Delia

/** * Copyright (C) 2023 the project owner * * This program is free software: you can redistribute it and/or modify * it under the terms of the GNU General Public License as published by * the Free Software Foundation, either version 3 of the License, or * (at your option) any later version. * * This program is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * * You should have received a copy of the GNU General Public License * along with this program. If not, see <http://www.gnu.org/licenses/>. */ import { NextResponse } from "next/server" import { readFile, writeFile } from "fs/promises" import { join } from "path" // Path to settings.json (always in project root) const PROJECT_ROOT = join(process.cwd(), "..") const SETTINGS_FILE = join(PROJECT_ROOT, "settings.json") interface BackendModels { quick: string coder: string moe: string thinking: string } interface BackendConfig { id: string name: string provider: "ollama" | "llamacpp" | "vllm" | "openai" | "gemini" | "custom" type: "local" | "remote" url: string enabled: boolean priority: number models: BackendModels health_endpoint?: string models_endpoint?: string context_limit?: number api_key?: string } interface BackendStatus extends BackendConfig { health: { available: boolean response_time_ms: number last_error: string loaded_models: string[] circuit_open: boolean } } // Full settings structure matches settings.json interface SettingsConfig { version: string system?: any backends: BackendConfig[] routing: { prefer_local: boolean fallback_enabled: boolean load_balance: boolean } models?: any generation?: any costs?: any auth?: any } // Provider-specific health check endpoints const PROVIDER_HEALTH_ENDPOINTS: Record<string, string> = { ollama: "/api/tags", llamacpp: "/health", vllm: "/health", openai: "", // No health check for OpenAI gemini: "", // Gemini uses API key validation instead custom: "/health", } // Provider-specific model list endpoints const PROVIDER_MODEL_ENDPOINTS: Record<string, string> = { ollama: "/api/tags", llamacpp: "/v1/models", vllm: "/v1/models", openai: "/v1/models", gemini: "/v1beta/models", custom: "/v1/models", } async function loadSettingsConfig(): Promise<SettingsConfig> { try { const content = await readFile(SETTINGS_FILE, "utf-8") return JSON.parse(content) } catch { // Return default config if file doesn't exist return { version: "1.0", system: { gpu_memory_limit_gb: 8, memory_buffer_gb: 1, max_concurrent_requests_per_backend: 1, }, backends: [], routing: { prefer_local: true, fallback_enabled: true, load_balance: false, }, models: {}, auth: { enabled: false, tracking_enabled: true } } } } async function saveSettingsConfig(config: SettingsConfig): Promise<void> { await writeFile(SETTINGS_FILE, JSON.stringify(config, null, 2), "utf-8") } async function checkBackendHealth(backend: BackendConfig): Promise<BackendStatus> { const healthEndpoint = backend.health_endpoint || PROVIDER_HEALTH_ENDPOINTS[backend.provider] || "/health" const startTime = Date.now() const status: BackendStatus = { ...backend, health: { available: false, response_time_ms: 0, last_error: "", loaded_models: [], circuit_open: false, } } if (!backend.enabled) { status.health.last_error = "Disabled" return status } // Special handling for Gemini - validate API key if (backend.provider === "gemini") { if (!backend.api_key) { status.health.last_error = "No API key" return status } try { const controller = new AbortController() const timeoutId = setTimeout(() => controller.abort(), 5000) const response = await fetch( `https://generativelanguage.googleapis.com/v1beta/models?key=${backend.api_key}`, { signal: controller.signal } ) clearTimeout(timeoutId) status.health.response_time_ms = Date.now() - startTime if (response.ok) { status.health.available = true const data = await response.json() if (data.models) { status.health.loaded_models = data.models .filter((m: { supportedGenerationMethods?: string[] }) => m.supportedGenerationMethods?.includes("generateContent")) .map((m: { name: string }) => m.name.replace("models/", "")) .slice(0, 10) // Limit to first 10 } } else { status.health.last_error = response.status === 400 ? "Invalid API key" : `HTTP ${response.status}` } } catch (error) { status.health.response_time_ms = Date.now() - startTime status.health.last_error = error instanceof Error ? error.message : "Connection failed" } return status } // Skip health check for OpenAI (no health endpoint) if (!healthEndpoint) { status.health.available = true return status } try { const controller = new AbortController() const timeoutId = setTimeout(() => controller.abort(), 5000) const response = await fetch(`${backend.url}${healthEndpoint}`, { signal: controller.signal }) clearTimeout(timeoutId) status.health.response_time_ms = Date.now() - startTime if (response.ok) { status.health.available = true // Parse loaded models from response try { const data = await response.json() if (backend.provider === "ollama" && data.models) { status.health.loaded_models = data.models.map((m: { name: string }) => m.name) } else if (data.data) { // OpenAI-compatible format status.health.loaded_models = data.data.map((m: { id: string }) => m.id) } } catch { // Ignore JSON parsing errors } } else { status.health.last_error = `HTTP ${response.status}` } } catch (error) { status.health.response_time_ms = Date.now() - startTime status.health.last_error = error instanceof Error ? error.message : "Connection failed" } return status } export async function GET() { try { const config = await loadSettingsConfig() // Check all backends in parallel const statusPromises = config.backends.map(backend => checkBackendHealth(backend)) const statuses = await Promise.all(statusPromises) // Count available backends const availableCount = statuses.filter(s => s.health.available).length const localBackends = statuses.filter(s => s.type === "local") const remoteBackends = statuses.filter(s => s.type === "remote") // Determine active backend (first available, prefer local if configured) let activeBackend = "" if (config.routing.prefer_local) { const availableLocal = localBackends.find(b => b.health.available) activeBackend = availableLocal?.id || statuses.find(s => s.health.available)?.id || "" } else { activeBackend = statuses.find(s => s.health.available)?.id || "" } return NextResponse.json({ backends: statuses, routing: config.routing, activeBackend, summary: { total: config.backends.length, enabled: config.backends.filter(b => b.enabled).length, available: availableCount, local: localBackends.length, remote: remoteBackends.length, }, timestamp: new Date().toISOString() }, { headers: { "Cache-Control": "max-age=5, stale-while-revalidate=10" } }) } catch (error) { return NextResponse.json( { error: "Failed to check backend status", details: error instanceof Error ? error.message : "Unknown error" }, { status: 500 } ) } } // POST: Add a new backend export async function POST(request: Request) { try { const body = await request.json() as Partial<BackendConfig> // Validate required fields if (!body.id || !body.name || !body.provider || !body.type || !body.url) { return NextResponse.json({ success: false, error: "Missing required fields: id, name, provider, type, url" }, { status: 400 }) } // Validate provider if (!["ollama", "llamacpp", "vllm", "openai", "gemini", "custom"].includes(body.provider)) { return NextResponse.json({ success: false, error: "Invalid provider. Must be: ollama, llamacpp, vllm, openai, gemini, or custom" }, { status: 400 }) } // Validate type if (!["local", "remote"].includes(body.type)) { return NextResponse.json({ success: false, error: "Invalid type. Must be: local or remote" }, { status: 400 }) } const config = await loadSettingsConfig() // Check for duplicate ID if (config.backends.some(b => b.id === body.id)) { return NextResponse.json({ success: false, error: `Backend with ID '${body.id}' already exists` }, { status: 409 }) } // Create new backend with defaults const newBackend: BackendConfig = { id: body.id, name: body.name, provider: body.provider, type: body.type, url: body.url, enabled: body.enabled ?? true, priority: body.priority ?? config.backends.length, models: body.models || { quick: "", coder: "", moe: "", thinking: "" }, health_endpoint: body.health_endpoint || PROVIDER_HEALTH_ENDPOINTS[body.provider], models_endpoint: body.models_endpoint || PROVIDER_MODEL_ENDPOINTS[body.provider], context_limit: body.context_limit || 32768, ...(body.api_key && { api_key: body.api_key }), } config.backends.push(newBackend) await saveSettingsConfig(config) return NextResponse.json({ success: true, message: `Backend '${newBackend.name}' added`, backend: newBackend }) } catch (error) { return NextResponse.json({ success: false, error: error instanceof Error ? error.message : "Unknown error" }, { status: 500 }) } } // PUT: Update an existing backend export async function PUT(request: Request) { try { const body = await request.json() as Partial<BackendConfig> & { id: string } if (!body.id) { return NextResponse.json({ success: false, error: "Missing required field: id" }, { status: 400 }) } const config = await loadSettingsConfig() const index = config.backends.findIndex(b => b.id === body.id) if (index === -1) { return NextResponse.json({ success: false, error: `Backend with ID '${body.id}' not found` }, { status: 404 }) } // Update fields (preserve existing values for undefined fields) const existing = config.backends[index] config.backends[index] = { ...existing, name: body.name ?? existing.name, provider: body.provider ?? existing.provider, type: body.type ?? existing.type, url: body.url ?? existing.url, enabled: body.enabled ?? existing.enabled, priority: body.priority ?? existing.priority, models: body.models ?? existing.models, health_endpoint: body.health_endpoint ?? existing.health_endpoint, models_endpoint: body.models_endpoint ?? existing.models_endpoint, context_limit: body.context_limit ?? existing.context_limit, api_key: body.api_key ?? existing.api_key, } await saveSettingsConfig(config) return NextResponse.json({ success: true, message: `Backend '${body.id}' updated`, backend: config.backends[index] }) } catch (error) { return NextResponse.json({ success: false, error: error instanceof Error ? error.message : "Unknown error" }, { status: 500 }) } } // DELETE: Remove a backend export async function DELETE(request: Request) { try { const { searchParams } = new URL(request.url) const id = searchParams.get("id") if (!id) { return NextResponse.json({ success: false, error: "Missing required query parameter: id" }, { status: 400 }) } const config = await loadSettingsConfig() const index = config.backends.findIndex(b => b.id === id) if (index === -1) { return NextResponse.json({ success: false, error: `Backend with ID '${id}' not found` }, { status: 404 }) } const removed = config.backends.splice(index, 1)[0] await saveSettingsConfig(config) return NextResponse.json({ success: true, message: `Backend '${removed.name}' removed`, backend: removed }) } catch (error) { return NextResponse.json({ success: false, error: error instanceof Error ? error.message : "Unknown error" }, { status: 500 }) } } // PATCH: Update routing settings export async function PATCH(request: Request) { try { const body = await request.json() as Partial<SettingsConfig["routing"]> const config = await loadSettingsConfig() // Update routing settings if (typeof body.prefer_local === "boolean") { config.routing.prefer_local = body.prefer_local } if (typeof body.fallback_enabled === "boolean") { config.routing.fallback_enabled = body.fallback_enabled } if (typeof body.load_balance === "boolean") { config.routing.load_balance = body.load_balance } await saveSettingsConfig(config) return NextResponse.json({ success: true, message: "Routing settings updated", routing: config.routing }) } catch (error) { return NextResponse.json({ success: false, error: error instanceof Error ? error.message : "Unknown error" }, { status: 500 }) } }