multiple-models
This commit is contained in:
214
index.html
214
index.html
@@ -215,11 +215,12 @@
|
||||
padding: 12px 15px;
|
||||
border: 1px solid #ddd;
|
||||
border-radius: 8px;
|
||||
font-size: 14px;
|
||||
font-size: 15px;
|
||||
resize: none;
|
||||
min-height: 44px;
|
||||
max-height: 120px;
|
||||
min-height: 60px;
|
||||
max-height: 200px;
|
||||
font-family: inherit;
|
||||
line-height: 1.5;
|
||||
}
|
||||
|
||||
#userInput:focus {
|
||||
@@ -344,7 +345,22 @@
|
||||
<div id="chatContainer">
|
||||
<div id="header">
|
||||
<div>🤖 LM Studio Chat with Streaming</div>
|
||||
<div style="font-size: 0.9em; opacity: 0.9;">Connected to: <code>/api (proxied to plato.lan:1234)</code></div>
|
||||
<div style="font-size: 0.9em; opacity: 0.9; display: flex; align-items: center; gap: 15px; flex-wrap: wrap;">
|
||||
<div style="display: flex; align-items: center; gap: 8px;">
|
||||
<span>Backend:</span>
|
||||
<select id="backendSelector" style="padding: 4px 8px; border-radius: 4px; border: 1px solid #444; background: #2a2a2a; color: #e0e0e0;">
|
||||
<option value="plato">Plato (192.168.1.74)</option>
|
||||
<option value="stoic">Stoic (192.168.1.158)</option>
|
||||
</select>
|
||||
</div>
|
||||
<div style="display: flex; align-items: center; gap: 8px;">
|
||||
<span>Model:</span>
|
||||
<select id="modelSelector" style="padding: 4px 8px; border-radius: 4px; border: 1px solid #444; background: #2a2a2a; color: #e0e0e0; min-width: 200px;">
|
||||
<option value="">Loading models...</option>
|
||||
</select>
|
||||
</div>
|
||||
<code id="backendDisplay" style="font-size: 0.85em;"></code>
|
||||
</div>
|
||||
</div>
|
||||
|
||||
<div id="chatLog"></div>
|
||||
@@ -375,16 +391,107 @@
|
||||
|
||||
<script>
|
||||
// Configuration
|
||||
// Use /api proxy when served from production domain, direct access for local dev
|
||||
const API_BASE_URL = window.location.hostname === 'jarvis-lan.appmodel.nl'
|
||||
? '/api'
|
||||
: 'http://192.168.1.74:1234/v1'
|
||||
const BACKENDS = {
|
||||
plato: {
|
||||
prod: '/api/plato',
|
||||
dev: 'http://192.168.1.74:1234/v1',
|
||||
name: 'Plato (192.168.1.74)'
|
||||
},
|
||||
stoic: {
|
||||
prod: '/api/stoic',
|
||||
dev: 'http://192.168.1.159:1234/v1',
|
||||
name: 'Stoic (192.168.1.159)'
|
||||
}
|
||||
}
|
||||
|
||||
const IS_PRODUCTION = window.location.hostname === 'jarvis-lan.appmodel.nl'
|
||||
const API_KEY = 'not-needed'
|
||||
|
||||
// Global state
|
||||
let currentBackend = 'plato'
|
||||
let currentModel = null
|
||||
let availableModels = []
|
||||
let currentStreamController = null
|
||||
let isStreaming = false
|
||||
|
||||
// Get current API URL based on selected backend
|
||||
function getApiUrl() {
|
||||
const backend = BACKENDS[currentBackend]
|
||||
return IS_PRODUCTION ? backend.prod : backend.dev
|
||||
}
|
||||
|
||||
// Update backend display
|
||||
function updateBackendDisplay() {
|
||||
const backend = BACKENDS[currentBackend]
|
||||
const displayText = IS_PRODUCTION
|
||||
? `${backend.prod} → ${backend.name}`
|
||||
: backend.dev
|
||||
document.getElementById('backendDisplay').textContent = displayText
|
||||
}
|
||||
|
||||
// Fetch available models from backend
|
||||
async function fetchModels() {
|
||||
try {
|
||||
const response = await fetch(`${getApiUrl()}/models`)
|
||||
if (response.ok) {
|
||||
const data = await response.json()
|
||||
availableModels = data.data || []
|
||||
populateModelSelector()
|
||||
|
||||
// Auto-select first model if none selected
|
||||
if (!currentModel && availableModels.length > 0) {
|
||||
currentModel = availableModels[0].id
|
||||
document.getElementById('modelSelector').value = currentModel
|
||||
}
|
||||
} else {
|
||||
console.error('Failed to fetch models:', response.statusText)
|
||||
document.getElementById('modelSelector').innerHTML = '<option value="">Error loading models</option>'
|
||||
}
|
||||
} catch (error) {
|
||||
console.error('Error fetching models:', error)
|
||||
document.getElementById('modelSelector').innerHTML = '<option value="">Error loading models</option>'
|
||||
}
|
||||
}
|
||||
|
||||
// Populate model selector dropdown
|
||||
function populateModelSelector() {
|
||||
const selector = document.getElementById('modelSelector')
|
||||
if (availableModels.length === 0) {
|
||||
selector.innerHTML = '<option value="">No models available</option>'
|
||||
return
|
||||
}
|
||||
|
||||
selector.innerHTML = availableModels.map(model =>
|
||||
`<option value="${model.id}">${model.id}</option>`
|
||||
).join('')
|
||||
|
||||
if (currentModel) {
|
||||
selector.value = currentModel
|
||||
}
|
||||
}
|
||||
|
||||
// Handle backend and model selection changes
|
||||
document.addEventListener('DOMContentLoaded', () => {
|
||||
const backendSelector = document.getElementById('backendSelector')
|
||||
const modelSelector = document.getElementById('modelSelector')
|
||||
|
||||
backendSelector.value = currentBackend
|
||||
updateBackendDisplay()
|
||||
fetchModels()
|
||||
|
||||
backendSelector.addEventListener('change', (e) => {
|
||||
currentBackend = e.target.value
|
||||
updateBackendDisplay()
|
||||
console.log('Backend switched to:', currentBackend, '→', getApiUrl())
|
||||
fetchModels() // Reload models for new backend
|
||||
})
|
||||
|
||||
modelSelector.addEventListener('change', (e) => {
|
||||
currentModel = e.target.value
|
||||
console.log('Model selected:', currentModel)
|
||||
})
|
||||
})
|
||||
|
||||
// DOM Elements
|
||||
const chatLog = document.getElementById('chatLog')
|
||||
const userInput = document.getElementById('userInput')
|
||||
@@ -409,23 +516,10 @@ marked.setOptions({
|
||||
// Auto-resize textarea
|
||||
function autoResize(textarea) {
|
||||
textarea.style.height = 'auto'
|
||||
textarea.style.height = Math.min(textarea.scrollHeight, 120) + 'px'
|
||||
textarea.style.height = Math.min(textarea.scrollHeight, 200) + 'px'
|
||||
}
|
||||
|
||||
// Fetch available models
|
||||
async function fetchModels() {
|
||||
try {
|
||||
const response = await fetch(`${ API_BASE_URL }/models`)
|
||||
if (response.ok) {
|
||||
const data = await response.json()
|
||||
if (data.data && data.data.length > 0) {
|
||||
modelNameSpan.textContent = data.data[0].id
|
||||
}
|
||||
}
|
||||
} catch (error) {
|
||||
console.log('Could not fetch models:', error)
|
||||
}
|
||||
}
|
||||
// Old fetchModels function removed - now handled in DOMContentLoaded
|
||||
|
||||
// Add message to chat
|
||||
function addMessage(role, content, markdown = false, messageId = null) {
|
||||
@@ -572,15 +666,21 @@ async function handleStreamingResponse(userMessage) {
|
||||
contentDiv.appendChild(cursorSpan)
|
||||
|
||||
try {
|
||||
// Auto-select first model if none selected
|
||||
if (!currentModel && availableModels.length > 0) {
|
||||
currentModel = availableModels[0].id
|
||||
document.getElementById('modelSelector').value = currentModel
|
||||
}
|
||||
|
||||
const requestBody = {
|
||||
model : 'local-model',
|
||||
model : currentModel || 'local-model',
|
||||
messages : [{ role: 'user', content: userMessage }],
|
||||
stream : true,
|
||||
temperature: 0.7,
|
||||
max_tokens : 2000
|
||||
}
|
||||
|
||||
const response = await fetch(`${ API_BASE_URL }/chat/completions`, {
|
||||
const response = await fetch(`${ getApiUrl() }/chat/completions`, {
|
||||
method : 'POST',
|
||||
headers: {
|
||||
'Content-Type' : 'application/json',
|
||||
@@ -599,6 +699,46 @@ async function handleStreamingResponse(userMessage) {
|
||||
isStreaming = true
|
||||
currentStreamController = new AbortController()
|
||||
|
||||
// Debounce DOM updates to reduce stutter
|
||||
let updateScheduled = false
|
||||
let lastUpdate = Date.now()
|
||||
const MIN_UPDATE_INTERVAL = 50 // ms, update at most every 50ms
|
||||
|
||||
function scheduleUpdate() {
|
||||
if (updateScheduled) return
|
||||
|
||||
const now = Date.now()
|
||||
const timeSinceLastUpdate = now - lastUpdate
|
||||
|
||||
if (timeSinceLastUpdate >= MIN_UPDATE_INTERVAL) {
|
||||
// Update immediately if enough time has passed
|
||||
updateScheduled = true
|
||||
requestAnimationFrame(() => {
|
||||
updateMessageContent(contentDiv, accumulatedContent, markdownToggle.checked)
|
||||
if (cursorSpan.parentNode) {
|
||||
contentDiv.appendChild(cursorSpan)
|
||||
}
|
||||
chatLog.scrollTop = chatLog.scrollHeight
|
||||
lastUpdate = Date.now()
|
||||
updateScheduled = false
|
||||
})
|
||||
} else {
|
||||
// Schedule update after remaining time
|
||||
updateScheduled = true
|
||||
setTimeout(() => {
|
||||
requestAnimationFrame(() => {
|
||||
updateMessageContent(contentDiv, accumulatedContent, markdownToggle.checked)
|
||||
if (cursorSpan.parentNode) {
|
||||
contentDiv.appendChild(cursorSpan)
|
||||
}
|
||||
chatLog.scrollTop = chatLog.scrollHeight
|
||||
lastUpdate = Date.now()
|
||||
updateScheduled = false
|
||||
})
|
||||
}, MIN_UPDATE_INTERVAL - timeSinceLastUpdate)
|
||||
}
|
||||
}
|
||||
|
||||
while (true) {
|
||||
const { done, value } = await reader.read()
|
||||
if (done) break
|
||||
@@ -613,10 +753,12 @@ async function handleStreamingResponse(userMessage) {
|
||||
if (data === '[DONE]') {
|
||||
isStreaming = false
|
||||
currentStreamController = null
|
||||
// Remove cursor when done
|
||||
// Final update and remove cursor
|
||||
updateMessageContent(contentDiv, accumulatedContent, markdownToggle.checked)
|
||||
if (cursorSpan.parentNode) {
|
||||
cursorSpan.parentNode.removeChild(cursorSpan)
|
||||
}
|
||||
chatLog.scrollTop = chatLog.scrollHeight
|
||||
return
|
||||
}
|
||||
|
||||
@@ -624,15 +766,7 @@ async function handleStreamingResponse(userMessage) {
|
||||
const parsed = JSON.parse(data)
|
||||
if (parsed.choices && parsed.choices[0].delta.content) {
|
||||
accumulatedContent += parsed.choices[0].delta.content
|
||||
|
||||
// Update content every 5 characters for smoother display
|
||||
if (accumulatedContent.length % 5 === 0 || parsed.choices[0].finish_reason) {
|
||||
updateMessageContent(contentDiv, accumulatedContent, markdownToggle.checked)
|
||||
if (cursorSpan.parentNode) {
|
||||
contentDiv.appendChild(cursorSpan)
|
||||
}
|
||||
chatLog.scrollTop = chatLog.scrollHeight
|
||||
}
|
||||
scheduleUpdate()
|
||||
}
|
||||
} catch (e) {
|
||||
console.log('Error parsing stream data:', e)
|
||||
@@ -668,15 +802,21 @@ async function handleNonStreamingResponse(userMessage) {
|
||||
const typingIndicator = showTypingIndicator()
|
||||
|
||||
try {
|
||||
// Auto-select first model if none selected
|
||||
if (!currentModel && availableModels.length > 0) {
|
||||
currentModel = availableModels[0].id
|
||||
document.getElementById('modelSelector').value = currentModel
|
||||
}
|
||||
|
||||
const requestBody = {
|
||||
model : 'local-model',
|
||||
model : currentModel || 'local-model',
|
||||
messages : [{ role: 'user', content: userMessage }],
|
||||
stream : false,
|
||||
temperature: 0.7,
|
||||
max_tokens : 2000
|
||||
}
|
||||
|
||||
const response = await fetch(`${ API_BASE_URL }/chat/completions`, {
|
||||
const response = await fetch(`${ getApiUrl() }/chat/completions`, {
|
||||
method : 'POST',
|
||||
headers: {
|
||||
'Content-Type' : 'application/json',
|
||||
|
||||
19
nginx.conf
19
nginx.conf
@@ -10,11 +10,30 @@ server {
|
||||
try_files $uri $uri.html $uri/ =404;
|
||||
}
|
||||
|
||||
|
||||
location = /index.html {
|
||||
try_files /index.html =404;
|
||||
}
|
||||
|
||||
# Proxy API requests to plato.lan (192.168.1.74)
|
||||
location /api/plato/ {
|
||||
proxy_pass http://192.168.1.74:1234/v1/;
|
||||
proxy_set_header Host $host;
|
||||
proxy_set_header X-Real-IP $remote_addr;
|
||||
proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
|
||||
proxy_set_header X-Forwarded-Proto $scheme;
|
||||
}
|
||||
|
||||
# Proxy API requests to stoic.lan (192.168.1.159)
|
||||
location /api/stoic/ {
|
||||
proxy_pass http://192.168.1.159:1234/v1/;
|
||||
proxy_set_header Host $host;
|
||||
proxy_set_header X-Real-IP $remote_addr;
|
||||
proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
|
||||
proxy_set_header X-Forwarded-Proto $scheme;
|
||||
}
|
||||
|
||||
# Default /api/ points to plato for backwards compatibility
|
||||
location /api/ {
|
||||
proxy_pass http://192.168.1.74:1234/v1/;
|
||||
proxy_set_header Host $host;
|
||||
|
||||
Reference in New Issue
Block a user