mirror of
https://github.com/LostRuins/koboldcpp.git
synced 2026-09-13 10:18:49 +02:00
Merge commit '5016b7286240d29f8f640039989b84ea3a854344' into concedo_experimental
# Conflicts: # .github/workflows/build.yml # docs/ops.md # docs/ops/SYCL.csv # ggml/src/ggml-cann/aclnn_ops.cpp # ggml/src/ggml-opencl/ggml-opencl.cpp # ggml/src/ggml-sycl/backend.hpp # ggml/src/ggml-sycl/binbcast.cpp # ggml/src/ggml-sycl/binbcast.hpp # ggml/src/ggml-sycl/common.hpp # ggml/src/ggml-sycl/element_wise.cpp # ggml/src/ggml-sycl/element_wise.hpp # ggml/src/ggml-sycl/ggml-sycl.cpp # tests/test-chat-parser.cpp # tests/test-json-partial.cpp
This commit is contained in:
Concedo
Binary file not shown.
Generated
+69
@@ -50,6 +50,7 @@
|
||||
"eslint-plugin-svelte": "^3.0.0",
|
||||
"fflate": "^0.8.2",
|
||||
"globals": "^16.0.0",
|
||||
"mdast": "^3.0.0",
|
||||
"mdsvex": "^0.12.3",
|
||||
"playwright": "^1.53.0",
|
||||
"prettier": "^3.4.2",
|
||||
@@ -66,6 +67,7 @@
|
||||
"tw-animate-css": "^1.3.5",
|
||||
"typescript": "^5.0.0",
|
||||
"typescript-eslint": "^8.20.0",
|
||||
"unified": "^11.0.5",
|
||||
"uuid": "^13.0.0",
|
||||
"vite": "^7.0.4",
|
||||
"vite-plugin-devtools-json": "^0.2.0",
|
||||
@@ -2128,6 +2130,66 @@
|
||||
"node": ">=14.0.0"
|
||||
}
|
||||
},
|
||||
"node_modules/@tailwindcss/oxide-wasm32-wasi/node_modules/@emnapi/core": {
|
||||
"version": "1.4.3",
|
||||
"dev": true,
|
||||
"inBundle": true,
|
||||
"license": "MIT",
|
||||
"optional": true,
|
||||
"dependencies": {
|
||||
"@emnapi/wasi-threads": "1.0.2",
|
||||
"tslib": "^2.4.0"
|
||||
}
|
||||
},
|
||||
"node_modules/@tailwindcss/oxide-wasm32-wasi/node_modules/@emnapi/runtime": {
|
||||
"version": "1.4.3",
|
||||
"dev": true,
|
||||
"inBundle": true,
|
||||
"license": "MIT",
|
||||
"optional": true,
|
||||
"dependencies": {
|
||||
"tslib": "^2.4.0"
|
||||
}
|
||||
},
|
||||
"node_modules/@tailwindcss/oxide-wasm32-wasi/node_modules/@emnapi/wasi-threads": {
|
||||
"version": "1.0.2",
|
||||
"dev": true,
|
||||
"inBundle": true,
|
||||
"license": "MIT",
|
||||
"optional": true,
|
||||
"dependencies": {
|
||||
"tslib": "^2.4.0"
|
||||
}
|
||||
},
|
||||
"node_modules/@tailwindcss/oxide-wasm32-wasi/node_modules/@napi-rs/wasm-runtime": {
|
||||
"version": "0.2.11",
|
||||
"dev": true,
|
||||
"inBundle": true,
|
||||
"license": "MIT",
|
||||
"optional": true,
|
||||
"dependencies": {
|
||||
"@emnapi/core": "^1.4.3",
|
||||
"@emnapi/runtime": "^1.4.3",
|
||||
"@tybys/wasm-util": "^0.9.0"
|
||||
}
|
||||
},
|
||||
"node_modules/@tailwindcss/oxide-wasm32-wasi/node_modules/@tybys/wasm-util": {
|
||||
"version": "0.9.0",
|
||||
"dev": true,
|
||||
"inBundle": true,
|
||||
"license": "MIT",
|
||||
"optional": true,
|
||||
"dependencies": {
|
||||
"tslib": "^2.4.0"
|
||||
}
|
||||
},
|
||||
"node_modules/@tailwindcss/oxide-wasm32-wasi/node_modules/tslib": {
|
||||
"version": "2.8.0",
|
||||
"dev": true,
|
||||
"inBundle": true,
|
||||
"license": "0BSD",
|
||||
"optional": true
|
||||
},
|
||||
"node_modules/@tailwindcss/oxide-win32-arm64-msvc": {
|
||||
"version": "4.1.11",
|
||||
"resolved": "https://registry.npmjs.org/@tailwindcss/oxide-win32-arm64-msvc/-/oxide-win32-arm64-msvc-4.1.11.tgz",
|
||||
@@ -4946,6 +5008,13 @@
|
||||
"url": "https://github.com/sponsors/wooorm"
|
||||
}
|
||||
},
|
||||
"node_modules/mdast": {
|
||||
"version": "3.0.0",
|
||||
"resolved": "https://registry.npmjs.org/mdast/-/mdast-3.0.0.tgz",
|
||||
"integrity": "sha512-xySmf8g4fPKMeC07jXGz971EkLbWAJ83s4US2Tj9lEdnZ142UP5grN73H1Xd3HzrdbU5o9GYYP/y8F9ZSwLE9g==",
|
||||
"dev": true,
|
||||
"license": "MIT"
|
||||
},
|
||||
"node_modules/mdast-util-find-and-replace": {
|
||||
"version": "3.0.2",
|
||||
"resolved": "https://registry.npmjs.org/mdast-util-find-and-replace/-/mdast-util-find-and-replace-3.0.2.tgz",
|
||||
|
||||
@@ -52,6 +52,7 @@
|
||||
"eslint-plugin-svelte": "^3.0.0",
|
||||
"fflate": "^0.8.2",
|
||||
"globals": "^16.0.0",
|
||||
"mdast": "^3.0.0",
|
||||
"mdsvex": "^0.12.3",
|
||||
"playwright": "^1.53.0",
|
||||
"prettier": "^3.4.2",
|
||||
@@ -68,6 +69,7 @@
|
||||
"tw-animate-css": "^1.3.5",
|
||||
"typescript": "^5.0.0",
|
||||
"typescript-eslint": "^8.20.0",
|
||||
"unified": "^11.0.5",
|
||||
"uuid": "^13.0.0",
|
||||
"vite": "^7.0.4",
|
||||
"vite-plugin-devtools-json": "^0.2.0",
|
||||
|
||||
@@ -7,6 +7,7 @@
|
||||
ChatMessages,
|
||||
ChatProcessingInfo,
|
||||
EmptyFileAlertDialog,
|
||||
ChatErrorDialog,
|
||||
ServerErrorSplash,
|
||||
ServerInfo,
|
||||
ServerLoadingSplash,
|
||||
@@ -22,10 +23,11 @@
|
||||
activeMessages,
|
||||
activeConversation,
|
||||
deleteConversation,
|
||||
dismissErrorDialog,
|
||||
errorDialog,
|
||||
isLoading,
|
||||
sendMessage,
|
||||
stopGeneration,
|
||||
setMaxContextError
|
||||
stopGeneration
|
||||
} from '$lib/stores/chat.svelte';
|
||||
import {
|
||||
supportsVision,
|
||||
@@ -34,7 +36,6 @@
|
||||
serverWarning,
|
||||
serverStore
|
||||
} from '$lib/stores/server.svelte';
|
||||
import { contextService } from '$lib/services';
|
||||
import { parseFilesToMessageExtras } from '$lib/utils/convert-files-to-extra';
|
||||
import { isFileTypeSupported } from '$lib/utils/file-type';
|
||||
import { filterFilesByModalities } from '$lib/utils/modality-file-validation';
|
||||
@@ -79,6 +80,7 @@
|
||||
showCenteredEmpty && !activeConversation() && activeMessages().length === 0 && !isLoading()
|
||||
);
|
||||
|
||||
let activeErrorDialog = $derived(errorDialog());
|
||||
let isServerLoading = $derived(serverLoading());
|
||||
|
||||
async function handleDeleteConfirm() {
|
||||
@@ -105,6 +107,12 @@
|
||||
}
|
||||
}
|
||||
|
||||
function handleErrorDialogOpenChange(open: boolean) {
|
||||
if (!open) {
|
||||
dismissErrorDialog();
|
||||
}
|
||||
}
|
||||
|
||||
function handleDragOver(event: DragEvent) {
|
||||
event.preventDefault();
|
||||
}
|
||||
@@ -183,21 +191,6 @@
|
||||
|
||||
const extras = result?.extras;
|
||||
|
||||
// Check context limit using real-time slots data
|
||||
const contextCheck = await contextService.checkContextLimit();
|
||||
|
||||
if (contextCheck && contextCheck.wouldExceed) {
|
||||
const errorMessage = contextService.getContextErrorMessage(contextCheck);
|
||||
|
||||
setMaxContextError({
|
||||
message: errorMessage,
|
||||
estimatedTokens: contextCheck.currentUsage,
|
||||
maxContext: contextCheck.maxContext
|
||||
});
|
||||
|
||||
return false;
|
||||
}
|
||||
|
||||
// Enable autoscroll for user-initiated message sending
|
||||
userScrolledUp = false;
|
||||
autoScrollEnabled = true;
|
||||
@@ -461,6 +454,13 @@
|
||||
}}
|
||||
/>
|
||||
|
||||
<ChatErrorDialog
|
||||
message={activeErrorDialog?.message ?? ''}
|
||||
onOpenChange={handleErrorDialogOpenChange}
|
||||
open={Boolean(activeErrorDialog)}
|
||||
type={activeErrorDialog?.type ?? 'server'}
|
||||
/>
|
||||
|
||||
<style>
|
||||
.conversation-chat-form {
|
||||
position: relative;
|
||||
|
||||
@@ -0,0 +1,60 @@
|
||||
<script lang="ts">
|
||||
import * as AlertDialog from '$lib/components/ui/alert-dialog';
|
||||
import { AlertTriangle, TimerOff } from '@lucide/svelte';
|
||||
|
||||
interface Props {
|
||||
open: boolean;
|
||||
type: 'timeout' | 'server';
|
||||
message: string;
|
||||
onOpenChange?: (open: boolean) => void;
|
||||
}
|
||||
|
||||
let { open = $bindable(), type, message, onOpenChange }: Props = $props();
|
||||
|
||||
const isTimeout = $derived(type === 'timeout');
|
||||
const title = $derived(isTimeout ? 'TCP Timeout' : 'Server Error');
|
||||
const description = $derived(
|
||||
isTimeout
|
||||
? 'The request did not receive a response from the server before timing out.'
|
||||
: 'The server responded with an error message. Review the details below.'
|
||||
);
|
||||
const iconClass = $derived(isTimeout ? 'text-destructive' : 'text-amber-500');
|
||||
const badgeClass = $derived(
|
||||
isTimeout
|
||||
? 'border-destructive/40 bg-destructive/10 text-destructive'
|
||||
: 'border-amber-500/40 bg-amber-500/10 text-amber-600 dark:text-amber-400'
|
||||
);
|
||||
|
||||
function handleOpenChange(newOpen: boolean) {
|
||||
open = newOpen;
|
||||
onOpenChange?.(newOpen);
|
||||
}
|
||||
</script>
|
||||
|
||||
<AlertDialog.Root {open} onOpenChange={handleOpenChange}>
|
||||
<AlertDialog.Content>
|
||||
<AlertDialog.Header>
|
||||
<AlertDialog.Title class="flex items-center gap-2">
|
||||
{#if isTimeout}
|
||||
<TimerOff class={`h-5 w-5 ${iconClass}`} />
|
||||
{:else}
|
||||
<AlertTriangle class={`h-5 w-5 ${iconClass}`} />
|
||||
{/if}
|
||||
|
||||
{title}
|
||||
</AlertDialog.Title>
|
||||
|
||||
<AlertDialog.Description>
|
||||
{description}
|
||||
</AlertDialog.Description>
|
||||
</AlertDialog.Header>
|
||||
|
||||
<div class={`rounded-lg border px-4 py-3 text-sm ${badgeClass}`}>
|
||||
<p class="font-medium">{message}</p>
|
||||
</div>
|
||||
|
||||
<AlertDialog.Footer>
|
||||
<AlertDialog.Action onclick={() => handleOpenChange(false)}>Close</AlertDialog.Action>
|
||||
</AlertDialog.Footer>
|
||||
</AlertDialog.Content>
|
||||
</AlertDialog.Root>
|
||||
@@ -1,66 +0,0 @@
|
||||
<script lang="ts">
|
||||
import { AlertTriangle } from '@lucide/svelte';
|
||||
import * as AlertDialog from '$lib/components/ui/alert-dialog';
|
||||
import { maxContextError, clearMaxContextError } from '$lib/stores/chat.svelte';
|
||||
</script>
|
||||
|
||||
<AlertDialog.Root
|
||||
open={maxContextError() !== null}
|
||||
onOpenChange={(open) => !open && clearMaxContextError()}
|
||||
>
|
||||
<AlertDialog.Content>
|
||||
<AlertDialog.Header>
|
||||
<AlertDialog.Title class="flex items-center gap-2">
|
||||
<AlertTriangle class="h-5 w-5 text-destructive" />
|
||||
|
||||
Message Too Long
|
||||
</AlertDialog.Title>
|
||||
|
||||
<AlertDialog.Description>
|
||||
Your message exceeds the model's context window and cannot be processed.
|
||||
</AlertDialog.Description>
|
||||
</AlertDialog.Header>
|
||||
|
||||
{#if maxContextError()}
|
||||
<div class="space-y-3 text-sm">
|
||||
<div class="rounded-lg bg-muted p-3">
|
||||
<div class="mb-2 font-medium">Token Usage:</div>
|
||||
|
||||
<div class="space-y-1 text-muted-foreground">
|
||||
<div>
|
||||
Estimated tokens:
|
||||
|
||||
<span class="font-mono">
|
||||
{maxContextError()?.estimatedTokens.toLocaleString()}
|
||||
</span>
|
||||
</div>
|
||||
|
||||
<div>
|
||||
Context window:
|
||||
|
||||
<span class="font-mono">
|
||||
{maxContextError()?.maxContext.toLocaleString()}
|
||||
</span>
|
||||
</div>
|
||||
</div>
|
||||
</div>
|
||||
|
||||
<div>
|
||||
<div class="mb-2 font-medium">Suggestions:</div>
|
||||
|
||||
<ul class="list-inside list-disc space-y-1 text-muted-foreground">
|
||||
<li>Shorten your message</li>
|
||||
|
||||
<li>Remove some file attachments</li>
|
||||
|
||||
<li>Start a new conversation</li>
|
||||
</ul>
|
||||
</div>
|
||||
</div>
|
||||
{/if}
|
||||
|
||||
<AlertDialog.Footer>
|
||||
<AlertDialog.Action onclick={() => clearMaxContextError()}>Got it</AlertDialog.Action>
|
||||
</AlertDialog.Footer>
|
||||
</AlertDialog.Content>
|
||||
</AlertDialog.Root>
|
||||
@@ -30,12 +30,11 @@ export { default as ChatSidebar } from './chat/ChatSidebar/ChatSidebar.svelte';
|
||||
export { default as ChatSidebarConversationItem } from './chat/ChatSidebar/ChatSidebarConversationItem.svelte';
|
||||
export { default as ChatSidebarSearch } from './chat/ChatSidebar/ChatSidebarSearch.svelte';
|
||||
|
||||
export { default as ChatErrorDialog } from './dialogs/ChatErrorDialog.svelte';
|
||||
export { default as EmptyFileAlertDialog } from './dialogs/EmptyFileAlertDialog.svelte';
|
||||
|
||||
export { default as ConversationTitleUpdateDialog } from './dialogs/ConversationTitleUpdateDialog.svelte';
|
||||
|
||||
export { default as MaximumContextAlertDialog } from './dialogs/MaximumContextAlertDialog.svelte';
|
||||
|
||||
export { default as KeyboardShortcutInfo } from './misc/KeyboardShortcutInfo.svelte';
|
||||
|
||||
export { default as MarkdownContent } from './misc/MarkdownContent.svelte';
|
||||
|
||||
@@ -14,6 +14,7 @@
|
||||
import githubDarkCss from 'highlight.js/styles/github-dark.css?inline';
|
||||
import githubLightCss from 'highlight.js/styles/github.css?inline';
|
||||
import { mode } from 'mode-watcher';
|
||||
import { remarkLiteralHtml } from '$lib/markdown/literal-html';
|
||||
|
||||
interface Props {
|
||||
content: string;
|
||||
@@ -50,36 +51,59 @@
|
||||
.use(remarkGfm) // GitHub Flavored Markdown
|
||||
.use(remarkMath) // Parse $inline$ and $$block$$ math
|
||||
.use(remarkBreaks) // Convert line breaks to <br>
|
||||
.use(remarkRehype) // Convert to rehype (HTML AST)
|
||||
.use(remarkLiteralHtml) // Treat raw HTML as literal text with preserved indentation
|
||||
.use(remarkRehype) // Convert Markdown AST to rehype
|
||||
.use(rehypeKatex) // Render math using KaTeX
|
||||
.use(rehypeHighlight) // Add syntax highlighting
|
||||
.use(rehypeStringify); // Convert to HTML string
|
||||
});
|
||||
|
||||
function enhanceLinks(html: string): string {
|
||||
if (!html.includes('<a')) {
|
||||
return html;
|
||||
}
|
||||
|
||||
const tempDiv = document.createElement('div');
|
||||
tempDiv.innerHTML = html;
|
||||
|
||||
// Make all links open in new tabs
|
||||
const linkElements = tempDiv.querySelectorAll('a[href]');
|
||||
let mutated = false;
|
||||
|
||||
for (const link of linkElements) {
|
||||
const target = link.getAttribute('target');
|
||||
const rel = link.getAttribute('rel');
|
||||
|
||||
if (target !== '_blank' || rel !== 'noopener noreferrer') {
|
||||
mutated = true;
|
||||
}
|
||||
|
||||
link.setAttribute('target', '_blank');
|
||||
link.setAttribute('rel', 'noopener noreferrer');
|
||||
}
|
||||
|
||||
return tempDiv.innerHTML;
|
||||
return mutated ? tempDiv.innerHTML : html;
|
||||
}
|
||||
|
||||
function enhanceCodeBlocks(html: string): string {
|
||||
if (!html.includes('<pre')) {
|
||||
return html;
|
||||
}
|
||||
|
||||
const tempDiv = document.createElement('div');
|
||||
tempDiv.innerHTML = html;
|
||||
|
||||
const preElements = tempDiv.querySelectorAll('pre');
|
||||
let mutated = false;
|
||||
|
||||
for (const [index, pre] of Array.from(preElements).entries()) {
|
||||
const codeElement = pre.querySelector('code');
|
||||
|
||||
if (!codeElement) continue;
|
||||
if (!codeElement) {
|
||||
continue;
|
||||
}
|
||||
|
||||
mutated = true;
|
||||
|
||||
let language = 'text';
|
||||
const classList = Array.from(codeElement.classList);
|
||||
@@ -127,7 +151,7 @@
|
||||
pre.parentNode?.replaceChild(wrapper, pre);
|
||||
}
|
||||
|
||||
return tempDiv.innerHTML;
|
||||
return mutated ? tempDiv.innerHTML : html;
|
||||
}
|
||||
|
||||
async function processMarkdown(text: string): Promise<string> {
|
||||
|
||||
@@ -0,0 +1,15 @@
|
||||
export const LINE_BREAK = /\r?\n/;
|
||||
|
||||
export const PHRASE_PARENTS = new Set([
|
||||
'paragraph',
|
||||
'heading',
|
||||
'emphasis',
|
||||
'strong',
|
||||
'delete',
|
||||
'link',
|
||||
'linkReference',
|
||||
'tableCell'
|
||||
]);
|
||||
|
||||
export const NBSP = '\u00a0';
|
||||
export const TAB_AS_SPACES = NBSP.repeat(4);
|
||||
@@ -0,0 +1,121 @@
|
||||
import type { Plugin } from 'unified';
|
||||
import { visit } from 'unist-util-visit';
|
||||
import type { Break, Content, Paragraph, PhrasingContent, Root, Text } from 'mdast';
|
||||
import { LINE_BREAK, NBSP, PHRASE_PARENTS, TAB_AS_SPACES } from '$lib/constants/literal-html';
|
||||
|
||||
/**
|
||||
* remark plugin that rewrites raw HTML nodes into plain-text equivalents.
|
||||
*
|
||||
* remark parses inline HTML into `html` nodes even when we do not want to render
|
||||
* them. We turn each of those nodes into regular text (plus `<br>` break markers)
|
||||
* so the downstream rehype pipeline escapes the characters instead of executing
|
||||
* them. Leading spaces and tab characters are converted to non‑breaking spaces to
|
||||
* keep indentation identical to the original author input.
|
||||
*/
|
||||
|
||||
function preserveIndent(line: string): string {
|
||||
let index = 0;
|
||||
let output = '';
|
||||
|
||||
while (index < line.length) {
|
||||
const char = line[index];
|
||||
|
||||
if (char === ' ') {
|
||||
output += NBSP;
|
||||
index += 1;
|
||||
continue;
|
||||
}
|
||||
|
||||
if (char === '\t') {
|
||||
output += TAB_AS_SPACES;
|
||||
index += 1;
|
||||
continue;
|
||||
}
|
||||
|
||||
break;
|
||||
}
|
||||
|
||||
return output + line.slice(index);
|
||||
}
|
||||
|
||||
function createLiteralChildren(value: string): PhrasingContent[] {
|
||||
const lines = value.split(LINE_BREAK);
|
||||
const nodes: PhrasingContent[] = [];
|
||||
|
||||
for (const [lineIndex, rawLine] of lines.entries()) {
|
||||
if (lineIndex > 0) {
|
||||
nodes.push({ type: 'break' } as Break as unknown as PhrasingContent);
|
||||
}
|
||||
|
||||
nodes.push({
|
||||
type: 'text',
|
||||
value: preserveIndent(rawLine)
|
||||
} as Text as unknown as PhrasingContent);
|
||||
}
|
||||
|
||||
if (!nodes.length) {
|
||||
nodes.push({ type: 'text', value: '' } as Text as unknown as PhrasingContent);
|
||||
}
|
||||
|
||||
return nodes;
|
||||
}
|
||||
|
||||
export const remarkLiteralHtml: Plugin<[], Root> = () => {
|
||||
return (tree) => {
|
||||
visit(tree, 'html', (node, index, parent) => {
|
||||
if (!parent || typeof index !== 'number') {
|
||||
return;
|
||||
}
|
||||
|
||||
const replacement = createLiteralChildren(node.value);
|
||||
|
||||
if (!PHRASE_PARENTS.has(parent.type as string)) {
|
||||
const paragraph: Paragraph = {
|
||||
type: 'paragraph',
|
||||
children: replacement as Paragraph['children'],
|
||||
data: { literalHtml: true }
|
||||
};
|
||||
|
||||
const siblings = parent.children as unknown as Content[];
|
||||
siblings.splice(index, 1, paragraph as unknown as Content);
|
||||
|
||||
if (index > 0) {
|
||||
const previous = siblings[index - 1] as Paragraph | undefined;
|
||||
|
||||
if (
|
||||
previous?.type === 'paragraph' &&
|
||||
(previous.data as { literalHtml?: boolean } | undefined)?.literalHtml
|
||||
) {
|
||||
const prevChildren = previous.children as unknown as PhrasingContent[];
|
||||
|
||||
if (prevChildren.length) {
|
||||
const lastChild = prevChildren[prevChildren.length - 1];
|
||||
|
||||
if (lastChild.type !== 'break') {
|
||||
prevChildren.push({
|
||||
type: 'break'
|
||||
} as Break as unknown as PhrasingContent);
|
||||
}
|
||||
}
|
||||
|
||||
prevChildren.push(...(paragraph.children as unknown as PhrasingContent[]));
|
||||
|
||||
siblings.splice(index, 1);
|
||||
|
||||
return index;
|
||||
}
|
||||
}
|
||||
|
||||
return index + 1;
|
||||
}
|
||||
|
||||
(parent.children as unknown as PhrasingContent[]).splice(
|
||||
index,
|
||||
1,
|
||||
...(replacement as unknown as PhrasingContent[])
|
||||
);
|
||||
|
||||
return index + replacement.length;
|
||||
});
|
||||
};
|
||||
};
|
||||
@@ -13,7 +13,7 @@ import { slotsService } from './slots';
|
||||
* - Manages streaming and non-streaming response parsing
|
||||
* - Provides request abortion capabilities
|
||||
* - Converts database messages to API format
|
||||
* - Handles error translation and context detection
|
||||
* - Handles error translation for server responses
|
||||
*
|
||||
* - **ChatStore**: Stateful orchestration and UI state management
|
||||
* - Uses ChatService for all AI model communication
|
||||
@@ -26,7 +26,6 @@ import { slotsService } from './slots';
|
||||
* - Streaming response handling with real-time callbacks
|
||||
* - Reasoning content extraction and processing
|
||||
* - File attachment processing (images, PDFs, audio, text)
|
||||
* - Context error detection and reporting
|
||||
* - Request lifecycle management (abort, cleanup)
|
||||
*/
|
||||
export class ChatService {
|
||||
@@ -209,10 +208,13 @@ export class ChatService {
|
||||
userFriendlyError = new Error(
|
||||
'Unable to connect to server - please check if the server is running'
|
||||
);
|
||||
userFriendlyError.name = 'NetworkError';
|
||||
} else if (error.message.includes('ECONNREFUSED')) {
|
||||
userFriendlyError = new Error('Connection refused - server may be offline');
|
||||
userFriendlyError.name = 'NetworkError';
|
||||
} else if (error.message.includes('ETIMEDOUT')) {
|
||||
userFriendlyError = new Error('Request timeout - server may be overloaded');
|
||||
userFriendlyError = new Error('Request timed out - the server took too long to respond');
|
||||
userFriendlyError.name = 'TimeoutError';
|
||||
} else {
|
||||
userFriendlyError = error;
|
||||
}
|
||||
@@ -262,6 +264,7 @@ export class ChatService {
|
||||
let fullReasoningContent = '';
|
||||
let hasReceivedData = false;
|
||||
let lastTimings: ChatMessageTimings | undefined;
|
||||
let streamFinished = false;
|
||||
|
||||
try {
|
||||
let chunk = '';
|
||||
@@ -277,18 +280,8 @@ export class ChatService {
|
||||
if (line.startsWith('data: ')) {
|
||||
const data = line.slice(6);
|
||||
if (data === '[DONE]') {
|
||||
if (!hasReceivedData && aggregatedContent.length === 0) {
|
||||
const contextError = new Error(
|
||||
'The request exceeds the available context size. Try increasing the context size or enable context shift.'
|
||||
);
|
||||
contextError.name = 'ContextError';
|
||||
onError?.(contextError);
|
||||
return;
|
||||
}
|
||||
|
||||
onComplete?.(aggregatedContent, fullReasoningContent || undefined, lastTimings);
|
||||
|
||||
return;
|
||||
streamFinished = true;
|
||||
continue;
|
||||
}
|
||||
|
||||
try {
|
||||
@@ -326,13 +319,13 @@ export class ChatService {
|
||||
}
|
||||
}
|
||||
|
||||
if (!hasReceivedData && aggregatedContent.length === 0) {
|
||||
const contextError = new Error(
|
||||
'The request exceeds the available context size. Try increasing the context size or enable context shift.'
|
||||
);
|
||||
contextError.name = 'ContextError';
|
||||
onError?.(contextError);
|
||||
return;
|
||||
if (streamFinished) {
|
||||
if (!hasReceivedData && aggregatedContent.length === 0) {
|
||||
const noResponseError = new Error('No response received from server. Please try again.');
|
||||
throw noResponseError;
|
||||
}
|
||||
|
||||
onComplete?.(aggregatedContent, fullReasoningContent || undefined, lastTimings);
|
||||
}
|
||||
} catch (error) {
|
||||
const err = error instanceof Error ? error : new Error('Stream error');
|
||||
@@ -368,12 +361,8 @@ export class ChatService {
|
||||
const responseText = await response.text();
|
||||
|
||||
if (!responseText.trim()) {
|
||||
const contextError = new Error(
|
||||
'The request exceeds the available context size. Try increasing the context size or enable context shift.'
|
||||
);
|
||||
contextError.name = 'ContextError';
|
||||
onError?.(contextError);
|
||||
throw contextError;
|
||||
const noResponseError = new Error('No response received from server. Please try again.');
|
||||
throw noResponseError;
|
||||
}
|
||||
|
||||
const data: ApiChatCompletionResponse = JSON.parse(responseText);
|
||||
@@ -385,22 +374,14 @@ export class ChatService {
|
||||
}
|
||||
|
||||
if (!content.trim()) {
|
||||
const contextError = new Error(
|
||||
'The request exceeds the available context size. Try increasing the context size or enable context shift.'
|
||||
);
|
||||
contextError.name = 'ContextError';
|
||||
onError?.(contextError);
|
||||
throw contextError;
|
||||
const noResponseError = new Error('No response received from server. Please try again.');
|
||||
throw noResponseError;
|
||||
}
|
||||
|
||||
onComplete?.(content, reasoningContent);
|
||||
|
||||
return content;
|
||||
} catch (error) {
|
||||
if (error instanceof Error && error.name === 'ContextError') {
|
||||
throw error;
|
||||
}
|
||||
|
||||
const err = error instanceof Error ? error : new Error('Parse error');
|
||||
|
||||
onError?.(err);
|
||||
@@ -594,37 +575,19 @@ export class ChatService {
|
||||
const errorText = await response.text();
|
||||
const errorData: ApiErrorResponse = JSON.parse(errorText);
|
||||
|
||||
if (errorData.error?.type === 'exceed_context_size_error') {
|
||||
const contextError = errorData.error as ApiContextSizeError;
|
||||
const error = new Error(contextError.message);
|
||||
error.name = 'ContextError';
|
||||
// Attach structured context information
|
||||
(
|
||||
error as Error & {
|
||||
contextInfo?: { promptTokens: number; maxContext: number; estimatedTokens: number };
|
||||
}
|
||||
).contextInfo = {
|
||||
promptTokens: contextError.n_prompt_tokens,
|
||||
maxContext: contextError.n_ctx,
|
||||
estimatedTokens: contextError.n_prompt_tokens
|
||||
};
|
||||
return error;
|
||||
}
|
||||
|
||||
// Fallback for other error types
|
||||
const message = errorData.error?.message || 'Unknown server error';
|
||||
return new Error(message);
|
||||
const error = new Error(message);
|
||||
error.name = response.status === 400 ? 'ServerError' : 'HttpError';
|
||||
|
||||
return error;
|
||||
} catch {
|
||||
// If we can't parse the error response, return a generic error
|
||||
return new Error(`Server error (${response.status}): ${response.statusText}`);
|
||||
const fallback = new Error(`Server error (${response.status}): ${response.statusText}`);
|
||||
fallback.name = 'HttpError';
|
||||
return fallback;
|
||||
}
|
||||
}
|
||||
|
||||
/**
|
||||
* Updates the processing state with timing information from the server response
|
||||
* @param timings - Timing data from the API response
|
||||
* @param promptProgress - Progress data from the API response
|
||||
*/
|
||||
private updateProcessingState(
|
||||
timings?: ChatMessageTimings,
|
||||
promptProgress?: ChatMessagePromptProgress
|
||||
|
||||
@@ -1,102 +0,0 @@
|
||||
import { slotsService } from './slots';
|
||||
|
||||
export interface ContextCheckResult {
|
||||
wouldExceed: boolean;
|
||||
currentUsage: number;
|
||||
maxContext: number;
|
||||
availableTokens: number;
|
||||
reservedTokens: number;
|
||||
}
|
||||
|
||||
/**
|
||||
* ContextService - Context window management and limit checking
|
||||
*
|
||||
* This service provides context window monitoring and limit checking using real-time
|
||||
* server data from the slots service. It helps prevent context overflow by tracking
|
||||
* current usage and calculating available space for new content.
|
||||
*
|
||||
* **Architecture & Relationships:**
|
||||
* - **ContextService** (this class): Context limit monitoring
|
||||
* - Uses SlotsService for real-time context usage data
|
||||
* - Calculates available tokens with configurable reserves
|
||||
* - Provides context limit checking and error messaging
|
||||
* - Helps prevent context window overflow
|
||||
*
|
||||
* - **SlotsService**: Provides current context usage from server slots
|
||||
* - **ChatStore**: Uses context checking before sending messages
|
||||
* - **UI Components**: Display context usage warnings and limits
|
||||
*
|
||||
* **Key Features:**
|
||||
* - **Real-time Context Checking**: Uses live server data for accuracy
|
||||
* - **Token Reservation**: Reserves tokens for response generation
|
||||
* - **Limit Detection**: Prevents context window overflow
|
||||
* - **Usage Reporting**: Detailed context usage statistics
|
||||
* - **Error Messaging**: User-friendly context limit messages
|
||||
* - **Configurable Reserves**: Adjustable token reservation for responses
|
||||
*
|
||||
* **Context Management:**
|
||||
* - Monitors current context usage from active slots
|
||||
* - Calculates available space considering reserved tokens
|
||||
* - Provides early warning before context limits are reached
|
||||
* - Helps optimize conversation length and content
|
||||
*/
|
||||
export class ContextService {
|
||||
private reserveTokens: number;
|
||||
|
||||
constructor(reserveTokens = 512) {
|
||||
this.reserveTokens = reserveTokens;
|
||||
}
|
||||
|
||||
/**
|
||||
* Checks if the context limit would be exceeded
|
||||
*
|
||||
* @returns {Promise<ContextCheckResult | null>} Promise that resolves to the context check result or null if an error occurs
|
||||
*/
|
||||
async checkContextLimit(): Promise<ContextCheckResult | null> {
|
||||
try {
|
||||
const currentState = await slotsService.getCurrentState();
|
||||
|
||||
if (!currentState) {
|
||||
return null;
|
||||
}
|
||||
|
||||
const maxContext = currentState.contextTotal;
|
||||
const currentUsage = currentState.contextUsed;
|
||||
const availableTokens = maxContext - currentUsage - this.reserveTokens;
|
||||
const wouldExceed = availableTokens <= 0;
|
||||
|
||||
return {
|
||||
wouldExceed,
|
||||
currentUsage,
|
||||
maxContext,
|
||||
availableTokens: Math.max(0, availableTokens),
|
||||
reservedTokens: this.reserveTokens
|
||||
};
|
||||
} catch (error) {
|
||||
console.warn('Error checking context limit:', error);
|
||||
return null;
|
||||
}
|
||||
}
|
||||
|
||||
/**
|
||||
* Returns a formatted error message for context limit exceeded
|
||||
*
|
||||
* @param {ContextCheckResult} result - Context check result
|
||||
* @returns {string} Formatted error message
|
||||
*/
|
||||
getContextErrorMessage(result: ContextCheckResult): string {
|
||||
const usagePercent = Math.round((result.currentUsage / result.maxContext) * 100);
|
||||
return `Context window is nearly full. Current usage: ${result.currentUsage.toLocaleString()}/${result.maxContext.toLocaleString()} tokens (${usagePercent}%). Available space: ${result.availableTokens.toLocaleString()} tokens (${result.reservedTokens} reserved for response).`;
|
||||
}
|
||||
|
||||
/**
|
||||
* Sets the number of tokens to reserve for response generation
|
||||
*
|
||||
* @param {number} tokens - Number of tokens to reserve
|
||||
*/
|
||||
setReserveTokens(tokens: number): void {
|
||||
this.reserveTokens = tokens;
|
||||
}
|
||||
}
|
||||
|
||||
export const contextService = new ContextService();
|
||||
@@ -1,3 +1,2 @@
|
||||
export { chatService } from './chat';
|
||||
export { contextService } from './context';
|
||||
export { slotsService } from './slots';
|
||||
|
||||
@@ -39,7 +39,6 @@ import type { ExportedConversations } from '$lib/types/database';
|
||||
* - Conversation branching for exploring different response paths
|
||||
* - Streaming AI responses with real-time content updates
|
||||
* - File attachment support (images, PDFs, text files, audio)
|
||||
* - Context window management with error recovery
|
||||
* - Partial response saving when generation is interrupted
|
||||
* - Message editing with automatic response regeneration
|
||||
*/
|
||||
@@ -48,11 +47,9 @@ class ChatStore {
|
||||
activeMessages = $state<DatabaseMessage[]>([]);
|
||||
conversations = $state<DatabaseConversation[]>([]);
|
||||
currentResponse = $state('');
|
||||
errorDialogState = $state<{ type: 'timeout' | 'server'; message: string } | null>(null);
|
||||
isInitialized = $state(false);
|
||||
isLoading = $state(false);
|
||||
maxContextError = $state<{ message: string; estimatedTokens: number; maxContext: number } | null>(
|
||||
null
|
||||
);
|
||||
titleUpdateConfirmationCallback?: (currentTitle: string, newTitle: string) => Promise<boolean>;
|
||||
|
||||
constructor() {
|
||||
@@ -69,8 +66,6 @@ class ChatStore {
|
||||
try {
|
||||
await this.loadConversations();
|
||||
|
||||
this.maxContextError = null;
|
||||
|
||||
this.isInitialized = true;
|
||||
} catch (error) {
|
||||
console.error('Failed to initialize chat store:', error);
|
||||
@@ -99,8 +94,6 @@ class ChatStore {
|
||||
this.activeConversation = conversation;
|
||||
this.activeMessages = [];
|
||||
|
||||
this.maxContextError = null;
|
||||
|
||||
await goto(`#/chat/${conversation.id}`);
|
||||
|
||||
return conversation.id;
|
||||
@@ -133,8 +126,6 @@ class ChatStore {
|
||||
this.activeMessages = await DatabaseStore.getConversationMessages(convId);
|
||||
}
|
||||
|
||||
this.maxContextError = null;
|
||||
|
||||
return true;
|
||||
} catch (error) {
|
||||
console.error('Failed to load conversation:', error);
|
||||
@@ -418,56 +409,6 @@ class ChatStore {
|
||||
return;
|
||||
}
|
||||
|
||||
if (error.name === 'ContextError') {
|
||||
console.warn('Context error detected:', error.message);
|
||||
this.isLoading = false;
|
||||
this.currentResponse = '';
|
||||
|
||||
const messageIndex = this.activeMessages.findIndex(
|
||||
(m: DatabaseMessage) => m.id === assistantMessage.id
|
||||
);
|
||||
|
||||
if (messageIndex !== -1) {
|
||||
this.activeMessages.splice(messageIndex, 1);
|
||||
DatabaseStore.deleteMessage(assistantMessage.id).catch(console.error);
|
||||
}
|
||||
|
||||
// Use structured context info from new exceed_context_size_error format if available
|
||||
const contextInfo = (
|
||||
error as Error & {
|
||||
contextInfo?: { promptTokens: number; maxContext: number; estimatedTokens: number };
|
||||
}
|
||||
).contextInfo;
|
||||
let estimatedTokens = 0;
|
||||
let maxContext = serverStore.serverProps?.default_generation_settings.n_ctx || 8192;
|
||||
|
||||
if (contextInfo) {
|
||||
// Use precise token counts from server response
|
||||
estimatedTokens = contextInfo.promptTokens;
|
||||
maxContext = contextInfo.maxContext;
|
||||
} else {
|
||||
// Fallback to estimation for older error format
|
||||
try {
|
||||
// Rough estimation: ~4 characters per token
|
||||
const messageContent = JSON.stringify(messages);
|
||||
estimatedTokens = Math.ceil(messageContent.length / 4);
|
||||
} catch {
|
||||
estimatedTokens = 0;
|
||||
}
|
||||
}
|
||||
|
||||
this.maxContextError = {
|
||||
message: error.message,
|
||||
estimatedTokens,
|
||||
maxContext
|
||||
};
|
||||
|
||||
if (onError) {
|
||||
onError(error);
|
||||
}
|
||||
return;
|
||||
}
|
||||
|
||||
console.error('Streaming error:', error);
|
||||
this.isLoading = false;
|
||||
this.currentResponse = '';
|
||||
@@ -477,9 +418,19 @@ class ChatStore {
|
||||
);
|
||||
|
||||
if (messageIndex !== -1) {
|
||||
this.activeMessages[messageIndex].content = `Error: ${error.message}`;
|
||||
const [failedMessage] = this.activeMessages.splice(messageIndex, 1);
|
||||
|
||||
if (failedMessage) {
|
||||
DatabaseStore.deleteMessage(failedMessage.id).catch((cleanupError) => {
|
||||
console.error('Failed to remove assistant message after error:', cleanupError);
|
||||
});
|
||||
}
|
||||
}
|
||||
|
||||
const dialogType = error.name === 'TimeoutError' ? 'timeout' : 'server';
|
||||
|
||||
this.showErrorDialog(dialogType, error.message);
|
||||
|
||||
if (onError) {
|
||||
onError(error);
|
||||
}
|
||||
@@ -487,6 +438,14 @@ class ChatStore {
|
||||
});
|
||||
}
|
||||
|
||||
private showErrorDialog(type: 'timeout' | 'server', message: string): void {
|
||||
this.errorDialogState = { type, message };
|
||||
}
|
||||
|
||||
dismissErrorDialog(): void {
|
||||
this.errorDialogState = null;
|
||||
}
|
||||
|
||||
/**
|
||||
* Checks if an error is an abort error (user cancelled operation)
|
||||
* @param error - The error to check
|
||||
@@ -574,6 +533,7 @@ class ChatStore {
|
||||
return;
|
||||
}
|
||||
|
||||
this.errorDialogState = null;
|
||||
this.isLoading = true;
|
||||
this.currentResponse = '';
|
||||
|
||||
@@ -603,37 +563,23 @@ class ChatStore {
|
||||
|
||||
const conversationContext = this.activeMessages.slice(0, -1);
|
||||
|
||||
await this.streamChatCompletion(
|
||||
conversationContext,
|
||||
assistantMessage,
|
||||
undefined,
|
||||
(error: Error) => {
|
||||
if (error.name === 'ContextError' && userMessage) {
|
||||
const userMessageIndex = this.findMessageIndex(userMessage.id);
|
||||
|
||||
if (userMessageIndex !== -1) {
|
||||
this.activeMessages.splice(userMessageIndex, 1);
|
||||
DatabaseStore.deleteMessage(userMessage.id).catch(console.error);
|
||||
}
|
||||
}
|
||||
}
|
||||
);
|
||||
await this.streamChatCompletion(conversationContext, assistantMessage);
|
||||
} catch (error) {
|
||||
if (this.isAbortError(error)) {
|
||||
this.isLoading = false;
|
||||
return;
|
||||
}
|
||||
|
||||
if (error instanceof Error && error.name === 'ContextError' && userMessage) {
|
||||
const userMessageIndex = this.findMessageIndex(userMessage.id);
|
||||
if (userMessageIndex !== -1) {
|
||||
this.activeMessages.splice(userMessageIndex, 1);
|
||||
DatabaseStore.deleteMessage(userMessage.id).catch(console.error);
|
||||
}
|
||||
}
|
||||
|
||||
console.error('Failed to send message:', error);
|
||||
this.isLoading = false;
|
||||
if (!this.errorDialogState) {
|
||||
if (error instanceof Error) {
|
||||
const dialogType = error.name === 'TimeoutError' ? 'timeout' : 'server';
|
||||
this.showErrorDialog(dialogType, error.message);
|
||||
} else {
|
||||
this.showErrorDialog('server', 'Unknown error occurred while sending message');
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
@@ -662,24 +608,6 @@ class ChatStore {
|
||||
this.currentResponse = '';
|
||||
}
|
||||
|
||||
/**
|
||||
* Clears the max context error state
|
||||
* Removes any displayed context limit warnings
|
||||
*/
|
||||
clearMaxContextError(): void {
|
||||
this.maxContextError = null;
|
||||
}
|
||||
|
||||
/**
|
||||
* Sets the max context error state
|
||||
* @param error - The context error details or null to clear
|
||||
*/
|
||||
setMaxContextError(
|
||||
error: { message: string; estimatedTokens: number; maxContext: number } | null
|
||||
): void {
|
||||
this.maxContextError = error;
|
||||
}
|
||||
|
||||
/**
|
||||
* Saves partial response if generation was interrupted
|
||||
* Preserves user's partial content and timing data when generation is stopped early
|
||||
@@ -1250,7 +1178,6 @@ class ChatStore {
|
||||
this.activeMessages = [];
|
||||
this.currentResponse = '';
|
||||
this.isLoading = false;
|
||||
this.maxContextError = null;
|
||||
}
|
||||
|
||||
/** Refreshes active messages based on currNode after branch navigation */
|
||||
@@ -1538,6 +1465,7 @@ class ChatStore {
|
||||
private async generateResponseForMessage(userMessageId: string): Promise<void> {
|
||||
if (!this.activeConversation) return;
|
||||
|
||||
this.errorDialogState = null;
|
||||
this.isLoading = true;
|
||||
this.currentResponse = '';
|
||||
|
||||
@@ -1584,7 +1512,7 @@ export const activeMessages = () => chatStore.activeMessages;
|
||||
export const isLoading = () => chatStore.isLoading;
|
||||
export const currentResponse = () => chatStore.currentResponse;
|
||||
export const isInitialized = () => chatStore.isInitialized;
|
||||
export const maxContextError = () => chatStore.maxContextError;
|
||||
export const errorDialog = () => chatStore.errorDialogState;
|
||||
|
||||
export const createConversation = chatStore.createConversation.bind(chatStore);
|
||||
export const downloadConversation = chatStore.downloadConversation.bind(chatStore);
|
||||
@@ -1592,9 +1520,9 @@ export const exportAllConversations = chatStore.exportAllConversations.bind(chat
|
||||
export const importConversations = chatStore.importConversations.bind(chatStore);
|
||||
export const deleteConversation = chatStore.deleteConversation.bind(chatStore);
|
||||
export const sendMessage = chatStore.sendMessage.bind(chatStore);
|
||||
export const dismissErrorDialog = chatStore.dismissErrorDialog.bind(chatStore);
|
||||
|
||||
export const gracefulStop = chatStore.gracefulStop.bind(chatStore);
|
||||
export const clearMaxContextError = chatStore.clearMaxContextError.bind(chatStore);
|
||||
export const setMaxContextError = chatStore.setMaxContextError.bind(chatStore);
|
||||
|
||||
// Branching operations
|
||||
export const refreshActiveMessages = chatStore.refreshActiveMessages.bind(chatStore);
|
||||
|
||||
@@ -197,7 +197,7 @@ class ServerStore {
|
||||
errorMessage = 'Server not found - check server address';
|
||||
isOfflineLikeError = true;
|
||||
} else if (error.message.includes('ETIMEDOUT')) {
|
||||
errorMessage = 'Connection timeout - server may be overloaded';
|
||||
errorMessage = 'Request timed out - the server took too long to respond';
|
||||
isOfflineLikeError = true;
|
||||
} else if (error.message.includes('503')) {
|
||||
errorMessage = 'Server temporarily unavailable - try again shortly';
|
||||
|
||||
@@ -1,11 +1,7 @@
|
||||
<script lang="ts">
|
||||
import '../app.css';
|
||||
import { page } from '$app/state';
|
||||
import {
|
||||
ChatSidebar,
|
||||
ConversationTitleUpdateDialog,
|
||||
MaximumContextAlertDialog
|
||||
} from '$lib/components/app';
|
||||
import { ChatSidebar, ConversationTitleUpdateDialog } from '$lib/components/app';
|
||||
import {
|
||||
activeMessages,
|
||||
isLoading,
|
||||
@@ -145,8 +141,6 @@
|
||||
|
||||
<Toaster richColors />
|
||||
|
||||
<MaximumContextAlertDialog />
|
||||
|
||||
<ConversationTitleUpdateDialog
|
||||
bind:open={titleUpdateDialogOpen}
|
||||
currentTitle={titleUpdateCurrentTitle}
|
||||
|
||||
Reference in New Issue
Block a user