Enhanced ChatGPT Clone: Features Agents, MCP, Skills, DeepSeek, Anthropic, AWS, OpenAI, Responses API, Azure, Groq, o1, GPT-5, Mistral, OpenRouter, Vertex AI, Gemini, Artifacts, AI model switching, message search, Code Interpreter, langchain, DALL-E-3, OpenAPI Actions, Functions, Secure Multi-User Auth, Presets, open-source for self-hosting. Active
TypeScript aianthropicartifactsawsazurechatgptchatgpt-cloneclaudeclonedeepseekgeminigooglegpt-5librechatmcpo1openairesponses-apivisionwebui
import { Constants, ErrorTypes, ContentTypes } from 'librechat-data-provider';
import type { TMessage } from 'librechat-data-provider';
import {
isAttachmentError,
mayBeAttachmentError,
findAttachmentRecovery,
} from '../Error/attachments';
describe('isAttachmentError', () => {
it.each([
['a code environment upload rate limit', 'Error uploading code environment file: 429'],
[
'the agent attachment limit counting resent files',
'An error occurred while processing the request: This turn exceeds the configured attachment count limit (12 > 10). Remove some attachments or use smaller files and try again.',
],
[
'duplicate document names on Bedrock',
JSON.stringify({
type: ErrorTypes.UPSTREAM_MODEL_ERROR,
message:
'Messages can’t contain duplicate document names. Rename the document and retry your request.',
}),
],
['an unrestorable file', JSON.stringify({ type: ErrorTypes.RESOURCE_RECOVERY_REQUIRED })],
['an over-long prompt', 'An error occurred while processing the request: prompt is too long'],
['a typed input length error', JSON.stringify({ type: ErrorTypes.INPUT_LENGTH, info: '9/5' })],
[
'a typed context overflow',
JSON.stringify({ type: ErrorTypes.FINAL_CONTEXT_OVERFLOW, projectedMessageTokens: 9 }),
],
['a missing attachment', 'An attached file is no longer available. Remove it and retry.'],
['an oversized image', 'messages.3.content.1.image.source.base64: image exceeds 5 MB maximum'],
])('recognizes %s', (_label, text) => {
expect(isAttachmentError(text)).toBe(true);
});
it.each([
['a model rate limit', JSON.stringify({ type: ErrorTypes.MODEL_RATE_LIMIT })],
['a dropped stream', 'Error connecting to server, try refreshing the page.'],
['an invalid key', JSON.stringify({ type: ErrorTypes.INVALID_USER_KEY })],
['provider prose', 'An error occurred while processing the request: 500 Internal Server Error'],
])('does not blame files for %s', (_label, text) => {
expect(isAttachmentError(text)).toBe(false);
});
});
describe('mayBeAttachmentError', () => {
it.each([
['a file failure', 'Error uploading code environment file: 429'],
['an unexplained provider failure', JSON.stringify({ type: ErrorTypes.UPSTREAM_MODEL_ERROR })],
[
'an Anthropic error body',
JSON.stringify({ type: 'error', error: { type: 'invalid_request_error', message: 'bad' } }),
],
['provider prose', 'An error occurred while processing the request: 400 Bad Request'],
])('offers the files as a way out of %s', (_label, text) => {
expect(mayBeAttachmentError(text)).toBe(true);
});
it.each([
['a missing key', JSON.stringify({ type: ErrorTypes.NO_USER_KEY })],
['a model rate limit', JSON.stringify({ type: ErrorTypes.MODEL_RATE_LIMIT })],
['a token balance', JSON.stringify({ type: 'token_balance', balance: 0 })],
[
'a legacy LangChain rate limit',
'An error occurred while processing the request: 429 Too many requests\n\nTroubleshooting URL: https://docs.langchain.com/oss/javascript/langchain/errors/MODEL_RATE_LIMIT/',
],
['a nested quota code', JSON.stringify({ error: { code: 'insufficient_quota' } })],
[
'an Anthropic rate limit inside its generic envelope',
JSON.stringify({ type: 'error', error: { type: 'rate_limit_error', message: 'slow down' } }),
],
])('does not point at files for %s', (_label, text) => {
expect(mayBeAttachmentError(text)).toBe(false);
});
});
const screenshot = { file_id: 'screenshot', filename: 'screenshot.png', type: 'image/png' };
const user = (
messageId: string,
parentMessageId: string,
fields: Partial<TMessage> = {},
): TMessage =>
({
messageId,
parentMessageId,
conversationId: 'convo',
isCreatedByUser: true,
text: `question ${messageId}`,
...fields,
}) as TMessage;
const answer = (
messageId: string,
parentMessageId: string,
fields: Partial<TMessage> = {},
): TMessage =>
({
messageId,
parentMessageId,
conversationId: 'convo',
isCreatedByUser: false,
text: `answer ${messageId}`,
...fields,
}) as TMessage;
describe('findAttachmentRecovery', () => {
it('offers only a retry when the failed turn alone carries files', () => {
const messages = [
user('u1', Constants.NO_PARENT),
answer('a1', 'u1'),
user('u2', 'a1', { files: [screenshot] }),
answer('a2', 'u2', { error: true }),
];
const recovery = findAttachmentRecovery(messages, messages[3]);
expect(recovery).toEqual({ parent: messages[2], canRetry: true, branchTargetId: undefined });
});
it('also offers a branch ending before the failed turn when earlier turns carry files', () => {
const messages = [
user('u1', Constants.NO_PARENT, { files: [screenshot] }),
answer('a1', 'u1'),
user('u2', 'a1', { files: [screenshot] }),
answer('a2', 'u2', { error: true }),
];
expect(findAttachmentRecovery(messages, messages[3])).toEqual({
parent: messages[2],
canRetry: true,
branchTargetId: 'a1',
});
});
it('counts code outputs on earlier answers as files', () => {
const messages = [
user('u1', Constants.NO_PARENT),
answer('a1', 'u1', {
attachments: [{ file_id: 'chart', toolCallId: 't' }] as TMessage['attachments'],
}),
user('u2', 'a1'),
answer('a2', 'u2', { error: true }),
];
expect(findAttachmentRecovery(messages, messages[3])).toEqual({
parent: messages[2],
canRetry: false,
branchTargetId: 'a1',
});
});
it('counts attachments steered into an earlier response as files', () => {
const steered = [
{ type: ContentTypes.TEXT, text: 'Working on it.' },
{ type: ContentTypes.STEER, steer: 'use this', steerId: 's1', files: [screenshot] },
] as TMessage['content'];
const messages = [
user('u1', Constants.NO_PARENT),
answer('a1', 'u1', { content: steered }),
user('u2', 'a1'),
answer('a2', 'u2', { error: true }),
];
expect(findAttachmentRecovery(messages, messages[3])).toEqual({
parent: messages[2],
canRetry: false,
branchTargetId: 'a1',
});
});
it('counts a generated image in an earlier response as a file', () => {
const messages = [
user('u1', Constants.NO_PARENT),
answer('a1', 'u1', {
content: [
{ type: ContentTypes.IMAGE_FILE, image_file: { file_id: 'generated' } },
] as TMessage['content'],
}),
user('u2', 'a1'),
answer('a2', 'u2', { error: true }),
];
expect(findAttachmentRecovery(messages, messages[3])?.branchTargetId).toBe('a1');
});
it('does not count search sources as files', () => {
const messages = [
user('u1', Constants.NO_PARENT),
answer('a1', 'u1', {
attachments: [{ type: 'web_search', toolCallId: 't' }] as TMessage['attachments'],
}),
user('u2', 'a1'),
answer('a2', 'u2', { error: true }),
];
expect(findAttachmentRecovery(messages, messages[3])).toBeNull();
});
it('ignores files on a sibling branch', () => {
const messages = [
user('u1', Constants.NO_PARENT),
answer('a1', 'u1'),
user('sibling', 'a1', { files: [screenshot] }),
user('u2', 'a1'),
answer('a2', 'u2', { error: true }),
];
expect(findAttachmentRecovery(messages, messages[4])).toBeNull();
});
it('offers nothing to retry for a files-only message, since there is no text to resend', () => {
const messages = [
user('u1', Constants.NO_PARENT, { text: '', files: [screenshot] }),
answer('a1', 'u1', { error: true }),
];
expect(findAttachmentRecovery(messages, messages[1])).toBeNull();
});
it('offers nothing on a user turn or a response with no user turn behind it', () => {
const messages = [
user('u1', Constants.NO_PARENT, { files: [screenshot] }),
answer('a1', 'u1'),
answer('a2', 'a1', { error: true }),
];
expect(findAttachmentRecovery(messages, messages[0])).toBeNull();
expect(findAttachmentRecovery(messages, messages[2])).toBeNull();
expect(findAttachmentRecovery(undefined, messages[1])).toBeNull();
});
});