happyz
/
llama.cpp
mirror of https://github.com/ggerganov/llama.cpp.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

refactor: Architecture cleanup

This commit is contained in:
Aleksander Grygier 2025-11-27 14:03:25 +01:00
parent 78ead49830
commit d73353732f
3 changed files with 39 additions and 36 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

56
tools/server/webui/src/lib/services/chat.ts
View File

@ -1,7 +1,4 @@
import { config } from '$lib/stores/settings.svelte';
import { getJsonHeaders } from '$lib/utils/api-headers';
import { selectedModelName } from '$lib/stores/models.svelte';
import { isRouterMode, serverStore } from '$lib/stores/server.svelte';
import type {
ApiChatCompletionRequest,
ApiChatCompletionResponse,
@ -106,11 +103,12 @@ export class ChatService {
// Other parameters
samplers,
custom,
timings_per_token
timings_per_token,
// Config options
systemMessage,
disableReasoningFormat
} = options;
const currentConfig = config();
const normalizedMessages: ApiChatMessageData[] = messages
.map((msg) => {
if ('id' in msg && 'convId' in msg && 'timestamp' in msg) {
@ -130,7 +128,7 @@ export class ChatService {
return true;
});
const processedMessages = ChatService.injectSystemMessage(normalizedMessages);
const processedMessages = ChatService.injectSystemMessage(normalizedMessages, systemMessage);
const requestBody: ApiChatCompletionRequest = {
messages: processedMessages.map((msg: ApiChatMessageData) => ({
@ -140,14 +138,12 @@ export class ChatService {
stream
};
const isRouter = isRouterMode();
const activeModel = isRouter ? options.model || selectedModelName() : null;
if (isRouter && activeModel) {
requestBody.model = activeModel;
// Include model in request if provided (required in ROUTER mode)
if (options.model) {
requestBody.model = options.model;
}
requestBody.reasoning_format = currentConfig.disableReasoningFormat ? 'none' : 'auto';
requestBody.reasoning_format = disableReasoningFormat ? 'none' : 'auto';
if (temperature !== undefined) requestBody.temperature = temperature;
if (max_tokens !== undefined) {
@ -728,28 +724,30 @@ export class ChatService {
}
/**
* Injects a system message at the beginning of the conversation if configured in settings.
* Checks for existing system messages to avoid duplication and retrieves the system message
* from the current configuration settings.
* Injects a system message at the beginning of the conversation if provided.
* Checks for existing system messages to avoid duplication.
*
* @param messages - Array of chat messages to process
* @returns Array of messages with system message injected at the beginning if configured
* @param systemMessage - Optional system message to inject
* @returns Array of messages with system message injected at the beginning if provided
* @private
*/
private static injectSystemMessage(messages: ApiChatMessageData[]): ApiChatMessageData[] {
const currentConfig = config();
const systemMessage = currentConfig.systemMessage?.toString().trim();
private static injectSystemMessage(
messages: ApiChatMessageData[],
systemMessage?: string
): ApiChatMessageData[] {
const trimmedSystemMessage = systemMessage?.trim();
if (!systemMessage) {
if (!trimmedSystemMessage) {
return messages;
}
if (messages.length > 0 && messages[0].role === 'system') {
if (messages[0].content !== systemMessage) {
if (messages[0].content !== trimmedSystemMessage) {
const updatedMessages = [...messages];
updatedMessages[0] = {
role: 'system',
content: systemMessage
content: trimmedSystemMessage
};
return updatedMessages;
}
@ -759,7 +757,7 @@ export class ChatService {
const systemMsg: ApiChatMessageData = {
role: 'system',
content: systemMessage
content: trimmedSystemMessage
};
return [systemMsg, ...messages];
@ -799,16 +797,6 @@ export class ChatService {
* @private
*/
private static extractModelName(data: unknown): string | undefined {
// WORKAROUND: In single model mode, use model name from props instead of API response
// because llama-server returns `gpt-3.5-turbo` value in the `model` field
const isRouter = isRouterMode();
if (!isRouter) {
const propsModelName = serverStore.modelName;
if (propsModelName) {
return propsModelName;
}
}
const asRecord = (value: unknown): Record<string, unknown> | undefined => {
return typeof value === 'object' && value !== null
? (value as Record<string, unknown>)

13
tools/server/webui/src/lib/stores/chat.svelte.ts
View File

@ -1,7 +1,8 @@
import { DatabaseService, ChatService } from '$lib/services';
import { conversationsStore } from '$lib/stores/conversations.svelte';
import { config } from '$lib/stores/settings.svelte';
import { contextSize } from '$lib/stores/server.svelte';
import { contextSize, isRouterMode } from '$lib/stores/server.svelte';
import { selectedModelName } from '$lib/stores/models.svelte';
import { normalizeModelName } from '$lib/utils/model-names';
import { filterByLeafNodeId, findDescendantMessages, findLeafNode } from '$lib/utils/branching';
import { SvelteMap } from 'svelte/reactivity';
@ -78,6 +79,16 @@ class ChatStore {
const apiOptions: Record<string, unknown> = { stream: true, timings_per_token: true };
// Model selection (required in ROUTER mode)
if (isRouterMode()) {
const modelName = selectedModelName();
if (modelName) apiOptions.model = modelName;
}
// Config options needed by ChatService
if (currentConfig.systemMessage) apiOptions.systemMessage = currentConfig.systemMessage;
if (currentConfig.disableReasoningFormat) apiOptions.disableReasoningFormat = true;
if (hasValue(currentConfig.temperature))
apiOptions.temperature = Number(currentConfig.temperature);
if (hasValue(currentConfig.max_tokens))

6
tools/server/webui/src/lib/types/settings.d.ts vendored
View File

@ -14,8 +14,12 @@ export interface SettingsFieldConfig {
export interface SettingsChatServiceOptions {
stream?: boolean;
// Model override (for regenerate with specific model)
// Model (required in ROUTER mode, optional in MODEL mode)
model?: string;
// System message to inject
systemMessage?: string;
// Disable reasoning format (use 'none' instead of 'auto')
disableReasoningFormat?: boolean;
// Generation parameters
temperature?: number;
max_tokens?: number;