

@@ -466,6 +466,100 @@
                 justify-content: center;
             }
         }
+        
+        /* Modern Media Player */
+        .modern-media-player {
+            background: #2a2a2a;
+            border-radius: 24px;
+            padding: 10px 14px;
+            display: flex;
+            align-items: center;
+            gap: 14px;
+            color: white;
+            position: relative;
+            margin-top: 1.5rem;
+            box-shadow: 0 10px 30px rgba(0,0,0,0.5);
+            font-family: -apple-system, BlinkMacSystemFont, "Segoe UI", Roboto, Helvetica, Arial, sans-serif;
+        }
+        .mmp-art {
+            width: 50px;
+            height: 50px;
+            border-radius: 12px;
+            overflow: hidden;
+            flex-shrink: 0;
+            box-shadow: 0 4px 10px rgba(0,0,0,0.3);
+        }
+        .mmp-art img {
+            width: 100%;
+            height: 100%;
+            object-fit: cover;
+        }
+        .mmp-info {
+            flex-grow: 1;
+            display: flex;
+            flex-direction: column;
+            justify-content: center;
+            min-width: 0;
+        }
+        .mmp-source {
+            font-size: 0.7rem;
+            color: #b3b3b3;
+            display: flex;
+            align-items: center;
+            gap: 5px;
+            margin-bottom: 2px;
+        }
+        .mmp-title {
+            font-size: 0.95rem;
+            font-weight: 600;
+            white-space: nowrap;
+            overflow: hidden;
+            text-overflow: ellipsis;
+            margin-bottom: 1px;
+            color: #ffffff;
+        }
+        .mmp-artist {
+            font-size: 0.8rem;
+            color: #b3b3b3;
+            white-space: nowrap;
+            overflow: hidden;
+            text-overflow: ellipsis;
+        }
+        .mmp-controls {
+            display: flex;
+            align-items: center;
+            gap: 12px;
+            margin-right: 18px;
+        }
+        .mmp-btn {
+            background: none;
+            border: none;
+            color: white;
+            font-size: 1.1rem;
+            cursor: pointer;
+            display: flex;
+            align-items: center;
+            justify-content: center;
+            padding: 4px;
+            transition: transform 0.1s;
+        }
+        .mmp-btn:active {
+            transform: scale(0.9);
+        }
+        .mmp-btn.play-pause {
+            font-size: 1.4rem;
+        }
+        .mmp-btn:hover {
+            color: #e0e0e0;
+        }
+        .mmp-expand {
+            position: absolute;
+            top: 14px;
+            right: 14px;
+            color: #b3b3b3;
+            font-size: 0.85rem;
+            cursor: pointer;
+        }
     </style>
 </head>
 
@@ -499,6 +593,28 @@
                     <strong id="drowsyEvents" style="color: var(--text-main);">0</strong>
                 </div>
             </div>
+
+            <!-- MODERN MEDIA PLAYER -->
+            <div class="modern-media-player" id="custom-media-player" style="display:none;">
+                <div class="mmp-art">
+                    <img src="https://via.placeholder.com/60" alt="Album Art" id="mmp-thumbnail">
+                </div>
+                <div class="mmp-info">
+                    <div class="mmp-source">
+                        <i class="fa-regular fa-circle-play"></i> This copilot
+                    </div>
+                    <div class="mmp-title" id="mmp-title">Unknown Song</div>
+                    <div class="mmp-artist" id="mmp-artist">Unknown Artist</div>
+                </div>
+                <div class="mmp-controls">
+                    <button class="mmp-btn" id="mmp-prev" onclick="if(ytPlayer && typeof ytPlayer.previousVideo === 'function') ytPlayer.previousVideo()"><i class="fa-solid fa-backward-step"></i></button>
+                    <button class="mmp-btn play-pause" id="mmp-play-pause" onclick="toggleMmpPlay()"><i class="fa-solid fa-pause"></i></button>
+                    <button class="mmp-btn" id="mmp-next" onclick="if(ytPlayer && typeof ytPlayer.nextVideo === 'function') ytPlayer.nextVideo()"><i class="fa-solid fa-forward-step"></i></button>
+                </div>
+                <div class="mmp-expand">
+                    <i class="fa-solid fa-chevron-down"></i>
+                </div>
+            </div>
         </aside>
 
         <!-- Main Content -->
@@ -520,6 +636,11 @@
                 </div>
             </div>
 
+
+
+            <!-- NATIVE AUDIO PLAYER -->
+            <audio id="native-audio" style="display:none;"></audio>
+
             <!-- Chat Tab -->
             <div class="tab-content" id="chat-tab">
                 <div class="chat-area" id="chatHistory">
@@ -540,6 +661,7 @@
         </main>
     </div>
 
+    <!-- No external iframe scripts needed -->
     <script>
         // Tab Navigation
         const tabBtns = document.querySelectorAll('.tab-btn');
@@ -584,6 +706,71 @@
         let isHardwareActive = false;
         let isSpeaking = false;
 
+
+        // --- NATIVE AUDIO API ---
+        const nativeAudio = document.getElementById('native-audio');
+        let isMmpPlaying = false;
+
+        nativeAudio.addEventListener('play', () => {
+            document.getElementById('mmp-play-pause').innerHTML = '<i class="fa-solid fa-pause"></i>';
+            isMmpPlaying = true;
+        });
+
+        nativeAudio.addEventListener('pause', () => {
+            document.getElementById('mmp-play-pause').innerHTML = '<i class="fa-solid fa-play"></i>';
+            isMmpPlaying = false;
+        });
+
+        nativeAudio.addEventListener('ended', () => {
+            document.getElementById('mmp-play-pause').innerHTML = '<i class="fa-solid fa-play"></i>';
+            isMmpPlaying = false;
+            if (currentQueue.length > currentQueueIndex + 1) {
+                currentQueueIndex++;
+                playNativeAudio(currentQueue[currentQueueIndex], "Up Next...");
+            }
+        });
+
+        let currentQueue = [];
+        let currentQueueIndex = 0;
+
+        function toggleMmpPlay() {
+            if (isMmpPlaying) {
+                nativeAudio.pause();
+            } else {
+                nativeAudio.play();
+            }
+        }
+
+        async function playNativeAudio(videoId, titleGuess) {
+            document.getElementById('custom-media-player').style.display = 'flex';
+            document.getElementById('mmp-thumbnail').src = `https://img.youtube.com/vi/${videoId}/hqdefault.jpg`;
+            document.getElementById('mmp-title').innerText = "Extracting Stream...";
+            document.getElementById('mmp-artist').innerText = titleGuess || "Loading";
+            document.getElementById('mmp-play-pause').innerHTML = '<i class="fa-solid fa-spinner fa-spin"></i>';
+            
+            try {
+                const res = await fetch(`/api/music_url?video_id=${videoId}`);
+                const data = await res.json();
+                if (data.url) {
+                    nativeAudio.src = data.url;
+                    nativeAudio.volume = 1.0;
+                    nativeAudio.play();
+                    document.getElementById('mmp-title').innerText = data.title || titleGuess;
+                    document.getElementById('mmp-artist').innerText = data.artist || "yt-dlp Engine";
+                    if (data.thumbnail) {
+                        document.getElementById('mmp-thumbnail').src = data.thumbnail;
+                    }
+                } else {
+                    document.getElementById('mmp-title').innerText = "Stream Failed";
+                    document.getElementById('mmp-play-pause').innerHTML = '<i class="fa-solid fa-xmark"></i>';
+                }
+            } catch (err) {
+                console.error("Audio fetch error:", err);
+                document.getElementById('mmp-title').innerText = "Network Error";
+                document.getElementById('mmp-play-pause').innerHTML = '<i class="fa-solid fa-xmark"></i>';
+            }
+        }
+
         // TTS Engine
         function speakOutLoud(text) {
             if (!text) return;
@@ -637,11 +824,45 @@
                     if (data.speak) {
                         speakOutLoud(data.speak);
                     }
-                    if (data.action === "open_url" && data.url) {
+                    if (data.action === "play_native" && data.video_ids && data.video_ids.length > 0) {
+                        currentQueue = data.video_ids;
+                        currentQueueIndex = 0;
+                        playNativeAudio(currentQueue[currentQueueIndex], data.title);
+                    } else if (data.action === "pause_native") {
+                        nativeAudio.pause();
+                    } else if (data.action === "resume_native") {
+                        nativeAudio.play();
+                    } else if (data.action === "stop_native") {
+                        nativeAudio.pause();
+                        nativeAudio.currentTime = 0;
+                        document.getElementById('custom-media-player').style.display = 'none';
+                    } else if (data.action === "next_native") {
+                        if (currentQueue.length > currentQueueIndex + 1) {
+                            currentQueueIndex++;
+                            playNativeAudio(currentQueue[currentQueueIndex], "Up Next...");
+                        } else {
+                            appendMessage('assistant', "End of queue reached.");
+                        }
+                    } else if (data.action === "prev_native") {
+                        if (currentQueueIndex > 0) {
+                            currentQueueIndex--;
+                            playNativeAudio(currentQueue[currentQueueIndex], "Going Back...");
+                        } else {
+                            appendMessage('assistant', "No previous song in queue.");
+                        }
+                    } else if (data.action === "vol_up_native") {
+                        nativeAudio.volume = Math.min(1.0, nativeAudio.volume + 0.2);
+                    } else if (data.action === "vol_down_native") {
+                        nativeAudio.volume = Math.max(0.0, nativeAudio.volume - 0.2);
+                    } else if (data.action === "mute_native") {
+                        nativeAudio.muted = true;
+                    } else if (data.action === "unmute_native") {
+                        nativeAudio.muted = false;
+                    } else if (data.action === "open_url" && data.url) {
                         let newWin = window.open(data.url, "_blank");
                         if (!newWin || newWin.closed || typeof newWin.closed == 'undefined') {
                             // Popup blocked
-                            appendMessage('system', `⚠️ **Popup Blocked**<br>I tried to open the song, but your browser blocked it.<br><a href="${data.url}" target="_blank" style="color:var(--primary); font-weight:bold; font-size:1.1rem;">👉 Click Here to Open YouTube Music</a><br>*(To make this automatic, click the popup-blocker icon in your browser's address bar and select "Always allow popups from this site")*`);
+                            appendMessage('system', `⚠️ **Popup Blocked**<br>I tried to open the song, but your browser blocked it.<br><a href="${data.url}" target="_blank" style="color:var(--primary); font-weight:bold; font-size:1.1rem;">👉 Click Here to Open YouTube Music</a>`);
                         }
                     }
                     if (data.text) {