606 lines
21 KiB
TypeScript
606 lines
21 KiB
TypeScript
/* This file is fucking horrendous, sorry */
|
|
import { Request, Response } from "express";
|
|
import * as http from "http";
|
|
import util from "util";
|
|
import zlib from "zlib";
|
|
import { enqueue, trackWaitTime } from "../../queue";
|
|
import { HttpError } from "../../../shared/errors";
|
|
import { keyPool } from "../../../shared/key-management";
|
|
import { getOpenAIModelFamily } from "../../../shared/models";
|
|
import { countTokens } from "../../../shared/tokenization";
|
|
import {
|
|
incrementPromptCount,
|
|
incrementTokenCount,
|
|
} from "../../../shared/users/user-store";
|
|
import { assertNever } from "../../../shared/utils";
|
|
import { refundLastAttempt } from "../../rate-limit";
|
|
import {
|
|
getCompletionFromBody,
|
|
isImageGenerationRequest,
|
|
isTextGenerationRequest,
|
|
writeErrorResponse,
|
|
} from "../common";
|
|
import { handleStreamedResponse } from "./handle-streamed-response";
|
|
import { logPrompt } from "./log-prompt";
|
|
import { saveImage } from "./save-image";
|
|
|
|
const DECODER_MAP = {
|
|
gzip: util.promisify(zlib.gunzip),
|
|
deflate: util.promisify(zlib.inflate),
|
|
br: util.promisify(zlib.brotliDecompress),
|
|
};
|
|
|
|
const isSupportedContentEncoding = (
|
|
contentEncoding: string
|
|
): contentEncoding is keyof typeof DECODER_MAP => {
|
|
return contentEncoding in DECODER_MAP;
|
|
};
|
|
|
|
export class RetryableError extends Error {
|
|
constructor(message: string) {
|
|
super(message);
|
|
this.name = "RetryableError";
|
|
}
|
|
}
|
|
|
|
/**
|
|
* Either decodes or streams the entire response body and then passes it as the
|
|
* last argument to the rest of the middleware stack.
|
|
*/
|
|
export type RawResponseBodyHandler = (
|
|
proxyRes: http.IncomingMessage,
|
|
req: Request,
|
|
res: Response
|
|
) => Promise<string | Record<string, any>>;
|
|
export type ProxyResHandlerWithBody = (
|
|
proxyRes: http.IncomingMessage,
|
|
req: Request,
|
|
res: Response,
|
|
/**
|
|
* This will be an object if the response content-type is application/json,
|
|
* or if the response is a streaming response. Otherwise it will be a string.
|
|
*/
|
|
body: string | Record<string, any>
|
|
) => Promise<void>;
|
|
export type ProxyResMiddleware = ProxyResHandlerWithBody[];
|
|
|
|
/**
|
|
* Returns a on.proxyRes handler that executes the given middleware stack after
|
|
* the common proxy response handlers have processed the response and decoded
|
|
* the body. Custom middleware won't execute if the response is determined to
|
|
* be an error from the upstream service as the response will be taken over by
|
|
* the common error handler.
|
|
*
|
|
* For streaming responses, the handleStream middleware will block remaining
|
|
* middleware from executing as it consumes the stream and forwards events to
|
|
* the client. Once the stream is closed, the finalized body will be attached
|
|
* to res.body and the remaining middleware will execute.
|
|
*/
|
|
export const createOnProxyResHandler = (apiMiddleware: ProxyResMiddleware) => {
|
|
return async (
|
|
proxyRes: http.IncomingMessage,
|
|
req: Request,
|
|
res: Response
|
|
) => {
|
|
const initialHandler = req.isStreaming
|
|
? handleStreamedResponse
|
|
: decodeResponseBody;
|
|
|
|
let lastMiddleware = initialHandler.name;
|
|
|
|
try {
|
|
const body = await initialHandler(proxyRes, req, res);
|
|
|
|
const middlewareStack: ProxyResMiddleware = [];
|
|
|
|
if (req.isStreaming) {
|
|
// `handleStreamedResponse` writes to the response and ends it, so
|
|
// we can only execute middleware that doesn't write to the response.
|
|
middlewareStack.push(
|
|
trackRateLimit,
|
|
countResponseTokens,
|
|
incrementUsage,
|
|
logPrompt
|
|
);
|
|
} else {
|
|
middlewareStack.push(
|
|
trackRateLimit,
|
|
handleUpstreamErrors,
|
|
countResponseTokens,
|
|
incrementUsage,
|
|
copyHttpHeaders,
|
|
saveImage,
|
|
logPrompt,
|
|
...apiMiddleware
|
|
);
|
|
}
|
|
|
|
for (const middleware of middlewareStack) {
|
|
lastMiddleware = middleware.name;
|
|
await middleware(proxyRes, req, res, body);
|
|
}
|
|
|
|
trackWaitTime(req);
|
|
} catch (error) {
|
|
// Hack: if the error is a retryable rate-limit error, the request has
|
|
// been re-enqueued and we can just return without doing anything else.
|
|
if (error instanceof RetryableError) {
|
|
return;
|
|
}
|
|
|
|
// Already logged and responded to the client by handleUpstreamErrors
|
|
if (error instanceof HttpError) {
|
|
if (!res.writableEnded) res.end();
|
|
return;
|
|
}
|
|
|
|
const { stack, message } = error;
|
|
const info = { stack, lastMiddleware, key: req.key?.hash };
|
|
const description = `Error while executing proxy response middleware: ${lastMiddleware} (${message})`;
|
|
|
|
if (res.headersSent) {
|
|
req.log.error(info, description);
|
|
if (!res.writableEnded) res.end();
|
|
return;
|
|
} else {
|
|
req.log.error(info, description);
|
|
res
|
|
.status(500)
|
|
.json({ error: "Internal server error", proxy_note: description });
|
|
}
|
|
}
|
|
};
|
|
};
|
|
|
|
function reenqueueRequest(req: Request) {
|
|
req.log.info(
|
|
{ key: req.key?.hash, retryCount: req.retryCount },
|
|
`Re-enqueueing request due to retryable error`
|
|
);
|
|
req.retryCount++;
|
|
enqueue(req);
|
|
}
|
|
|
|
/**
|
|
* Handles the response from the upstream service and decodes the body if
|
|
* necessary. If the response is JSON, it will be parsed and returned as an
|
|
* object. Otherwise, it will be returned as a string.
|
|
* @throws {Error} Unsupported content-encoding or invalid application/json body
|
|
*/
|
|
export const decodeResponseBody: RawResponseBodyHandler = async (
|
|
proxyRes,
|
|
req,
|
|
res
|
|
) => {
|
|
if (req.isStreaming) {
|
|
const err = new Error("decodeResponseBody called for a streaming request.");
|
|
req.log.error({ stack: err.stack, api: req.inboundApi }, err.message);
|
|
throw err;
|
|
}
|
|
|
|
return new Promise<string>((resolve, reject) => {
|
|
let chunks: Buffer[] = [];
|
|
proxyRes.on("data", (chunk) => chunks.push(chunk));
|
|
proxyRes.on("end", async () => {
|
|
let body = Buffer.concat(chunks);
|
|
|
|
const contentEncoding = proxyRes.headers["content-encoding"];
|
|
if (contentEncoding) {
|
|
if (isSupportedContentEncoding(contentEncoding)) {
|
|
const decoder = DECODER_MAP[contentEncoding];
|
|
body = await decoder(body);
|
|
} else {
|
|
const errorMessage = `Proxy received response with unsupported content-encoding: ${contentEncoding}`;
|
|
req.log.warn({ contentEncoding, key: req.key?.hash }, errorMessage);
|
|
writeErrorResponse(req, res, 500, {
|
|
error: errorMessage,
|
|
contentEncoding,
|
|
});
|
|
return reject(errorMessage);
|
|
}
|
|
}
|
|
|
|
try {
|
|
if (proxyRes.headers["content-type"]?.includes("application/json")) {
|
|
const json = JSON.parse(body.toString());
|
|
return resolve(json);
|
|
}
|
|
return resolve(body.toString());
|
|
} catch (error: any) {
|
|
const errorMessage = `Proxy received response with invalid JSON: ${error.message}`;
|
|
req.log.warn({ error: error.stack, key: req.key?.hash }, errorMessage);
|
|
writeErrorResponse(req, res, 500, { error: errorMessage });
|
|
return reject(errorMessage);
|
|
}
|
|
});
|
|
});
|
|
};
|
|
|
|
type ProxiedErrorPayload = {
|
|
error?: Record<string, any>;
|
|
message?: string;
|
|
proxy_note?: string;
|
|
};
|
|
|
|
/**
|
|
* Handles non-2xx responses from the upstream service. If the proxied response
|
|
* is an error, this will respond to the client with an error payload and throw
|
|
* an error to stop the middleware stack.
|
|
* On 429 errors, if request queueing is enabled, the request will be silently
|
|
* re-enqueued. Otherwise, the request will be rejected with an error payload.
|
|
* @throws {HttpError} On HTTP error status code from upstream service
|
|
*/
|
|
const handleUpstreamErrors: ProxyResHandlerWithBody = async (
|
|
proxyRes,
|
|
req,
|
|
res,
|
|
body
|
|
) => {
|
|
const statusCode = proxyRes.statusCode || 500;
|
|
|
|
if (statusCode < 400) {
|
|
return;
|
|
}
|
|
|
|
let errorPayload: ProxiedErrorPayload;
|
|
const tryAgainMessage = keyPool.available(req.body?.model)
|
|
? `There may be more keys available for this model; try again in a few seconds.`
|
|
: "There are no more keys available for this model.";
|
|
|
|
try {
|
|
assertJsonResponse(body);
|
|
errorPayload = body;
|
|
} catch (parseError) {
|
|
// Likely Bad Gateway or Gateway Timeout from upstream's reverse proxy
|
|
const hash = req.key?.hash;
|
|
const statusMessage = proxyRes.statusMessage || "Unknown error";
|
|
req.log.warn({ statusCode, statusMessage, key: hash }, parseError.message);
|
|
|
|
const errorObject = {
|
|
statusCode,
|
|
statusMessage: proxyRes.statusMessage,
|
|
error: parseError.message,
|
|
proxy_note: `This is likely a temporary error with the upstream service.`,
|
|
};
|
|
writeErrorResponse(req, res, statusCode, errorObject);
|
|
throw new HttpError(statusCode, parseError.message);
|
|
}
|
|
|
|
const errorType =
|
|
errorPayload.error?.code ||
|
|
errorPayload.error?.type ||
|
|
getAwsErrorType(proxyRes.headers["x-amzn-errortype"]);
|
|
|
|
req.log.warn(
|
|
{ statusCode, type: errorType, errorPayload, key: req.key?.hash },
|
|
`Received error response from upstream. (${proxyRes.statusMessage})`
|
|
);
|
|
|
|
const service = req.key!.service;
|
|
if (service === "aws") {
|
|
// Try to standardize the error format for AWS
|
|
errorPayload.error = { message: errorPayload.message, type: errorType };
|
|
delete errorPayload.message;
|
|
}
|
|
|
|
if (statusCode === 400) {
|
|
// Bad request. For OpenAI, this is usually due to prompt length.
|
|
// For Anthropic, this is usually due to missing preamble.
|
|
switch (service) {
|
|
case "openai":
|
|
case "google-palm":
|
|
case "azure":
|
|
const filteredCodes = ["content_policy_violation", "content_filter"];
|
|
if (filteredCodes.includes(errorPayload.error?.code)) {
|
|
errorPayload.proxy_note = `Request was filtered by the upstream API's content moderation system. Modify your prompt and try again.`;
|
|
refundLastAttempt(req);
|
|
} else if (errorPayload.error?.code === "billing_hard_limit_reached") {
|
|
// For some reason, some models return this 400 error instead of the
|
|
// same 429 billing error that other models return.
|
|
handleOpenAIRateLimitError(req, tryAgainMessage, errorPayload);
|
|
} else {
|
|
errorPayload.proxy_note = `The upstream API rejected the request. Your prompt may be too long for ${req.body?.model}.`;
|
|
}
|
|
break;
|
|
case "anthropic":
|
|
case "aws":
|
|
maybeHandleMissingPreambleError(req, errorPayload);
|
|
break;
|
|
default:
|
|
assertNever(service);
|
|
}
|
|
} else if (statusCode === 401) {
|
|
// Key is invalid or was revoked
|
|
keyPool.disable(req.key!, "revoked");
|
|
errorPayload.proxy_note = `API key is invalid or revoked. ${tryAgainMessage}`;
|
|
} else if (statusCode === 403) {
|
|
// Amazon is the only service that returns 403.
|
|
switch (errorType) {
|
|
case "UnrecognizedClientException":
|
|
// Key is invalid.
|
|
keyPool.disable(req.key!, "revoked");
|
|
errorPayload.proxy_note = `API key is invalid or revoked. ${tryAgainMessage}`;
|
|
break;
|
|
case "AccessDeniedException":
|
|
req.log.error(
|
|
{ key: req.key?.hash, model: req.body?.model },
|
|
"Disabling key due to AccessDeniedException when invoking model. If credentials are valid, check IAM permissions."
|
|
);
|
|
keyPool.disable(req.key!, "revoked");
|
|
errorPayload.proxy_note = `API key doesn't have access to the requested resource.`;
|
|
break;
|
|
default:
|
|
errorPayload.proxy_note = `Received 403 error. Key may be invalid.`;
|
|
}
|
|
} else if (statusCode === 429) {
|
|
switch (service) {
|
|
case "openai":
|
|
handleOpenAIRateLimitError(req, tryAgainMessage, errorPayload);
|
|
break;
|
|
case "anthropic":
|
|
handleAnthropicRateLimitError(req, errorPayload);
|
|
break;
|
|
case "aws":
|
|
handleAwsRateLimitError(req, errorPayload);
|
|
break;
|
|
case "google-palm":
|
|
case "azure":
|
|
errorPayload.proxy_note = `Automatic rate limit retries are not supported for this service. Try again in a few seconds.`;
|
|
break;
|
|
default:
|
|
assertNever(service);
|
|
}
|
|
} else if (statusCode === 404) {
|
|
// Most likely model not found
|
|
switch (service) {
|
|
case "openai":
|
|
if (errorPayload.error?.code === "model_not_found") {
|
|
const requestedModel = req.body.model;
|
|
const modelFamily = getOpenAIModelFamily(requestedModel);
|
|
errorPayload.proxy_note = `The key assigned to your prompt does not support the requested model (${requestedModel}, family: ${modelFamily}).`;
|
|
req.log.error(
|
|
{ key: req.key?.hash, model: requestedModel, modelFamily },
|
|
"Prompt was routed to a key that does not support the requested model."
|
|
);
|
|
}
|
|
break;
|
|
case "anthropic":
|
|
errorPayload.proxy_note = `The requested Claude model might not exist, or the key might not be provisioned for it.`;
|
|
break;
|
|
case "google-palm":
|
|
errorPayload.proxy_note = `The requested Google PaLM model might not exist, or the key might not be provisioned for it.`;
|
|
break;
|
|
case "aws":
|
|
errorPayload.proxy_note = `The requested AWS resource might not exist, or the key might not have access to it.`;
|
|
break;
|
|
case "azure":
|
|
errorPayload.proxy_note = `The assigned Azure deployment does not support the requested model.`;
|
|
break;
|
|
default:
|
|
assertNever(service);
|
|
}
|
|
} else {
|
|
errorPayload.proxy_note = `Unrecognized error from upstream service.`;
|
|
}
|
|
|
|
// Some OAI errors contain the organization ID, which we don't want to reveal.
|
|
if (errorPayload.error?.message) {
|
|
errorPayload.error.message = errorPayload.error.message.replace(
|
|
/org-.{24}/gm,
|
|
"org-xxxxxxxxxxxxxxxxxxx"
|
|
);
|
|
}
|
|
|
|
writeErrorResponse(req, res, statusCode, errorPayload);
|
|
throw new HttpError(statusCode, errorPayload.error?.message);
|
|
};
|
|
|
|
/**
|
|
* This is a workaround for a very strange issue where certain API keys seem to
|
|
* enforce more strict input validation than others -- specifically, they will
|
|
* require a `\n\nHuman:` prefix on the prompt, perhaps to prevent the key from
|
|
* being used as a generic text completion service and to enforce the use of
|
|
* the chat RLHF. This is not documented anywhere, and it's not clear why some
|
|
* keys enforce this and others don't.
|
|
* This middleware checks for that specific error and marks the key as being
|
|
* one that requires the prefix, and then re-enqueues the request.
|
|
* The exact error is:
|
|
* ```
|
|
* {
|
|
* "error": {
|
|
* "type": "invalid_request_error",
|
|
* "message": "prompt must start with \"\n\nHuman:\" turn"
|
|
* }
|
|
* }
|
|
* ```
|
|
*/
|
|
function maybeHandleMissingPreambleError(
|
|
req: Request,
|
|
errorPayload: ProxiedErrorPayload
|
|
) {
|
|
if (
|
|
errorPayload.error?.type === "invalid_request_error" &&
|
|
errorPayload.error?.message === 'prompt must start with "\n\nHuman:" turn'
|
|
) {
|
|
req.log.warn(
|
|
{ key: req.key?.hash },
|
|
"Request failed due to missing preamble. Key will be marked as such for subsequent requests."
|
|
);
|
|
keyPool.update(req.key!, { requiresPreamble: true });
|
|
reenqueueRequest(req);
|
|
throw new RetryableError("Claude request re-enqueued to add preamble.");
|
|
} else {
|
|
errorPayload.proxy_note = `Proxy received unrecognized error from Anthropic. Check the specific error for more information.`;
|
|
}
|
|
}
|
|
|
|
function handleAnthropicRateLimitError(
|
|
req: Request,
|
|
errorPayload: ProxiedErrorPayload
|
|
) {
|
|
if (errorPayload.error?.type === "rate_limit_error") {
|
|
keyPool.markRateLimited(req.key!);
|
|
reenqueueRequest(req);
|
|
throw new RetryableError("Claude rate-limited request re-enqueued.");
|
|
} else {
|
|
errorPayload.proxy_note = `Unrecognized rate limit error from Anthropic. Key may be over quota.`;
|
|
}
|
|
}
|
|
|
|
function handleAwsRateLimitError(
|
|
req: Request,
|
|
errorPayload: ProxiedErrorPayload
|
|
) {
|
|
const errorType = errorPayload.error?.type;
|
|
switch (errorType) {
|
|
case "ThrottlingException":
|
|
keyPool.markRateLimited(req.key!);
|
|
reenqueueRequest(req);
|
|
throw new RetryableError("AWS rate-limited request re-enqueued.");
|
|
case "ModelNotReadyException":
|
|
errorPayload.proxy_note = `The requested model is overloaded. Try again in a few seconds.`;
|
|
break;
|
|
default:
|
|
errorPayload.proxy_note = `Unrecognized rate limit error from AWS. (${errorType})`;
|
|
}
|
|
}
|
|
|
|
function handleOpenAIRateLimitError(
|
|
req: Request,
|
|
tryAgainMessage: string,
|
|
errorPayload: ProxiedErrorPayload
|
|
): Record<string, any> {
|
|
const type = errorPayload.error?.type;
|
|
switch (type) {
|
|
case "insufficient_quota":
|
|
case "invalid_request_error": // this is the billing_hard_limit_reached error seen in some cases
|
|
// Billing quota exceeded (key is dead, disable it)
|
|
keyPool.disable(req.key!, "quota");
|
|
errorPayload.proxy_note = `Assigned key's quota has been exceeded. ${tryAgainMessage}`;
|
|
break;
|
|
case "access_terminated":
|
|
// Account banned (key is dead, disable it)
|
|
keyPool.disable(req.key!, "revoked");
|
|
errorPayload.proxy_note = `Assigned key has been banned by OpenAI for policy violations. ${tryAgainMessage}`;
|
|
break;
|
|
case "billing_not_active":
|
|
// Key valid but account billing is delinquent
|
|
keyPool.disable(req.key!, "quota");
|
|
errorPayload.proxy_note = `Assigned key has been disabled due to delinquent billing. ${tryAgainMessage}`;
|
|
break;
|
|
case "requests":
|
|
case "tokens":
|
|
keyPool.markRateLimited(req.key!);
|
|
if (errorPayload.error?.message?.match(/on requests per day/)) {
|
|
// This key has a very low rate limit, so we can't re-enqueue it.
|
|
errorPayload.proxy_note = `Assigned key has reached its per-day request limit for this model. Try another model.`;
|
|
break;
|
|
}
|
|
|
|
// Per-minute request or token rate limit is exceeded, which we can retry
|
|
reenqueueRequest(req);
|
|
throw new RetryableError("Rate-limited request re-enqueued.");
|
|
default:
|
|
errorPayload.proxy_note = `This is likely a temporary error with OpenAI. Try again in a few seconds.`;
|
|
break;
|
|
}
|
|
return errorPayload;
|
|
}
|
|
|
|
const incrementUsage: ProxyResHandlerWithBody = async (_proxyRes, req) => {
|
|
if (isTextGenerationRequest(req) || isImageGenerationRequest(req)) {
|
|
const model = req.body.model;
|
|
const tokensUsed = req.promptTokens! + req.outputTokens!;
|
|
req.log.debug(
|
|
{
|
|
model,
|
|
tokensUsed,
|
|
promptTokens: req.promptTokens,
|
|
outputTokens: req.outputTokens,
|
|
},
|
|
`Incrementing usage for model`
|
|
);
|
|
keyPool.incrementUsage(req.key!, model, tokensUsed);
|
|
if (req.user) {
|
|
incrementPromptCount(req.user.token);
|
|
incrementTokenCount(req.user.token, model, req.outboundApi, tokensUsed);
|
|
}
|
|
}
|
|
};
|
|
|
|
const countResponseTokens: ProxyResHandlerWithBody = async (
|
|
_proxyRes,
|
|
req,
|
|
_res,
|
|
body
|
|
) => {
|
|
if (req.outboundApi === "openai-image") {
|
|
req.outputTokens = req.promptTokens;
|
|
req.promptTokens = 0;
|
|
return;
|
|
}
|
|
|
|
// This function is prone to breaking if the upstream API makes even minor
|
|
// changes to the response format, especially for SSE responses. If you're
|
|
// seeing errors in this function, check the reassembled response body from
|
|
// handleStreamedResponse to see if the upstream API has changed.
|
|
try {
|
|
assertJsonResponse(body);
|
|
const service = req.outboundApi;
|
|
const completion = getCompletionFromBody(req, body);
|
|
const tokens = await countTokens({ req, completion, service });
|
|
|
|
req.log.debug(
|
|
{ service, tokens, prevOutputTokens: req.outputTokens },
|
|
`Counted tokens for completion`
|
|
);
|
|
if (req.tokenizerInfo) {
|
|
req.tokenizerInfo.completion_tokens = tokens;
|
|
}
|
|
|
|
req.outputTokens = tokens.token_count;
|
|
} catch (error) {
|
|
req.log.warn(
|
|
error,
|
|
"Error while counting completion tokens; assuming `max_output_tokens`"
|
|
);
|
|
// req.outputTokens will already be set to `max_output_tokens` from the
|
|
// prompt counting middleware, so we don't need to do anything here.
|
|
}
|
|
};
|
|
|
|
const trackRateLimit: ProxyResHandlerWithBody = async (proxyRes, req) => {
|
|
keyPool.updateRateLimits(req.key!, proxyRes.headers);
|
|
};
|
|
|
|
const copyHttpHeaders: ProxyResHandlerWithBody = async (
|
|
proxyRes,
|
|
_req,
|
|
res
|
|
) => {
|
|
Object.keys(proxyRes.headers).forEach((key) => {
|
|
// Omit content-encoding because we will always decode the response body
|
|
if (key === "content-encoding") {
|
|
return;
|
|
}
|
|
// We're usually using res.json() to send the response, which causes express
|
|
// to set content-length. That's not valid for chunked responses and some
|
|
// clients will reject it so we need to omit it.
|
|
if (key === "transfer-encoding") {
|
|
return;
|
|
}
|
|
res.setHeader(key, proxyRes.headers[key] as string);
|
|
});
|
|
};
|
|
|
|
function getAwsErrorType(header: string | string[] | undefined) {
|
|
const val = String(header).match(/^(\w+):?/)?.[1];
|
|
return val || String(header);
|
|
}
|
|
|
|
function assertJsonResponse(body: any): asserts body is Record<string, any> {
|
|
if (typeof body !== "object") {
|
|
throw new Error("Expected response to be an object");
|
|
}
|
|
}
|