Files
next-ai-draw-io/lib/llm-errors.ts
T
dayuan.jiang ab1a58f999 feat(errors): classify provider errors and show a hint the user can act on
- lib/llm-errors.ts sorts an error into about a dozen kinds (key
  rejected, no access, unknown model, no credit, rate limited, context
  too long, no image input, no tool calls, output cut off, provider down,
  cannot connect, timeout): first texts that name the cause precisely,
  then the HTTP status code, then general texts. It unwraps RetryError and
  hides keys and Bearer tokens in the provider's message
- The chat route uses it for errors before the stream and, through
  toUIMessageStreamResponse's onError, for errors in the stream. Errors
  of the model's own tool call stay as they are: the same text goes back
  to the model so it can fix the call
- The chat shows the hint in the user's language, then the provider's
  message; a rejected key, missing access or unknown model adds an "Open
  model settings" button. The Test button shows the same hints
- Fixes: our message "API key is required when using a custom base URL"
  was replaced by "Authentication failed" because it contains "key"; a
  provider's "Rate limit exceeded" opened this site's quota toast; an
  error body like {"error": ...} was shown as raw JSON; the Test button
  matched "401" in the message, where providers rarely put it
- Remove the string matching fallbacks in the chat panel
2026-10-04 13:49:34 +09:00

129 lines
4.2 KiB
TypeScript

import {
APICallError,
InvalidToolInputError,
LoadAPIKeyError,
NoSuchToolError,
RetryError,
ToolCallRepairError,
} from "ai"
/**
* What went wrong with a model call, for a hint the user can act on. The
* provider's own message always goes along, because a guess can be wrong.
*/
export type LLMErrorCode =
| "invalid_api_key"
| "forbidden"
| "model_not_found"
| "insufficient_quota"
| "rate_limited"
| "context_too_long"
| "images_unsupported"
| "tools_unsupported"
| "output_truncated"
| "provider_unavailable"
| "cannot_connect"
| "timeout"
| "unknown"
export interface LLMError {
type: "provider"
code: LLMErrorCode
message: string
}
// Texts that name the cause more precisely than the status code: a quota
// error can come as 403 or 429, a context or image error as a plain 400
const SPECIFIC_TEXTS: Array<[RegExp, LLMErrorCode]> = [
[
/context length|context window|maximum context|prompt is too long|input is too long|too many (?:input )?tokens/i,
"context_too_long",
],
[
/image content block|image_url|does not support image|image input is not supported/i,
"images_unsupported",
],
[
/does not support tools|tool use is not supported|tools? (?:are|is) not supported|function calling is not supported/i,
"tools_unsupported",
],
// Bedrock, when the output limit cut the tool call's JSON short
[/toolUse\.input is invalid/i, "output_truncated"],
[
/insufficient[_ ]quota|insufficient balance|exceeded your current quota|credit balance is too low|余额不足/i,
"insufficient_quota",
],
]
const STATUS_CODES: Record<number, LLMErrorCode> = {
401: "invalid_api_key",
402: "insufficient_quota",
// Not "invalid key": a valid key can lack access to a model or region
403: "forbidden",
404: "model_not_found",
408: "timeout",
413: "context_too_long",
429: "rate_limited",
}
const GENERAL_TEXTS: Array<[RegExp, LLMErrorCode]> = [
[
/model[_ ]not[_ ]found|model .*does not exist|unknown model|no such model/i,
"model_not_found",
],
[
/invalid[_ ]api[_ ]key|incorrect api key|unauthorized/i,
"invalid_api_key",
],
[/rate limit|too many requests/i, "rate_limited"],
[/ECONNREFUSED|ENOTFOUND|ECONNRESET|fetch failed/i, "cannot_connect"],
]
/** Secrets a provider may echo back: API keys, Bearer tokens, key=value */
function redact(text: string): string {
return text
.replace(/\b(sk|pk|rk|ak)-[A-Za-z0-9_-]{8,}/g, "$1-[redacted]")
.replace(/\bBearer\s+[A-Za-z0-9._~+/-]+=*/gi, "Bearer [redacted]")
.replace(/\bAKIA[0-9A-Z]{16}\b/g, "[redacted]")
.replace(
/\b(api[_-]?key|access[_-]?key|secret|token|password|signature)(["']?\s*[:=]\s*["']?)[^\s"',&}]+/gi,
"$1$2[redacted]",
)
}
/**
* Model and tool errors the SDK sends back to the model as the tool result,
* so it can fix its call. Their text has to stay as it is.
*/
export function isToolCallError(error: unknown): boolean {
return (
InvalidToolInputError.isInstance(error) ||
NoSuchToolError.isInstance(error) ||
ToolCallRepairError.isInstance(error)
)
}
export function classifyLLMError(error: unknown): LLMError {
// After the SDK's retries, the last attempt says what happened
const e = RetryError.isInstance(error) ? error.lastError : error
const raw = e instanceof Error ? e.message : String(e)
const message = redact(raw).slice(0, 500)
const body = APICallError.isInstance(e) ? (e.responseBody ?? "") : ""
const text = `${raw} ${body}`
const status = APICallError.isInstance(e)
? e.statusCode
: (e as { statusCode?: number })?.statusCode
const find = (rules: Array<[RegExp, LLMErrorCode]>) =>
rules.find(([pattern]) => pattern.test(text))?.[1]
const code =
(e instanceof Error && e.name === "TimeoutError" && "timeout") ||
(LoadAPIKeyError.isInstance(e) && "invalid_api_key") ||
find(SPECIFIC_TEXTS) ||
(status && STATUS_CODES[status]) ||
(status && status >= 500 && "provider_unavailable") ||
find(GENERAL_TEXTS) ||
"unknown"
return { type: "provider", code, message }
}