Upload Linny-Web-Server.py

Linny-Web-Server.py

@@ -0,0 +1,501 @@
+import os
+import sys
+import time
+import uuid
+import torch
+import torch.nn as nn
+from flask import Flask, request, Response, stream_with_context, jsonify
+from werkzeug.utils import secure_filename
+
+# ==========================================
+# ⚙️ GLOBAL CONFIGURATION
+# ==========================================
+ADMIN_PASSWORD = "admin123" # Change this to a secure password
+UPLOAD_FOLDER = "./models"
+DEFAULT_CHECKPOINT = "/Users/lucienkachadoorian/Downloads/Linny_Speed_Chatbot-Pro-custom-Gen1.pt" # App tries to load this on startup if it exists
+
+# Constants for Generation
+USER_TAG = "### Instruction:"
+BOT_TAG = "### Response:"
+EOS_TOKEN = "<|end|>"
+
+os.makedirs(UPLOAD_FOLDER, exist_ok=True)
+
+# Hardware Setup
+if torch.backends.mps.is_available():
+    device = torch.device("mps")
+    print("🚀 Powered by: Apple Metal (MPS)")
+elif torch.cuda.is_available():
+    device = torch.device("cuda")
+    print("🚀 Powered by: NVIDIA CUDA")
+else:
+    device = torch.device("cpu")
+    print("⚠️ GPU not available, using CPU.")
+
+# ==========================================
+# 🧠 MODEL ARCHITECTURE (From your code)
+# ==========================================
+class LSTMCharLM(nn.Module):
+    def __init__(self, vocab_size, embed_size, hidden_size, num_layers, dropout=0.2):
+        super().__init__()
+        self.embed = nn.Embedding(vocab_size, embed_size)
+        self.lstm = nn.LSTM(embed_size, hidden_size, num_layers=num_layers, 
+                            batch_first=True, dropout=dropout)
+        self.fc = nn.Linear(hidden_size, vocab_size)
+
+    def forward(self, x, hidden=None):
+        e = self.embed(x)
+        out, hidden = self.lstm(e, hidden)
+        logits = self.fc(out)
+        return logits, hidden
+
+# ==========================================
+# 🛠️ INFERENCE ENGINE
+# ==========================================
+class LinnyChat:
+    def __init__(self, checkpoint_path, embed_size, neurons, hidden_layers, dropout=0.2):
+        print(f"🧠 Loading Linny's brain from: {checkpoint_path}")
+        try:
+            ckpt = torch.load(checkpoint_path, map_location=device)
+            self.chars = ckpt['chars']
+            self.stoi = {ch: i for i, ch in enumerate(self.chars)}
+            self.itos = {i: ch for i, ch in enumerate(self.chars)}
+            self.vocab_size = len(self.chars)
+
+            self.model = LSTMCharLM(self.vocab_size, embed_size, neurons, hidden_layers, dropout).to(device)
+            self.model.load_state_dict(ckpt['model_state'])
+            self.model.eval()
+            self.ready = True
+            print("✅ Model Online and ready.")
+        except Exception as e:
+            print(f"❌ Failed to load model: {e}")
+            self.ready = False
+
+    def stream_generate(self, prompt, temperature=0.7, max_len=1575, penalty=1.2, top_p=0.9, top_k=50):
+        if not self.ready:
+            yield "Model not properly loaded."
+            return
+
+        formatted = f"{USER_TAG}\n{prompt}\n\n{BOT_TAG}\n"
+        input_ids = [self.stoi.get(c, 0) for c in formatted]
+        input_tensor = torch.tensor([input_ids], dtype=torch.long).to(device)
+        
+        hidden = None
+        generated_chars = ""
+        
+        with torch.no_grad():
+            _, hidden = self.model(input_tensor, hidden)
+            input_token = input_tensor[:, -1:]
+            
+            for _ in range(max_len):
+                logits, hidden = self.model(input_token, hidden)
+                logits = logits[0, -1] / temperature
+                
+                # Repetition Penalty
+                recent_chars = generated_chars[-30:]
+                for char in set(recent_chars):
+                    char_idx = self.stoi.get(char, 0)
+                    if logits[char_idx] > 0:
+                        logits[char_idx] /= penalty
+                    else:
+                        logits[char_idx] *= penalty
+
+                # Top-K
+                if top_k > 0:
+                    top_k_values, _ = torch.topk(logits, min(top_k, len(logits)))
+                    logits[logits < top_k_values[-1]] = float('-inf')
+                
+                # Top-P
+                if top_p < 1.0:
+                    sorted_logits, sorted_indices = torch.sort(logits, descending=True)
+                    cumulative_probs = torch.cumsum(torch.softmax(sorted_logits, dim=-1), dim=-1)
+                    sorted_indices_to_remove = cumulative_probs > top_p
+                    sorted_indices_to_remove[..., 1:] = sorted_indices_to_remove[..., :-1].clone()
+                    sorted_indices_to_remove[..., 0] = 0
+                    indices_to_remove = sorted_indices[sorted_indices_to_remove]
+                    logits[indices_to_remove] = float('-inf')
+                
+                probs = torch.softmax(logits, dim=0)
+                idx = torch.multinomial(probs, 1).item()
+                char = self.itos[idx]
+                
+                if char == EOS_TOKEN:
+                    break
+                
+                if char == "#" and len(generated_chars) >= 2:
+                    if generated_chars[-1] == "#" and generated_chars[-2] == "#":
+                        break # Prevent "###" bleeding
+                
+                yield char
+                generated_chars += char
+                input_token = torch.tensor([[idx]], dtype=torch.long).to(device)
+
+
+# ==========================================
+# 🌐 FLASK WEB SERVER
+# ==========================================
+app = Flask(__name__)
+app.config['MAX_CONTENT_LENGTH'] = 1024 * 1024 * 1024 # 1GB max upload
+
+# Dictionary to hold active models in memory mapping model_id -> LinnyChat instance
+active_models = {}
+
+# Try to load the default model if it exists
+if os.path.exists(os.path.join(UPLOAD_FOLDER, DEFAULT_CHECKPOINT)):
+    active_models["default"] = LinnyChat(os.path.join(UPLOAD_FOLDER, DEFAULT_CHECKPOINT), 384, 768, 5, 0.2)
+else:
+    active_models["default"] = None
+    print("⚠️ No default model found. Admin needs to upload one.")
+
+
+HTML_TEMPLATE = """
+<!DOCTYPE html>
+<html lang="en">
+<head>
+    <meta charset="UTF-8">
+    <meta name="viewport" content="width=device-width, initial-scale=1.0">
+    <title>Linny AI</title>
+    <script src="https://cdn.tailwindcss.com"></script>
+    <style>
+        body { background-color: #0f172a; color: #f8fafc; font-family: -apple-system, BlinkMacSystemFont, "Segoe UI", Roboto, sans-serif; }
+        .chat-container { height: calc(100vh - 160px); overflow-y: auto; scroll-behavior: smooth; }
+        .message { max-width: 85%; line-height: 1.6; }
+        .msg-user { background-color: #3b82f6; border-radius: 1rem 1rem 0 1rem; margin-left: auto; }
+        .msg-bot { background-color: #1e293b; border-radius: 1rem 1rem 1rem 0; margin-right: auto; }
+        
+        /* The Magic Reasoning Block Styles */
+        .reasoning-block {
+            background-color: #020617;
+            color: #94a3b8;
+            border-left: 4px solid #475569;
+            padding: 12px 16px;
+            margin: 12px 0;
+            border-radius: 0 8px 8px 0;
+            font-size: 0.9em;
+            font-family: ui-monospace, SFMono-Regular, Menlo, Monaco, Consolas, monospace;
+            white-space: pre-wrap;
+        }
+        .reasoning-header { font-weight: bold; margin-bottom: 8px; color: #cbd5e1; display: flex; align-items: center; gap: 6px; }
+        
+        ::-webkit-scrollbar { width: 8px; }
+        ::-webkit-scrollbar-track { background: #0f172a; }
+        ::-webkit-scrollbar-thumb { background: #334155; border-radius: 4px; }
+        .loader { border: 2px solid #334155; border-top: 2px solid #3b82f6; border-radius: 50%; width: 14px; height: 14px; animation: spin 1s linear infinite; display: inline-block; }
+        @keyframes spin { 0% { transform: rotate(0deg); } 100% { transform: rotate(360deg); } }
+    </style>
+</head>
+<body class="flex flex-col h-screen">
+
+    <!-- Navbar -->
+    <header class="bg-slate-900 border-b border-slate-800 p-4 flex justify-between items-center shadow-md">
+        <div class="flex items-center gap-3">
+            <div class="w-8 h-8 rounded-full bg-blue-500 flex items-center justify-center font-bold text-white shadow-lg shadow-blue-500/50">L</div>
+            <h1 class="font-bold text-xl tracking-wide">Linny AI</h1>
+            <span id="activeModelBadge" class="ml-2 text-xs bg-slate-800 px-2 py-1 rounded text-slate-400">Model: Default</span>
+        </div>
+        <div class="flex gap-2">
+            <button onclick="openModal('uploadModal')" class="bg-slate-800 hover:bg-slate-700 text-sm px-4 py-2 rounded-lg transition-colors font-medium text-slate-200 shadow">Upload Model</button>
+            <button onclick="openModal('adminModal')" class="text-slate-400 hover:text-white px-3 py-2 text-sm transition-colors">Admin</button>
+        </div>
+    </header>
+
+    <!-- Chat Area -->
+    <main class="flex-1 max-w-4xl w-full mx-auto p-4 flex flex-col relative w-full">
+        <div id="chatbox" class="chat-container w-full flex flex-col gap-6 p-2 mb-4">
+            <div class="text-center text-slate-500 mt-10">
+                <div class="w-16 h-16 rounded-full bg-slate-800 flex items-center justify-center mx-auto mb-4 text-2xl">👋</div>
+                <h2 class="text-xl font-medium text-slate-300">Hello, I'm Linny.</h2>
+                <p class="text-sm mt-2">I am ready to chat. Type a message below.</p>
+            </div>
+        </div>
+        
+        <!-- Input Area -->
+        <div class="bg-slate-800 rounded-2xl p-2 flex items-end gap-2 shadow-xl border border-slate-700 shrink-0">
+            <textarea id="prompt" class="bg-transparent text-white w-full max-h-48 resize-none outline-none p-3 placeholder-slate-400" 
+                rows="1" placeholder="Message Linny..." oninput="autoGrow(this)" onkeydown="checkEnter(event)"></textarea>
+            <button id="sendBtn" onclick="sendMessage()" class="bg-blue-600 hover:bg-blue-500 text-white p-3 rounded-xl mb-1 transition-colors shadow-lg shadow-blue-600/30">
+                <svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round"><line x1="22" y1="2" x2="11" y2="13"></line><polygon points="22 2 15 22 11 13 2 9 22 2"></polygon></svg>
+            </button>
+        </div>
+    </main>
+
+    <!-- Modals Background -->
+    <div id="modalBackdrop" class="fixed inset-0 bg-black/60 backdrop-blur-sm hidden flex justify-center items-center z-50 transition-opacity">
+        
+        <!-- Upload Modal -->
+        <div id="uploadModal" class="hidden bg-slate-900 border border-slate-700 p-6 rounded-2xl max-w-md w-full shadow-2xl">
+            <h2 class="text-xl font-bold mb-4 flex items-center gap-2"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round"><path d="M21 15v4a2 2 0 0 1-2 2H5a2 2 0 0 1-2-2v-4"></path><polyline points="17 8 12 3 7 8"></polyline><line x1="12" y1="3" x2="12" y2="15"></line></svg> Upload .pt Model</h2>
+            <p class="text-sm text-slate-400 mb-4">Configure your architecture settings to match your training exactly.</p>
+            
+            <input type="file" id="modelFile" accept=".pt" class="block w-full text-sm text-slate-400 file:mr-4 file:py-2 file:px-4 file:rounded-full file:border-0 file:text-sm file:font-semibold file:bg-blue-600 file:text-white hover:file:bg-blue-500 mb-4 bg-slate-800 rounded-lg p-2"/>
+            
+            <div class="grid grid-cols-2 gap-4 mb-4">
+                <div>
+                    <label class="text-xs text-slate-400 font-bold uppercase block mb-1">Hidden Layers</label>
+                    <input type="number" id="cfgLayers" value="5" class="w-full bg-slate-800 border border-slate-700 rounded p-2 text-white">
+                </div>
+                <div>
+                    <label class="text-xs text-slate-400 font-bold uppercase block mb-1">Neurons</label>
+                    <input type="number" id="cfgNeurons" value="768" class="w-full bg-slate-800 border border-slate-700 rounded p-2 text-white">
+                </div>
+                <div>
+                    <label class="text-xs text-slate-400 font-bold uppercase block mb-1">Embed Size</label>
+                    <input type="number" id="cfgEmbed" value="384" class="w-full bg-slate-800 border border-slate-700 rounded p-2 text-white">
+                </div>
+                <div>
+                    <label class="text-xs text-slate-400 font-bold uppercase block mb-1">Dropout</label>
+                    <input type="number" step="0.1" id="cfgDropout" value="0.2" class="w-full bg-slate-800 border border-slate-700 rounded p-2 text-white">
+                </div>
+            </div>
+            
+            <div class="flex gap-2 justify-end mt-6">
+                <button onclick="closeModals()" class="px-4 py-2 rounded text-slate-400 hover:text-white">Cancel</button>
+                <button onclick="uploadModel()" id="uploadBtn" class="px-4 py-2 bg-blue-600 hover:bg-blue-500 rounded text-white shadow font-medium">Load & Use Model</button>
+            </div>
+        </div>
+
+        <!-- Admin Modal -->
+        <div id="adminModal" class="hidden bg-slate-900 border border-slate-700 p-6 rounded-2xl max-w-sm w-full shadow-2xl">
+            <h2 class="text-xl font-bold mb-4 text-red-400">Admin Control</h2>
+            <p class="text-sm text-slate-400 mb-4">Set your currently loaded model as the global default for all new visitors.</p>
+            <input type="password" id="adminPass" placeholder="Admin Password" class="w-full bg-slate-800 border border-slate-700 rounded p-2 text-white mb-4">
+            <div class="flex gap-2 justify-end">
+                <button onclick="closeModals()" class="px-4 py-2 rounded text-slate-400 hover:text-white">Cancel</button>
+                <button onclick="setAsDefault()" class="px-4 py-2 bg-red-600 hover:bg-red-500 rounded text-white shadow font-medium">Set Global Default</button>
+            </div>
+        </div>
+    </div>
+
+    <script>
+        let currentModelId = "default";
+        
+        function autoGrow(element) {
+            element.style.height = "5px";
+            element.style.height = (element.scrollHeight) + "px";
+        }
+
+        function checkEnter(e) {
+            if (e.key === 'Enter' && !e.shiftKey) {
+                e.preventDefault();
+                sendMessage();
+            }
+        }
+
+        function openModal(id) {
+            document.getElementById('modalBackdrop').classList.remove('hidden');
+            document.getElementById('uploadModal').classList.add('hidden');
+            document.getElementById('adminModal').classList.add('hidden');
+            document.getElementById(id).classList.remove('hidden');
+        }
+
+        function closeModals() {
+            document.getElementById('modalBackdrop').classList.add('hidden');
+        }
+
+        async function uploadModel() {
+            const fileInput = document.getElementById('modelFile');
+            if(!fileInput.files.length) return alert("Please select a .pt file");
+
+            const btn = document.getElementById('uploadBtn');
+            btn.innerHTML = `<span class="loader mr-2"></span> Loading...`;
+            btn.disabled = true;
+
+            const formData = new FormData();
+            formData.append('file', fileInput.files[0]);
+            formData.append('layers', document.getElementById('cfgLayers').value);
+            formData.append('neurons', document.getElementById('cfgNeurons').value);
+            formData.append('embed', document.getElementById('cfgEmbed').value);
+            formData.append('dropout', document.getElementById('cfgDropout').value);
+
+            try {
+                const res = await fetch('/api/upload', { method: 'POST', body: formData });
+                const data = await res.json();
+                if(data.success) {
+                    currentModelId = data.model_id;
+                    document.getElementById('activeModelBadge').innerText = "Model: Custom Session";
+                    document.getElementById('activeModelBadge').classList.replace('text-slate-400', 'text-blue-400');
+                    closeModals();
+                    alert("Model successfully loaded into memory!");
+                } else {
+                    alert("Error loading model: " + data.error);
+                }
+            } catch(e) {
+                alert("Upload failed.");
+            }
+            btn.innerHTML = "Load & Use Model";
+            btn.disabled = false;
+        }
+
+        async function setAsDefault() {
+            const pass = document.getElementById('adminPass').value;
+            const res = await fetch('/api/set_default', {
+                method: 'POST',
+                headers: {'Content-Type': 'application/json'},
+                body: JSON.stringify({ password: pass, model_id: currentModelId })
+            });
+            const data = await res.json();
+            if(data.success) {
+                alert("Global default successfully updated!");
+                closeModals();
+            } else {
+                alert("Failed: " + data.error);
+            }
+        }
+
+        // --- Core Chat & Formatting Logic ---
+        
+        function formatMessage(text) {
+            // Escape HTML safely
+            let safeText = text.replace(/</g, "&lt;").replace(/>/g, "&gt;");
+            
+            // Dynamic parsing of <think> tags for UI
+            let html = safeText
+                .replace(/&lt;think&gt;/g, '<div class="reasoning-block"><div class="reasoning-header"><svg xmlns="http://www.w3.org/2000/svg" width="16" height="16" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round"><circle cx="12" cy="12" r="10"></circle><path d="M9.09 9a3 3 0 0 1 5.83 1c0 2-3 3-3 3"></path><line x1="12" y1="17" x2="12.01" y2="17"></line></svg> Reasoning Process</div>')
+                .replace(/&lt;\/think&gt;/g, '</div>');
+            
+            // Auto-close unclosed think tags during streaming
+            let openTags = (html.match(/<div class="reasoning-block">/g) || []).length;
+            let closeTags = (html.match(/<\/div>/g) || []).length; // (Simplified, relies on structure)
+            if (openTags > closeTags) {
+                html += '</div>';
+            }
+            
+            // Fix formatting for line breaks
+            return html;
+        }
+
+        async function sendMessage() {
+            const promptEl = document.getElementById('prompt');
+            const promptText = promptEl.value.trim();
+            if(!promptText) return;
+
+            promptEl.value = "";
+            promptEl.style.height = "auto";
+            
+            const chatbox = document.getElementById('chatbox');
+            
+            // Remove welcome message if exists
+            const textCenter = chatbox.querySelector('.text-center.mt-10');
+            if(textCenter) textCenter.remove();
+
+            // Append User Message
+            chatbox.innerHTML += `<div class="message msg-user p-4 shadow"><p class="whitespace-pre-wrap">${promptText.replace(/</g,"&lt;")}</p></div>`;
+            
+            // Create Bot Message container
+            const botId = 'bot-' + Date.now();
+            chatbox.innerHTML += `<div class="message msg-bot p-4 shadow border border-slate-700" id="${botId}"><span class="loader"></span></div>`;
+            chatbox.scrollTop = chatbox.scrollHeight;
+
+            try {
+                const response = await fetch('/api/chat', {
+                    method: 'POST',
+                    headers: { 'Content-Type': 'application/json' },
+                    body: JSON.stringify({ prompt: promptText, model_id: currentModelId })
+                });
+
+                if (!response.ok) throw new Error("Server error");
+
+                const reader = response.body.getReader();
+                const decoder = new TextDecoder("utf-8");
+                let fullText = "";
+                const botEl = document.getElementById(botId);
+
+                while (true) {
+                    const { done, value } = await reader.read();
+                    if (done) break;
+                    
+                    fullText += decoder.decode(value, { stream: true });
+                    botEl.innerHTML = formatMessage(fullText);
+                    chatbox.scrollTop = chatbox.scrollHeight;
+                }
+            } catch (err) {
+                document.getElementById(botId).innerHTML = `<span class="text-red-400">Error connecting to server. Is the Python script running?</span>`;
+            }
+        }
+    </script>
+</body>
+</html>
+"""
+
+# ==========================================
+# 🛣️ ROUTES
+# ==========================================
+@app.route('/')
+def index():
+    return HTML_TEMPLATE
+
+@app.route('/api/chat', methods=['POST'])
+def chat():
+    data = request.json
+    prompt = data.get('prompt', '')
+    model_id = data.get('model_id', 'default')
+
+    chatbot = active_models.get(model_id) or active_models.get('default')
+    
+    if chatbot is None or not chatbot.ready:
+        def err():
+            yield "Server error: Model is not loaded properly. Upload a model first."
+        return Response(stream_with_context(err()), mimetype='text/plain')
+
+    def generate():
+        for char in chatbot.stream_generate(prompt):
+            yield char
+
+    return Response(stream_with_context(generate()), mimetype='text/plain')
+
+@app.route('/api/upload', methods=['POST'])
+def upload_model():
+    if 'file' not in request.files:
+        return jsonify({"success": False, "error": "No file uploaded"})
+    
+    file = request.files['file']
+    if file.filename == '':
+        return jsonify({"success": False, "error": "Empty filename"})
+
+    if file and file.filename.endswith('.pt'):
+        try:
+            # Generate unique ID for this session's model
+            model_id = str(uuid.uuid4())
+            filename = secure_filename(f"{model_id}.pt")
+            filepath = os.path.join(UPLOAD_FOLDER, filename)
+            file.save(filepath)
+
+            # Get configuration
+            layers = int(request.form.get('layers', 5))
+            neurons = int(request.form.get('neurons', 768))
+            embed = int(request.form.get('embed', 384))
+            dropout = float(request.form.get('dropout', 0.2))
+
+            # Load into memory
+            chatbot = LinnyChat(filepath, embed, neurons, layers, dropout)
+            if chatbot.ready:
+                active_models[model_id] = chatbot
+                return jsonify({"success": True, "model_id": model_id})
+            else:
+                return jsonify({"success": False, "error": "Model failed to initialize. Check config."})
+        except Exception as e:
+            return jsonify({"success": False, "error": str(e)})
+
+    return jsonify({"success": False, "error": "Invalid file type. Only .pt allowed."})
+
+@app.route('/api/set_default', methods=['POST'])
+def set_default():
+    data = request.json
+    if data.get('password') != ADMIN_PASSWORD:
+        return jsonify({"success": False, "error": "Invalid Admin Password"})
+    
+    target_id = data.get('model_id')
+    if target_id in active_models:
+        # Promote session model to default global model
+        active_models['default'] = active_models[target_id]
+        return jsonify({"success": True})
+    
+    return jsonify({"success": False, "error": "Model not found in memory"})
+
+if __name__ == "__main__":
+    print("\n" + "="*50)
+    print("🌍 LINNY WEB SERVER STARTING...")
+    print("="*50)
+    print("Open your browser to: http://127.0.0.1:6000\n")
+    # Setting threaded=True handles multiple users hitting the chat endpoint concurrently
+    app.run(host='0.0.0.0', port=6000, threaded=True)