Autopoll Ollama models for inference endpoints
This commit is contained in:
@@ -0,0 +1,86 @@
|
||||
import { NextResponse } from "next/server"
|
||||
|
||||
const DEFAULT_OLLAMA_URL = "http://127.0.0.1:11434"
|
||||
const OLLAMA_BASE_URL = (process.env.OLLAMA_BASE_URL || DEFAULT_OLLAMA_URL).replace(/\/+$/, "")
|
||||
|
||||
type OllamaTag = {
|
||||
name?: string
|
||||
model?: string
|
||||
modified_at?: string
|
||||
size?: number
|
||||
digest?: string
|
||||
details?: {
|
||||
family?: string
|
||||
parameter_size?: string
|
||||
quantization_level?: string
|
||||
}
|
||||
}
|
||||
|
||||
function formatBytes(size: unknown) {
|
||||
if (typeof size !== "number" || !Number.isFinite(size) || size <= 0) return null
|
||||
const units = ["B", "KB", "MB", "GB", "TB"]
|
||||
let value = size
|
||||
let unit = 0
|
||||
while (value >= 1024 && unit < units.length - 1) {
|
||||
value /= 1024
|
||||
unit += 1
|
||||
}
|
||||
return `${value >= 10 || unit === 0 ? value.toFixed(0) : value.toFixed(1)} ${units[unit]}`
|
||||
}
|
||||
|
||||
export async function GET() {
|
||||
const startedAt = Date.now()
|
||||
const controller = new AbortController()
|
||||
const timeout = setTimeout(() => controller.abort(), 2500)
|
||||
|
||||
try {
|
||||
const response = await fetch(`${OLLAMA_BASE_URL}/api/tags`, {
|
||||
cache: "no-store",
|
||||
signal: controller.signal,
|
||||
})
|
||||
const data = await response.json()
|
||||
if (!response.ok) {
|
||||
throw new Error(data?.error || `Ollama returned HTTP ${response.status}`)
|
||||
}
|
||||
|
||||
const models = Array.isArray(data?.models)
|
||||
? data.models.map((item: OllamaTag) => ({
|
||||
name: item.model || item.name || "",
|
||||
model: item.model || item.name || "",
|
||||
modifiedAt: item.modified_at || null,
|
||||
size: typeof item.size === "number" ? item.size : null,
|
||||
sizeLabel: formatBytes(item.size),
|
||||
digest: item.digest || null,
|
||||
family: item.details?.family || null,
|
||||
parameterSize: item.details?.parameter_size || null,
|
||||
quantization: item.details?.quantization_level || null,
|
||||
})).filter((item: { name: string }) => item.name)
|
||||
: []
|
||||
|
||||
return NextResponse.json({
|
||||
ok: true,
|
||||
available: true,
|
||||
baseUrl: OLLAMA_BASE_URL,
|
||||
models,
|
||||
count: models.length,
|
||||
elapsedMs: Date.now() - startedAt,
|
||||
})
|
||||
} catch (error) {
|
||||
const message = error instanceof Error && error.name === "AbortError"
|
||||
? "Ollama probe timed out"
|
||||
: error instanceof Error
|
||||
? error.message
|
||||
: "Failed to fetch Ollama models"
|
||||
return NextResponse.json({
|
||||
ok: false,
|
||||
available: false,
|
||||
baseUrl: OLLAMA_BASE_URL,
|
||||
models: [],
|
||||
count: 0,
|
||||
elapsedMs: Date.now() - startedAt,
|
||||
error: message,
|
||||
}, { status: 200 })
|
||||
} finally {
|
||||
clearTimeout(timeout)
|
||||
}
|
||||
}
|
||||
Reference in New Issue
Block a user