Evenfire is an intelligence layer platform for running LLM agents that take real actions, on your infrastructure
TypeScript
import { describe, expect, it } from 'vitest'
import { LlmErrorCode } from '../../core/errors'
import { OpenAICompatibleProvider } from '../openaiCompatible'
import { openaiApiError } from './sdkErrorFixtures'
const zai = new OpenAICompatibleProvider(
{ id: 'zai', baseURL: 'https://example.invalid', defaultModel: 'glm-4' },
'fake-key'
)
const bailian = new OpenAICompatibleProvider(
{ id: 'bailian', baseURL: 'https://example.invalid', defaultModel: 'qwen-max' },
'fake-key'
)
// Fixtures derived from OpenAI.APIError.generate (T1): the provider-native code
// lands at `err.error.code`, the level the override reads. Deriving guards
// against the same nesting mismatch that broke the Claude arm.
describe('OpenAICompatibleProvider.classifyError', () => {
it('maps z.ai model-not-available codes (1211/1220) to ModelNotAvailable', () => {
for (const code of ['1211', '1220']) {
const c = zai.classifyError(openaiApiError(404, { code, message: 'model not available' }))
expect(c.code).toBe(LlmErrorCode.ModelNotAvailable)
expect(c.retryable).toBe(false)
expect(c.providerCode).toBe(code)
}
})
it('maps z.ai billing code 1113 inside a 429 to InsufficientQuota (not a retryable rate-limit)', () => {
const c = zai.classifyError(
openaiApiError(429, { code: '1113', message: 'insufficient balance' })
)
expect(c.code).toBe(LlmErrorCode.InsufficientQuota)
expect(c.retryable).toBe(false)
})
it('maps Bailian ModelNotFound / Model.AccessDenied to ModelNotAvailable', () => {
for (const code of ['ModelNotFound', 'Model.AccessDenied']) {
const c = bailian.classifyError(openaiApiError(404, { code, message: 'no access to model' }))
expect(c.code).toBe(LlmErrorCode.ModelNotAvailable)
expect(c.retryable).toBe(false)
}
})
it('maps Bailian Arrearage (billing) to InsufficientQuota, not retryable', () => {
const c = bailian.classifyError(openaiApiError(429, { code: 'Arrearage', message: 'arrears' }))
expect(c.code).toBe(LlmErrorCode.InsufficientQuota)
expect(c.retryable).toBe(false)
})
it('falls through to the shared HTTP classifier for generic 404', () => {
const c = zai.classifyError(openaiApiError(404, { message: 'not found' }))
expect(c.code).toBe(LlmErrorCode.ModelNotAvailable)
expect(c.retryable).toBe(false)
})
it('falls through to the shared HTTP classifier for a plain rate-limit 429', () => {
const c = zai.classifyError(openaiApiError(429, { message: 'rate limited' }))
expect(c.code).toBe(LlmErrorCode.RateLimited)
expect(c.retryable).toBe(true)
})
it('falls back to unknown classification for a plain Error (does not throw)', () => {
const c = zai.classifyError(new Error('ECONNRESET'))
expect(c.code).toBe(LlmErrorCode.ApiCallFailed)
expect(c.retryable).toBe(true)
})
})