Spaces:
Sleeping
Sleeping
File size: 25,627 Bytes
c40c75a |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 |
import React, { useState, useEffect, useRef } from "react";
import ReactMarkdown from "react-markdown";
import {
Card,
Title,
Table,
TableHead,
TableRow,
TableCell,
TableBody,
Grid,
Tab,
TabGroup,
TabList,
TabPanel,
TabPanels,
Metric,
Col,
Text,
SelectItem,
TextInput,
Button,
Divider,
} from "@tremor/react";
import { v4 as uuidv4 } from 'uuid';
import { message, Select, Spin, Typography, Tooltip, Input } from "antd";
import { makeOpenAIChatCompletionRequest } from "./chat_ui/llm_calls/chat_completion";
import { makeOpenAIImageGenerationRequest } from "./chat_ui/llm_calls/image_generation";
import { makeOpenAIResponsesRequest } from "./chat_ui/llm_calls/responses_api";
import { fetchAvailableModels, ModelGroup } from "./chat_ui/llm_calls/fetch_models";
import { litellmModeMapping, ModelMode, EndpointType, getEndpointType } from "./chat_ui/mode_endpoint_mapping";
import { Prism as SyntaxHighlighter } from "react-syntax-highlighter";
import { coy } from 'react-syntax-highlighter/dist/esm/styles/prism';
import EndpointSelector from "./chat_ui/EndpointSelector";
import TagSelector from "./tag_management/TagSelector";
import VectorStoreSelector from "./vector_store_management/VectorStoreSelector";
import { determineEndpointType } from "./chat_ui/EndpointUtils";
import { MessageType } from "./chat_ui/types";
import ReasoningContent from "./chat_ui/ReasoningContent";
import ResponseMetrics, { TokenUsage } from "./chat_ui/ResponseMetrics";
import {
SendOutlined,
ApiOutlined,
KeyOutlined,
ClearOutlined,
RobotOutlined,
UserOutlined,
DeleteOutlined,
LoadingOutlined,
TagsOutlined,
DatabaseOutlined,
InfoCircleOutlined
} from "@ant-design/icons";
const { TextArea } = Input;
interface ChatUIProps {
accessToken: string | null;
token: string | null;
userRole: string | null;
userID: string | null;
disabledPersonalKeyCreation: boolean;
}
const ChatUI: React.FC<ChatUIProps> = ({
accessToken,
token,
userRole,
userID,
disabledPersonalKeyCreation,
}) => {
const [apiKeySource, setApiKeySource] = useState<'session' | 'custom'>(
disabledPersonalKeyCreation ? 'custom' : 'session'
);
const [apiKey, setApiKey] = useState("");
const [inputMessage, setInputMessage] = useState("");
const [chatHistory, setChatHistory] = useState<MessageType[]>([]);
const [selectedModel, setSelectedModel] = useState<string | undefined>(
undefined
);
const [showCustomModelInput, setShowCustomModelInput] = useState<boolean>(false);
const [modelInfo, setModelInfo] = useState<ModelGroup[]>([]);
const customModelTimeout = useRef<NodeJS.Timeout | null>(null);
const [endpointType, setEndpointType] = useState<string>(EndpointType.CHAT);
const [isLoading, setIsLoading] = useState<boolean>(false);
const abortControllerRef = useRef<AbortController | null>(null);
const [selectedTags, setSelectedTags] = useState<string[]>([]);
const [selectedVectorStores, setSelectedVectorStores] = useState<string[]>([]);
const [messageTraceId, setMessageTraceId] = useState<string | null>(null);
const chatEndRef = useRef<HTMLDivElement>(null);
useEffect(() => {
let userApiKey = apiKeySource === 'session' ? accessToken : apiKey;
if (!userApiKey || !token || !userRole || !userID) {
console.log("userApiKey or token or userRole or userID is missing = ", userApiKey, token, userRole, userID);
return;
}
// Fetch model info and set the default selected model
const loadModels = async () => {
try {
if (!userApiKey) {
console.log("userApiKey is missing");
return;
}
const uniqueModels = await fetchAvailableModels(
userApiKey,
);
console.log("Fetched models:", uniqueModels);
if (uniqueModels.length > 0) {
setModelInfo(uniqueModels);
setSelectedModel(uniqueModels[0].model_group);
// Auto-set endpoint based on the first model's mode
if (uniqueModels[0].mode) {
const initialEndpointType = determineEndpointType(uniqueModels[0].model_group, uniqueModels);
setEndpointType(initialEndpointType);
}
}
} catch (error) {
console.error("Error fetching model info:", error);
}
};
loadModels();
}, [accessToken, userID, userRole, apiKeySource, apiKey]);
useEffect(() => {
// Scroll to the bottom of the chat whenever chatHistory updates
if (chatEndRef.current) {
// Add a small delay to ensure content is rendered
setTimeout(() => {
chatEndRef.current?.scrollIntoView({
behavior: "smooth",
block: "end" // Keep the scroll position at the end
});
}, 100);
}
}, [chatHistory]);
const updateTextUI = (role: string, chunk: string, model?: string) => {
console.log("updateTextUI called with:", role, chunk, model);
setChatHistory((prev) => {
const last = prev[prev.length - 1];
// if the last message is already from this same role, append
if (last && last.role === role && !last.isImage) {
// build a new object, but only set `model` if it wasn't there already
const updated: MessageType = {
...last,
content: last.content + chunk,
model: last.model ?? model, // ← only use the passed‐in model on the first chunk
};
return [...prev.slice(0, -1), updated];
} else {
// otherwise start a brand new assistant bubble
return [
...prev,
{
role,
content: chunk,
model, // model set exactly once here
},
];
}
});
};
const updateReasoningContent = (chunk: string) => {
setChatHistory((prevHistory) => {
const lastMessage = prevHistory[prevHistory.length - 1];
if (lastMessage && lastMessage.role === "assistant" && !lastMessage.isImage) {
return [
...prevHistory.slice(0, prevHistory.length - 1),
{
...lastMessage,
reasoningContent: (lastMessage.reasoningContent || "") + chunk
},
];
} else {
// If there's no assistant message yet, we'll create one with empty content
// but with reasoning content
if (prevHistory.length > 0 && prevHistory[prevHistory.length - 1].role === "user") {
return [
...prevHistory,
{
role: "assistant",
content: "",
reasoningContent: chunk
}
];
}
return prevHistory;
}
});
};
const updateTimingData = (timeToFirstToken: number) => {
console.log("updateTimingData called with:", timeToFirstToken);
setChatHistory((prevHistory) => {
const lastMessage = prevHistory[prevHistory.length - 1];
console.log("Current last message:", lastMessage);
if (lastMessage && lastMessage.role === "assistant") {
console.log("Updating assistant message with timeToFirstToken:", timeToFirstToken);
const updatedHistory = [
...prevHistory.slice(0, prevHistory.length - 1),
{
...lastMessage,
timeToFirstToken
},
];
console.log("Updated chat history:", updatedHistory);
return updatedHistory;
}
// If the last message is a user message and no assistant message exists yet,
// create a new assistant message with empty content
else if (lastMessage && lastMessage.role === "user") {
console.log("Creating new assistant message with timeToFirstToken:", timeToFirstToken);
return [
...prevHistory,
{
role: "assistant",
content: "",
timeToFirstToken
}
];
}
console.log("No appropriate message found to update timing");
return prevHistory;
});
};
const updateUsageData = (usage: TokenUsage) => {
console.log("Received usage data:", usage);
setChatHistory((prevHistory) => {
const lastMessage = prevHistory[prevHistory.length - 1];
if (lastMessage && lastMessage.role === "assistant") {
console.log("Updating message with usage data:", usage);
const updatedMessage = {
...lastMessage,
usage
};
console.log("Updated message:", updatedMessage);
return [
...prevHistory.slice(0, prevHistory.length - 1),
updatedMessage
];
}
return prevHistory;
});
};
const updateImageUI = (imageUrl: string, model: string) => {
setChatHistory((prevHistory) => [
...prevHistory,
{ role: "assistant", content: imageUrl, model, isImage: true }
]);
};
const handleKeyDown = (event: React.KeyboardEvent<HTMLTextAreaElement>) => {
if (event.key === 'Enter' && !event.shiftKey) {
event.preventDefault(); // Prevent default to avoid newline
handleSendMessage();
}
// If Shift+Enter is pressed, the default behavior (inserting a newline) will occur
};
const handleCancelRequest = () => {
if (abortControllerRef.current) {
abortControllerRef.current.abort();
abortControllerRef.current = null;
setIsLoading(false);
message.info("Request cancelled");
}
};
const handleSendMessage = async () => {
if (inputMessage.trim() === "") return;
if (!token || !userRole || !userID) {
return;
}
const effectiveApiKey = apiKeySource === 'session' ? accessToken : apiKey;
if (!effectiveApiKey) {
message.error("Please provide an API key or select Current UI Session");
return;
}
// Create new abort controller for this request
abortControllerRef.current = new AbortController();
const signal = abortControllerRef.current.signal;
// Create message object without model field for API call
const newUserMessage = { role: "user", content: inputMessage };
// Generate new trace ID for a new conversation or use existing one
const traceId = messageTraceId || uuidv4();
if (!messageTraceId) {
setMessageTraceId(traceId);
}
// Update UI with full message object
setChatHistory([...chatHistory, newUserMessage]);
setIsLoading(true);
try {
if (selectedModel) {
if (endpointType === EndpointType.CHAT) {
// Create chat history for API call - strip out model field and isImage field
const apiChatHistory = [...chatHistory.filter(msg => !msg.isImage).map(({ role, content }) => ({ role, content })), newUserMessage];
await makeOpenAIChatCompletionRequest(
apiChatHistory,
(chunk, model) => updateTextUI("assistant", chunk, model),
selectedModel,
effectiveApiKey,
selectedTags,
signal,
updateReasoningContent,
updateTimingData,
updateUsageData,
traceId,
selectedVectorStores.length > 0 ? selectedVectorStores : undefined
);
} else if (endpointType === EndpointType.IMAGE) {
// For image generation
await makeOpenAIImageGenerationRequest(
inputMessage,
(imageUrl, model) => updateImageUI(imageUrl, model),
selectedModel,
effectiveApiKey,
selectedTags,
signal
);
} else if (endpointType === EndpointType.RESPONSES) {
// Create chat history for API call - strip out model field and isImage field
const apiChatHistory = [...chatHistory.filter(msg => !msg.isImage).map(({ role, content }) => ({ role, content })), newUserMessage];
await makeOpenAIResponsesRequest(
apiChatHistory,
(role, delta, model) => updateTextUI(role, delta, model),
selectedModel,
effectiveApiKey,
selectedTags,
signal,
updateReasoningContent,
updateTimingData,
updateUsageData,
traceId,
selectedVectorStores.length > 0 ? selectedVectorStores : undefined
);
}
}
} catch (error) {
if (signal.aborted) {
console.log("Request was cancelled");
} else {
console.error("Error fetching response", error);
updateTextUI("assistant", "Error fetching response");
}
} finally {
setIsLoading(false);
abortControllerRef.current = null;
}
setInputMessage("");
};
const clearChatHistory = () => {
setChatHistory([]);
setMessageTraceId(null);
message.success("Chat history cleared.");
};
if (userRole && userRole === "Admin Viewer") {
const { Title, Paragraph } = Typography;
return (
<div>
<Title level={1}>Access Denied</Title>
<Paragraph>Ask your proxy admin for access to test models</Paragraph>
</div>
);
}
const onModelChange = (value: string) => {
console.log(`selected ${value}`);
setSelectedModel(value);
// Use the utility function to determine the endpoint type
if (value !== 'custom') {
const newEndpointType = determineEndpointType(value, modelInfo);
setEndpointType(newEndpointType);
}
setShowCustomModelInput(value === 'custom');
};
const handleEndpointChange = (value: string) => {
setEndpointType(value);
};
const antIcon = <LoadingOutlined style={{ fontSize: 24 }} spin />;
return (
<div className="w-full h-screen p-4 bg-white">
<Card className="w-full rounded-xl shadow-md overflow-hidden">
<div className="flex h-[80vh] w-full">
{/* Left Sidebar with Controls */}
<div className="w-1/4 p-4 border-r border-gray-200 bg-gray-50">
<div className="mb-6">
<div className="space-y-6">
<div>
<Text className="font-medium block mb-2 text-gray-700 flex items-center">
<KeyOutlined className="mr-2" /> API Key Source
</Text>
<Select
disabled={disabledPersonalKeyCreation}
defaultValue="session"
style={{ width: "100%" }}
onChange={(value) => setApiKeySource(value as "session" | "custom")}
options={[
{ value: 'session', label: 'Current UI Session' },
{ value: 'custom', label: 'Virtual Key' },
]}
className="rounded-md"
/>
{apiKeySource === 'custom' && (
<TextInput
className="mt-2"
placeholder="Enter custom API key"
type="password"
onValueChange={setApiKey}
value={apiKey}
icon={KeyOutlined}
/>
)}
</div>
<div>
<Text className="font-medium block mb-2 text-gray-700 flex items-center">
<RobotOutlined className="mr-2" /> Select Model
</Text>
<Select
placeholder="Select a Model"
onChange={onModelChange}
options={[
...Array.from(new Set(modelInfo.map(option => option.model_group)))
.map((model_group, index) => ({
value: model_group,
label: model_group,
key: index
})),
{ value: 'custom', label: 'Enter custom model', key: 'custom' }
]}
style={{ width: "100%" }}
showSearch={true}
className="rounded-md"
/>
{showCustomModelInput && (
<TextInput
className="mt-2"
placeholder="Enter custom model name"
onValueChange={(value) => {
// Using setTimeout to create a simple debounce effect
if (customModelTimeout.current) {
clearTimeout(customModelTimeout.current);
}
customModelTimeout.current = setTimeout(() => {
setSelectedModel(value);
}, 500); // 500ms delay after typing stops
}}
/>
)}
</div>
<div>
<Text className="font-medium block mb-2 text-gray-700 flex items-center">
<ApiOutlined className="mr-2" /> Endpoint Type
</Text>
<EndpointSelector
endpointType={endpointType}
onEndpointChange={handleEndpointChange}
className="mb-4"
/>
</div>
<div>
<Text className="font-medium block mb-2 text-gray-700 flex items-center">
<TagsOutlined className="mr-2" /> Tags
</Text>
<TagSelector
value={selectedTags}
onChange={setSelectedTags}
className="mb-4"
accessToken={accessToken || ""}
/>
</div>
<div>
<Text className="font-medium block mb-2 text-gray-700 flex items-center">
<DatabaseOutlined className="mr-2" /> Vector Store
<Tooltip
className="ml-1"
title={
<span>
Select vector store(s) to use for this LLM API call. You can set up your vector store <a href="?page=vector-stores" style={{ color: '#1890ff' }}>here</a>.
</span>
}>
<InfoCircleOutlined />
</Tooltip>
</Text>
<VectorStoreSelector
value={selectedVectorStores}
onChange={setSelectedVectorStores}
className="mb-4"
accessToken={accessToken || ""}
/>
</div>
<Button
onClick={clearChatHistory}
className="w-full bg-gray-100 hover:bg-gray-200 text-gray-700 border-gray-300 mt-4"
icon={ClearOutlined}
>
Clear Chat
</Button>
</div>
</div>
</div>
{/* Main Chat Area */}
<div className="w-3/4 flex flex-col bg-white">
<div className="flex-1 overflow-auto p-4 pb-0">
{chatHistory.length === 0 && (
<div className="h-full flex flex-col items-center justify-center text-gray-400">
<RobotOutlined style={{ fontSize: '48px', marginBottom: '16px' }} />
<Text>Start a conversation or generate an image</Text>
</div>
)}
{chatHistory.map((message, index) => (
<div
key={index}
className={`mb-4 ${message.role === "user" ? "text-right" : "text-left"}`}
>
<div className="inline-block max-w-[80%] rounded-lg shadow-sm p-3.5 px-4" style={{
backgroundColor: message.role === "user" ? '#f0f8ff' : '#ffffff',
border: message.role === "user" ? '1px solid #e6f0fa' : '1px solid #f0f0f0',
textAlign: 'left'
}}>
<div className="flex items-center gap-2 mb-1.5">
<div className="flex items-center justify-center w-6 h-6 rounded-full mr-1" style={{
backgroundColor: message.role === "user" ? '#e6f0fa' : '#f5f5f5',
}}>
{message.role === "user" ?
<UserOutlined style={{ fontSize: '12px', color: '#2563eb' }} /> :
<RobotOutlined style={{ fontSize: '12px', color: '#4b5563' }} />
}
</div>
<strong className="text-sm capitalize">{message.role}</strong>
{message.role === "assistant" && message.model && (
<span className="text-xs px-2 py-0.5 rounded bg-gray-100 text-gray-600 font-normal">
{message.model}
</span>
)}
</div>
{message.reasoningContent && (
<ReasoningContent reasoningContent={message.reasoningContent} />
)}
<div className="whitespace-pre-wrap break-words max-w-full message-content"
style={{
wordWrap: 'break-word',
overflowWrap: 'break-word',
wordBreak: 'break-word',
hyphens: 'auto'
}}>
{message.isImage ? (
<img
src={message.content}
alt="Generated image"
className="max-w-full rounded-md border border-gray-200 shadow-sm"
style={{ maxHeight: '500px' }}
/>
) : (
<ReactMarkdown
components={{
code({node, inline, className, children, ...props}: React.ComponentPropsWithoutRef<'code'> & {
inline?: boolean;
node?: any;
}) {
const match = /language-(\w+)/.exec(className || '');
return !inline && match ? (
<SyntaxHighlighter
style={coy as any}
language={match[1]}
PreTag="div"
className="rounded-md my-2"
wrapLines={true}
wrapLongLines={true}
{...props}
>
{String(children).replace(/\n$/, '')}
</SyntaxHighlighter>
) : (
<code className={`${className} px-1.5 py-0.5 rounded bg-gray-100 text-sm font-mono`} style={{ wordBreak: 'break-word' }} {...props}>
{children}
</code>
);
},
pre: ({ node, ...props }) => (
<pre style={{ overflowX: 'auto', maxWidth: '100%' }} {...props} />
)
}}
>
{message.content}
</ReactMarkdown>
)}
{message.role === "assistant" && (message.timeToFirstToken || message.usage) && (
<ResponseMetrics
timeToFirstToken={message.timeToFirstToken}
usage={message.usage}
/>
)}
</div>
</div>
</div>
))}
{isLoading && (
<div className="flex justify-center items-center my-4">
<Spin indicator={antIcon} />
</div>
)}
<div ref={chatEndRef} style={{ height: "1px" }} />
</div>
<div className="p-4 border-t border-gray-200 bg-white">
<div className="flex items-center">
<TextArea
value={inputMessage}
onChange={(e) => setInputMessage(e.target.value)}
onKeyDown={handleKeyDown}
placeholder={
endpointType === EndpointType.CHAT || endpointType === EndpointType.RESPONSES
? "Type your message... (Shift+Enter for new line)"
: "Describe the image you want to generate..."
}
disabled={isLoading}
className="flex-1"
autoSize={{ minRows: 1, maxRows: 6 }}
style={{ resize: 'none', paddingRight: '10px', paddingLeft: '10px' }}
/>
{isLoading ? (
<Button
onClick={handleCancelRequest}
className="ml-2 bg-red-50 hover:bg-red-100 text-red-600 border-red-200"
icon={DeleteOutlined}
>
Cancel
</Button>
) : (
<Button
onClick={handleSendMessage}
className="ml-2 text-white"
icon={endpointType === EndpointType.CHAT ? SendOutlined : RobotOutlined}
>
{endpointType === EndpointType.CHAT ? "Send" : "Generate"}
</Button>
)}
</div>
</div>
</div>
</div>
</Card>
</div>
);
};
export default ChatUI;
|