mirror of
https://github.com/danny-avila/LibreChat.git
synced 2025-09-22 06:00:56 +02:00
👤 feat: User Placeholder Variables for Custom Endpoint Headers (#7993)
* 🔧 refactor: move `processMCPEnv` from `librechat-data-provider` and move to `@librechat/api` * 🔧 refactor: Update resolveHeaders import paths * 🔧 refactor: Enhance resolveHeaders to support user and custom variables - Updated resolveHeaders function to accept user and custom user variables for placeholder replacement. - Modified header resolution in multiple client and controller files to utilize the enhanced resolveHeaders functionality. - Added comprehensive tests for resolveHeaders to ensure correct processing of user and custom variables. * 🔧 fix: Update user ID placeholder processing in env.ts * 🔧 fix: Remove arguments passing this.user rather than req.user - Updated multiple client and controller files to call resolveHeaders without the user parameter * 🔧 refactor: Enhance processUserPlaceholders to be more readable / less nested * 🔧 refactor: Update processUserPlaceholders to pass all tests in mpc.spec.ts and env.spec.ts * chore: remove legacy ChatGPTClient * chore: remove LLM initialization code * chore: initial deprecation removal of `gptPlugins` * chore: remove cohere-ai dependency from package.json and package-lock.json * chore: update brace-expansion to version 2.0.2 and add license information * chore: remove PluginsClient test file * chore: remove legacy * ci: remove deprecated sendMessage/getCompletion/chatCompletion tests --------- Co-authored-by: Dustin Healy <54083382+dustinhealy@users.noreply.github.com>
This commit is contained in:
Danny Avila
GitHub
parent
01e9b196bc
commit
a058963a9f
30 changed files with 542 additions and 2844 deletions
|
|
@ -1,804 +0,0 @@
|
|||
const { Keyv } = require('keyv');
|
||||
const crypto = require('crypto');
|
||||
const { CohereClient } = require('cohere-ai');
|
||||
const { fetchEventSource } = require('@waylaidwanderer/fetch-event-source');
|
||||
const { constructAzureURL, genAzureChatCompletion } = require('@librechat/api');
|
||||
const { encoding_for_model: encodingForModel, get_encoding: getEncoding } = require('tiktoken');
|
||||
const {
|
||||
ImageDetail,
|
||||
EModelEndpoint,
|
||||
resolveHeaders,
|
||||
CohereConstants,
|
||||
mapModelToAzureConfig,
|
||||
} = require('librechat-data-provider');
|
||||
const { createContextHandlers } = require('./prompts');
|
||||
const { createCoherePayload } = require('./llm');
|
||||
const { extractBaseURL } = require('~/utils');
|
||||
const BaseClient = require('./BaseClient');
|
||||
const { logger } = require('~/config');
|
||||
|
||||
const CHATGPT_MODEL = 'gpt-3.5-turbo';
|
||||
const tokenizersCache = {};
|
||||
|
||||
class ChatGPTClient extends BaseClient {
|
||||
constructor(apiKey, options = {}, cacheOptions = {}) {
|
||||
super(apiKey, options, cacheOptions);
|
||||
|
||||
cacheOptions.namespace = cacheOptions.namespace || 'chatgpt';
|
||||
this.conversationsCache = new Keyv(cacheOptions);
|
||||
this.setOptions(options);
|
||||
}
|
||||
|
||||
setOptions(options) {
|
||||
if (this.options && !this.options.replaceOptions) {
|
||||
// nested options aren't spread properly, so we need to do this manually
|
||||
this.options.modelOptions = {
|
||||
...this.options.modelOptions,
|
||||
...options.modelOptions,
|
||||
};
|
||||
delete options.modelOptions;
|
||||
// now we can merge options
|
||||
this.options = {
|
||||
...this.options,
|
||||
...options,
|
||||
};
|
||||
} else {
|
||||
this.options = options;
|
||||
}
|
||||
|
||||
if (this.options.openaiApiKey) {
|
||||
this.apiKey = this.options.openaiApiKey;
|
||||
}
|
||||
|
||||
const modelOptions = this.options.modelOptions || {};
|
||||
this.modelOptions = {
|
||||
...modelOptions,
|
||||
// set some good defaults (check for undefined in some cases because they may be 0)
|
||||
model: modelOptions.model || CHATGPT_MODEL,
|
||||
temperature: typeof modelOptions.temperature === 'undefined' ? 0.8 : modelOptions.temperature,
|
||||
top_p: typeof modelOptions.top_p === 'undefined' ? 1 : modelOptions.top_p,
|
||||
presence_penalty:
|
||||
typeof modelOptions.presence_penalty === 'undefined' ? 1 : modelOptions.presence_penalty,
|
||||
stop: modelOptions.stop,
|
||||
};
|
||||
|
||||
this.isChatGptModel = this.modelOptions.model.includes('gpt-');
|
||||
const { isChatGptModel } = this;
|
||||
this.isUnofficialChatGptModel =
|
||||
this.modelOptions.model.startsWith('text-chat') ||
|
||||
this.modelOptions.model.startsWith('text-davinci-002-render');
|
||||
const { isUnofficialChatGptModel } = this;
|
||||
|
||||
// Davinci models have a max context length of 4097 tokens.
|
||||
this.maxContextTokens = this.options.maxContextTokens || (isChatGptModel ? 4095 : 4097);
|
||||
// I decided to reserve 1024 tokens for the response.
|
||||
// The max prompt tokens is determined by the max context tokens minus the max response tokens.
|
||||
// Earlier messages will be dropped until the prompt is within the limit.
|
||||
this.maxResponseTokens = this.modelOptions.max_tokens || 1024;
|
||||
this.maxPromptTokens =
|
||||
this.options.maxPromptTokens || this.maxContextTokens - this.maxResponseTokens;
|
||||
|
||||
if (this.maxPromptTokens + this.maxResponseTokens > this.maxContextTokens) {
|
||||
throw new Error(
|
||||
`maxPromptTokens + max_tokens (${this.maxPromptTokens} + ${this.maxResponseTokens} = ${
|
||||
this.maxPromptTokens + this.maxResponseTokens
|
||||
}) must be less than or equal to maxContextTokens (${this.maxContextTokens})`,
|
||||
);
|
||||
}
|
||||
|
||||
this.userLabel = this.options.userLabel || 'User';
|
||||
this.chatGptLabel = this.options.chatGptLabel || 'ChatGPT';
|
||||
|
||||
if (isChatGptModel) {
|
||||
// Use these faux tokens to help the AI understand the context since we are building the chat log ourselves.
|
||||
// Trying to use "<|im_start|>" causes the AI to still generate "<" or "<|" at the end sometimes for some reason,
|
||||
// without tripping the stop sequences, so I'm using "||>" instead.
|
||||
this.startToken = '||>';
|
||||
this.endToken = '';
|
||||
this.gptEncoder = this.constructor.getTokenizer('cl100k_base');
|
||||
} else if (isUnofficialChatGptModel) {
|
||||
this.startToken = '<|im_start|>';
|
||||
this.endToken = '<|im_end|>';
|
||||
this.gptEncoder = this.constructor.getTokenizer('text-davinci-003', true, {
|
||||
'<|im_start|>': 100264,
|
||||
'<|im_end|>': 100265,
|
||||
});
|
||||
} else {
|
||||
// Previously I was trying to use "<|endoftext|>" but there seems to be some bug with OpenAI's token counting
|
||||
// system that causes only the first "<|endoftext|>" to be counted as 1 token, and the rest are not treated
|
||||
// as a single token. So we're using this instead.
|
||||
this.startToken = '||>';
|
||||
this.endToken = '';
|
||||
try {
|
||||
this.gptEncoder = this.constructor.getTokenizer(this.modelOptions.model, true);
|
||||
} catch {
|
||||
this.gptEncoder = this.constructor.getTokenizer('text-davinci-003', true);
|
||||
}
|
||||
}
|
||||
|
||||
if (!this.modelOptions.stop) {
|
||||
const stopTokens = [this.startToken];
|
||||
if (this.endToken && this.endToken !== this.startToken) {
|
||||
stopTokens.push(this.endToken);
|
||||
}
|
||||
stopTokens.push(`\n${this.userLabel}:`);
|
||||
stopTokens.push('<|diff_marker|>');
|
||||
// I chose not to do one for `chatGptLabel` because I've never seen it happen
|
||||
this.modelOptions.stop = stopTokens;
|
||||
}
|
||||
|
||||
if (this.options.reverseProxyUrl) {
|
||||
this.completionsUrl = this.options.reverseProxyUrl;
|
||||
} else if (isChatGptModel) {
|
||||
this.completionsUrl = 'https://api.openai.com/v1/chat/completions';
|
||||
} else {
|
||||
this.completionsUrl = 'https://api.openai.com/v1/completions';
|
||||
}
|
||||
|
||||
return this;
|
||||
}
|
||||
|
||||
static getTokenizer(encoding, isModelName = false, extendSpecialTokens = {}) {
|
||||
if (tokenizersCache[encoding]) {
|
||||
return tokenizersCache[encoding];
|
||||
}
|
||||
let tokenizer;
|
||||
if (isModelName) {
|
||||
tokenizer = encodingForModel(encoding, extendSpecialTokens);
|
||||
} else {
|
||||
tokenizer = getEncoding(encoding, extendSpecialTokens);
|
||||
}
|
||||
tokenizersCache[encoding] = tokenizer;
|
||||
return tokenizer;
|
||||
}
|
||||
|
||||
/** @type {getCompletion} */
|
||||
async getCompletion(input, onProgress, onTokenProgress, abortController = null) {
|
||||
if (!abortController) {
|
||||
abortController = new AbortController();
|
||||
}
|
||||
|
||||
let modelOptions = { ...this.modelOptions };
|
||||
if (typeof onProgress === 'function') {
|
||||
modelOptions.stream = true;
|
||||
}
|
||||
if (this.isChatGptModel) {
|
||||
modelOptions.messages = input;
|
||||
} else {
|
||||
modelOptions.prompt = input;
|
||||
}
|
||||
|
||||
if (this.useOpenRouter && modelOptions.prompt) {
|
||||
delete modelOptions.stop;
|
||||
}
|
||||
|
||||
const { debug } = this.options;
|
||||
let baseURL = this.completionsUrl;
|
||||
if (debug) {
|
||||
console.debug();
|
||||
console.debug(baseURL);
|
||||
console.debug(modelOptions);
|
||||
console.debug();
|
||||
}
|
||||
|
||||
const opts = {
|
||||
method: 'POST',
|
||||
headers: {
|
||||
'Content-Type': 'application/json',
|
||||
},
|
||||
};
|
||||
|
||||
if (this.isVisionModel) {
|
||||
modelOptions.max_tokens = 4000;
|
||||
}
|
||||
|
||||
/** @type {TAzureConfig | undefined} */
|
||||
const azureConfig = this.options?.req?.app?.locals?.[EModelEndpoint.azureOpenAI];
|
||||
|
||||
const isAzure = this.azure || this.options.azure;
|
||||
if (
|
||||
(isAzure && this.isVisionModel && azureConfig) ||
|
||||
(azureConfig && this.isVisionModel && this.options.endpoint === EModelEndpoint.azureOpenAI)
|
||||
) {
|
||||
const { modelGroupMap, groupMap } = azureConfig;
|
||||
const {
|
||||
azureOptions,
|
||||
baseURL,
|
||||
headers = {},
|
||||
serverless,
|
||||
} = mapModelToAzureConfig({
|
||||
modelName: modelOptions.model,
|
||||
modelGroupMap,
|
||||
groupMap,
|
||||
});
|
||||
opts.headers = resolveHeaders(headers);
|
||||
this.langchainProxy = extractBaseURL(baseURL);
|
||||
this.apiKey = azureOptions.azureOpenAIApiKey;
|
||||
|
||||
const groupName = modelGroupMap[modelOptions.model].group;
|
||||
this.options.addParams = azureConfig.groupMap[groupName].addParams;
|
||||
this.options.dropParams = azureConfig.groupMap[groupName].dropParams;
|
||||
// Note: `forcePrompt` not re-assigned as only chat models are vision models
|
||||
|
||||
this.azure = !serverless && azureOptions;
|
||||
this.azureEndpoint =
|
||||
!serverless && genAzureChatCompletion(this.azure, modelOptions.model, this);
|
||||
if (serverless === true) {
|
||||
this.options.defaultQuery = azureOptions.azureOpenAIApiVersion
|
||||
? { 'api-version': azureOptions.azureOpenAIApiVersion }
|
||||
: undefined;
|
||||
this.options.headers['api-key'] = this.apiKey;
|
||||
}
|
||||
}
|
||||
|
||||
if (this.options.defaultQuery) {
|
||||
opts.defaultQuery = this.options.defaultQuery;
|
||||
}
|
||||
|
||||
if (this.options.headers) {
|
||||
opts.headers = { ...opts.headers, ...this.options.headers };
|
||||
}
|
||||
|
||||
if (isAzure) {
|
||||
// Azure does not accept `model` in the body, so we need to remove it.
|
||||
delete modelOptions.model;
|
||||
|
||||
baseURL = this.langchainProxy
|
||||
? constructAzureURL({
|
||||
baseURL: this.langchainProxy,
|
||||
azureOptions: this.azure,
|
||||
})
|
||||
: this.azureEndpoint.split(/(?<!\/)\/(chat|completion)\//)[0];
|
||||
|
||||
if (this.options.forcePrompt) {
|
||||
baseURL += '/completions';
|
||||
} else {
|
||||
baseURL += '/chat/completions';
|
||||
}
|
||||
|
||||
opts.defaultQuery = { 'api-version': this.azure.azureOpenAIApiVersion };
|
||||
opts.headers = { ...opts.headers, 'api-key': this.apiKey };
|
||||
} else if (this.apiKey) {
|
||||
opts.headers.Authorization = `Bearer ${this.apiKey}`;
|
||||
}
|
||||
|
||||
if (process.env.OPENAI_ORGANIZATION) {
|
||||
opts.headers['OpenAI-Organization'] = process.env.OPENAI_ORGANIZATION;
|
||||
}
|
||||
|
||||
if (this.useOpenRouter) {
|
||||
opts.headers['HTTP-Referer'] = 'https://librechat.ai';
|
||||
opts.headers['X-Title'] = 'LibreChat';
|
||||
}
|
||||
|
||||
/* hacky fixes for Mistral AI API:
|
||||
- Re-orders system message to the top of the messages payload, as not allowed anywhere else
|
||||
- If there is only one message and it's a system message, change the role to user
|
||||
*/
|
||||
if (baseURL.includes('https://api.mistral.ai/v1') && modelOptions.messages) {
|
||||
const { messages } = modelOptions;
|
||||
|
||||
const systemMessageIndex = messages.findIndex((msg) => msg.role === 'system');
|
||||
|
||||
if (systemMessageIndex > 0) {
|
||||
const [systemMessage] = messages.splice(systemMessageIndex, 1);
|
||||
messages.unshift(systemMessage);
|
||||
}
|
||||
|
||||
modelOptions.messages = messages;
|
||||
|
||||
if (messages.length === 1 && messages[0].role === 'system') {
|
||||
modelOptions.messages[0].role = 'user';
|
||||
}
|
||||
}
|
||||
|
||||
if (this.options.addParams && typeof this.options.addParams === 'object') {
|
||||
modelOptions = {
|
||||
...modelOptions,
|
||||
...this.options.addParams,
|
||||
};
|
||||
logger.debug('[ChatGPTClient] chatCompletion: added params', {
|
||||
addParams: this.options.addParams,
|
||||
modelOptions,
|
||||
});
|
||||
}
|
||||
|
||||
if (this.options.dropParams && Array.isArray(this.options.dropParams)) {
|
||||
this.options.dropParams.forEach((param) => {
|
||||
delete modelOptions[param];
|
||||
});
|
||||
logger.debug('[ChatGPTClient] chatCompletion: dropped params', {
|
||||
dropParams: this.options.dropParams,
|
||||
modelOptions,
|
||||
});
|
||||
}
|
||||
|
||||
if (baseURL.startsWith(CohereConstants.API_URL)) {
|
||||
const payload = createCoherePayload({ modelOptions });
|
||||
return await this.cohereChatCompletion({ payload, onTokenProgress });
|
||||
}
|
||||
|
||||
if (baseURL.includes('v1') && !baseURL.includes('/completions') && !this.isChatCompletion) {
|
||||
baseURL = baseURL.split('v1')[0] + 'v1/completions';
|
||||
} else if (
|
||||
baseURL.includes('v1') &&
|
||||
!baseURL.includes('/chat/completions') &&
|
||||
this.isChatCompletion
|
||||
) {
|
||||
baseURL = baseURL.split('v1')[0] + 'v1/chat/completions';
|
||||
}
|
||||
|
||||
const BASE_URL = new URL(baseURL);
|
||||
if (opts.defaultQuery) {
|
||||
Object.entries(opts.defaultQuery).forEach(([key, value]) => {
|
||||
BASE_URL.searchParams.append(key, value);
|
||||
});
|
||||
delete opts.defaultQuery;
|
||||
}
|
||||
|
||||
const completionsURL = BASE_URL.toString();
|
||||
opts.body = JSON.stringify(modelOptions);
|
||||
|
||||
if (modelOptions.stream) {
|
||||
return new Promise(async (resolve, reject) => {
|
||||
try {
|
||||
let done = false;
|
||||
await fetchEventSource(completionsURL, {
|
||||
...opts,
|
||||
signal: abortController.signal,
|
||||
async onopen(response) {
|
||||
if (response.status === 200) {
|
||||
return;
|
||||
}
|
||||
if (debug) {
|
||||
console.debug(response);
|
||||
}
|
||||
let error;
|
||||
try {
|
||||
const body = await response.text();
|
||||
error = new Error(`Failed to send message. HTTP ${response.status} - ${body}`);
|
||||
error.status = response.status;
|
||||
error.json = JSON.parse(body);
|
||||
} catch {
|
||||
error = error || new Error(`Failed to send message. HTTP ${response.status}`);
|
||||
}
|
||||
throw error;
|
||||
},
|
||||
onclose() {
|
||||
if (debug) {
|
||||
console.debug('Server closed the connection unexpectedly, returning...');
|
||||
}
|
||||
// workaround for private API not sending [DONE] event
|
||||
if (!done) {
|
||||
onProgress('[DONE]');
|
||||
resolve();
|
||||
}
|
||||
},
|
||||
onerror(err) {
|
||||
if (debug) {
|
||||
console.debug(err);
|
||||
}
|
||||
// rethrow to stop the operation
|
||||
throw err;
|
||||
},
|
||||
onmessage(message) {
|
||||
if (debug) {
|
||||
console.debug(message);
|
||||
}
|
||||
if (!message.data || message.event === 'ping') {
|
||||
return;
|
||||
}
|
||||
if (message.data === '[DONE]') {
|
||||
onProgress('[DONE]');
|
||||
resolve();
|
||||
done = true;
|
||||
return;
|
||||
}
|
||||
onProgress(JSON.parse(message.data));
|
||||
},
|
||||
});
|
||||
} catch (err) {
|
||||
reject(err);
|
||||
}
|
||||
});
|
||||
}
|
||||
const response = await fetch(completionsURL, {
|
||||
...opts,
|
||||
signal: abortController.signal,
|
||||
});
|
||||
if (response.status !== 200) {
|
||||
const body = await response.text();
|
||||
const error = new Error(`Failed to send message. HTTP ${response.status} - ${body}`);
|
||||
error.status = response.status;
|
||||
try {
|
||||
error.json = JSON.parse(body);
|
||||
} catch {
|
||||
error.body = body;
|
||||
}
|
||||
throw error;
|
||||
}
|
||||
return response.json();
|
||||
}
|
||||
|
||||
/** @type {cohereChatCompletion} */
|
||||
async cohereChatCompletion({ payload, onTokenProgress }) {
|
||||
const cohere = new CohereClient({
|
||||
token: this.apiKey,
|
||||
environment: this.completionsUrl,
|
||||
});
|
||||
|
||||
if (!payload.stream) {
|
||||
const chatResponse = await cohere.chat(payload);
|
||||
return chatResponse.text;
|
||||
}
|
||||
|
||||
const chatStream = await cohere.chatStream(payload);
|
||||
let reply = '';
|
||||
for await (const message of chatStream) {
|
||||
if (!message) {
|
||||
continue;
|
||||
}
|
||||
|
||||
if (message.eventType === 'text-generation' && message.text) {
|
||||
onTokenProgress(message.text);
|
||||
reply += message.text;
|
||||
}
|
||||
/*
|
||||
Cohere API Chinese Unicode character replacement hotfix.
|
||||
Should be un-commented when the following issue is resolved:
|
||||
https://github.com/cohere-ai/cohere-typescript/issues/151
|
||||
|
||||
else if (message.eventType === 'stream-end' && message.response) {
|
||||
reply = message.response.text;
|
||||
}
|
||||
*/
|
||||
}
|
||||
|
||||
return reply;
|
||||
}
|
||||
|
||||
async generateTitle(userMessage, botMessage) {
|
||||
const instructionsPayload = {
|
||||
role: 'system',
|
||||
content: `Write an extremely concise subtitle for this conversation with no more than a few words. All words should be capitalized. Exclude punctuation.
|
||||
|
||||
||>Message:
|
||||
${userMessage.message}
|
||||
||>Response:
|
||||
${botMessage.message}
|
||||
|
||||
||>Title:`,
|
||||
};
|
||||
|
||||
const titleGenClientOptions = JSON.parse(JSON.stringify(this.options));
|
||||
titleGenClientOptions.modelOptions = {
|
||||
model: 'gpt-3.5-turbo',
|
||||
temperature: 0,
|
||||
presence_penalty: 0,
|
||||
frequency_penalty: 0,
|
||||
};
|
||||
const titleGenClient = new ChatGPTClient(this.apiKey, titleGenClientOptions);
|
||||
const result = await titleGenClient.getCompletion([instructionsPayload], null);
|
||||
// remove any non-alphanumeric characters, replace multiple spaces with 1, and then trim
|
||||
return result.choices[0].message.content
|
||||
.replace(/[^a-zA-Z0-9' ]/g, '')
|
||||
.replace(/\s+/g, ' ')
|
||||
.trim();
|
||||
}
|
||||
|
||||
async sendMessage(message, opts = {}) {
|
||||
if (opts.clientOptions && typeof opts.clientOptions === 'object') {
|
||||
this.setOptions(opts.clientOptions);
|
||||
}
|
||||
|
||||
const conversationId = opts.conversationId || crypto.randomUUID();
|
||||
const parentMessageId = opts.parentMessageId || crypto.randomUUID();
|
||||
|
||||
let conversation =
|
||||
typeof opts.conversation === 'object'
|
||||
? opts.conversation
|
||||
: await this.conversationsCache.get(conversationId);
|
||||
|
||||
let isNewConversation = false;
|
||||
if (!conversation) {
|
||||
conversation = {
|
||||
messages: [],
|
||||
createdAt: Date.now(),
|
||||
};
|
||||
isNewConversation = true;
|
||||
}
|
||||
|
||||
const shouldGenerateTitle = opts.shouldGenerateTitle && isNewConversation;
|
||||
|
||||
const userMessage = {
|
||||
id: crypto.randomUUID(),
|
||||
parentMessageId,
|
||||
role: 'User',
|
||||
message,
|
||||
};
|
||||
conversation.messages.push(userMessage);
|
||||
|
||||
// Doing it this way instead of having each message be a separate element in the array seems to be more reliable,
|
||||
// especially when it comes to keeping the AI in character. It also seems to improve coherency and context retention.
|
||||
const { prompt: payload, context } = await this.buildPrompt(
|
||||
conversation.messages,
|
||||
userMessage.id,
|
||||
{
|
||||
isChatGptModel: this.isChatGptModel,
|
||||
promptPrefix: opts.promptPrefix,
|
||||
},
|
||||
);
|
||||
|
||||
if (this.options.keepNecessaryMessagesOnly) {
|
||||
conversation.messages = context;
|
||||
}
|
||||
|
||||
let reply = '';
|
||||
let result = null;
|
||||
if (typeof opts.onProgress === 'function') {
|
||||
await this.getCompletion(
|
||||
payload,
|
||||
(progressMessage) => {
|
||||
if (progressMessage === '[DONE]') {
|
||||
return;
|
||||
}
|
||||
const token = this.isChatGptModel
|
||||
? progressMessage.choices[0].delta.content
|
||||
: progressMessage.choices[0].text;
|
||||
// first event's delta content is always undefined
|
||||
if (!token) {
|
||||
return;
|
||||
}
|
||||
if (this.options.debug) {
|
||||
console.debug(token);
|
||||
}
|
||||
if (token === this.endToken) {
|
||||
return;
|
||||
}
|
||||
opts.onProgress(token);
|
||||
reply += token;
|
||||
},
|
||||
opts.abortController || new AbortController(),
|
||||
);
|
||||
} else {
|
||||
result = await this.getCompletion(
|
||||
payload,
|
||||
null,
|
||||
opts.abortController || new AbortController(),
|
||||
);
|
||||
if (this.options.debug) {
|
||||
console.debug(JSON.stringify(result));
|
||||
}
|
||||
if (this.isChatGptModel) {
|
||||
reply = result.choices[0].message.content;
|
||||
} else {
|
||||
reply = result.choices[0].text.replace(this.endToken, '');
|
||||
}
|
||||
}
|
||||
|
||||
// avoids some rendering issues when using the CLI app
|
||||
if (this.options.debug) {
|
||||
console.debug();
|
||||
}
|
||||
|
||||
reply = reply.trim();
|
||||
|
||||
const replyMessage = {
|
||||
id: crypto.randomUUID(),
|
||||
parentMessageId: userMessage.id,
|
||||
role: 'ChatGPT',
|
||||
message: reply,
|
||||
};
|
||||
conversation.messages.push(replyMessage);
|
||||
|
||||
const returnData = {
|
||||
response: replyMessage.message,
|
||||
conversationId,
|
||||
parentMessageId: replyMessage.parentMessageId,
|
||||
messageId: replyMessage.id,
|
||||
details: result || {},
|
||||
};
|
||||
|
||||
if (shouldGenerateTitle) {
|
||||
conversation.title = await this.generateTitle(userMessage, replyMessage);
|
||||
returnData.title = conversation.title;
|
||||
}
|
||||
|
||||
await this.conversationsCache.set(conversationId, conversation);
|
||||
|
||||
if (this.options.returnConversation) {
|
||||
returnData.conversation = conversation;
|
||||
}
|
||||
|
||||
return returnData;
|
||||
}
|
||||
|
||||
async buildPrompt(messages, { isChatGptModel = false, promptPrefix = null }) {
|
||||
promptPrefix = (promptPrefix || this.options.promptPrefix || '').trim();
|
||||
|
||||
// Handle attachments and create augmentedPrompt
|
||||
if (this.options.attachments) {
|
||||
const attachments = await this.options.attachments;
|
||||
const lastMessage = messages[messages.length - 1];
|
||||
|
||||
if (this.message_file_map) {
|
||||
this.message_file_map[lastMessage.messageId] = attachments;
|
||||
} else {
|
||||
this.message_file_map = {
|
||||
[lastMessage.messageId]: attachments,
|
||||
};
|
||||
}
|
||||
|
||||
const files = await this.addImageURLs(lastMessage, attachments);
|
||||
this.options.attachments = files;
|
||||
|
||||
this.contextHandlers = createContextHandlers(this.options.req, lastMessage.text);
|
||||
}
|
||||
|
||||
if (this.message_file_map) {
|
||||
this.contextHandlers = createContextHandlers(
|
||||
this.options.req,
|
||||
messages[messages.length - 1].text,
|
||||
);
|
||||
}
|
||||
|
||||
// Calculate image token cost and process embedded files
|
||||
messages.forEach((message, i) => {
|
||||
if (this.message_file_map && this.message_file_map[message.messageId]) {
|
||||
const attachments = this.message_file_map[message.messageId];
|
||||
for (const file of attachments) {
|
||||
if (file.embedded) {
|
||||
this.contextHandlers?.processFile(file);
|
||||
continue;
|
||||
}
|
||||
|
||||
messages[i].tokenCount =
|
||||
(messages[i].tokenCount || 0) +
|
||||
this.calculateImageTokenCost({
|
||||
width: file.width,
|
||||
height: file.height,
|
||||
detail: this.options.imageDetail ?? ImageDetail.auto,
|
||||
});
|
||||
}
|
||||
}
|
||||
});
|
||||
|
||||
if (this.contextHandlers) {
|
||||
this.augmentedPrompt = await this.contextHandlers.createContext();
|
||||
promptPrefix = this.augmentedPrompt + promptPrefix;
|
||||
}
|
||||
|
||||
if (promptPrefix) {
|
||||
// If the prompt prefix doesn't end with the end token, add it.
|
||||
if (!promptPrefix.endsWith(`${this.endToken}`)) {
|
||||
promptPrefix = `${promptPrefix.trim()}${this.endToken}\n\n`;
|
||||
}
|
||||
promptPrefix = `${this.startToken}Instructions:\n${promptPrefix}`;
|
||||
}
|
||||
const promptSuffix = `${this.startToken}${this.chatGptLabel}:\n`; // Prompt ChatGPT to respond.
|
||||
|
||||
const instructionsPayload = {
|
||||
role: 'system',
|
||||
content: promptPrefix,
|
||||
};
|
||||
|
||||
const messagePayload = {
|
||||
role: 'system',
|
||||
content: promptSuffix,
|
||||
};
|
||||
|
||||
let currentTokenCount;
|
||||
if (isChatGptModel) {
|
||||
currentTokenCount =
|
||||
this.getTokenCountForMessage(instructionsPayload) +
|
||||
this.getTokenCountForMessage(messagePayload);
|
||||
} else {
|
||||
currentTokenCount = this.getTokenCount(`${promptPrefix}${promptSuffix}`);
|
||||
}
|
||||
let promptBody = '';
|
||||
const maxTokenCount = this.maxPromptTokens;
|
||||
|
||||
const context = [];
|
||||
|
||||
// Iterate backwards through the messages, adding them to the prompt until we reach the max token count.
|
||||
// Do this within a recursive async function so that it doesn't block the event loop for too long.
|
||||
const buildPromptBody = async () => {
|
||||
if (currentTokenCount < maxTokenCount && messages.length > 0) {
|
||||
const message = messages.pop();
|
||||
const roleLabel =
|
||||
message?.isCreatedByUser || message?.role?.toLowerCase() === 'user'
|
||||
? this.userLabel
|
||||
: this.chatGptLabel;
|
||||
const messageString = `${this.startToken}${roleLabel}:\n${
|
||||
message?.text ?? message?.message
|
||||
}${this.endToken}\n`;
|
||||
let newPromptBody;
|
||||
if (promptBody || isChatGptModel) {
|
||||
newPromptBody = `${messageString}${promptBody}`;
|
||||
} else {
|
||||
// Always insert prompt prefix before the last user message, if not gpt-3.5-turbo.
|
||||
// This makes the AI obey the prompt instructions better, which is important for custom instructions.
|
||||
// After a bunch of testing, it doesn't seem to cause the AI any confusion, even if you ask it things
|
||||
// like "what's the last thing I wrote?".
|
||||
newPromptBody = `${promptPrefix}${messageString}${promptBody}`;
|
||||
}
|
||||
|
||||
context.unshift(message);
|
||||
|
||||
const tokenCountForMessage = this.getTokenCount(messageString);
|
||||
const newTokenCount = currentTokenCount + tokenCountForMessage;
|
||||
if (newTokenCount > maxTokenCount) {
|
||||
if (promptBody) {
|
||||
// This message would put us over the token limit, so don't add it.
|
||||
return false;
|
||||
}
|
||||
// This is the first message, so we can't add it. Just throw an error.
|
||||
throw new Error(
|
||||
`Prompt is too long. Max token count is ${maxTokenCount}, but prompt is ${newTokenCount} tokens long.`,
|
||||
);
|
||||
}
|
||||
promptBody = newPromptBody;
|
||||
currentTokenCount = newTokenCount;
|
||||
// wait for next tick to avoid blocking the event loop
|
||||
await new Promise((resolve) => setImmediate(resolve));
|
||||
return buildPromptBody();
|
||||
}
|
||||
return true;
|
||||
};
|
||||
|
||||
await buildPromptBody();
|
||||
|
||||
const prompt = `${promptBody}${promptSuffix}`;
|
||||
if (isChatGptModel) {
|
||||
messagePayload.content = prompt;
|
||||
// Add 3 tokens for Assistant Label priming after all messages have been counted.
|
||||
currentTokenCount += 3;
|
||||
}
|
||||
|
||||
// Use up to `this.maxContextTokens` tokens (prompt + response), but try to leave `this.maxTokens` tokens for the response.
|
||||
this.modelOptions.max_tokens = Math.min(
|
||||
this.maxContextTokens - currentTokenCount,
|
||||
this.maxResponseTokens,
|
||||
);
|
||||
|
||||
if (isChatGptModel) {
|
||||
return { prompt: [instructionsPayload, messagePayload], context };
|
||||
}
|
||||
return { prompt, context, promptTokens: currentTokenCount };
|
||||
}
|
||||
|
||||
getTokenCount(text) {
|
||||
return this.gptEncoder.encode(text, 'all').length;
|
||||
}
|
||||
|
||||
/**
|
||||
* Algorithm adapted from "6. Counting tokens for chat API calls" of
|
||||
* https://github.com/openai/openai-cookbook/blob/main/examples/How_to_count_tokens_with_tiktoken.ipynb
|
||||
*
|
||||
* An additional 3 tokens need to be added for assistant label priming after all messages have been counted.
|
||||
*
|
||||
* @param {Object} message
|
||||
*/
|
||||
getTokenCountForMessage(message) {
|
||||
// Note: gpt-3.5-turbo and gpt-4 may update over time. Use default for these as well as for unknown models
|
||||
let tokensPerMessage = 3;
|
||||
let tokensPerName = 1;
|
||||
|
||||
if (this.modelOptions.model === 'gpt-3.5-turbo-0301') {
|
||||
tokensPerMessage = 4;
|
||||
tokensPerName = -1;
|
||||
}
|
||||
|
||||
let numTokens = tokensPerMessage;
|
||||
for (let [key, value] of Object.entries(message)) {
|
||||
numTokens += this.getTokenCount(value);
|
||||
if (key === 'name') {
|
||||
numTokens += tokensPerName;
|
||||
}
|
||||
}
|
||||
|
||||
return numTokens;
|
||||
}
|
||||
}
|
||||
|
||||
module.exports = ChatGPTClient;
|
||||
|
|
@ -5,6 +5,7 @@ const {
|
|||
isEnabled,
|
||||
Tokenizer,
|
||||
createFetch,
|
||||
resolveHeaders,
|
||||
constructAzureURL,
|
||||
genAzureChatCompletion,
|
||||
createStreamEventHandlers,
|
||||
|
|
@ -15,7 +16,6 @@ const {
|
|||
ContentTypes,
|
||||
parseTextParts,
|
||||
EModelEndpoint,
|
||||
resolveHeaders,
|
||||
KnownEndpoints,
|
||||
openAISettings,
|
||||
ImageDetailCost,
|
||||
|
|
@ -37,7 +37,6 @@ const { addSpaceIfNeeded, sleep } = require('~/server/utils');
|
|||
const { spendTokens } = require('~/models/spendTokens');
|
||||
const { handleOpenAIErrors } = require('./tools/util');
|
||||
const { createLLM, RunManager } = require('./llm');
|
||||
const ChatGPTClient = require('./ChatGPTClient');
|
||||
const { summaryBuffer } = require('./memory');
|
||||
const { runTitleChain } = require('./chains');
|
||||
const { tokenSplit } = require('./document');
|
||||
|
|
@ -47,12 +46,6 @@ const { logger } = require('~/config');
|
|||
class OpenAIClient extends BaseClient {
|
||||
constructor(apiKey, options = {}) {
|
||||
super(apiKey, options);
|
||||
this.ChatGPTClient = new ChatGPTClient();
|
||||
this.buildPrompt = this.ChatGPTClient.buildPrompt.bind(this);
|
||||
/** @type {getCompletion} */
|
||||
this.getCompletion = this.ChatGPTClient.getCompletion.bind(this);
|
||||
/** @type {cohereChatCompletion} */
|
||||
this.cohereChatCompletion = this.ChatGPTClient.cohereChatCompletion.bind(this);
|
||||
this.contextStrategy = options.contextStrategy
|
||||
? options.contextStrategy.toLowerCase()
|
||||
: 'discard';
|
||||
|
|
@ -379,23 +372,12 @@ class OpenAIClient extends BaseClient {
|
|||
return files;
|
||||
}
|
||||
|
||||
async buildMessages(
|
||||
messages,
|
||||
parentMessageId,
|
||||
{ isChatCompletion = false, promptPrefix = null },
|
||||
opts,
|
||||
) {
|
||||
async buildMessages(messages, parentMessageId, { promptPrefix = null }, opts) {
|
||||
let orderedMessages = this.constructor.getMessagesForConversation({
|
||||
messages,
|
||||
parentMessageId,
|
||||
summary: this.shouldSummarize,
|
||||
});
|
||||
if (!isChatCompletion) {
|
||||
return await this.buildPrompt(orderedMessages, {
|
||||
isChatGptModel: isChatCompletion,
|
||||
promptPrefix,
|
||||
});
|
||||
}
|
||||
|
||||
let payload;
|
||||
let instructions;
|
||||
|
|
|
|||
|
|
@ -1,542 +0,0 @@
|
|||
const OpenAIClient = require('./OpenAIClient');
|
||||
const { CallbackManager } = require('@langchain/core/callbacks/manager');
|
||||
const { BufferMemory, ChatMessageHistory } = require('langchain/memory');
|
||||
const { addImages, buildErrorInput, buildPromptPrefix } = require('./output_parsers');
|
||||
const { initializeCustomAgent, initializeFunctionsAgent } = require('./agents');
|
||||
const { processFileURL } = require('~/server/services/Files/process');
|
||||
const { EModelEndpoint } = require('librechat-data-provider');
|
||||
const { checkBalance } = require('~/models/balanceMethods');
|
||||
const { formatLangChainMessages } = require('./prompts');
|
||||
const { extractBaseURL } = require('~/utils');
|
||||
const { loadTools } = require('./tools/util');
|
||||
const { logger } = require('~/config');
|
||||
|
||||
class PluginsClient extends OpenAIClient {
|
||||
constructor(apiKey, options = {}) {
|
||||
super(apiKey, options);
|
||||
this.sender = options.sender ?? 'Assistant';
|
||||
this.tools = [];
|
||||
this.actions = [];
|
||||
this.setOptions(options);
|
||||
this.openAIApiKey = this.apiKey;
|
||||
this.executor = null;
|
||||
}
|
||||
|
||||
setOptions(options) {
|
||||
this.agentOptions = { ...options.agentOptions };
|
||||
this.functionsAgent = this.agentOptions?.agent === 'functions';
|
||||
this.agentIsGpt3 = this.agentOptions?.model?.includes('gpt-3');
|
||||
|
||||
super.setOptions(options);
|
||||
|
||||
this.isGpt3 = this.modelOptions?.model?.includes('gpt-3');
|
||||
|
||||
if (this.options.reverseProxyUrl) {
|
||||
this.langchainProxy = extractBaseURL(this.options.reverseProxyUrl);
|
||||
}
|
||||
}
|
||||
|
||||
getSaveOptions() {
|
||||
return {
|
||||
artifacts: this.options.artifacts,
|
||||
chatGptLabel: this.options.chatGptLabel,
|
||||
modelLabel: this.options.modelLabel,
|
||||
promptPrefix: this.options.promptPrefix,
|
||||
tools: this.options.tools,
|
||||
...this.modelOptions,
|
||||
agentOptions: this.agentOptions,
|
||||
iconURL: this.options.iconURL,
|
||||
greeting: this.options.greeting,
|
||||
spec: this.options.spec,
|
||||
};
|
||||
}
|
||||
|
||||
saveLatestAction(action) {
|
||||
this.actions.push(action);
|
||||
}
|
||||
|
||||
getFunctionModelName(input) {
|
||||
if (/-(?!0314)\d{4}/.test(input)) {
|
||||
return input;
|
||||
} else if (input.includes('gpt-3.5-turbo')) {
|
||||
return 'gpt-3.5-turbo';
|
||||
} else if (input.includes('gpt-4')) {
|
||||
return 'gpt-4';
|
||||
} else {
|
||||
return 'gpt-3.5-turbo';
|
||||
}
|
||||
}
|
||||
|
||||
getBuildMessagesOptions(opts) {
|
||||
return {
|
||||
isChatCompletion: true,
|
||||
promptPrefix: opts.promptPrefix,
|
||||
abortController: opts.abortController,
|
||||
};
|
||||
}
|
||||
|
||||
async initialize({ user, message, onAgentAction, onChainEnd, signal }) {
|
||||
const modelOptions = {
|
||||
modelName: this.agentOptions.model,
|
||||
temperature: this.agentOptions.temperature,
|
||||
};
|
||||
|
||||
const model = this.initializeLLM({
|
||||
...modelOptions,
|
||||
context: 'plugins',
|
||||
initialMessageCount: this.currentMessages.length + 1,
|
||||
});
|
||||
|
||||
logger.debug(
|
||||
`[PluginsClient] Agent Model: ${model.modelName} | Temp: ${model.temperature} | Functions: ${this.functionsAgent}`,
|
||||
);
|
||||
|
||||
// Map Messages to Langchain format
|
||||
const pastMessages = formatLangChainMessages(this.currentMessages.slice(0, -1), {
|
||||
userName: this.options?.name,
|
||||
});
|
||||
logger.debug('[PluginsClient] pastMessages: ' + pastMessages.length);
|
||||
|
||||
// TODO: use readOnly memory, TokenBufferMemory? (both unavailable in LangChainJS)
|
||||
const memory = new BufferMemory({
|
||||
llm: model,
|
||||
chatHistory: new ChatMessageHistory(pastMessages),
|
||||
});
|
||||
|
||||
const { loadedTools } = await loadTools({
|
||||
user,
|
||||
model,
|
||||
tools: this.options.tools,
|
||||
functions: this.functionsAgent,
|
||||
options: {
|
||||
memory,
|
||||
signal: this.abortController.signal,
|
||||
openAIApiKey: this.openAIApiKey,
|
||||
conversationId: this.conversationId,
|
||||
fileStrategy: this.options.req.app.locals.fileStrategy,
|
||||
processFileURL,
|
||||
message,
|
||||
},
|
||||
useSpecs: true,
|
||||
});
|
||||
|
||||
if (loadedTools.length === 0) {
|
||||
return;
|
||||
}
|
||||
|
||||
this.tools = loadedTools;
|
||||
|
||||
logger.debug('[PluginsClient] Requested Tools', this.options.tools);
|
||||
logger.debug(
|
||||
'[PluginsClient] Loaded Tools',
|
||||
this.tools.map((tool) => tool.name),
|
||||
);
|
||||
|
||||
const handleAction = (action, runId, callback = null) => {
|
||||
this.saveLatestAction(action);
|
||||
|
||||
logger.debug('[PluginsClient] Latest Agent Action ', this.actions[this.actions.length - 1]);
|
||||
|
||||
if (typeof callback === 'function') {
|
||||
callback(action, runId);
|
||||
}
|
||||
};
|
||||
|
||||
// initialize agent
|
||||
const initializer = this.functionsAgent ? initializeFunctionsAgent : initializeCustomAgent;
|
||||
|
||||
let customInstructions = (this.options.promptPrefix ?? '').trim();
|
||||
if (typeof this.options.artifactsPrompt === 'string' && this.options.artifactsPrompt) {
|
||||
customInstructions = `${customInstructions ?? ''}\n${this.options.artifactsPrompt}`.trim();
|
||||
}
|
||||
|
||||
this.executor = await initializer({
|
||||
model,
|
||||
signal,
|
||||
pastMessages,
|
||||
tools: this.tools,
|
||||
customInstructions,
|
||||
verbose: this.options.debug,
|
||||
returnIntermediateSteps: true,
|
||||
customName: this.options.chatGptLabel,
|
||||
currentDateString: this.currentDateString,
|
||||
callbackManager: CallbackManager.fromHandlers({
|
||||
async handleAgentAction(action, runId) {
|
||||
handleAction(action, runId, onAgentAction);
|
||||
},
|
||||
async handleChainEnd(action) {
|
||||
if (typeof onChainEnd === 'function') {
|
||||
onChainEnd(action);
|
||||
}
|
||||
},
|
||||
}),
|
||||
});
|
||||
|
||||
logger.debug('[PluginsClient] Loaded agent.');
|
||||
}
|
||||
|
||||
async executorCall(message, { signal, stream, onToolStart, onToolEnd }) {
|
||||
let errorMessage = '';
|
||||
const maxAttempts = 1;
|
||||
|
||||
for (let attempts = 1; attempts <= maxAttempts; attempts++) {
|
||||
const errorInput = buildErrorInput({
|
||||
message,
|
||||
errorMessage,
|
||||
actions: this.actions,
|
||||
functionsAgent: this.functionsAgent,
|
||||
});
|
||||
const input = attempts > 1 ? errorInput : message;
|
||||
|
||||
logger.debug(`[PluginsClient] Attempt ${attempts} of ${maxAttempts}`);
|
||||
|
||||
if (errorMessage.length > 0) {
|
||||
logger.debug('[PluginsClient] Caught error, input: ' + JSON.stringify(input));
|
||||
}
|
||||
|
||||
try {
|
||||
this.result = await this.executor.call({ input, signal }, [
|
||||
{
|
||||
async handleToolStart(...args) {
|
||||
await onToolStart(...args);
|
||||
},
|
||||
async handleToolEnd(...args) {
|
||||
await onToolEnd(...args);
|
||||
},
|
||||
async handleLLMEnd(output) {
|
||||
const { generations } = output;
|
||||
const { text } = generations[0][0];
|
||||
if (text && typeof stream === 'function') {
|
||||
await stream(text);
|
||||
}
|
||||
},
|
||||
},
|
||||
]);
|
||||
break; // Exit the loop if the function call is successful
|
||||
} catch (err) {
|
||||
logger.error('[PluginsClient] executorCall error:', err);
|
||||
if (attempts === maxAttempts) {
|
||||
const { run } = this.runManager.getRunByConversationId(this.conversationId);
|
||||
const defaultOutput = `Encountered an error while attempting to respond: ${err.message}`;
|
||||
this.result.output = run && run.error ? run.error : defaultOutput;
|
||||
this.result.errorMessage = run && run.error ? run.error : err.message;
|
||||
this.result.intermediateSteps = this.actions;
|
||||
break;
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
/**
|
||||
*
|
||||
* @param {TMessage} responseMessage
|
||||
* @param {Partial<TMessage>} saveOptions
|
||||
* @param {string} user
|
||||
* @returns
|
||||
*/
|
||||
async handleResponseMessage(responseMessage, saveOptions, user) {
|
||||
const { output, errorMessage, ...result } = this.result;
|
||||
logger.debug('[PluginsClient][handleResponseMessage] Output:', {
|
||||
output,
|
||||
errorMessage,
|
||||
...result,
|
||||
});
|
||||
const { error } = responseMessage;
|
||||
if (!error) {
|
||||
responseMessage.tokenCount = this.getTokenCountForResponse(responseMessage);
|
||||
responseMessage.completionTokens = this.getTokenCount(responseMessage.text);
|
||||
}
|
||||
|
||||
// Record usage only when completion is skipped as it is already recorded in the agent phase.
|
||||
if (!this.agentOptions.skipCompletion && !error) {
|
||||
await this.recordTokenUsage(responseMessage);
|
||||
}
|
||||
|
||||
const databasePromise = this.saveMessageToDatabase(responseMessage, saveOptions, user);
|
||||
delete responseMessage.tokenCount;
|
||||
return { ...responseMessage, ...result, databasePromise };
|
||||
}
|
||||
|
||||
async sendMessage(message, opts = {}) {
|
||||
/** @type {Promise<TMessage>} */
|
||||
let userMessagePromise;
|
||||
/** @type {{ filteredTools: string[], includedTools: string[] }} */
|
||||
const { filteredTools = [], includedTools = [] } = this.options.req.app.locals;
|
||||
|
||||
if (includedTools.length > 0) {
|
||||
const tools = this.options.tools.filter((plugin) => includedTools.includes(plugin));
|
||||
this.options.tools = tools;
|
||||
} else {
|
||||
const tools = this.options.tools.filter((plugin) => !filteredTools.includes(plugin));
|
||||
this.options.tools = tools;
|
||||
}
|
||||
|
||||
// If a message is edited, no tools can be used.
|
||||
const completionMode = this.options.tools.length === 0 || opts.isEdited;
|
||||
if (completionMode) {
|
||||
this.setOptions(opts);
|
||||
return super.sendMessage(message, opts);
|
||||
}
|
||||
|
||||
logger.debug('[PluginsClient] sendMessage', { userMessageText: message, opts });
|
||||
const {
|
||||
user,
|
||||
conversationId,
|
||||
responseMessageId,
|
||||
saveOptions,
|
||||
userMessage,
|
||||
onAgentAction,
|
||||
onChainEnd,
|
||||
onToolStart,
|
||||
onToolEnd,
|
||||
} = await this.handleStartMethods(message, opts);
|
||||
|
||||
if (opts.progressCallback) {
|
||||
opts.onProgress = opts.progressCallback.call(null, {
|
||||
...(opts.progressOptions ?? {}),
|
||||
parentMessageId: userMessage.messageId,
|
||||
messageId: responseMessageId,
|
||||
});
|
||||
}
|
||||
|
||||
this.currentMessages.push(userMessage);
|
||||
|
||||
let {
|
||||
prompt: payload,
|
||||
tokenCountMap,
|
||||
promptTokens,
|
||||
} = await this.buildMessages(
|
||||
this.currentMessages,
|
||||
userMessage.messageId,
|
||||
this.getBuildMessagesOptions({
|
||||
promptPrefix: null,
|
||||
abortController: this.abortController,
|
||||
}),
|
||||
);
|
||||
|
||||
if (tokenCountMap) {
|
||||
logger.debug('[PluginsClient] tokenCountMap', { tokenCountMap });
|
||||
if (tokenCountMap[userMessage.messageId]) {
|
||||
userMessage.tokenCount = tokenCountMap[userMessage.messageId];
|
||||
logger.debug('[PluginsClient] userMessage.tokenCount', userMessage.tokenCount);
|
||||
}
|
||||
this.handleTokenCountMap(tokenCountMap);
|
||||
}
|
||||
|
||||
this.result = {};
|
||||
if (payload) {
|
||||
this.currentMessages = payload;
|
||||
}
|
||||
|
||||
if (!this.skipSaveUserMessage) {
|
||||
userMessagePromise = this.saveMessageToDatabase(userMessage, saveOptions, user);
|
||||
if (typeof opts?.getReqData === 'function') {
|
||||
opts.getReqData({
|
||||
userMessagePromise,
|
||||
});
|
||||
}
|
||||
}
|
||||
|
||||
const balance = this.options.req?.app?.locals?.balance;
|
||||
if (balance?.enabled) {
|
||||
await checkBalance({
|
||||
req: this.options.req,
|
||||
res: this.options.res,
|
||||
txData: {
|
||||
user: this.user,
|
||||
tokenType: 'prompt',
|
||||
amount: promptTokens,
|
||||
debug: this.options.debug,
|
||||
model: this.modelOptions.model,
|
||||
endpoint: EModelEndpoint.openAI,
|
||||
},
|
||||
});
|
||||
}
|
||||
|
||||
const responseMessage = {
|
||||
endpoint: EModelEndpoint.gptPlugins,
|
||||
iconURL: this.options.iconURL,
|
||||
messageId: responseMessageId,
|
||||
conversationId,
|
||||
parentMessageId: userMessage.messageId,
|
||||
isCreatedByUser: false,
|
||||
model: this.modelOptions.model,
|
||||
sender: this.sender,
|
||||
promptTokens,
|
||||
};
|
||||
|
||||
await this.initialize({
|
||||
user,
|
||||
message,
|
||||
onAgentAction,
|
||||
onChainEnd,
|
||||
signal: this.abortController.signal,
|
||||
onProgress: opts.onProgress,
|
||||
});
|
||||
|
||||
// const stream = async (text) => {
|
||||
// await this.generateTextStream.call(this, text, opts.onProgress, { delay: 1 });
|
||||
// };
|
||||
await this.executorCall(message, {
|
||||
signal: this.abortController.signal,
|
||||
// stream,
|
||||
onToolStart,
|
||||
onToolEnd,
|
||||
});
|
||||
|
||||
// If message was aborted mid-generation
|
||||
if (this.result?.errorMessage?.length > 0 && this.result?.errorMessage?.includes('cancel')) {
|
||||
responseMessage.text = 'Cancelled.';
|
||||
return await this.handleResponseMessage(responseMessage, saveOptions, user);
|
||||
}
|
||||
|
||||
// If error occurred during generation (likely token_balance)
|
||||
if (this.result?.errorMessage?.length > 0) {
|
||||
responseMessage.error = true;
|
||||
responseMessage.text = this.result.output;
|
||||
return await this.handleResponseMessage(responseMessage, saveOptions, user);
|
||||
}
|
||||
|
||||
if (this.agentOptions.skipCompletion && this.result.output && this.functionsAgent) {
|
||||
const partialText = opts.getPartialText();
|
||||
const trimmedPartial = opts.getPartialText().replaceAll(':::plugin:::\n', '');
|
||||
responseMessage.text =
|
||||
trimmedPartial.length === 0 ? `${partialText}${this.result.output}` : partialText;
|
||||
addImages(this.result.intermediateSteps, responseMessage);
|
||||
await this.generateTextStream(this.result.output, opts.onProgress, { delay: 5 });
|
||||
return await this.handleResponseMessage(responseMessage, saveOptions, user);
|
||||
}
|
||||
|
||||
if (this.agentOptions.skipCompletion && this.result.output) {
|
||||
responseMessage.text = this.result.output;
|
||||
addImages(this.result.intermediateSteps, responseMessage);
|
||||
await this.generateTextStream(this.result.output, opts.onProgress, { delay: 5 });
|
||||
return await this.handleResponseMessage(responseMessage, saveOptions, user);
|
||||
}
|
||||
|
||||
logger.debug('[PluginsClient] Completion phase: this.result', this.result);
|
||||
|
||||
const promptPrefix = buildPromptPrefix({
|
||||
result: this.result,
|
||||
message,
|
||||
functionsAgent: this.functionsAgent,
|
||||
});
|
||||
|
||||
logger.debug('[PluginsClient]', { promptPrefix });
|
||||
|
||||
payload = await this.buildCompletionPrompt({
|
||||
messages: this.currentMessages,
|
||||
promptPrefix,
|
||||
});
|
||||
|
||||
logger.debug('[PluginsClient] buildCompletionPrompt Payload', payload);
|
||||
responseMessage.text = await this.sendCompletion(payload, opts);
|
||||
return await this.handleResponseMessage(responseMessage, saveOptions, user);
|
||||
}
|
||||
|
||||
async buildCompletionPrompt({ messages, promptPrefix: _promptPrefix }) {
|
||||
logger.debug('[PluginsClient] buildCompletionPrompt messages', messages);
|
||||
|
||||
const orderedMessages = messages;
|
||||
let promptPrefix = _promptPrefix.trim();
|
||||
// If the prompt prefix doesn't end with the end token, add it.
|
||||
if (!promptPrefix.endsWith(`${this.endToken}`)) {
|
||||
promptPrefix = `${promptPrefix.trim()}${this.endToken}\n\n`;
|
||||
}
|
||||
promptPrefix = `${this.startToken}Instructions:\n${promptPrefix}`;
|
||||
const promptSuffix = `${this.startToken}${this.chatGptLabel ?? 'Assistant'}:\n`;
|
||||
|
||||
const instructionsPayload = {
|
||||
role: 'system',
|
||||
content: promptPrefix,
|
||||
};
|
||||
|
||||
const messagePayload = {
|
||||
role: 'system',
|
||||
content: promptSuffix,
|
||||
};
|
||||
|
||||
if (this.isGpt3) {
|
||||
instructionsPayload.role = 'user';
|
||||
messagePayload.role = 'user';
|
||||
instructionsPayload.content += `\n${promptSuffix}`;
|
||||
}
|
||||
|
||||
// testing if this works with browser endpoint
|
||||
if (!this.isGpt3 && this.options.reverseProxyUrl) {
|
||||
instructionsPayload.role = 'user';
|
||||
}
|
||||
|
||||
let currentTokenCount =
|
||||
this.getTokenCountForMessage(instructionsPayload) +
|
||||
this.getTokenCountForMessage(messagePayload);
|
||||
|
||||
let promptBody = '';
|
||||
const maxTokenCount = this.maxPromptTokens;
|
||||
// Iterate backwards through the messages, adding them to the prompt until we reach the max token count.
|
||||
// Do this within a recursive async function so that it doesn't block the event loop for too long.
|
||||
const buildPromptBody = async () => {
|
||||
if (currentTokenCount < maxTokenCount && orderedMessages.length > 0) {
|
||||
const message = orderedMessages.pop();
|
||||
const isCreatedByUser = message.isCreatedByUser || message.role?.toLowerCase() === 'user';
|
||||
const roleLabel = isCreatedByUser ? this.userLabel : this.chatGptLabel;
|
||||
let messageString = `${this.startToken}${roleLabel}:\n${
|
||||
message.text ?? message.content ?? ''
|
||||
}${this.endToken}\n`;
|
||||
let newPromptBody = `${messageString}${promptBody}`;
|
||||
|
||||
const tokenCountForMessage = this.getTokenCount(messageString);
|
||||
const newTokenCount = currentTokenCount + tokenCountForMessage;
|
||||
if (newTokenCount > maxTokenCount) {
|
||||
if (promptBody) {
|
||||
// This message would put us over the token limit, so don't add it.
|
||||
return false;
|
||||
}
|
||||
// This is the first message, so we can't add it. Just throw an error.
|
||||
throw new Error(
|
||||
`Prompt is too long. Max token count is ${maxTokenCount}, but prompt is ${newTokenCount} tokens long.`,
|
||||
);
|
||||
}
|
||||
promptBody = newPromptBody;
|
||||
currentTokenCount = newTokenCount;
|
||||
// wait for next tick to avoid blocking the event loop
|
||||
await new Promise((resolve) => setTimeout(resolve, 0));
|
||||
return buildPromptBody();
|
||||
}
|
||||
return true;
|
||||
};
|
||||
|
||||
await buildPromptBody();
|
||||
const prompt = promptBody;
|
||||
messagePayload.content = prompt;
|
||||
// Add 2 tokens for metadata after all messages have been counted.
|
||||
currentTokenCount += 2;
|
||||
|
||||
if (this.isGpt3 && messagePayload.content.length > 0) {
|
||||
const context = 'Chat History:\n';
|
||||
messagePayload.content = `${context}${prompt}`;
|
||||
currentTokenCount += this.getTokenCount(context);
|
||||
}
|
||||
|
||||
// Use up to `this.maxContextTokens` tokens (prompt + response), but try to leave `this.maxTokens` tokens for the response.
|
||||
this.modelOptions.max_tokens = Math.min(
|
||||
this.maxContextTokens - currentTokenCount,
|
||||
this.maxResponseTokens,
|
||||
);
|
||||
|
||||
if (this.isGpt3) {
|
||||
messagePayload.content += promptSuffix;
|
||||
return [instructionsPayload, messagePayload];
|
||||
}
|
||||
|
||||
const result = [messagePayload, instructionsPayload];
|
||||
|
||||
if (this.functionsAgent && !this.isGpt3) {
|
||||
result[1].content = `${result[1].content}\n${this.startToken}${this.chatGptLabel}:\nSure thing! Here is the output you requested:\n`;
|
||||
}
|
||||
|
||||
return result.filter((message) => message.content.length > 0);
|
||||
}
|
||||
}
|
||||
|
||||
module.exports = PluginsClient;
|
||||
|
|
@ -1,15 +1,11 @@
|
|||
const ChatGPTClient = require('./ChatGPTClient');
|
||||
const OpenAIClient = require('./OpenAIClient');
|
||||
const PluginsClient = require('./PluginsClient');
|
||||
const GoogleClient = require('./GoogleClient');
|
||||
const TextStream = require('./TextStream');
|
||||
const AnthropicClient = require('./AnthropicClient');
|
||||
const toolUtils = require('./tools/util');
|
||||
|
||||
module.exports = {
|
||||
ChatGPTClient,
|
||||
OpenAIClient,
|
||||
PluginsClient,
|
||||
GoogleClient,
|
||||
TextStream,
|
||||
AnthropicClient,
|
||||
|
|
|
|||
|
|
@ -531,44 +531,6 @@ describe('OpenAIClient', () => {
|
|||
});
|
||||
});
|
||||
|
||||
describe('sendMessage/getCompletion/chatCompletion', () => {
|
||||
afterEach(() => {
|
||||
delete process.env.AZURE_OPENAI_DEFAULT_MODEL;
|
||||
delete process.env.AZURE_USE_MODEL_AS_DEPLOYMENT_NAME;
|
||||
});
|
||||
|
||||
it('should call getCompletion and fetchEventSource when using a text/instruct model', async () => {
|
||||
const model = 'text-davinci-003';
|
||||
const onProgress = jest.fn().mockImplementation(() => ({}));
|
||||
|
||||
const testClient = new OpenAIClient('test-api-key', {
|
||||
...defaultOptions,
|
||||
modelOptions: { model },
|
||||
});
|
||||
|
||||
const getCompletion = jest.spyOn(testClient, 'getCompletion');
|
||||
await testClient.sendMessage('Hi mom!', { onProgress });
|
||||
|
||||
expect(getCompletion).toHaveBeenCalled();
|
||||
expect(getCompletion.mock.calls.length).toBe(1);
|
||||
|
||||
expect(getCompletion.mock.calls[0][0]).toBe('||>User:\nHi mom!\n||>Assistant:\n');
|
||||
|
||||
expect(fetchEventSource).toHaveBeenCalled();
|
||||
expect(fetchEventSource.mock.calls.length).toBe(1);
|
||||
|
||||
// Check if the first argument (url) is correct
|
||||
const firstCallArgs = fetchEventSource.mock.calls[0];
|
||||
|
||||
const expectedURL = 'https://api.openai.com/v1/completions';
|
||||
expect(firstCallArgs[0]).toBe(expectedURL);
|
||||
|
||||
const requestBody = JSON.parse(firstCallArgs[1].body);
|
||||
expect(requestBody).toHaveProperty('model');
|
||||
expect(requestBody.model).toBe(model);
|
||||
});
|
||||
});
|
||||
|
||||
describe('checkVisionRequest functionality', () => {
|
||||
let client;
|
||||
const attachments = [{ type: 'image/png' }];
|
||||
|
|
|
|||
|
|
@ -1,314 +0,0 @@
|
|||
const crypto = require('crypto');
|
||||
const { Constants } = require('librechat-data-provider');
|
||||
const { HumanMessage, AIMessage } = require('@langchain/core/messages');
|
||||
const PluginsClient = require('../PluginsClient');
|
||||
|
||||
jest.mock('~/db/connect');
|
||||
jest.mock('~/models/Conversation', () => {
|
||||
return function () {
|
||||
return {
|
||||
save: jest.fn(),
|
||||
deleteConvos: jest.fn(),
|
||||
};
|
||||
};
|
||||
});
|
||||
|
||||
const defaultAzureOptions = {
|
||||
azureOpenAIApiInstanceName: 'your-instance-name',
|
||||
azureOpenAIApiDeploymentName: 'your-deployment-name',
|
||||
azureOpenAIApiVersion: '2020-07-01-preview',
|
||||
};
|
||||
|
||||
describe('PluginsClient', () => {
|
||||
let TestAgent;
|
||||
let options = {
|
||||
tools: [],
|
||||
modelOptions: {
|
||||
model: 'gpt-3.5-turbo',
|
||||
temperature: 0,
|
||||
max_tokens: 2,
|
||||
},
|
||||
agentOptions: {
|
||||
model: 'gpt-3.5-turbo',
|
||||
},
|
||||
};
|
||||
let parentMessageId;
|
||||
let conversationId;
|
||||
const fakeMessages = [];
|
||||
const userMessage = 'Hello, ChatGPT!';
|
||||
const apiKey = 'fake-api-key';
|
||||
|
||||
beforeEach(() => {
|
||||
TestAgent = new PluginsClient(apiKey, options);
|
||||
TestAgent.loadHistory = jest
|
||||
.fn()
|
||||
.mockImplementation((conversationId, parentMessageId = null) => {
|
||||
if (!conversationId) {
|
||||
TestAgent.currentMessages = [];
|
||||
return Promise.resolve([]);
|
||||
}
|
||||
|
||||
const orderedMessages = TestAgent.constructor.getMessagesForConversation({
|
||||
messages: fakeMessages,
|
||||
parentMessageId,
|
||||
});
|
||||
|
||||
const chatMessages = orderedMessages.map((msg) =>
|
||||
msg?.isCreatedByUser || msg?.role?.toLowerCase() === 'user'
|
||||
? new HumanMessage(msg.text)
|
||||
: new AIMessage(msg.text),
|
||||
);
|
||||
|
||||
TestAgent.currentMessages = orderedMessages;
|
||||
return Promise.resolve(chatMessages);
|
||||
});
|
||||
TestAgent.sendMessage = jest.fn().mockImplementation(async (message, opts = {}) => {
|
||||
if (opts && typeof opts === 'object') {
|
||||
TestAgent.setOptions(opts);
|
||||
}
|
||||
const conversationId = opts.conversationId || crypto.randomUUID();
|
||||
const parentMessageId = opts.parentMessageId || Constants.NO_PARENT;
|
||||
const userMessageId = opts.overrideParentMessageId || crypto.randomUUID();
|
||||
this.pastMessages = await TestAgent.loadHistory(
|
||||
conversationId,
|
||||
TestAgent.options?.parentMessageId,
|
||||
);
|
||||
|
||||
const userMessage = {
|
||||
text: message,
|
||||
sender: 'ChatGPT',
|
||||
isCreatedByUser: true,
|
||||
messageId: userMessageId,
|
||||
parentMessageId,
|
||||
conversationId,
|
||||
};
|
||||
|
||||
const response = {
|
||||
sender: 'ChatGPT',
|
||||
text: 'Hello, User!',
|
||||
isCreatedByUser: false,
|
||||
messageId: crypto.randomUUID(),
|
||||
parentMessageId: userMessage.messageId,
|
||||
conversationId,
|
||||
};
|
||||
|
||||
fakeMessages.push(userMessage);
|
||||
fakeMessages.push(response);
|
||||
return response;
|
||||
});
|
||||
});
|
||||
|
||||
test('initializes PluginsClient without crashing', () => {
|
||||
expect(TestAgent).toBeInstanceOf(PluginsClient);
|
||||
});
|
||||
|
||||
test('check setOptions function', () => {
|
||||
expect(TestAgent.agentIsGpt3).toBe(true);
|
||||
});
|
||||
|
||||
describe('sendMessage', () => {
|
||||
test('sendMessage should return a response message', async () => {
|
||||
const expectedResult = expect.objectContaining({
|
||||
sender: 'ChatGPT',
|
||||
text: expect.any(String),
|
||||
isCreatedByUser: false,
|
||||
messageId: expect.any(String),
|
||||
parentMessageId: expect.any(String),
|
||||
conversationId: expect.any(String),
|
||||
});
|
||||
|
||||
const response = await TestAgent.sendMessage(userMessage);
|
||||
parentMessageId = response.messageId;
|
||||
conversationId = response.conversationId;
|
||||
expect(response).toEqual(expectedResult);
|
||||
});
|
||||
|
||||
test('sendMessage should work with provided conversationId and parentMessageId', async () => {
|
||||
const userMessage = 'Second message in the conversation';
|
||||
const opts = {
|
||||
conversationId,
|
||||
parentMessageId,
|
||||
};
|
||||
|
||||
const expectedResult = expect.objectContaining({
|
||||
sender: 'ChatGPT',
|
||||
text: expect.any(String),
|
||||
isCreatedByUser: false,
|
||||
messageId: expect.any(String),
|
||||
parentMessageId: expect.any(String),
|
||||
conversationId: opts.conversationId,
|
||||
});
|
||||
|
||||
const response = await TestAgent.sendMessage(userMessage, opts);
|
||||
parentMessageId = response.messageId;
|
||||
expect(response.conversationId).toEqual(conversationId);
|
||||
expect(response).toEqual(expectedResult);
|
||||
});
|
||||
|
||||
test('should return chat history', async () => {
|
||||
const chatMessages = await TestAgent.loadHistory(conversationId, parentMessageId);
|
||||
expect(TestAgent.currentMessages).toHaveLength(4);
|
||||
expect(chatMessages[0].text).toEqual(userMessage);
|
||||
});
|
||||
});
|
||||
|
||||
describe('getFunctionModelName', () => {
|
||||
let client;
|
||||
|
||||
beforeEach(() => {
|
||||
client = new PluginsClient('dummy_api_key');
|
||||
});
|
||||
|
||||
test('should return the input when it includes a dash followed by four digits', () => {
|
||||
expect(client.getFunctionModelName('-1234')).toBe('-1234');
|
||||
expect(client.getFunctionModelName('gpt-4-5678-preview')).toBe('gpt-4-5678-preview');
|
||||
});
|
||||
|
||||
test('should return the input for all function-capable models (`0613` models and above)', () => {
|
||||
expect(client.getFunctionModelName('gpt-4-0613')).toBe('gpt-4-0613');
|
||||
expect(client.getFunctionModelName('gpt-4-32k-0613')).toBe('gpt-4-32k-0613');
|
||||
expect(client.getFunctionModelName('gpt-3.5-turbo-0613')).toBe('gpt-3.5-turbo-0613');
|
||||
expect(client.getFunctionModelName('gpt-3.5-turbo-16k-0613')).toBe('gpt-3.5-turbo-16k-0613');
|
||||
expect(client.getFunctionModelName('gpt-3.5-turbo-1106')).toBe('gpt-3.5-turbo-1106');
|
||||
expect(client.getFunctionModelName('gpt-4-1106-preview')).toBe('gpt-4-1106-preview');
|
||||
expect(client.getFunctionModelName('gpt-4-1106')).toBe('gpt-4-1106');
|
||||
});
|
||||
|
||||
test('should return the corresponding model if input is non-function capable (`0314` models)', () => {
|
||||
expect(client.getFunctionModelName('gpt-4-0314')).toBe('gpt-4');
|
||||
expect(client.getFunctionModelName('gpt-4-32k-0314')).toBe('gpt-4');
|
||||
expect(client.getFunctionModelName('gpt-3.5-turbo-0314')).toBe('gpt-3.5-turbo');
|
||||
expect(client.getFunctionModelName('gpt-3.5-turbo-16k-0314')).toBe('gpt-3.5-turbo');
|
||||
});
|
||||
|
||||
test('should return "gpt-3.5-turbo" when the input includes "gpt-3.5-turbo"', () => {
|
||||
expect(client.getFunctionModelName('test gpt-3.5-turbo model')).toBe('gpt-3.5-turbo');
|
||||
});
|
||||
|
||||
test('should return "gpt-4" when the input includes "gpt-4"', () => {
|
||||
expect(client.getFunctionModelName('testing gpt-4')).toBe('gpt-4');
|
||||
});
|
||||
|
||||
test('should return "gpt-3.5-turbo" for input that does not meet any specific condition', () => {
|
||||
expect(client.getFunctionModelName('random string')).toBe('gpt-3.5-turbo');
|
||||
expect(client.getFunctionModelName('')).toBe('gpt-3.5-turbo');
|
||||
});
|
||||
});
|
||||
|
||||
describe('Azure OpenAI tests specific to Plugins', () => {
|
||||
// TODO: add more tests for Azure OpenAI integration with Plugins
|
||||
// let client;
|
||||
// beforeEach(() => {
|
||||
// client = new PluginsClient('dummy_api_key');
|
||||
// });
|
||||
|
||||
test('should not call getFunctionModelName when azure options are set', () => {
|
||||
const spy = jest.spyOn(PluginsClient.prototype, 'getFunctionModelName');
|
||||
const model = 'gpt-4-turbo';
|
||||
|
||||
// note, without the azure change in PR #1766, `getFunctionModelName` is called twice
|
||||
const testClient = new PluginsClient('dummy_api_key', {
|
||||
agentOptions: {
|
||||
model,
|
||||
agent: 'functions',
|
||||
},
|
||||
azure: defaultAzureOptions,
|
||||
});
|
||||
|
||||
expect(spy).not.toHaveBeenCalled();
|
||||
expect(testClient.agentOptions.model).toBe(model);
|
||||
|
||||
spy.mockRestore();
|
||||
});
|
||||
});
|
||||
|
||||
describe('sendMessage with filtered tools', () => {
|
||||
let TestAgent;
|
||||
const apiKey = 'fake-api-key';
|
||||
const mockTools = [{ name: 'tool1' }, { name: 'tool2' }, { name: 'tool3' }, { name: 'tool4' }];
|
||||
|
||||
beforeEach(() => {
|
||||
TestAgent = new PluginsClient(apiKey, {
|
||||
tools: mockTools,
|
||||
modelOptions: {
|
||||
model: 'gpt-3.5-turbo',
|
||||
temperature: 0,
|
||||
max_tokens: 2,
|
||||
},
|
||||
agentOptions: {
|
||||
model: 'gpt-3.5-turbo',
|
||||
},
|
||||
});
|
||||
|
||||
TestAgent.options.req = {
|
||||
app: {
|
||||
locals: {},
|
||||
},
|
||||
};
|
||||
|
||||
TestAgent.sendMessage = jest.fn().mockImplementation(async () => {
|
||||
const { filteredTools = [], includedTools = [] } = TestAgent.options.req.app.locals;
|
||||
|
||||
if (includedTools.length > 0) {
|
||||
const tools = TestAgent.options.tools.filter((plugin) =>
|
||||
includedTools.includes(plugin.name),
|
||||
);
|
||||
TestAgent.options.tools = tools;
|
||||
} else {
|
||||
const tools = TestAgent.options.tools.filter(
|
||||
(plugin) => !filteredTools.includes(plugin.name),
|
||||
);
|
||||
TestAgent.options.tools = tools;
|
||||
}
|
||||
|
||||
return {
|
||||
text: 'Mocked response',
|
||||
tools: TestAgent.options.tools,
|
||||
};
|
||||
});
|
||||
});
|
||||
|
||||
test('should filter out tools when filteredTools is provided', async () => {
|
||||
TestAgent.options.req.app.locals.filteredTools = ['tool1', 'tool3'];
|
||||
const response = await TestAgent.sendMessage('Test message');
|
||||
expect(response.tools).toHaveLength(2);
|
||||
expect(response.tools).toEqual(
|
||||
expect.arrayContaining([
|
||||
expect.objectContaining({ name: 'tool2' }),
|
||||
expect.objectContaining({ name: 'tool4' }),
|
||||
]),
|
||||
);
|
||||
});
|
||||
|
||||
test('should only include specified tools when includedTools is provided', async () => {
|
||||
TestAgent.options.req.app.locals.includedTools = ['tool2', 'tool4'];
|
||||
const response = await TestAgent.sendMessage('Test message');
|
||||
expect(response.tools).toHaveLength(2);
|
||||
expect(response.tools).toEqual(
|
||||
expect.arrayContaining([
|
||||
expect.objectContaining({ name: 'tool2' }),
|
||||
expect.objectContaining({ name: 'tool4' }),
|
||||
]),
|
||||
);
|
||||
});
|
||||
|
||||
test('should prioritize includedTools over filteredTools', async () => {
|
||||
TestAgent.options.req.app.locals.filteredTools = ['tool1', 'tool3'];
|
||||
TestAgent.options.req.app.locals.includedTools = ['tool1', 'tool2'];
|
||||
const response = await TestAgent.sendMessage('Test message');
|
||||
expect(response.tools).toHaveLength(2);
|
||||
expect(response.tools).toEqual(
|
||||
expect.arrayContaining([
|
||||
expect.objectContaining({ name: 'tool1' }),
|
||||
expect.objectContaining({ name: 'tool2' }),
|
||||
]),
|
||||
);
|
||||
});
|
||||
|
||||
test('should not modify tools when no filters are provided', async () => {
|
||||
const response = await TestAgent.sendMessage('Test message');
|
||||
expect(response.tools).toHaveLength(4);
|
||||
expect(response.tools).toEqual(expect.arrayContaining(mockTools));
|
||||
});
|
||||
});
|
||||
});
|
||||
Loading…
Add table
Add a link
Reference in a new issue