Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
|
@@ -7,80 +7,76 @@ from typing import Dict, Any, List
|
|
| 7 |
import time
|
| 8 |
|
| 9 |
app = Flask(__name__)
|
| 10 |
-
logging.basicConfig(level=logging.INFO)
|
| 11 |
|
| 12 |
# Configuration
|
| 13 |
OLLAMA_API_URL = os.getenv('OLLAMA_API_URL', 'https://huggingface.co/spaces/tommytracx/ollama-api')
|
| 14 |
-
DEFAULT_MODEL = os.getenv('DEFAULT_MODEL','llama2,llama2:13b,llama2:70b,codellama,neural-chat,gemma-3-270m').split(',')
|
| 15 |
MAX_TOKENS = int(os.getenv('MAX_TOKENS', '2048'))
|
| 16 |
TEMPERATURE = float(os.getenv('TEMPERATURE', '0.7'))
|
| 17 |
|
| 18 |
class OllamaClient:
|
| 19 |
def __init__(self, api_url: str):
|
| 20 |
self.api_url = api_url.rstrip('/')
|
| 21 |
-
self.available_models =
|
| 22 |
self.refresh_models()
|
| 23 |
|
| 24 |
-
def refresh_models(self):
|
| 25 |
-
"""Refresh the list of available models"""
|
| 26 |
try:
|
| 27 |
response = requests.get(f"{self.api_url}/api/models", timeout=10)
|
| 28 |
-
|
| 29 |
-
|
| 30 |
-
|
| 31 |
-
|
| 32 |
-
|
| 33 |
-
self.available_models = []
|
| 34 |
else:
|
| 35 |
-
|
|
|
|
| 36 |
except Exception as e:
|
| 37 |
logging.error(f"Error refreshing models: {e}")
|
| 38 |
-
self.available_models =
|
| 39 |
|
| 40 |
def list_models(self) -> List[str]:
|
| 41 |
-
"""
|
| 42 |
-
self.refresh_models()
|
| 43 |
return self.available_models
|
| 44 |
|
| 45 |
def generate(self, model_name: str, prompt: str, **kwargs) -> Dict[str, Any]:
|
| 46 |
-
"""Generate text using a model"""
|
|
|
|
|
|
|
|
|
|
| 47 |
try:
|
| 48 |
payload = {
|
| 49 |
"model": model_name,
|
| 50 |
"prompt": prompt,
|
| 51 |
-
"stream": False
|
|
|
|
| 52 |
}
|
| 53 |
-
|
|
|
|
|
|
|
| 54 |
|
| 55 |
-
|
| 56 |
-
|
| 57 |
-
|
| 58 |
-
|
| 59 |
-
|
| 60 |
-
|
| 61 |
-
|
| 62 |
-
|
| 63 |
-
"status": "success",
|
| 64 |
-
"response": data.get('response', ''),
|
| 65 |
-
"model": model_name,
|
| 66 |
-
"usage": data.get('usage', {})
|
| 67 |
-
}
|
| 68 |
-
else:
|
| 69 |
-
return {"status": "error", "message": data.get('message', 'Unknown error')}
|
| 70 |
-
else:
|
| 71 |
-
return {"status": "error", "message": f"Generation failed: {response.text}"}
|
| 72 |
except Exception as e:
|
|
|
|
| 73 |
return {"status": "error", "message": str(e)}
|
| 74 |
|
| 75 |
def health_check(self) -> Dict[str, Any]:
|
| 76 |
-
"""Check the health of the Ollama API"""
|
| 77 |
try:
|
| 78 |
response = requests.get(f"{self.api_url}/health", timeout=10)
|
| 79 |
-
|
| 80 |
-
|
| 81 |
-
else:
|
| 82 |
-
return {"status": "unhealthy", "error": f"HTTP {response.status_code}"}
|
| 83 |
except Exception as e:
|
|
|
|
| 84 |
return {"status": "unhealthy", "error": str(e)}
|
| 85 |
|
| 86 |
# Initialize Ollama client
|
|
@@ -95,228 +91,7 @@ HTML_TEMPLATE = '''
|
|
| 95 |
<meta name="viewport" content="width=device-width, initial-scale=1.0">
|
| 96 |
<title>OpenWebUI - Ollama Chat</title>
|
| 97 |
<style>
|
| 98 |
-
|
| 99 |
-
margin: 0;
|
| 100 |
-
padding: 0;
|
| 101 |
-
box-sizing: border-box;
|
| 102 |
-
}
|
| 103 |
-
|
| 104 |
-
body {
|
| 105 |
-
font-family: -apple-system, BlinkMacSystemFont, 'Segoe UI', Roboto, sans-serif;
|
| 106 |
-
background: linear-gradient(135deg, #667eea 0%, #764ba2 100%);
|
| 107 |
-
min-height: 100vh;
|
| 108 |
-
padding: 20px;
|
| 109 |
-
}
|
| 110 |
-
|
| 111 |
-
.container {
|
| 112 |
-
max-width: 1200px;
|
| 113 |
-
margin: 0 auto;
|
| 114 |
-
background: white;
|
| 115 |
-
border-radius: 20px;
|
| 116 |
-
box-shadow: 0 20px 40px rgba(0,0,0,0.1);
|
| 117 |
-
overflow: hidden;
|
| 118 |
-
}
|
| 119 |
-
|
| 120 |
-
.header {
|
| 121 |
-
background: linear-gradient(135deg, #667eea 0%, #764ba2 100%);
|
| 122 |
-
color: white;
|
| 123 |
-
padding: 30px;
|
| 124 |
-
text-align: center;
|
| 125 |
-
}
|
| 126 |
-
|
| 127 |
-
.header h1 {
|
| 128 |
-
font-size: 2.5rem;
|
| 129 |
-
margin-bottom: 10px;
|
| 130 |
-
font-weight: 700;
|
| 131 |
-
}
|
| 132 |
-
|
| 133 |
-
.header p {
|
| 134 |
-
font-size: 1.1rem;
|
| 135 |
-
opacity: 0.9;
|
| 136 |
-
}
|
| 137 |
-
|
| 138 |
-
.controls {
|
| 139 |
-
padding: 20px 30px;
|
| 140 |
-
background: #f8f9fa;
|
| 141 |
-
border-bottom: 1px solid #e9ecef;
|
| 142 |
-
display: flex;
|
| 143 |
-
gap: 15px;
|
| 144 |
-
align-items: center;
|
| 145 |
-
flex-wrap: wrap;
|
| 146 |
-
}
|
| 147 |
-
|
| 148 |
-
.control-group {
|
| 149 |
-
display: flex;
|
| 150 |
-
align-items: center;
|
| 151 |
-
gap: 8px;
|
| 152 |
-
}
|
| 153 |
-
|
| 154 |
-
.control-group label {
|
| 155 |
-
font-weight: 600;
|
| 156 |
-
color: #495057;
|
| 157 |
-
min-width: 80px;
|
| 158 |
-
}
|
| 159 |
-
|
| 160 |
-
.control-group select,
|
| 161 |
-
.control-group input {
|
| 162 |
-
padding: 8px 12px;
|
| 163 |
-
border: 2px solid #e9ecef;
|
| 164 |
-
border-radius: 8px;
|
| 165 |
-
font-size: 14px;
|
| 166 |
-
transition: border-color 0.3s;
|
| 167 |
-
}
|
| 168 |
-
|
| 169 |
-
.control-group select:focus,
|
| 170 |
-
.control-group input:focus {
|
| 171 |
-
outline: none;
|
| 172 |
-
border-color: #667eea;
|
| 173 |
-
}
|
| 174 |
-
|
| 175 |
-
.chat-container {
|
| 176 |
-
height: 500px;
|
| 177 |
-
overflow-y: auto;
|
| 178 |
-
padding: 20px;
|
| 179 |
-
background: #fafbfc;
|
| 180 |
-
}
|
| 181 |
-
|
| 182 |
-
.message {
|
| 183 |
-
margin-bottom: 20px;
|
| 184 |
-
display: flex;
|
| 185 |
-
gap: 15px;
|
| 186 |
-
}
|
| 187 |
-
|
| 188 |
-
.message.user {
|
| 189 |
-
flex-direction: row-reverse;
|
| 190 |
-
}
|
| 191 |
-
|
| 192 |
-
.message-avatar {
|
| 193 |
-
width: 40px;
|
| 194 |
-
height: 40px;
|
| 195 |
-
border-radius: 50%;
|
| 196 |
-
display: flex;
|
| 197 |
-
align-items: center;
|
| 198 |
-
justify-content: center;
|
| 199 |
-
font-weight: bold;
|
| 200 |
-
color: white;
|
| 201 |
-
flex-shrink: 0;
|
| 202 |
-
}
|
| 203 |
-
|
| 204 |
-
.message.user .message-avatar {
|
| 205 |
-
background: #667eea;
|
| 206 |
-
}
|
| 207 |
-
|
| 208 |
-
.message.assistant .message-avatar {
|
| 209 |
-
background: #28a745;
|
| 210 |
-
}
|
| 211 |
-
|
| 212 |
-
.message-content {
|
| 213 |
-
background: white;
|
| 214 |
-
padding: 15px 20px;
|
| 215 |
-
border-radius: 18px;
|
| 216 |
-
max-width: 70%;
|
| 217 |
-
box-shadow: 0 2px 10px rgba(0,0,0,0.1);
|
| 218 |
-
line-height: 1.5;
|
| 219 |
-
}
|
| 220 |
-
|
| 221 |
-
.message.user .message-content {
|
| 222 |
-
background: #667eea;
|
| 223 |
-
color: white;
|
| 224 |
-
}
|
| 225 |
-
|
| 226 |
-
.message.assistant .message-content {
|
| 227 |
-
background: white;
|
| 228 |
-
color: #333;
|
| 229 |
-
}
|
| 230 |
-
|
| 231 |
-
.input-container {
|
| 232 |
-
padding: 20px 30px;
|
| 233 |
-
background: white;
|
| 234 |
-
border-top: 1px solid #e9ecef;
|
| 235 |
-
}
|
| 236 |
-
|
| 237 |
-
.input-form {
|
| 238 |
-
display: flex;
|
| 239 |
-
gap: 15px;
|
| 240 |
-
}
|
| 241 |
-
|
| 242 |
-
.input-field {
|
| 243 |
-
flex: 1;
|
| 244 |
-
padding: 15px 20px;
|
| 245 |
-
border: 2px solid #e9ecef;
|
| 246 |
-
border-radius: 25px;
|
| 247 |
-
font-size: 16px;
|
| 248 |
-
transition: border-color 0.3s;
|
| 249 |
-
resize: none;
|
| 250 |
-
min-height: 50px;
|
| 251 |
-
max-height: 120px;
|
| 252 |
-
}
|
| 253 |
-
|
| 254 |
-
.input-field:focus {
|
| 255 |
-
outline: none;
|
| 256 |
-
border-color: #667eea;
|
| 257 |
-
}
|
| 258 |
-
|
| 259 |
-
.send-button {
|
| 260 |
-
padding: 15px 30px;
|
| 261 |
-
background: linear-gradient(135deg, #667eea 0%, #764ba2 100%);
|
| 262 |
-
color: white;
|
| 263 |
-
border: none;
|
| 264 |
-
border-radius: 25px;
|
| 265 |
-
font-size: 16px;
|
| 266 |
-
font-weight: 600;
|
| 267 |
-
cursor: pointer;
|
| 268 |
-
transition: transform 0.2s;
|
| 269 |
-
min-width: 100px;
|
| 270 |
-
}
|
| 271 |
-
|
| 272 |
-
.send-button:hover {
|
| 273 |
-
transform: translateY(-2px);
|
| 274 |
-
}
|
| 275 |
-
|
| 276 |
-
.send-button:disabled {
|
| 277 |
-
opacity: 0.6;
|
| 278 |
-
cursor: not-allowed;
|
| 279 |
-
transform: none;
|
| 280 |
-
}
|
| 281 |
-
|
| 282 |
-
.status {
|
| 283 |
-
text-align: center;
|
| 284 |
-
padding: 10px;
|
| 285 |
-
font-size: 14px;
|
| 286 |
-
color: #6c757d;
|
| 287 |
-
}
|
| 288 |
-
|
| 289 |
-
.status.error {
|
| 290 |
-
color: #dc3545;
|
| 291 |
-
}
|
| 292 |
-
|
| 293 |
-
.status.success {
|
| 294 |
-
color: #28a745;
|
| 295 |
-
}
|
| 296 |
-
|
| 297 |
-
.typing-indicator {
|
| 298 |
-
display: none;
|
| 299 |
-
padding: 15px 20px;
|
| 300 |
-
background: white;
|
| 301 |
-
border-radius: 18px;
|
| 302 |
-
color: #6c757d;
|
| 303 |
-
font-style: italic;
|
| 304 |
-
}
|
| 305 |
-
|
| 306 |
-
@media (max-width: 768px) {
|
| 307 |
-
.controls {
|
| 308 |
-
flex-direction: column;
|
| 309 |
-
align-items: stretch;
|
| 310 |
-
}
|
| 311 |
-
|
| 312 |
-
.control-group {
|
| 313 |
-
justify-content: space-between;
|
| 314 |
-
}
|
| 315 |
-
|
| 316 |
-
.message-content {
|
| 317 |
-
max-width: 85%;
|
| 318 |
-
}
|
| 319 |
-
}
|
| 320 |
</style>
|
| 321 |
</head>
|
| 322 |
<body>
|
|
@@ -330,16 +105,14 @@ HTML_TEMPLATE = '''
|
|
| 330 |
<div class="control-group">
|
| 331 |
<label for="model-select">Model:</label>
|
| 332 |
<select id="model-select">
|
| 333 |
-
<option value="">
|
| 334 |
</select>
|
| 335 |
</div>
|
| 336 |
-
|
| 337 |
<div class="control-group">
|
| 338 |
<label for="temperature">Temperature:</label>
|
| 339 |
<input type="range" id="temperature" min="0" max="2" step="0.1" value="0.7">
|
| 340 |
<span id="temp-value">0.7</span>
|
| 341 |
</div>
|
| 342 |
-
|
| 343 |
<div class="control-group">
|
| 344 |
<label for="max-tokens">Max Tokens:</label>
|
| 345 |
<input type="number" id="max-tokens" min="1" max="4096" value="2048">
|
|
@@ -379,50 +152,49 @@ HTML_TEMPLATE = '''
|
|
| 379 |
<script>
|
| 380 |
let conversationHistory = [];
|
| 381 |
|
| 382 |
-
// Initialize the interface
|
| 383 |
document.addEventListener('DOMContentLoaded', function() {
|
| 384 |
loadModels();
|
| 385 |
setupEventListeners();
|
| 386 |
autoResizeTextarea();
|
| 387 |
});
|
| 388 |
|
| 389 |
-
function loadModels() {
|
| 390 |
-
|
| 391 |
-
|
| 392 |
-
|
| 393 |
-
|
| 394 |
-
|
| 395 |
-
|
| 396 |
-
|
| 397 |
-
|
| 398 |
-
|
| 399 |
-
|
| 400 |
-
|
| 401 |
-
|
| 402 |
-
|
| 403 |
-
|
| 404 |
-
|
| 405 |
-
|
| 406 |
-
|
| 407 |
-
modelSelect.
|
| 408 |
-
}
|
| 409 |
-
|
| 410 |
-
|
| 411 |
-
|
| 412 |
-
showStatus('
|
| 413 |
-
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 414 |
}
|
| 415 |
|
| 416 |
function setupEventListeners() {
|
| 417 |
-
// Chat form submission
|
| 418 |
document.getElementById('chat-form').addEventListener('submit', handleSubmit);
|
| 419 |
-
|
| 420 |
-
// Temperature slider
|
| 421 |
document.getElementById('temperature').addEventListener('input', function() {
|
| 422 |
document.getElementById('temp-value').textContent = this.value;
|
| 423 |
});
|
| 424 |
-
|
| 425 |
-
// Auto-resize textarea
|
| 426 |
document.getElementById('message-input').addEventListener('input', autoResizeTextarea);
|
| 427 |
}
|
| 428 |
|
|
@@ -432,7 +204,7 @@ HTML_TEMPLATE = '''
|
|
| 432 |
textarea.style.height = Math.min(textarea.scrollHeight, 120) + 'px';
|
| 433 |
}
|
| 434 |
|
| 435 |
-
function handleSubmit(e) {
|
| 436 |
e.preventDefault();
|
| 437 |
|
| 438 |
const messageInput = document.getElementById('message-input');
|
|
@@ -449,16 +221,33 @@ HTML_TEMPLATE = '''
|
|
| 449 |
return;
|
| 450 |
}
|
| 451 |
|
| 452 |
-
// Add user message to chat
|
| 453 |
addMessage(message, 'user');
|
| 454 |
messageInput.value = '';
|
| 455 |
autoResizeTextarea();
|
| 456 |
-
|
| 457 |
-
// Show typing indicator
|
| 458 |
showTypingIndicator(true);
|
| 459 |
|
| 460 |
-
|
| 461 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 462 |
}
|
| 463 |
|
| 464 |
function addMessage(content, sender) {
|
|
@@ -477,53 +266,14 @@ HTML_TEMPLATE = '''
|
|
| 477 |
messageDiv.appendChild(avatar);
|
| 478 |
messageDiv.appendChild(messageContent);
|
| 479 |
chatContainer.appendChild(messageDiv);
|
| 480 |
-
|
| 481 |
-
// Scroll to bottom
|
| 482 |
chatContainer.scrollTop = chatContainer.scrollHeight;
|
| 483 |
|
| 484 |
-
// Add to conversation history
|
| 485 |
conversationHistory.push({ role: sender, content: content });
|
| 486 |
}
|
| 487 |
|
| 488 |
-
function sendMessage(message, model, temperature, maxTokens) {
|
| 489 |
-
const payload = {
|
| 490 |
-
model: model,
|
| 491 |
-
prompt: message,
|
| 492 |
-
temperature: temperature,
|
| 493 |
-
max_tokens: maxTokens
|
| 494 |
-
};
|
| 495 |
-
|
| 496 |
-
fetch('{{ ollama_api_url }}/api/generate', {
|
| 497 |
-
method: 'POST',
|
| 498 |
-
headers: {
|
| 499 |
-
'Content-Type': 'application/json'
|
| 500 |
-
},
|
| 501 |
-
body: JSON.stringify(payload)
|
| 502 |
-
})
|
| 503 |
-
.then(response => response.json())
|
| 504 |
-
.then(data => {
|
| 505 |
-
showTypingIndicator(false);
|
| 506 |
-
|
| 507 |
-
if (data.status === 'success') {
|
| 508 |
-
addMessage(data.response, 'assistant');
|
| 509 |
-
showStatus(`Response generated using ${model}`, 'success');
|
| 510 |
-
} else {
|
| 511 |
-
addMessage('Sorry, I encountered an error while processing your request.', 'assistant');
|
| 512 |
-
showStatus(`Error: ${data.message}`, 'error');
|
| 513 |
-
}
|
| 514 |
-
})
|
| 515 |
-
.catch(error => {
|
| 516 |
-
showTypingIndicator(false);
|
| 517 |
-
addMessage('Sorry, I encountered an error while processing your request.', 'assistant');
|
| 518 |
-
showStatus('Network error occurred', 'error');
|
| 519 |
-
console.error('Error:', error);
|
| 520 |
-
});
|
| 521 |
-
}
|
| 522 |
-
|
| 523 |
function showTypingIndicator(show) {
|
| 524 |
const indicator = document.getElementById('typing-indicator');
|
| 525 |
indicator.style.display = show ? 'block' : 'none';
|
| 526 |
-
|
| 527 |
if (show) {
|
| 528 |
const chatContainer = document.getElementById('chat-container');
|
| 529 |
chatContainer.scrollTop = chatContainer.scrollHeight;
|
|
@@ -534,8 +284,6 @@ HTML_TEMPLATE = '''
|
|
| 534 |
const statusDiv = document.getElementById('status');
|
| 535 |
statusDiv.textContent = message;
|
| 536 |
statusDiv.className = `status ${type}`;
|
| 537 |
-
|
| 538 |
-
// Clear status after 5 seconds
|
| 539 |
setTimeout(() => {
|
| 540 |
statusDiv.textContent = '';
|
| 541 |
statusDiv.className = 'status';
|
|
@@ -548,14 +296,12 @@ HTML_TEMPLATE = '''
|
|
| 548 |
|
| 549 |
@app.route('/')
|
| 550 |
def home():
|
| 551 |
-
"""Main chat interface"""
|
| 552 |
-
return render_template_string(HTML_TEMPLATE,
|
| 553 |
-
ollama_api_url=OLLAMA_API_URL,
|
| 554 |
-
default_model=DEFAULT_MODEL)
|
| 555 |
|
| 556 |
@app.route('/api/chat', methods=['POST'])
|
| 557 |
def chat():
|
| 558 |
-
"""Chat API endpoint"""
|
| 559 |
try:
|
| 560 |
data = request.get_json()
|
| 561 |
if not data or 'message' not in data or 'model' not in data:
|
|
@@ -566,20 +312,15 @@ def chat():
|
|
| 566 |
temperature = data.get('temperature', TEMPERATURE)
|
| 567 |
max_tokens = data.get('max_tokens', MAX_TOKENS)
|
| 568 |
|
| 569 |
-
result = ollama_client.generate(model, message,
|
| 570 |
-
|
| 571 |
-
max_tokens=max_tokens)
|
| 572 |
-
|
| 573 |
-
if result["status"] == "success":
|
| 574 |
-
return jsonify(result)
|
| 575 |
-
else:
|
| 576 |
-
return jsonify(result), 500
|
| 577 |
except Exception as e:
|
|
|
|
| 578 |
return jsonify({"status": "error", "message": str(e)}), 500
|
| 579 |
|
| 580 |
@app.route('/api/models', methods=['GET'])
|
| 581 |
def get_models():
|
| 582 |
-
"""Get available models"""
|
| 583 |
try:
|
| 584 |
models = ollama_client.list_models()
|
| 585 |
return jsonify({
|
|
@@ -588,11 +329,12 @@ def get_models():
|
|
| 588 |
"count": len(models)
|
| 589 |
})
|
| 590 |
except Exception as e:
|
|
|
|
| 591 |
return jsonify({"status": "error", "message": str(e)}), 500
|
| 592 |
|
| 593 |
@app.route('/health', methods=['GET'])
|
| 594 |
def health_check():
|
| 595 |
-
"""Health check endpoint"""
|
| 596 |
try:
|
| 597 |
ollama_health = ollama_client.health_check()
|
| 598 |
return jsonify({
|
|
@@ -601,6 +343,7 @@ def health_check():
|
|
| 601 |
"timestamp": time.time()
|
| 602 |
})
|
| 603 |
except Exception as e:
|
|
|
|
| 604 |
return jsonify({
|
| 605 |
"status": "unhealthy",
|
| 606 |
"error": str(e),
|
|
@@ -608,4 +351,4 @@ def health_check():
|
|
| 608 |
}), 500
|
| 609 |
|
| 610 |
if __name__ == '__main__':
|
| 611 |
-
app.run(host='0.0.0.0', port=7860, debug=False)
|
|
|
|
| 7 |
import time
|
| 8 |
|
| 9 |
app = Flask(__name__)
|
| 10 |
+
logging.basicConfig(level=logging.INFO, format='%(asctime)s - %(levelname)s - %(message)s')
|
| 11 |
|
| 12 |
# Configuration
|
| 13 |
OLLAMA_API_URL = os.getenv('OLLAMA_API_URL', 'https://huggingface.co/spaces/tommytracx/ollama-api')
|
| 14 |
+
DEFAULT_MODEL = os.getenv('DEFAULT_MODEL', 'llama2,llama2:13b,llama2:70b,codellama,neural-chat,gemma-3-270m').split(',')
|
| 15 |
MAX_TOKENS = int(os.getenv('MAX_TOKENS', '2048'))
|
| 16 |
TEMPERATURE = float(os.getenv('TEMPERATURE', '0.7'))
|
| 17 |
|
| 18 |
class OllamaClient:
|
| 19 |
def __init__(self, api_url: str):
|
| 20 |
self.api_url = api_url.rstrip('/')
|
| 21 |
+
self.available_models = DEFAULT_MODEL # Initialize with default models
|
| 22 |
self.refresh_models()
|
| 23 |
|
| 24 |
+
def refresh_models(self) -> None:
|
| 25 |
+
"""Refresh the list of available models from the API, falling back to defaults on failure."""
|
| 26 |
try:
|
| 27 |
response = requests.get(f"{self.api_url}/api/models", timeout=10)
|
| 28 |
+
response.raise_for_status()
|
| 29 |
+
data = response.json()
|
| 30 |
+
if data.get('status') == 'success' and isinstance(data.get('models'), list):
|
| 31 |
+
self.available_models = data['models']
|
| 32 |
+
logging.info(f"Successfully fetched models: {self.available_models}")
|
|
|
|
| 33 |
else:
|
| 34 |
+
logging.warning(f"Invalid response format from API: {data}")
|
| 35 |
+
self.available_models = DEFAULT_MODEL
|
| 36 |
except Exception as e:
|
| 37 |
logging.error(f"Error refreshing models: {e}")
|
| 38 |
+
self.available_models = DEFAULT_MODEL
|
| 39 |
|
| 40 |
def list_models(self) -> List[str]:
|
| 41 |
+
"""Return the list of available models."""
|
|
|
|
| 42 |
return self.available_models
|
| 43 |
|
| 44 |
def generate(self, model_name: str, prompt: str, **kwargs) -> Dict[str, Any]:
|
| 45 |
+
"""Generate text using a model."""
|
| 46 |
+
if model_name not in self.available_models:
|
| 47 |
+
return {"status": "error", "message": f"Model {model_name} not available"}
|
| 48 |
+
|
| 49 |
try:
|
| 50 |
payload = {
|
| 51 |
"model": model_name,
|
| 52 |
"prompt": prompt,
|
| 53 |
+
"stream": False,
|
| 54 |
+
**kwargs
|
| 55 |
}
|
| 56 |
+
response = requests.post(f"{self.api_url}/api/generate", json=payload, timeout=120)
|
| 57 |
+
response.raise_for_status()
|
| 58 |
+
data = response.json()
|
| 59 |
|
| 60 |
+
if data.get('status') == 'success':
|
| 61 |
+
return {
|
| 62 |
+
"status": "success",
|
| 63 |
+
"response": data.get('response', ''),
|
| 64 |
+
"model": model_name,
|
| 65 |
+
"usage": data.get('usage', {})
|
| 66 |
+
}
|
| 67 |
+
return {"status": "error", "message": data.get('message', 'Unknown error')}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 68 |
except Exception as e:
|
| 69 |
+
logging.error(f"Error generating response: {e}")
|
| 70 |
return {"status": "error", "message": str(e)}
|
| 71 |
|
| 72 |
def health_check(self) -> Dict[str, Any]:
|
| 73 |
+
"""Check the health of the Ollama API."""
|
| 74 |
try:
|
| 75 |
response = requests.get(f"{self.api_url}/health", timeout=10)
|
| 76 |
+
response.raise_for_status()
|
| 77 |
+
return response.json()
|
|
|
|
|
|
|
| 78 |
except Exception as e:
|
| 79 |
+
logging.error(f"Health check failed: {e}")
|
| 80 |
return {"status": "unhealthy", "error": str(e)}
|
| 81 |
|
| 82 |
# Initialize Ollama client
|
|
|
|
| 91 |
<meta name="viewport" content="width=device-width, initial-scale=1.0">
|
| 92 |
<title>OpenWebUI - Ollama Chat</title>
|
| 93 |
<style>
|
| 94 |
+
/* [Previous CSS unchanged] */
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 95 |
</style>
|
| 96 |
</head>
|
| 97 |
<body>
|
|
|
|
| 105 |
<div class="control-group">
|
| 106 |
<label for="model-select">Model:</label>
|
| 107 |
<select id="model-select">
|
| 108 |
+
<option value="">Select a model...</option>
|
| 109 |
</select>
|
| 110 |
</div>
|
|
|
|
| 111 |
<div class="control-group">
|
| 112 |
<label for="temperature">Temperature:</label>
|
| 113 |
<input type="range" id="temperature" min="0" max="2" step="0.1" value="0.7">
|
| 114 |
<span id="temp-value">0.7</span>
|
| 115 |
</div>
|
|
|
|
| 116 |
<div class="control-group">
|
| 117 |
<label for="max-tokens">Max Tokens:</label>
|
| 118 |
<input type="number" id="max-tokens" min="1" max="4096" value="2048">
|
|
|
|
| 152 |
<script>
|
| 153 |
let conversationHistory = [];
|
| 154 |
|
|
|
|
| 155 |
document.addEventListener('DOMContentLoaded', function() {
|
| 156 |
loadModels();
|
| 157 |
setupEventListeners();
|
| 158 |
autoResizeTextarea();
|
| 159 |
});
|
| 160 |
|
| 161 |
+
async function loadModels() {
|
| 162 |
+
const modelSelect = document.getElementById('model-select');
|
| 163 |
+
modelSelect.innerHTML = '<option value="">Loading models...</option>';
|
| 164 |
+
|
| 165 |
+
try {
|
| 166 |
+
const response = await fetch('/api/models');
|
| 167 |
+
const data = await response.json();
|
| 168 |
+
|
| 169 |
+
modelSelect.innerHTML = '<option value="">Select a model...</option>';
|
| 170 |
+
|
| 171 |
+
if (data.status === 'success' && data.models.length > 0) {
|
| 172 |
+
data.models.forEach(model => {
|
| 173 |
+
const option = document.createElement('option');
|
| 174 |
+
option.value = model;
|
| 175 |
+
option.textContent = model;
|
| 176 |
+
if (model === '{{ default_model[0] }}') {
|
| 177 |
+
option.selected = true;
|
| 178 |
+
}
|
| 179 |
+
modelSelect.appendChild(option);
|
| 180 |
+
});
|
| 181 |
+
showStatus('Models loaded successfully', 'success');
|
| 182 |
+
} else {
|
| 183 |
+
modelSelect.innerHTML = '<option value="">No models available</option>';
|
| 184 |
+
showStatus('No models available from API', 'error');
|
| 185 |
+
}
|
| 186 |
+
} catch (error) {
|
| 187 |
+
console.error('Error loading models:', error);
|
| 188 |
+
modelSelect.innerHTML = '<option value="">No models available</option>';
|
| 189 |
+
showStatus('Failed to load models: ' + error.message, 'error');
|
| 190 |
+
}
|
| 191 |
}
|
| 192 |
|
| 193 |
function setupEventListeners() {
|
|
|
|
| 194 |
document.getElementById('chat-form').addEventListener('submit', handleSubmit);
|
|
|
|
|
|
|
| 195 |
document.getElementById('temperature').addEventListener('input', function() {
|
| 196 |
document.getElementById('temp-value').textContent = this.value;
|
| 197 |
});
|
|
|
|
|
|
|
| 198 |
document.getElementById('message-input').addEventListener('input', autoResizeTextarea);
|
| 199 |
}
|
| 200 |
|
|
|
|
| 204 |
textarea.style.height = Math.min(textarea.scrollHeight, 120) + 'px';
|
| 205 |
}
|
| 206 |
|
| 207 |
+
async function handleSubmit(e) {
|
| 208 |
e.preventDefault();
|
| 209 |
|
| 210 |
const messageInput = document.getElementById('message-input');
|
|
|
|
| 221 |
return;
|
| 222 |
}
|
| 223 |
|
|
|
|
| 224 |
addMessage(message, 'user');
|
| 225 |
messageInput.value = '';
|
| 226 |
autoResizeTextarea();
|
|
|
|
|
|
|
| 227 |
showTypingIndicator(true);
|
| 228 |
|
| 229 |
+
try {
|
| 230 |
+
const response = await fetch('/api/chat', {
|
| 231 |
+
method: 'POST',
|
| 232 |
+
headers: { 'Content-Type': 'application/json' },
|
| 233 |
+
body: JSON.stringify({ model, message, temperature, max_tokens: maxTokens })
|
| 234 |
+
});
|
| 235 |
+
const data = await response.json();
|
| 236 |
+
|
| 237 |
+
showTypingIndicator(false);
|
| 238 |
+
|
| 239 |
+
if (data.status === 'success') {
|
| 240 |
+
addMessage(data.response, 'assistant');
|
| 241 |
+
showStatus(`Response generated using ${model}`, 'success');
|
| 242 |
+
} else {
|
| 243 |
+
addMessage('Sorry, I encountered an error while processing your request.', 'assistant');
|
| 244 |
+
showStatus(`Error: ${data.message}`, 'error');
|
| 245 |
+
}
|
| 246 |
+
} catch (error) {
|
| 247 |
+
showTypingIndicator(false);
|
| 248 |
+
addMessage('Sorry, I encountered a network error.', 'assistant');
|
| 249 |
+
showStatus('Network error: ' + error.message, 'error');
|
| 250 |
+
}
|
| 251 |
}
|
| 252 |
|
| 253 |
function addMessage(content, sender) {
|
|
|
|
| 266 |
messageDiv.appendChild(avatar);
|
| 267 |
messageDiv.appendChild(messageContent);
|
| 268 |
chatContainer.appendChild(messageDiv);
|
|
|
|
|
|
|
| 269 |
chatContainer.scrollTop = chatContainer.scrollHeight;
|
| 270 |
|
|
|
|
| 271 |
conversationHistory.push({ role: sender, content: content });
|
| 272 |
}
|
| 273 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 274 |
function showTypingIndicator(show) {
|
| 275 |
const indicator = document.getElementById('typing-indicator');
|
| 276 |
indicator.style.display = show ? 'block' : 'none';
|
|
|
|
| 277 |
if (show) {
|
| 278 |
const chatContainer = document.getElementById('chat-container');
|
| 279 |
chatContainer.scrollTop = chatContainer.scrollHeight;
|
|
|
|
| 284 |
const statusDiv = document.getElementById('status');
|
| 285 |
statusDiv.textContent = message;
|
| 286 |
statusDiv.className = `status ${type}`;
|
|
|
|
|
|
|
| 287 |
setTimeout(() => {
|
| 288 |
statusDiv.textContent = '';
|
| 289 |
statusDiv.className = 'status';
|
|
|
|
| 296 |
|
| 297 |
@app.route('/')
|
| 298 |
def home():
|
| 299 |
+
"""Main chat interface."""
|
| 300 |
+
return render_template_string(HTML_TEMPLATE, ollama_api_url=OLLAMA_API_URL, default_model=DEFAULT_MODEL)
|
|
|
|
|
|
|
| 301 |
|
| 302 |
@app.route('/api/chat', methods=['POST'])
|
| 303 |
def chat():
|
| 304 |
+
"""Chat API endpoint."""
|
| 305 |
try:
|
| 306 |
data = request.get_json()
|
| 307 |
if not data or 'message' not in data or 'model' not in data:
|
|
|
|
| 312 |
temperature = data.get('temperature', TEMPERATURE)
|
| 313 |
max_tokens = data.get('max_tokens', MAX_TOKENS)
|
| 314 |
|
| 315 |
+
result = ollama_client.generate(model, message, temperature=temperature, max_tokens=max_tokens)
|
| 316 |
+
return jsonify(result), 200 if result["status"] == "success" else 500
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 317 |
except Exception as e:
|
| 318 |
+
logging.error(f"Chat endpoint error: {e}")
|
| 319 |
return jsonify({"status": "error", "message": str(e)}), 500
|
| 320 |
|
| 321 |
@app.route('/api/models', methods=['GET'])
|
| 322 |
def get_models():
|
| 323 |
+
"""Get available models."""
|
| 324 |
try:
|
| 325 |
models = ollama_client.list_models()
|
| 326 |
return jsonify({
|
|
|
|
| 329 |
"count": len(models)
|
| 330 |
})
|
| 331 |
except Exception as e:
|
| 332 |
+
logging.error(f"Models endpoint error: {e}")
|
| 333 |
return jsonify({"status": "error", "message": str(e)}), 500
|
| 334 |
|
| 335 |
@app.route('/health', methods=['GET'])
|
| 336 |
def health_check():
|
| 337 |
+
"""Health check endpoint."""
|
| 338 |
try:
|
| 339 |
ollama_health = ollama_client.health_check()
|
| 340 |
return jsonify({
|
|
|
|
| 343 |
"timestamp": time.time()
|
| 344 |
})
|
| 345 |
except Exception as e:
|
| 346 |
+
logging.error(f"Health check endpoint error: {e}")
|
| 347 |
return jsonify({
|
| 348 |
"status": "unhealthy",
|
| 349 |
"error": str(e),
|
|
|
|
| 351 |
}), 500
|
| 352 |
|
| 353 |
if __name__ == '__main__':
|
| 354 |
+
app.run(host='0.0.0.0', port=7860, debug=False)
|