317 lines
8.5 KiB
TypeScript
317 lines
8.5 KiB
TypeScript
import { chats2GPTMessages } from '@fastgpt/global/core/chat/adapt';
|
||
import { filterGPTMessageByMaxTokens } from '@fastgpt/service/core/chat/utils';
|
||
import {
|
||
countGptMessagesTokens,
|
||
countMessagesTokens
|
||
} from '@fastgpt/global/common/string/tiktoken';
|
||
import type { ChatItemType } from '@fastgpt/global/core/chat/type.d';
|
||
import { ChatItemValueTypeEnum, ChatRoleEnum } from '@fastgpt/global/core/chat/constants';
|
||
import { getAIApi } from '@fastgpt/service/core/ai/config';
|
||
import type { ClassifyQuestionAgentItemType } from '@fastgpt/global/core/module/type.d';
|
||
import { ModuleInputKeyEnum } from '@fastgpt/global/core/module/constants';
|
||
import { DispatchNodeResponseKeyEnum } from '@fastgpt/global/core/module/runtime/constants';
|
||
import type { ModuleDispatchProps } from '@fastgpt/global/core/module/type.d';
|
||
import { replaceVariable } from '@fastgpt/global/common/string/tools';
|
||
import { Prompt_CQJson } from '@/global/core/prompt/agent';
|
||
import { LLMModelItemType } from '@fastgpt/global/core/ai/model.d';
|
||
import { ModelTypeEnum, getLLMModel } from '@fastgpt/service/core/ai/model';
|
||
import { getHistories } from '../utils';
|
||
import { formatModelChars2Points } from '@fastgpt/service/support/wallet/usage/utils';
|
||
import { ChatCompletionRequestMessageRoleEnum } from '@fastgpt/global/core/ai/constants';
|
||
import {
|
||
ChatCompletionCreateParams,
|
||
ChatCompletionMessageParam,
|
||
ChatCompletionTool
|
||
} from '@fastgpt/global/core/ai/type';
|
||
import { DispatchNodeResultType } from '@fastgpt/global/core/module/runtime/type';
|
||
|
||
type Props = ModuleDispatchProps<{
|
||
[ModuleInputKeyEnum.aiModel]: string;
|
||
[ModuleInputKeyEnum.aiSystemPrompt]?: string;
|
||
[ModuleInputKeyEnum.history]?: ChatItemType[] | number;
|
||
[ModuleInputKeyEnum.userChatInput]: string;
|
||
[ModuleInputKeyEnum.agents]: ClassifyQuestionAgentItemType[];
|
||
}>;
|
||
type CQResponse = DispatchNodeResultType<{
|
||
[key: string]: any;
|
||
}>;
|
||
type ActionProps = Props & { cqModel: LLMModelItemType };
|
||
|
||
const agentFunName = 'classify_question';
|
||
|
||
/* request openai chat */
|
||
export const dispatchClassifyQuestion = async (props: Props): Promise<CQResponse> => {
|
||
const {
|
||
user,
|
||
module: { name },
|
||
histories,
|
||
params: { model, history = 6, agents, userChatInput }
|
||
} = props as Props;
|
||
|
||
if (!userChatInput) {
|
||
return Promise.reject('Input is empty');
|
||
}
|
||
|
||
const cqModel = getLLMModel(model);
|
||
|
||
const chatHistories = getHistories(history, histories);
|
||
|
||
const { arg, tokens } = await (async () => {
|
||
if (cqModel.toolChoice) {
|
||
return toolChoice({
|
||
...props,
|
||
histories: chatHistories,
|
||
cqModel
|
||
});
|
||
}
|
||
if (cqModel.functionCall) {
|
||
return functionCall({
|
||
...props,
|
||
histories: chatHistories,
|
||
cqModel
|
||
});
|
||
}
|
||
return completions({
|
||
...props,
|
||
histories: chatHistories,
|
||
cqModel
|
||
});
|
||
})();
|
||
|
||
const result = agents.find((item) => item.key === arg?.type) || agents[agents.length - 1];
|
||
|
||
const { totalPoints, modelName } = formatModelChars2Points({
|
||
model: cqModel.model,
|
||
tokens,
|
||
modelType: ModelTypeEnum.llm
|
||
});
|
||
|
||
return {
|
||
[result.key]: true,
|
||
[DispatchNodeResponseKeyEnum.nodeResponse]: {
|
||
totalPoints: user.openaiAccount?.key ? 0 : totalPoints,
|
||
model: modelName,
|
||
query: userChatInput,
|
||
tokens,
|
||
cqList: agents,
|
||
cqResult: result.value,
|
||
contextTotalLen: chatHistories.length + 2
|
||
},
|
||
[DispatchNodeResponseKeyEnum.nodeDispatchUsages]: [
|
||
{
|
||
moduleName: name,
|
||
totalPoints: user.openaiAccount?.key ? 0 : totalPoints,
|
||
model: modelName,
|
||
tokens
|
||
}
|
||
]
|
||
};
|
||
};
|
||
|
||
const getFunctionCallSchema = ({
|
||
cqModel,
|
||
histories,
|
||
params: { agents, systemPrompt, userChatInput }
|
||
}: ActionProps) => {
|
||
const messages: ChatItemType[] = [
|
||
...histories,
|
||
{
|
||
obj: ChatRoleEnum.Human,
|
||
value: [
|
||
{
|
||
type: ChatItemValueTypeEnum.text,
|
||
text: {
|
||
content: systemPrompt
|
||
? `<背景知识>
|
||
${systemPrompt}
|
||
</背景知识>
|
||
|
||
问题: "${userChatInput}"
|
||
`
|
||
: userChatInput
|
||
}
|
||
}
|
||
]
|
||
}
|
||
];
|
||
|
||
const adaptMessages = chats2GPTMessages({ messages, reserveId: false });
|
||
const filterMessages = filterGPTMessageByMaxTokens({
|
||
messages: adaptMessages,
|
||
maxTokens: cqModel.maxContext
|
||
});
|
||
|
||
// function body
|
||
const agentFunction = {
|
||
name: agentFunName,
|
||
description: '结合对话记录及背景知识，对问题进行分类，并返回对应的类型字段',
|
||
parameters: {
|
||
type: 'object',
|
||
properties: {
|
||
type: {
|
||
type: 'string',
|
||
description: `问题类型。下面是几种可选的问题类型: ${agents
|
||
.map((item) => `${item.value}，返回：'${item.key}'`)
|
||
.join('；')}`,
|
||
enum: agents.map((item) => item.key)
|
||
}
|
||
},
|
||
required: ['type']
|
||
}
|
||
};
|
||
|
||
return {
|
||
agentFunction,
|
||
filterMessages
|
||
};
|
||
};
|
||
|
||
const toolChoice = async (props: ActionProps) => {
|
||
const { user, cqModel } = props;
|
||
|
||
const { agentFunction, filterMessages } = getFunctionCallSchema(props);
|
||
// function body
|
||
const tools: ChatCompletionTool[] = [
|
||
{
|
||
type: 'function',
|
||
function: agentFunction
|
||
}
|
||
];
|
||
|
||
const ai = getAIApi({
|
||
userKey: user.openaiAccount,
|
||
timeout: 480000
|
||
});
|
||
|
||
const response = await ai.chat.completions.create({
|
||
model: cqModel.model,
|
||
temperature: 0,
|
||
messages: filterMessages,
|
||
tools,
|
||
tool_choice: { type: 'function', function: { name: agentFunName } }
|
||
});
|
||
|
||
try {
|
||
const arg = JSON.parse(
|
||
response?.choices?.[0]?.message?.tool_calls?.[0]?.function?.arguments || ''
|
||
);
|
||
const completeMessages: ChatCompletionMessageParam[] = [
|
||
...filterMessages,
|
||
{
|
||
role: ChatCompletionRequestMessageRoleEnum.Assistant,
|
||
tool_calls: response.choices?.[0]?.message?.tool_calls
|
||
}
|
||
];
|
||
|
||
return {
|
||
arg,
|
||
tokens: countGptMessagesTokens(completeMessages, tools)
|
||
};
|
||
} catch (error) {
|
||
console.log(response.choices?.[0]?.message);
|
||
|
||
console.log('Your model may not support toll_call', error);
|
||
|
||
return {
|
||
arg: {},
|
||
tokens: 0
|
||
};
|
||
}
|
||
};
|
||
|
||
const functionCall = async (props: ActionProps) => {
|
||
const { user, cqModel } = props;
|
||
|
||
const { agentFunction, filterMessages } = getFunctionCallSchema(props);
|
||
const functions: ChatCompletionCreateParams.Function[] = [agentFunction];
|
||
|
||
const ai = getAIApi({
|
||
userKey: user.openaiAccount,
|
||
timeout: 480000
|
||
});
|
||
|
||
const response = await ai.chat.completions.create({
|
||
model: cqModel.model,
|
||
temperature: 0,
|
||
messages: filterMessages,
|
||
function_call: {
|
||
name: agentFunName
|
||
},
|
||
functions
|
||
});
|
||
|
||
try {
|
||
const arg = JSON.parse(response?.choices?.[0]?.message?.function_call?.arguments || '');
|
||
const completeMessages: ChatCompletionMessageParam[] = [
|
||
...filterMessages,
|
||
{
|
||
role: ChatCompletionRequestMessageRoleEnum.Assistant,
|
||
function_call: response.choices?.[0]?.message?.function_call
|
||
}
|
||
];
|
||
|
||
return {
|
||
arg,
|
||
tokens: countGptMessagesTokens(completeMessages, undefined, functions)
|
||
};
|
||
} catch (error) {
|
||
console.log(response.choices?.[0]?.message);
|
||
|
||
console.log('Your model may not support toll_call', error);
|
||
|
||
return {
|
||
arg: {},
|
||
tokens: 0
|
||
};
|
||
}
|
||
};
|
||
|
||
const completions = async ({
|
||
cqModel,
|
||
user,
|
||
histories,
|
||
params: { agents, systemPrompt = '', userChatInput }
|
||
}: ActionProps) => {
|
||
const messages: ChatItemType[] = [
|
||
{
|
||
obj: ChatRoleEnum.Human,
|
||
value: [
|
||
{
|
||
type: ChatItemValueTypeEnum.text,
|
||
text: {
|
||
content: replaceVariable(cqModel.customCQPrompt || Prompt_CQJson, {
|
||
systemPrompt: systemPrompt || 'null',
|
||
typeList: agents
|
||
.map((item) => `{"questionType": "${item.value}", "typeId": "${item.key}"}`)
|
||
.join('\n'),
|
||
history: histories.map((item) => `${item.obj}:${item.value}`).join('\n'),
|
||
question: userChatInput
|
||
})
|
||
}
|
||
}
|
||
]
|
||
}
|
||
];
|
||
|
||
const ai = getAIApi({
|
||
userKey: user.openaiAccount,
|
||
timeout: 480000
|
||
});
|
||
|
||
const data = await ai.chat.completions.create({
|
||
model: cqModel.model,
|
||
temperature: 0.01,
|
||
messages: chats2GPTMessages({ messages, reserveId: false }),
|
||
stream: false
|
||
});
|
||
const answer = data.choices?.[0].message?.content || '';
|
||
|
||
const id =
|
||
agents.find((item) => answer.includes(item.key) || answer.includes(item.value))?.key || '';
|
||
|
||
return {
|
||
tokens: countMessagesTokens(messages),
|
||
arg: { type: id }
|
||
};
|
||
};
|