Files
risuai/src/ts/process/request.ts
2025-02-26 12:22:10 +09:00

3164 lines
101 KiB
TypeScript

import type { MultiModal, OpenAIChat, OpenAIChatFull } from "./index.svelte";
import { getCurrentCharacter, getCurrentChat, getDatabase, setDatabase, type character } from "../storage/database.svelte";
import { pluginProcess, pluginV2 } from "../plugins/plugins";
import { language } from "../../lang";
import { stringlizeAINChat, getStopStrings, unstringlizeAIN, unstringlizeChat } from "./stringlize";
import { addFetchLog, fetchNative, globalFetch, isNodeServer, isTauri, textifyReadableStream } from "../globalApi.svelte";
import { sleep } from "../util";
import { NovelAIBadWordIds, stringlizeNAIChat } from "./models/nai";
import { strongBan, tokenize, tokenizeNum } from "../tokenizer";
import { risuChatParser } from "../parser.svelte";
import { SignatureV4 } from "@smithy/signature-v4";
import { HttpRequest } from "@smithy/protocol-http";
import { Sha256 } from "@aws-crypto/sha256-js";
import { supportsInlayImage } from "./files/inlays";
import { Capacitor } from "@capacitor/core";
import { getFreeOpenRouterModel } from "../model/openrouter";
import { runTransformers } from "./transformers";
import {Ollama} from 'ollama/dist/browser.mjs'
import { applyChatTemplate } from "./templates/chatTemplate";
import { OobaParams } from "./prompt";
import { extractJSON, getGeneralJSONSchema, getOpenAIJSONSchema } from "./templates/jsonSchema";
import { getModelInfo, LLMFlags, LLMFormat, type LLMModel } from "../model/modellist";
import { runTrigger } from "./triggers";
interface requestDataArgument{
formated: OpenAIChat[]
bias: {[key:number]:number}
biasString?: [string,number][]
currentChar?: character
temperature?: number
maxTokens?:number
PresensePenalty?: number
frequencyPenalty?: number,
useStreaming?:boolean
isGroupChat?:boolean
useEmotion?:boolean
continue?:boolean
chatId?:string
noMultiGen?:boolean
schema?:string
extractJson?:string
}
interface RequestDataArgumentExtended extends requestDataArgument{
aiModel?:string
multiGen?:boolean
abortSignal?:AbortSignal
modelInfo?:LLMModel
customURL?:string
mode?:ModelModeExtended
}
type requestDataResponse = {
type: 'success'|'fail'
result: string
noRetry?: boolean,
special?: {
emotion?: string
},
failByServerError?: boolean
}|{
type: "streaming",
result: ReadableStream<StreamResponseChunk>,
special?: {
emotion?: string
}
}|{
type: "multiline",
result: ['user'|'char',string][],
special?: {
emotion?: string
}
}
interface StreamResponseChunk{[key:string]:string}
interface OaiFunctions {
name: string;
description: string;
parameters: {
type: string;
properties: {
[key:string]: {
type: string;
enum: string[]
};
};
required: string[];
};
}
export type Parameter = 'temperature'|'top_k'|'repetition_penalty'|'min_p'|'top_a'|'top_p'|'frequency_penalty'|'presence_penalty'|'reasoning_effort'|'thinking_tokens'
export type ModelModeExtended = 'model'|'submodel'|'memory'|'emotion'|'otherAx'|'translate'
type ParameterMap = {
[key in Parameter]?: string;
};
function setObjectValue<T>(obj: T, key: string, value: any): T {
const splitKey = key.split('.');
if(splitKey.length > 1){
const firstKey = splitKey.shift()
if(!obj[firstKey]){
obj[firstKey] = {};
}
obj[firstKey] = setObjectValue(obj[firstKey], splitKey.join('.'), value);
return obj;
}
obj[key] = value;
return obj;
}
function applyParameters(data: { [key: string]: any }, parameters: Parameter[], rename: ParameterMap, ModelMode:ModelModeExtended, arg:{
ignoreTopKIfZero?:boolean
} = {}): { [key: string]: any } {
const db = getDatabase()
function getEffort(effort:number){
switch(effort){
case 0:{
return 'low'
}
case 1:{
return 'medium'
}
case 2:{
return 'high'
}
default:{
return 'medium'
}
}
}
if(db.seperateParametersEnabled && ModelMode !== 'model'){
if(ModelMode === 'submodel'){
ModelMode = 'otherAx'
}
for(const parameter of parameters){
let value:number|string = 0
if(parameter === 'top_k' && arg.ignoreTopKIfZero && db.seperateParameters[ModelMode][parameter] === 0){
continue
}
switch(parameter){
case 'temperature':{
value = db.seperateParameters[ModelMode].temperature === -1000 ? -1000 : (db.seperateParameters[ModelMode].temperature / 100)
break
}
case 'top_k':{
value = db.seperateParameters[ModelMode].top_k
break
}
case 'repetition_penalty':{
value = db.seperateParameters[ModelMode].repetition_penalty
break
}
case 'min_p':{
value = db.seperateParameters[ModelMode].min_p
break
}
case 'top_a':{
value = db.seperateParameters[ModelMode].top_a
break
}
case 'top_p':{
value = db.seperateParameters[ModelMode].top_p
break
}
case 'thinking_tokens':{
value = db.seperateParameters[ModelMode].thinking_tokens
break
}
case 'frequency_penalty':{
value = db.seperateParameters[ModelMode].frequency_penalty === -1000 ? -1000 : (db.seperateParameters[ModelMode].frequency_penalty / 100)
break
}
case 'presence_penalty':{
value = db.seperateParameters[ModelMode].presence_penalty === -1000 ? -1000 : (db.seperateParameters[ModelMode].presence_penalty / 100)
break
}
case 'reasoning_effort':{
value = getEffort(db.seperateParameters[ModelMode].reasoning_effort)
break
}
}
if(value === -1000 || value === undefined){
continue
}
data = setObjectValue(data, rename[parameter] ?? parameter, value)
}
return data
}
for(const parameter of parameters){
let value:number|string = 0
if(parameter === 'top_k' && arg.ignoreTopKIfZero && db.top_k === 0){
continue
}
switch(parameter){
case 'temperature':{
value = db.temperature === -1000 ? -1000 : (db.temperature / 100)
break
}
case 'top_k':{
value = db.top_k
break
}
case 'repetition_penalty':{
value = db.repetition_penalty
break
}
case 'min_p':{
value = db.min_p
break
}
case 'top_a':{
value = db.top_a
break
}
case 'top_p':{
value = db.top_p
break
}
case 'reasoning_effort':{
value = getEffort(db.reasoningEffort)
break
}
case 'frequency_penalty':{
value = db.frequencyPenalty === -1000 ? -1000 : (db.frequencyPenalty / 100)
break
}
case 'presence_penalty':{
value = db.PresensePenalty === -1000 ? -1000 : (db.PresensePenalty / 100)
break
}
case 'thinking_tokens':{
value = db.thinkingTokens
break
}
}
if(value === -1000){
continue
}
data = setObjectValue(data, rename[parameter] ?? parameter, value)
}
return data
}
export async function requestChatData(arg:requestDataArgument, model:ModelModeExtended, abortSignal:AbortSignal=null):Promise<requestDataResponse> {
const db = getDatabase()
let trys = 0
while(true){
if(pluginV2.replacerbeforeRequest.size > 0){
for(const replacer of pluginV2.replacerbeforeRequest){
arg.formated = await replacer(arg.formated, model)
}
}
try{
const currentChar = getCurrentCharacter()
if(currentChar?.type !== 'group'){
const perf = performance.now()
const d = await runTrigger(currentChar, 'request', {
chat: getCurrentChat(),
displayMode: true,
displayData: JSON.stringify(arg.formated)
})
const got = JSON.parse(d.displayData)
if(!got || !Array.isArray(got)){
throw new Error('Invalid return')
}
arg.formated = got
console.log('Trigger time', performance.now() - perf)
}
}
catch(e){
console.error(e)
}
const da = await requestChatDataMain(arg, model, abortSignal)
if(da.type === 'success' && pluginV2.replacerafterRequest.size > 0){
for(const replacer of pluginV2.replacerafterRequest){
da.result = await replacer(da.result, model)
}
}
if(da.type === 'success' && db.banCharacterset?.length > 0){
let failed = false
for(const set of db.banCharacterset){
console.log(set)
const checkRegex = new RegExp(`\\p{Script=${set}}`, 'gu')
if(checkRegex.test(da.result)){
trys += 1
if(trys > db.requestRetrys){
return {
type: 'fail',
result: 'Banned character found, retry limit reached'
}
}
failed = true
break
}
}
if(failed){
continue
}
}
if(da.type !== 'fail' || da.noRetry){
return da
}
if(da.failByServerError){
await sleep(1000)
if(db.antiServerOverloads){
trys -= 0.5 // reduce trys by 0.5, so that it will retry twice as much
}
}
trys += 1
if(trys > db.requestRetrys){
return da
}
}
}
interface OpenAITextContents {
type: 'text'
text: string
}
interface OpenAIImageContents {
type: 'image'|'image_url'
image_url: {
url: string
detail: string
}
}
type OpenAIContents = OpenAITextContents|OpenAIImageContents
export interface OpenAIChatExtra {
role: 'system'|'user'|'assistant'|'function'|'developer'
content: string|OpenAIContents[]
memo?:string
name?:string
removable?:boolean
attr?:string[]
multimodals?:MultiModal[]
thoughts?:string[]
prefix?:boolean
reasoning_content?:string
}
function reformater(formated:OpenAIChat[],modelInfo:LLMModel){
const db = getDatabase()
let systemPrompt:OpenAIChat|null = null
if(!modelInfo.flags.includes(LLMFlags.hasFullSystemPrompt)){
if(modelInfo.flags.includes(LLMFlags.hasFirstSystemPrompt)){
while(formated[0].role === 'system'){
if(systemPrompt){
systemPrompt.content += '\n\n' + formated[0].content
}
else{
systemPrompt = formated[0]
}
formated = formated.slice(1)
}
}
for(let i=0;i<formated.length;i++){
if(formated[i].role === 'system'){
formated[i].content = db.systemContentReplacement ? db.systemContentReplacement.replace('{{slot}}', formated[i].content) : `system: ${formated[i].content}`
formated[i].role = db.systemRoleReplacement
}
}
}
if(modelInfo.flags.includes(LLMFlags.requiresAlternateRole)){
let newFormated:OpenAIChat[] = []
for(let i=0;i<formated.length;i++){
const m = formated[i]
if(newFormated.length === 0){
newFormated.push(m)
continue
}
if(newFormated[newFormated.length-1].role === m.role){
newFormated[newFormated.length-1].content += '\n' + m.content
if(m.multimodals){
if(!newFormated[newFormated.length-1].multimodals){
newFormated[newFormated.length-1].multimodals = []
}
newFormated[newFormated.length-1].multimodals.push(...m.multimodals)
}
if(m.thoughts){
if(!newFormated[newFormated.length-1].thoughts){
newFormated[newFormated.length-1].thoughts = []
}
newFormated[newFormated.length-1].thoughts.push(...m.thoughts)
}
continue
}
else{
newFormated.push(m)
}
}
formated = newFormated
}
if(modelInfo.flags.includes(LLMFlags.mustStartWithUserInput)){
if(formated.length === 0 || formated[0].role !== 'user'){
formated.unshift({
role: 'user',
content: ' '
})
}
}
if(systemPrompt){
formated.unshift(systemPrompt)
}
return formated
}
export async function requestChatDataMain(arg:requestDataArgument, model:ModelModeExtended, abortSignal:AbortSignal=null):Promise<requestDataResponse> {
const db = getDatabase()
const targ:RequestDataArgumentExtended = arg
targ.formated = safeStructuredClone(arg.formated)
targ.maxTokens = arg.maxTokens ??db.maxResponse
targ.temperature = arg.temperature ?? (db.temperature / 100)
targ.bias = arg.bias
targ.currentChar = arg.currentChar
targ.useStreaming = db.useStreaming && arg.useStreaming
targ.continue = arg.continue ?? false
targ.biasString = arg.biasString ?? []
targ.aiModel = (model === 'model' ? db.aiModel : db.subModel)
targ.multiGen = ((db.genTime > 1 && targ.aiModel.startsWith('gpt') && (!arg.continue)) && (!arg.noMultiGen))
targ.abortSignal = abortSignal
targ.modelInfo = getModelInfo(targ.aiModel)
targ.mode = model
targ.extractJson = arg.extractJson ?? db.extractJson
if(targ.aiModel === 'reverse_proxy'){
targ.modelInfo.internalID = db.customProxyRequestModel
targ.modelInfo.format = db.customAPIFormat
targ.customURL = db.forceReplaceUrl
}
const format = targ.modelInfo.format
targ.formated = reformater(targ.formated, targ.modelInfo)
switch(format){
case LLMFormat.OpenAICompatible:
case LLMFormat.Mistral:
return requestOpenAI(targ)
case LLMFormat.OpenAILegacyInstruct:
return requestOpenAILegacyInstruct(targ)
case LLMFormat.NovelAI:
return requestNovelAI(targ)
case LLMFormat.OobaLegacy:
return requestOobaLegacy(targ)
case LLMFormat.Plugin:
return requestPlugin(targ)
case LLMFormat.Ooba:
return requestOoba(targ)
case LLMFormat.GoogleCloud:
return requestGoogleCloudVertex(targ)
case LLMFormat.Kobold:
return requestKobold(targ)
case LLMFormat.NovelList:
return requestNovelList(targ)
case LLMFormat.Ollama:
return requestOllama(targ)
case LLMFormat.Cohere:
return requestCohere(targ)
case LLMFormat.Anthropic:
case LLMFormat.AnthropicLegacy:
case LLMFormat.AWSBedrockClaude:
return requestClaude(targ)
case LLMFormat.Horde:
return requestHorde(targ)
case LLMFormat.WebLLM:
return requestWebLLM(targ)
}
return {
type: 'fail',
result: (language.errors.unknownModel)
}
}
async function requestOpenAI(arg:RequestDataArgumentExtended):Promise<requestDataResponse>{
let formatedChat:OpenAIChatExtra[] = []
const formated = arg.formated
const db = getDatabase()
const aiModel = arg.aiModel
for(let i=0;i<formated.length;i++){
const m = formated[i]
if(m.multimodals && m.multimodals.length > 0 && m.role === 'user'){
let v:OpenAIChatExtra = safeStructuredClone(m)
let contents:OpenAIContents[] = []
for(let j=0;j<m.multimodals.length;j++){
contents.push({
"type": "image_url",
"image_url": {
"url": m.multimodals[j].base64,
"detail": db.gptVisionQuality
}
})
}
contents.push({
"type": "text",
"text": m.content
})
v.content = contents
formatedChat.push(v)
}
else{
formatedChat.push(m)
}
}
let oobaSystemPrompts:string[] = []
for(let i=0;i<formatedChat.length;i++){
if(formatedChat[i].role !== 'function'){
if(!(formatedChat[i].name && formatedChat[i].name.startsWith('example_') && db.newOAIHandle)){
formatedChat[i].name = undefined
}
if(db.newOAIHandle && formatedChat[i].memo && formatedChat[i].memo.startsWith('NewChat')){
formatedChat[i].content = ''
}
if(arg.modelInfo.flags.includes(LLMFlags.deepSeekPrefix) && i === formatedChat.length-1 && formatedChat[i].role === 'assistant'){
formatedChat[i].prefix = true
}
if(arg.modelInfo.flags.includes(LLMFlags.deepSeekThinkingInput) && i === formatedChat.length-1 && formatedChat[i].thoughts && formatedChat[i].thoughts.length > 0 && formatedChat[i].role === 'assistant'){
formatedChat[i].reasoning_content = formatedChat[i].thoughts.join('\n')
}
delete formatedChat[i].memo
delete formatedChat[i].removable
delete formatedChat[i].attr
delete formatedChat[i].multimodals
delete formatedChat[i].thoughts
}
if(aiModel === 'reverse_proxy' && db.reverseProxyOobaMode && formatedChat[i].role === 'system'){
const cont = formatedChat[i].content
if(typeof(cont) === 'string'){
oobaSystemPrompts.push(cont)
formatedChat[i].content = ''
}
}
}
if(oobaSystemPrompts.length > 0){
formatedChat.push({
role: 'system',
content: oobaSystemPrompts.join('\n')
})
}
if(db.newOAIHandle){
formatedChat = formatedChat.filter(m => {
return m.content !== '' || (m.multimodals && m.multimodals.length > 0)
})
}
for(let i=0;i<arg.biasString.length;i++){
const bia = arg.biasString[i]
if(bia[0].startsWith('[[') && bia[0].endsWith(']]')){
const num = parseInt(bia[0].replace('[[', '').replace(']]', ''))
arg.bias[num] = bia[1]
continue
}
if(bia[1] === -101){
arg.bias = await strongBan(bia[0], arg.bias)
continue
}
const tokens = await tokenizeNum(bia[0])
for(const token of tokens){
arg.bias[token] = bia[1]
}
}
let oaiFunctions:OaiFunctions[] = []
if(arg.useEmotion){
oaiFunctions.push(
{
"name": "set_emotion",
"description": "sets a role playing character's emotion display. must be called one time at the end of response.",
"parameters": {
"type": "object",
"properties": {
"emotion": {
"type": "string", "enum": []
},
},
"required": ["emotion"],
},
}
)
}
if(oaiFunctions.length === 0){
oaiFunctions = undefined
}
const oaiFunctionCall = oaiFunctions ? (arg.useEmotion ? {"name": "set_emotion"} : "auto") : undefined
let requestModel = (aiModel === 'reverse_proxy' || aiModel === 'openrouter') ? db.proxyRequestModel : aiModel
let openrouterRequestModel = db.openrouterRequestModel
if(aiModel === 'reverse_proxy'){
requestModel = db.customProxyRequestModel
}
if(aiModel === 'openrouter' && db.openrouterRequestModel === 'risu/free'){
openrouterRequestModel = await getFreeOpenRouterModel()
}
if(arg.modelInfo.flags.includes(LLMFlags.DeveloperRole)){
formatedChat = formatedChat.map((v) => {
if(v.role === 'system'){
v.role = 'developer'
}
return v
})
}
console.log(formatedChat)
if(arg.modelInfo.format === LLMFormat.Mistral){
requestModel = aiModel
let reformatedChat:OpenAIChatExtra[] = []
for(let i=0;i<formatedChat.length;i++){
const chat = formatedChat[i]
if(i === 0){
if(chat.role === 'user' || chat.role === 'system'){
reformatedChat.push({
role: chat.role,
content: chat.content
})
}
else{
reformatedChat.push({
role: 'system',
content: chat.role + ':' + chat.content
})
}
}
else{
const prevChat = reformatedChat[reformatedChat.length-1]
if(prevChat?.role === chat.role){
reformatedChat[reformatedChat.length-1].content += '\n' + chat.content
continue
}
else if(chat.role === 'system'){
if(prevChat?.role === 'user'){
reformatedChat[reformatedChat.length-1].content += '\nSystem:' + chat.content
}
else{
reformatedChat.push({
role: 'user',
content: 'System:' + chat.content
})
}
}
else if(chat.role === 'function'){
reformatedChat.push({
role: 'user',
content: chat.content
})
}
else{
reformatedChat.push({
role: chat.role,
content: chat.content
})
}
}
}
const res = await globalFetch(arg.customURL ?? "https://api.mistral.ai/v1/chat/completions", {
body: applyParameters({
model: requestModel,
messages: reformatedChat,
safe_prompt: false,
max_tokens: arg.maxTokens,
}, ['temperature', 'presence_penalty', 'frequency_penalty', 'top_p'], {}, arg.mode ),
headers: {
"Authorization": "Bearer " + db.mistralKey,
},
abortSignal: arg.abortSignal,
chatId: arg.chatId
})
const dat = res.data as any
if(res.ok){
try {
const msg:OpenAIChatFull = (dat.choices[0].message)
return {
type: 'success',
result: msg.content
}
} catch (error) {
return {
type: 'fail',
result: (language.errors.httpError + `${JSON.stringify(dat)}`)
}
}
}
else{
if(dat.error && dat.error.message){
return {
type: 'fail',
result: (language.errors.httpError + `${dat.error.message}`)
}
}
else{
return {
type: 'fail',
result: (language.errors.httpError + `${JSON.stringify(res.data)}`)
}
}
}
}
db.cipherChat = false
let body:{
[key:string]:any
} = ({
model: aiModel === 'openrouter' ? openrouterRequestModel :
requestModel === 'gpt35' ? 'gpt-3.5-turbo'
: requestModel === 'gpt35_0613' ? 'gpt-3.5-turbo-0613'
: requestModel === 'gpt35_16k' ? 'gpt-3.5-turbo-16k'
: requestModel === 'gpt35_16k_0613' ? 'gpt-3.5-turbo-16k-0613'
: requestModel === 'gpt4' ? 'gpt-4'
: requestModel === 'gpt45' ? 'gpt-4.5-preview'
: requestModel === 'gpt4_32k' ? 'gpt-4-32k'
: requestModel === "gpt4_0613" ? 'gpt-4-0613'
: requestModel === "gpt4_32k_0613" ? 'gpt-4-32k-0613'
: requestModel === "gpt4_1106" ? 'gpt-4-1106-preview'
: requestModel === 'gpt4_0125' ? 'gpt-4-0125-preview'
: requestModel === "gptvi4_1106" ? 'gpt-4-vision-preview'
: requestModel === "gpt35_0125" ? 'gpt-3.5-turbo-0125'
: requestModel === "gpt35_1106" ? 'gpt-3.5-turbo-1106'
: requestModel === 'gpt35_0301' ? 'gpt-3.5-turbo-0301'
: requestModel === 'gpt4_0314' ? 'gpt-4-0314'
: requestModel === 'gpt4_turbo_20240409' ? 'gpt-4-turbo-2024-04-09'
: requestModel === 'gpt4_turbo' ? 'gpt-4-turbo'
: requestModel === 'gpt4o' ? 'gpt-4o'
: requestModel === 'gpt4o-2024-05-13' ? 'gpt-4o-2024-05-13'
: requestModel === 'gpt4om' ? 'gpt-4o-mini'
: requestModel === 'gpt4om-2024-07-18' ? 'gpt-4o-mini-2024-07-18'
: requestModel === 'gpt4o-2024-08-06' ? 'gpt-4o-2024-08-06'
: requestModel === 'gpt4o-2024-11-20' ? 'gpt-4o-2024-11-20'
: requestModel === 'gpt4o-chatgpt' ? 'chatgpt-4o-latest'
: requestModel === 'gpt4o1-preview' ? 'o1-preview'
: requestModel === 'gpt4o1-mini' ? 'o1-mini'
: arg.modelInfo.internalID ? arg.modelInfo.internalID
: (!requestModel) ? 'gpt-3.5-turbo'
: requestModel,
messages: formatedChat,
max_tokens: arg.maxTokens,
logit_bias: arg.bias,
stream: false,
})
if(Object.keys(body.logit_bias).length === 0){
delete body.logit_bias
}
if(aiModel.startsWith('gpt4o1') || arg.modelInfo.flags.includes(LLMFlags.OAICompletionTokens)){
body.max_completion_tokens = body.max_tokens
delete body.max_tokens
}
if(db.generationSeed > 0){
body.seed = db.generationSeed
}
if(db.jsonSchemaEnabled || arg.schema){
body.response_format = {
"type": "json_schema",
"json_schema": getOpenAIJSONSchema(arg.schema)
}
}
if(db.OAIPrediction){
body.prediction = {
type: "content",
content: db.OAIPrediction
}
}
if(aiModel === 'openrouter'){
if(db.openrouterFallback){
body.route = "fallback"
}
body.transforms = db.openrouterMiddleOut ? ['middle-out'] : []
if(db.openrouterProvider){
body.provider = {
order: [db.openrouterProvider]
}
}
if(db.useInstructPrompt){
delete body.messages
const prompt = applyChatTemplate(formated)
body.prompt = prompt
}
}
body = applyParameters(
body,
arg.modelInfo.parameters,
{},
arg.mode
)
if(aiModel === 'reverse_proxy' && db.reverseProxyOobaMode){
const OobaBodyTemplate = db.reverseProxyOobaArgs
const keys = Object.keys(OobaBodyTemplate)
for(const key of keys){
if(OobaBodyTemplate[key] !== undefined && OobaBodyTemplate[key] !== null){
// @ts-ignore
body[key] = OobaBodyTemplate[key]
}
}
}
if(supportsInlayImage()){
// inlay models doesn't support logit_bias
// OpenAI's gpt based llm model supports both logit_bias and inlay image
if(!(
aiModel.startsWith('gpt') ||
(aiModel == 'reverse_proxy' && (
db.proxyRequestModel?.startsWith('gpt') ||
(db.proxyRequestModel === 'custom' && db.customProxyRequestModel.startsWith('gpt'))
)))){
// @ts-ignore
delete body.logit_bias
}
}
let replacerURL = aiModel === 'openrouter' ? "https://openrouter.ai/api/v1/chat/completions" :
(aiModel === 'reverse_proxy') ? (arg.customURL) : ('https://api.openai.com/v1/chat/completions')
if(arg.modelInfo?.endpoint){
replacerURL = arg.modelInfo.endpoint
}
let risuIdentify = false
if(replacerURL.startsWith("risu::")){
risuIdentify = true
replacerURL = replacerURL.replace("risu::", '')
}
if(aiModel === 'reverse_proxy' && db.autofillRequestUrl){
if(replacerURL.endsWith('v1')){
replacerURL += '/chat/completions'
}
else if(replacerURL.endsWith('v1/')){
replacerURL += 'chat/completions'
}
else if(!(replacerURL.endsWith('completions') || replacerURL.endsWith('completions/'))){
if(replacerURL.endsWith('/')){
replacerURL += 'v1/chat/completions'
}
else{
replacerURL += '/v1/chat/completions'
}
}
}
let headers = {
"Authorization": "Bearer " + (aiModel === 'reverse_proxy' ? db.proxyKey : (aiModel === 'openrouter' ? db.openrouterKey : db.openAIKey)),
"Content-Type": "application/json"
}
if(arg.modelInfo?.keyIdentifier){
headers["Authorization"] = "Bearer " + db.OaiCompAPIKeys[arg.modelInfo.keyIdentifier]
}
if(aiModel === 'openrouter'){
headers["X-Title"] = 'RisuAI'
headers["HTTP-Referer"] = 'https://risuai.xyz'
}
if(risuIdentify){
headers["X-Proxy-Risu"] = 'RisuAI'
}
if(aiModel.startsWith('jamba')){
headers['Authorization'] = 'Bearer ' + db.ai21Key
replacerURL = 'https://api.ai21.com/studio/v1/chat/completions'
}
if(arg.multiGen){
// @ts-ignore
body.n = db.genTime
}
let throughProxi = (!isTauri) && (!isNodeServer) && (!db.usePlainFetch) && (!Capacitor.isNativePlatform())
if(arg.useStreaming){
body.stream = true
let urlHost = new URL(replacerURL).host
if(urlHost.includes("localhost") || urlHost.includes("172.0.0.1") || urlHost.includes("0.0.0.0")){
if(!isTauri){
return {
type: 'fail',
result: 'You are trying local request on streaming. this is not allowed dude to browser/os security policy. turn off streaming.',
}
}
}
const da = await fetchNative(replacerURL, {
body: JSON.stringify(body),
method: "POST",
headers: headers,
signal: arg.abortSignal,
chatId: arg.chatId
})
if(da.status !== 200){
return {
type: "fail",
result: await textifyReadableStream(da.body)
}
}
if (!da.headers.get('Content-Type').includes('text/event-stream')){
return {
type: "fail",
result: await textifyReadableStream(da.body)
}
}
addFetchLog({
body: body,
response: "Streaming",
success: true,
url: replacerURL,
})
let dataUint:Uint8Array|Buffer = new Uint8Array([])
let reasoningContent = ""
const transtream = new TransformStream<Uint8Array, StreamResponseChunk>( {
async transform(chunk, control) {
dataUint = Buffer.from(new Uint8Array([...dataUint, ...chunk]))
let JSONreaded:{[key:string]:string} = {}
try {
const datas = dataUint.toString().split('\n')
let readed:{[key:string]:string} = {}
for(const data of datas){
if(data.startsWith("data: ")){
try {
const rawChunk = data.replace("data: ", "")
if(rawChunk === "[DONE]"){
if(arg.modelInfo.flags.includes(LLMFlags.deepSeekThinkingOutput)){
readed["0"] = readed["0"].replace(/(.*)\<\/think\>/gms, (m, p1) => {
reasoningContent = p1
return ""
})
if(reasoningContent){
reasoningContent = reasoningContent.replace(/\<think\>/gm, '')
}
}
if(arg.extractJson && (db.jsonSchemaEnabled || arg.schema)){
for(const key in readed){
const extracted = extractJSON(readed[key], arg.extractJson)
JSONreaded[key] = extracted
}
console.log(JSONreaded)
control.enqueue(JSONreaded)
}
else if(reasoningContent){
control.enqueue({
"0": `<Thoughts>\n${reasoningContent}\n</Thoughts>\n${readed["0"]}`
})
}
else{
control.enqueue(readed)
}
return
}
const choices = JSON.parse(rawChunk).choices
for(const choice of choices){
const chunk = choice.delta.content ?? choices.text
if(chunk){
if(arg.multiGen){
const ind = choice.index.toString()
if(!readed[ind]){
readed[ind] = ""
}
readed[ind] += chunk
}
else{
if(!readed["0"]){
readed["0"] = ""
}
readed["0"] += chunk
}
}
if(choice?.delta?.reasoning_content){
reasoningContent += choice.delta.reasoning_content
}
}
} catch (error) {}
}
}
if(arg.modelInfo.flags.includes(LLMFlags.deepSeekThinkingOutput)){
readed["0"] = readed["0"].replace(/(.*)\<\/think\>/gms, (m, p1) => {
reasoningContent = p1
return ""
})
if(reasoningContent){
reasoningContent = reasoningContent.replace(/\<think\>/gm, '')
}
}
if(arg.extractJson && (db.jsonSchemaEnabled || arg.schema)){
for(const key in readed){
const extracted = extractJSON(readed[key], arg.extractJson)
JSONreaded[key] = extracted
}
console.log(JSONreaded)
control.enqueue(JSONreaded)
}
else if(reasoningContent){
control.enqueue({
"0": `<Thoughts>\n${reasoningContent}\n</Thoughts>\n${readed["0"]}`
})
}
else{
control.enqueue(readed)
}
} catch (error) {
}
}
},)
da.body.pipeTo(transtream.writable)
return {
type: 'streaming',
result: transtream.readable
}
}
if(aiModel === 'reverse_proxy'){
const additionalParams = db.additionalParams
for(let i=0;i<additionalParams.length;i++){
let key = additionalParams[i][0]
let value = additionalParams[i][1]
if(!key || !value){
continue
}
if(value === '{{none}}'){
if(key.startsWith('header::')){
key = key.replace('header::', '')
delete headers[key]
}
else{
delete body[key]
}
continue
}
if(key.startsWith('header::')){
key = key.replace('header::', '')
headers[key] = value
}
else if(value.startsWith('json::')){
value = value.replace('json::', '')
try {
body[key] = JSON.parse(value)
} catch (error) {}
}
else if(isNaN(parseFloat(value))){
body[key] = value
}
else{
body[key] = parseFloat(value)
}
}
}
const res = await globalFetch(replacerURL, {
body: body,
headers: headers,
abortSignal: arg.abortSignal,
useRisuToken:throughProxi,
chatId: arg.chatId
})
const dat = res.data as any
if(res.ok){
try {
if(arg.multiGen && dat.choices){
if(arg.extractJson && (db.jsonSchemaEnabled || arg.schema)){
const c = dat.choices.map((v:{
message:{content:string}
}) => {
const extracted = extractJSON(v.message.content, arg.extractJson)
return ["char",extracted]
})
return {
type: 'multiline',
result: c
}
}
return {
type: 'multiline',
result: dat.choices.map((v) => {
return ["char",v.message.content]
})
}
}
if(dat?.choices[0]?.text){
let text = dat.choices[0].text as string
if(arg.extractJson && (db.jsonSchemaEnabled || arg.schema)){
try {
const parsed = JSON.parse(text)
const extracted = extractJSON(parsed, arg.extractJson)
return {
type: 'success',
result: extracted
}
} catch (error) {
console.log(error)
return {
type: 'success',
result: text
}
}
}
return {
type: 'success',
result: text
}
}
if(arg.extractJson && (db.jsonSchemaEnabled || arg.schema)){
return {
type: 'success',
result: extractJSON(dat.choices[0].message.content, arg.extractJson)
}
}
const msg:OpenAIChatFull = (dat.choices[0].message)
let result = msg.content
if(arg.modelInfo.flags.includes(LLMFlags.deepSeekThinkingOutput)){
console.log("Checking for reasoning content")
let reasoningContent = ""
result = result.replace(/(.*)\<\/think\>/gms, (m, p1) => {
reasoningContent = p1
return ""
})
console.log(`Reasoning Content: ${reasoningContent}`)
if(reasoningContent){
reasoningContent = reasoningContent.replace(/\<think\>/gms, '')
result = `<Thoughts>\n${reasoningContent}\n</Thoughts>\n${result}`
}
}
if(dat?.choices[0]?.reasoning_content){
result = `<Thoughts>\n${dat.choices[0].reasoning_content}\n</Thoughts>\n${result}`
}
return {
type: 'success',
result: result
}
} catch (error) {
return {
type: 'fail',
result: (language.errors.httpError + `${JSON.stringify(dat)}`)
}
}
}
else{
if(dat.error && dat.error.message){
return {
type: 'fail',
result: (language.errors.httpError + `${dat.error.message}`)
}
}
else{
return {
type: 'fail',
result: (language.errors.httpError + `${JSON.stringify(res.data)}`)
}
}
}
}
async function requestOpenAILegacyInstruct(arg:RequestDataArgumentExtended):Promise<requestDataResponse>{
const formated = arg.formated
const db = getDatabase()
const maxTokens = arg.maxTokens
const temperature = arg.temperature
const prompt = formated.filter(m => m.content?.trim()).map(m => {
let author = '';
if(m.role == 'system'){
m.content = m.content.trim();
}
console.log(m.role +":"+m.content);
switch (m.role) {
case 'user': author = 'User'; break;
case 'assistant': author = 'Assistant'; break;
case 'system': author = 'Instruction'; break;
default: author = m.role; break;
}
return `\n## ${author}\n${m.content.trim()}`;
//return `\n\n${author}: ${m.content.trim()}`;
}).join("") + `\n## Response\n`;
const response = await globalFetch(arg.customURL ?? "https://api.openai.com/v1/completions", {
body: {
model: "gpt-3.5-turbo-instruct",
prompt: prompt,
max_tokens: maxTokens,
temperature: temperature,
top_p: 1,
stop:["User:"," User:", "user:", " user:"],
presence_penalty: arg.PresensePenalty || (db.PresensePenalty / 100),
frequency_penalty: arg.frequencyPenalty || (db.frequencyPenalty / 100),
},
headers: {
"Content-Type": "application/json",
"Authorization": "Bearer " + db.openAIKey,
},
chatId: arg.chatId
});
if(!response.ok){
return {
type: 'fail',
result: (language.errors.httpError + `${JSON.stringify(response.data)}`)
}
}
const text:string = response.data.choices[0].text
return {
type: 'success',
result: text.replace(/##\n/g, '')
}
}
async function requestNovelAI(arg:RequestDataArgumentExtended):Promise<requestDataResponse>{
const formated = arg.formated
const db = getDatabase()
const aiModel = arg.aiModel
const temperature = arg.temperature
const maxTokens = arg.maxTokens
const biasString = arg.biasString
const currentChar = getCurrentCharacter()
const prompt = stringlizeNAIChat(formated, currentChar?.name ?? '', arg.continue)
const abortSignal = arg.abortSignal
let logit_bias_exp:{
sequence: number[], bias: number, ensure_sequence_finish: false, generate_once: true
}[] = []
for(let i=0;i<biasString.length;i++){
const bia = biasString[i]
const tokens = await tokenizeNum(bia[0])
const tokensInNumberArray:number[] = []
for(const token of tokens){
tokensInNumberArray.push(token)
}
logit_bias_exp.push({
sequence: tokensInNumberArray,
bias: bia[1],
ensure_sequence_finish: false,
generate_once: true
})
}
let prefix = 'vanilla'
if(db.NAIadventure){
prefix = 'theme_textadventure'
}
const gen = db.NAIsettings
const payload = {
temperature:temperature,
max_length: maxTokens,
min_length: 1,
top_k: gen.topK,
top_p: gen.topP,
top_a: gen.topA,
tail_free_sampling: gen.tailFreeSampling,
repetition_penalty: gen.repetitionPenalty,
repetition_penalty_range: gen.repetitionPenaltyRange,
repetition_penalty_slope: gen.repetitionPenaltySlope,
repetition_penalty_frequency: gen.frequencyPenalty,
repetition_penalty_presence: gen.presencePenalty,
generate_until_sentence: true,
use_cache: false,
use_string: true,
return_full_text: false,
prefix: prefix,
order: [6, 2, 3, 0, 4, 1, 5, 8],
typical_p: gen.typicalp,
repetition_penalty_whitelist:[49256,49264,49231,49230,49287,85,49255,49399,49262,336,333,432,363,468,492,745,401,426,623,794,1096,2919,2072,7379,1259,2110,620,526,487,16562,603,805,761,2681,942,8917,653,3513,506,5301,562,5010,614,10942,539,2976,462,5189,567,2032,123,124,125,126,127,128,129,130,131,132,588,803,1040,49209,4,5,6,7,8,9,10,11,12],
stop_sequences: [[49287], [49405]],
bad_words_ids: NovelAIBadWordIds,
logit_bias_exp: logit_bias_exp,
mirostat_lr: gen.mirostat_lr ?? 1,
mirostat_tau: gen.mirostat_tau ?? 0,
cfg_scale: gen.cfg_scale ?? 1,
cfg_uc: ""
}
const body = {
"input": prompt,
"model": aiModel === 'novelai_kayra' ? 'kayra-v1' : 'clio-v1',
"parameters":payload
}
const da = await globalFetch(aiModel === 'novelai_kayra' ? "https://text.novelai.net/ai/generate" : "https://api.novelai.net/ai/generate", {
body: body,
headers: {
"Authorization": "Bearer " + db.novelai.token
},
abortSignal,
chatId: arg.chatId
})
if((!da.ok )|| (!da.data.output)){
return {
type: 'fail',
result: (language.errors.httpError + `${JSON.stringify(da.data)}`)
}
}
return {
type: "success",
result: unstringlizeChat(da.data.output, formated, currentChar?.name ?? '')
}
}
async function requestOobaLegacy(arg:RequestDataArgumentExtended):Promise<requestDataResponse> {
const formated = arg.formated
const db = getDatabase()
const aiModel = arg.aiModel
const maxTokens = arg.maxTokens
const currentChar = getCurrentCharacter()
const useStreaming = arg.useStreaming
const abortSignal = arg.abortSignal
let streamUrl = db.textgenWebUIStreamURL.replace(/\/api.*/, "/api/v1/stream")
let blockingUrl = db.textgenWebUIBlockingURL.replace(/\/api.*/, "/api/v1/generate")
let bodyTemplate:{[key:string]:any} = {}
const prompt = applyChatTemplate(formated)
let stopStrings = getStopStrings(false)
if(db.localStopStrings){
stopStrings = db.localStopStrings.map((v) => {
return risuChatParser(v.replace(/\\n/g, "\n"))
})
}
bodyTemplate = {
'max_new_tokens': db.maxResponse,
'do_sample': db.ooba.do_sample,
'temperature': (db.temperature / 100),
'top_p': db.ooba.top_p,
'typical_p': db.ooba.typical_p,
'repetition_penalty': db.ooba.repetition_penalty,
'encoder_repetition_penalty': db.ooba.encoder_repetition_penalty,
'top_k': db.ooba.top_k,
'min_length': db.ooba.min_length,
'no_repeat_ngram_size': db.ooba.no_repeat_ngram_size,
'num_beams': db.ooba.num_beams,
'penalty_alpha': db.ooba.penalty_alpha,
'length_penalty': db.ooba.length_penalty,
'early_stopping': false,
'truncation_length': maxTokens,
'ban_eos_token': db.ooba.ban_eos_token,
'stopping_strings': stopStrings,
'seed': -1,
add_bos_token: db.ooba.add_bos_token,
topP: db.top_p,
prompt: prompt
}
const headers = (aiModel === 'textgen_webui') ? {} : {
'X-API-KEY': db.mancerHeader
}
if(useStreaming){
const oobaboogaSocket = new WebSocket(streamUrl);
const statusCode = await new Promise((resolve) => {
oobaboogaSocket.onopen = () => resolve(0)
oobaboogaSocket.onerror = () => resolve(1001)
oobaboogaSocket.onclose = ({ code }) => resolve(code)
})
if(abortSignal.aborted || statusCode !== 0) {
oobaboogaSocket.close()
return ({
type: "fail",
result: abortSignal.reason || `WebSocket connection failed to '${streamUrl}' failed!`,
})
}
const close = () => {
oobaboogaSocket.close()
}
const stream = new ReadableStream({
start(controller){
let readed = "";
oobaboogaSocket.onmessage = async (event) => {
const json = JSON.parse(event.data);
if (json.event === "stream_end") {
close()
controller.close()
return
}
if (json.event !== "text_stream") return
readed += json.text
controller.enqueue(readed)
};
oobaboogaSocket.send(JSON.stringify(bodyTemplate));
},
cancel(){
close()
}
})
oobaboogaSocket.onerror = close
oobaboogaSocket.onclose = close
abortSignal.addEventListener("abort", close)
return {
type: 'streaming',
result: stream
}
}
const res = await globalFetch(blockingUrl, {
body: bodyTemplate,
headers: headers,
abortSignal,
chatId: arg.chatId
})
const dat = res.data as any
if(res.ok){
try {
let result:string = dat.results[0].text
return {
type: 'success',
result: unstringlizeChat(result, formated, currentChar?.name ?? '')
}
} catch (error) {
return {
type: 'fail',
result: (language.errors.httpError + `${error}`)
}
}
}
else{
return {
type: 'fail',
result: (language.errors.httpError + `${JSON.stringify(res.data)}`)
}
}
}
async function requestOoba(arg:RequestDataArgumentExtended):Promise<requestDataResponse> {
const formated = arg.formated
const db = getDatabase()
const aiModel = arg.aiModel
const maxTokens = arg.maxTokens
const temperature = arg.temperature
const prompt = applyChatTemplate(formated)
let stopStrings = getStopStrings(false)
if(db.localStopStrings){
stopStrings = db.localStopStrings.map((v) => {
return risuChatParser(v.replace(/\\n/g, "\n"))
})
}
let bodyTemplate:Record<string, any> = {
'prompt': prompt,
presence_penalty: arg.PresensePenalty || (db.PresensePenalty / 100),
frequency_penalty: arg.frequencyPenalty || (db.frequencyPenalty / 100),
logit_bias: {},
max_tokens: maxTokens,
stop: stopStrings,
temperature: temperature,
top_p: db.top_p,
}
const url = new URL(db.textgenWebUIBlockingURL)
url.pathname = "/v1/completions"
const urlStr = url.toString()
const OobaBodyTemplate = db.reverseProxyOobaArgs
const keys = Object.keys(OobaBodyTemplate)
for(const key of keys){
if(OobaBodyTemplate[key] !== undefined && OobaBodyTemplate[key] !== null && OobaParams.includes(key)){
bodyTemplate[key] = OobaBodyTemplate[key]
}
else if(bodyTemplate[key]){
delete bodyTemplate[key]
}
}
const response = await globalFetch(urlStr, {
body: bodyTemplate,
chatId: arg.chatId
})
if(!response.ok){
return {
type: 'fail',
result: (language.errors.httpError + `${JSON.stringify(response.data)}`)
}
}
const text:string = response.data.choices[0].text
return {
type: 'success',
result: text.replace(/##\n/g, '')
}
}
async function requestPlugin(arg:RequestDataArgumentExtended):Promise<requestDataResponse> {
const db = getDatabase()
try {
const formated = arg.formated
const maxTokens = arg.maxTokens
const bias = arg.biasString
const v2Function = pluginV2.providers.get(db.currentPluginProvider)
const d = v2Function ? (await v2Function(applyParameters({
prompt_chat: formated,
mode: arg.mode,
bias: [],
max_tokens: maxTokens,
}, [
'frequency_penalty','min_p','presence_penalty','repetition_penalty','top_k','top_p','temperature'
], {}, arg.mode) as any)) : await pluginProcess({
bias: bias,
prompt_chat: formated,
temperature: (db.temperature / 100),
max_tokens: maxTokens,
presence_penalty: (db.PresensePenalty / 100),
frequency_penalty: (db.frequencyPenalty / 100)
})
if(!d){
return {
type: 'fail',
result: (language.errors.unknownModel)
}
}
else if(!d.success){
return {
type: 'fail',
result: d.content instanceof ReadableStream ? await (new Response(d.content)).text() : d.content
}
}
else if(d.content instanceof ReadableStream){
let fullText = ''
const piper = new TransformStream<string, StreamResponseChunk>( {
transform(chunk, control) {
fullText += chunk
control.enqueue({
"0": fullText
})
}
})
return {
type: 'streaming',
result: d.content.pipeThrough(piper)
}
}
else{
return {
type: 'success',
result: d.content
}
}
} catch (error) {
console.error(error)
return {
type: 'fail',
result: `Plugin Error from ${db.currentPluginProvider}: ` + JSON.stringify(error)
}
}
}
async function requestGoogleCloudVertex(arg:RequestDataArgumentExtended):Promise<requestDataResponse> {
const formated = arg.formated
const db = getDatabase()
const maxTokens = arg.maxTokens
interface GeminiPart{
text?:string
"inlineData"?: {
"mimeType": string,
"data": string
},
}
interface GeminiChat {
role: "USER"|"MODEL"
parts:|GeminiPart[]
}
let reformatedChat:GeminiChat[] = []
let systemPrompt = ''
if(formated[0].role === 'system'){
systemPrompt = formated[0].content
formated.shift()
}
for(let i=0;i<formated.length;i++){
const chat = formated[i]
const prevChat = reformatedChat[reformatedChat.length-1]
const qRole =
chat.role === 'user' ? 'USER' :
chat.role === 'assistant' ? 'MODEL' :
chat.role
if (chat.multimodals && chat.multimodals.length > 0 && chat.role === "user") {
let geminiParts: GeminiPart[] = [];
geminiParts.push({
text: chat.content,
});
for (const modal of chat.multimodals) {
if (
(modal.type === "image" && arg.modelInfo.flags.includes(LLMFlags.hasImageInput)) ||
(modal.type === "audio" && arg.modelInfo.flags.includes(LLMFlags.hasAudioInput)) ||
(modal.type === "video" && arg.modelInfo.flags.includes(LLMFlags.hasVideoInput))
) {
const dataurl = modal.base64;
const base64 = dataurl.split(",")[1];
const mediaType = dataurl.split(";")[0].split(":")[1];
geminiParts.push({
inlineData: {
mimeType: mediaType,
data: base64,
}
});
}
}
reformatedChat.push({
role: "USER",
parts: geminiParts,
});
} else if (prevChat?.role === qRole) {
if (reformatedChat[reformatedChat.length-1].parts[
reformatedChat[reformatedChat.length-1].parts.length-1
].inlineData) {
reformatedChat[reformatedChat.length-1].parts.push({
text: chat.content,
})
} else {
reformatedChat[reformatedChat.length-1].parts[
reformatedChat[reformatedChat.length-1].parts.length-1
].text += '\n' + chat.content
}
continue
}
else if(chat.role === 'system'){
if(prevChat?.role === 'USER'){
reformatedChat[reformatedChat.length-1].parts[0].text += '\nsystem:' + chat.content
}
else{
reformatedChat.push({
role: "USER",
parts: [{
text: chat.role + ':' + chat.content
}]
})
}
}
else if(chat.role === 'assistant' || chat.role === 'user'){
reformatedChat.push({
role: chat.role === 'user' ? 'USER' : 'MODEL',
parts: [{
text: chat.content
}]
})
}
else{
reformatedChat.push({
role: "USER",
parts: [{
text: chat.role + ':' + chat.content
}]
})
}
}
const uncensoredCatagory = [
{
"category": "HARM_CATEGORY_SEXUALLY_EXPLICIT",
"threshold": "BLOCK_NONE"
},
{
"category": "HARM_CATEGORY_HATE_SPEECH",
"threshold": "BLOCK_NONE"
},
{
"category": "HARM_CATEGORY_HARASSMENT",
"threshold": "BLOCK_NONE"
},
{
"category": "HARM_CATEGORY_DANGEROUS_CONTENT",
"threshold": "BLOCK_NONE"
},
{
"category": "HARM_CATEGORY_CIVIC_INTEGRITY",
"threshold": "BLOCK_NONE"
}
]
if(arg.modelInfo.flags.includes(LLMFlags.noCivilIntegrity)){
uncensoredCatagory.splice(4, 1)
}
if(arg.modelInfo.flags.includes(LLMFlags.geminiBlockOff)){
for(let i=0;i<uncensoredCatagory.length;i++){
uncensoredCatagory[i].threshold = "OFF"
}
}
let para:Parameter[] = ['temperature', 'top_p', 'top_k', 'presence_penalty', 'frequency_penalty']
para = para.filter((v) => {
return arg.modelInfo.parameters.includes(v)
})
const body = {
contents: reformatedChat,
generation_config: applyParameters({
"maxOutputTokens": maxTokens,
}, para, {
'top_p': "topP",
'top_k': "topK",
'presence_penalty': "presencePenalty",
'frequency_penalty': "frequencyPenalty"
}, arg.mode, {
ignoreTopKIfZero: true
}),
safetySettings: uncensoredCatagory,
systemInstruction: {
parts: [
{
"text": systemPrompt
}
]
},
}
let headers:{[key:string]:string} = {}
const PROJECT_ID=db.google.projectId
const REGION="us-central1"
console.log(arg.modelInfo)
async function generateToken(email:string,key:string){
key = key.replace('-----BEGIN PRIVATE KEY-----','').replace('-----END PRIVATE KEY-----','').replace(/\n/g, '').replace(/\r/g, '').trim()
const time = Math.floor(Date.now() / 1000);
const header = Buffer.from(JSON.stringify({
alg: "RS256",
typ: "JWT",
}))
const set = Buffer.from(JSON.stringify({
iss: email,
iat: time,
exp: time + 3600,
scope: "https://www.googleapis.com/auth/cloud-platform",
aud: "https://oauth2.googleapis.com/token",
})).toString('base64url');
const cryptokey = await crypto.subtle.importKey(
"pkcs8",
this.str2ab(key),
{
name: "RSASSA-PKCS1-v1_5",
hash: { name: "SHA-256" },
},
false,
["sign"]
);
const signature = Buffer.from(await crypto.subtle.sign(
"RSASSA-PKCS1-v1_5",
cryptokey,
Buffer.from(`${header}.${set}`)
)).toString('base64url');
const jwt = `${header}.${set}.${signature}`;
const response = await fetch("https://oauth2.googleapis.com/token", {
method: "POST",
body: `grant_type=urn%3Aietf%3Aparams%3Aoauth%3Agrant-type%3Ajwt-bearer&assertion=${jwt}`,
headers: {
"Content-Type": "application/x-www-form-urlencoded",
},
});
const data = await response.json();
const token = data.access_token;
const db2 = getDatabase()
db2.vertexAccessToken = token
db2.vertexAccessTokenExpires = Date.now() + 3500 * 1000
setDatabase(db2)
return token;
}
if(arg.modelInfo.format === LLMFormat.VertexAIGemini){
if(db.vertexAccessTokenExpires < Date.now()){
headers['Authorization'] = "Bearer " + generateToken(db.vertexClientEmail, db.vertexPrivateKey)
}
else{
headers['Authorization'] = "Bearer " + db.vertexAccessToken
}
}
if(db.jsonSchemaEnabled || arg.schema){
body.generation_config.response_mime_type = "application/json"
body.generation_config.response_schema = getGeneralJSONSchema(arg.schema, ['$schema','additionalProperties'])
console.log(body.generation_config.response_schema)
}
let url = ''
if(arg.customURL){
const u = new URL(arg.customURL)
u.searchParams.set('key', db.proxyKey)
url = u.toString()
}
else if(arg.modelInfo.format === LLMFormat.VertexAIGemini){
url =`https://${REGION}-aiplatform.googleapis.com/v1/projects/${PROJECT_ID}/locations/us-central1/publishers/google/models/${arg.modelInfo.internalID}:streamGenerateContent`
}
else if(arg.modelInfo.format === LLMFormat.GoogleCloud && arg.useStreaming){
url = `https://generativelanguage.googleapis.com/v1beta/models/${arg.modelInfo.internalID}:streamGenerateContent?key=${db.google.accessToken}`
}
else{
url = `https://generativelanguage.googleapis.com/v1beta/models/${arg.modelInfo.internalID}:generateContent?key=${db.google.accessToken}`
}
const fallBackGemini = async (originalError:string):Promise<requestDataResponse> => {
if(!db.antiServerOverloads){
return {
type: 'fail',
result: originalError,
failByServerError: true
}
}
if(arg?.abortSignal?.aborted){
return {
type: 'fail',
result: originalError,
failByServerError: true
}
}
if(arg.modelInfo.format === LLMFormat.VertexAIGemini){
return {
type: 'fail',
result: originalError,
failByServerError: true
}
}
try {
const OAIMessages:OpenAIChat[] = body.contents.map((v) => {
return {
role: v.role === 'USER' ? 'user' : 'assistant',
content: v.parts.map((v) => {
return v.text ?? ''
}).join('\n')
}
})
if(body?.systemInstruction?.parts?.[0]?.text){
OAIMessages.unshift({
role: 'system',
content: body.systemInstruction.parts[0].text
})
}
await sleep(2000)
const res = await fetch('https://generativelanguage.googleapis.com/v1beta/openai/chat/completions', {
body: JSON.stringify({
model: arg.modelInfo.internalID,
messages: OAIMessages,
max_tokens: maxTokens,
temperature: body.generation_config?.temperature,
top_p: body.generation_config?.topP,
presence_penalty: body.generation_config?.presencePenalty,
frequency_penalty: body.generation_config?.frequencyPenalty,
}),
method: 'POST',
headers: {
'Content-Type': 'application/json',
'Authorization': `Bearer ${db.google.accessToken}`
},
signal: arg.abortSignal
})
if(!res.ok){
return {
type: 'fail',
result: originalError,
failByServerError: true
}
}
if(arg?.abortSignal?.aborted){
return {
type: 'fail',
result: originalError
}
}
const d = await res.json()
if(d?.choices?.[0]?.message?.content){
return {
type: 'success',
result: d.choices[0].message.content
}
}
else{
return {
type: 'fail',
result: originalError,
failByServerError: true
}
}
} catch (error) {
return {
type: 'fail',
result: originalError,
failByServerError: true
}
}
}
if(arg.modelInfo.format === LLMFormat.GoogleCloud && arg.useStreaming){
headers['Content-Type'] = 'application/json'
const f = await fetchNative(url, {
headers: headers,
body: JSON.stringify(body),
method: 'POST',
chatId: arg.chatId,
})
if(f.status !== 200){
const text = await textifyReadableStream(f.body)
if(text.includes('RESOURCE_EXHAUSTED')){
return fallBackGemini(text)
}
return {
type: 'fail',
result: text
}
}
let fullResult:string = ''
const stream = new TransformStream<Uint8Array, StreamResponseChunk>( {
async transform(chunk, control) {
fullResult += new TextDecoder().decode(chunk)
try {
let reformatted = fullResult
if(reformatted.endsWith(',')){
reformatted = fullResult.slice(0, -1) + ']'
}
if(!reformatted.endsWith(']')){
reformatted = fullResult + ']'
}
const data = JSON.parse(reformatted)
let rDatas:string[] = ['']
for(const d of data){
const parts = d.candidates[0].content?.parts
for(let i=0;i<parts.length;i++){
const part = parts[i]
if(i > 0){
rDatas.push('')
}
rDatas[rDatas.length-1] += part.text
}
}
if(arg.extractJson && (db.jsonSchemaEnabled || arg.schema)){
for(let i=0;i<rDatas.length;i++){
const extracted = extractJSON(rDatas[i], arg.extractJson)
rDatas[i] = extracted
}
}
if(rDatas.length > 1){
const thought = rDatas.splice(rDatas.length-2, 1)[0]
rDatas[rDatas.length-1] = `<Thoughts>${thought}</Thoughts>\n\n${rDatas.join('\n\n')}`
}
control.enqueue({
'0': rDatas[rDatas.length-1],
})
} catch (error) {
console.log(error)
}
}
},)
return {
type: 'streaming',
result: f.body.pipeThrough(stream)
}
}
const res = await globalFetch(url, {
headers: headers,
body: body,
chatId: arg.chatId,
abortSignal: arg.abortSignal,
})
if(!res.ok){
const text = JSON.stringify(res.data)
if(text.includes('RESOURCE_EXHAUSTED')){
return fallBackGemini(text)
}
return {
type: 'fail',
result: `${JSON.stringify(res.data)}`
}
}
let rDatas:string[] = ['']
const processDataItem = (data:any) => {
const parts = data?.candidates?.[0]?.content?.parts
if(parts){
for(let i=0;i<parts.length;i++){
const part = parts[i]
if(i > 0){
rDatas.push('')
}
rDatas[rDatas.length-1] += part.text
}
}
if(data?.errors){
return {
type: 'fail',
result: `${JSON.stringify(data.errors)}`
}
}
else{
return {
type: 'fail',
result: `${JSON.stringify(data)}`
}
}
}
// traverse responded data if it contains multipart contents
if (typeof (res.data)[Symbol.iterator] === 'function') {
for(const data of res.data){
processDataItem(data)
}
} else {
processDataItem(res.data)
}
if(arg.extractJson && (db.jsonSchemaEnabled || arg.schema)){
for(let i=0;i<rDatas.length;i++){
const extracted = extractJSON(rDatas[i], arg.extractJson)
rDatas[i] = extracted
}
}
if(rDatas.length > 1){
const thought = rDatas.splice(rDatas.length-2, 1)[0]
rDatas[rDatas.length-1] = `<Thoughts>${thought}</Thoughts>\n\n${rDatas.join('\n\n')}`
}
return {
type: 'success',
result: rDatas[rDatas.length-1]
}
}
async function requestKobold(arg:RequestDataArgumentExtended):Promise<requestDataResponse> {
const formated = arg.formated
const db = getDatabase()
const maxTokens = arg.maxTokens
const abortSignal = arg.abortSignal
const prompt = applyChatTemplate(formated)
const url = new URL(db.koboldURL)
if(url.pathname.length < 3){
url.pathname = 'api/v1/generate'
}
const body = applyParameters({
"prompt": prompt,
max_length: maxTokens,
max_context_length: db.maxContext,
n: 1
}, [
'temperature',
'top_p',
'repetition_penalty',
'top_k',
'top_a'
], {
'repetition_penalty': 'rep_pen'
}, arg.mode) as KoboldGenerationInputSchema
const da = await globalFetch(url.toString(), {
method: "POST",
body: body,
headers: {
"content-type": "application/json",
},
abortSignal,
chatId: arg.chatId
})
if(!da.ok){
return {
type: "fail",
result: da.data,
noRetry: true
}
}
const data = da.data
return {
type: 'success',
result: data.results[0].text
}
}
async function requestNovelList(arg:RequestDataArgumentExtended):Promise<requestDataResponse> {
const formated = arg.formated
const db = getDatabase()
const maxTokens = arg.maxTokens
const temperature = arg.temperature
const biasString = arg.biasString
const currentChar = getCurrentCharacter()
const aiModel = arg.aiModel
const auth_key = db.novellistAPI;
const api_server_url = 'https://api.tringpt.com/';
const logit_bias:string[] = []
const logit_bias_values:string[] = []
for(let i=0;i<biasString.length;i++){
const bia = biasString[i]
logit_bias.push(bia[0])
logit_bias_values.push(bia[1].toString())
}
const headers = {
'Authorization': `Bearer ${auth_key}`,
'Content-Type': 'application/json'
};
const send_body = {
text: stringlizeAINChat(formated, currentChar?.name ?? '', arg.continue),
length: maxTokens,
temperature: temperature,
top_p: db.ainconfig.top_p,
top_k: db.ainconfig.top_k,
rep_pen: db.ainconfig.rep_pen,
top_a: db.ainconfig.top_a,
rep_pen_slope: db.ainconfig.rep_pen_slope,
rep_pen_range: db.ainconfig.rep_pen_range,
typical_p: db.ainconfig.typical_p,
badwords: db.ainconfig.badwords,
model: aiModel === 'novellist_damsel' ? 'damsel' : 'supertrin',
stoptokens: ["「"].join("<<|>>") + db.ainconfig.stoptokens,
logit_bias: (logit_bias.length > 0) ? logit_bias.join("<<|>>") : undefined,
logit_bias_values: (logit_bias_values.length > 0) ? logit_bias_values.join("|") : undefined,
};
const response = await globalFetch(arg.customURL ?? api_server_url + '/api', {
method: 'POST',
headers: headers,
body: send_body,
chatId: arg.chatId
});
if(!response.ok){
return {
type: 'fail',
result: response.data
}
}
if(response.data.error){
return {
'type': 'fail',
'result': `${response.data.error.replace("token", "api key")}`
}
}
const result = response.data.data[0];
const unstr = unstringlizeAIN(result, formated, currentChar?.name ?? '')
return {
'type': 'multiline',
'result': unstr
}
}
async function requestOllama(arg:RequestDataArgumentExtended):Promise<requestDataResponse> {
const formated = arg.formated
const db = getDatabase()
const ollama = new Ollama({host: db.ollamaURL})
const response = await ollama.chat({
model: db.ollamaModel,
messages: formated.map((v) => {
return {
role: v.role,
content: v.content
}
}).filter((v) => {
return v.role === 'assistant' || v.role === 'user' || v.role === 'system'
}),
stream: true
})
const readableStream = new ReadableStream<StreamResponseChunk>({
async start(controller){
for await(const chunk of response){
controller.enqueue({
"0": chunk.message.content
})
}
controller.close()
}
})
return {
type: 'streaming',
result: readableStream
}
}
async function requestCohere(arg:RequestDataArgumentExtended):Promise<requestDataResponse> {
const formated = arg.formated
const db = getDatabase()
const aiModel = arg.aiModel
let lastChatPrompt = ''
let preamble = ''
let lastChat = formated[formated.length-1]
if(lastChat.role === 'user'){
lastChatPrompt = lastChat.content
formated.pop()
}
else{
while(lastChat.role !== 'user'){
lastChat = formated.pop()
if(!lastChat){
return {
type: 'fail',
result: 'Cohere requires a user message to generate a response'
}
}
lastChatPrompt = (lastChat.role === 'user' ? '' : `${lastChat.role}: `) + '\n' + lastChat.content + lastChatPrompt
}
}
const firstChat = formated[0]
if(firstChat.role === 'system'){
preamble = firstChat.content
formated.shift()
}
//reformat chat
let body = applyParameters({
message: lastChatPrompt,
chat_history: formated.map((v) => {
if(v.role === 'assistant'){
return {
role: 'CHATBOT',
message: v.content
}
}
if(v.role === 'system'){
return {
role: 'SYSTEM',
message: v.content
}
}
if(v.role === 'user'){
return {
role: 'USER',
message: v.content
}
}
return null
}).filter((v) => v !== null).filter((v) => {
return v.message
}),
}, [
'temperature', 'top_k', 'top_p', 'presence_penalty', 'frequency_penalty'
], {
'top_k': 'k',
'top_p': 'p',
}, arg.mode)
if(aiModel !== 'cohere-command-r-03-2024' && aiModel !== 'cohere-command-r-plus-04-2024'){
body.safety_mode = "NONE"
}
if(preamble){
if(body.chat_history.length > 0){
// @ts-ignore
body.preamble = preamble
}
else{
body.message = `system: ${preamble}`
}
}
console.log(body)
const res = await globalFetch(arg.customURL ?? 'https://api.cohere.com/v1/chat', {
method: "POST",
headers: {
"Authorization": "Bearer " + db.cohereAPIKey,
"Content-Type": "application/json"
},
body: body
})
if(!res.ok){
return {
type: 'fail',
result: JSON.stringify(res.data)
}
}
const result = res.data.text
if(!result){
return {
type: 'fail',
result: JSON.stringify(res.data)
}
}
return {
type: 'success',
result: result
}
}
async function requestClaude(arg:RequestDataArgumentExtended):Promise<requestDataResponse> {
const formated = arg.formated
const db = getDatabase()
const aiModel = arg.aiModel
const useStreaming = arg.useStreaming
let replacerURL = (aiModel === 'reverse_proxy') ? (arg.customURL) : ('https://api.anthropic.com/v1/messages')
let apiKey = (aiModel === 'reverse_proxy') ? db.proxyKey : db.claudeAPIKey
const maxTokens = arg.maxTokens
if(aiModel === 'reverse_proxy' && db.autofillRequestUrl){
if(replacerURL.endsWith('v1')){
replacerURL += '/messages'
}
else if(replacerURL.endsWith('v1/')){
replacerURL += 'messages'
}
else if(!(replacerURL.endsWith('messages') || replacerURL.endsWith('messages/'))){
if(replacerURL.endsWith('/')){
replacerURL += 'v1/messages'
}
else{
replacerURL += '/v1/messages'
}
}
}
interface Claude3TextBlock {
type: 'text',
text: string,
cache_control?: {"type": "ephemeral"}
}
interface Claude3ImageBlock {
type: 'image',
source: {
type: 'base64'
media_type: string,
data: string
}
cache_control?: {"type": "ephemeral"}
}
type Claude3ContentBlock = Claude3TextBlock|Claude3ImageBlock
interface Claude3Chat {
role: 'user'|'assistant'
content: Claude3ContentBlock[]
}
interface Claude3ExtendedChat {
role: 'user'|'assistant'
content: Claude3ContentBlock[]|string
}
let claudeChat: Claude3Chat[] = []
let systemPrompt:string = ''
const addClaudeChat = (chat:{
role: 'user'|'assistant'
content: string
}, multimodals?:MultiModal[]) => {
if(claudeChat.length > 0 && claudeChat[claudeChat.length-1].role === chat.role){
let content = claudeChat[claudeChat.length-1].content
if(multimodals && multimodals.length > 0 && !Array.isArray(content)){
content = [{
type: 'text',
text: content
}]
}
if(Array.isArray(content)){
let lastContent = content[content.length-1]
if( lastContent?.type === 'text'){
lastContent.text += "\n\n" + chat.content
content[content.length-1] = lastContent
}
else{
content.push({
type: 'text',
text: chat.content
})
}
if(multimodals && multimodals.length > 0){
for(const modal of multimodals){
if(modal.type === 'image'){
const dataurl = modal.base64
const base64 = dataurl.split(',')[1]
const mediaType = dataurl.split(';')[0].split(':')[1]
content.unshift({
type: 'image',
source: {
type: 'base64',
media_type: mediaType,
data: base64
}
})
}
}
}
}
claudeChat[claudeChat.length-1].content = content
}
else{
let formatedChat:Claude3Chat = {
role: chat.role,
content: [{
type: 'text',
text: chat.content
}]
}
if(multimodals && multimodals.length > 0){
formatedChat.content = [{
type: 'text',
text: chat.content
}]
for(const modal of multimodals){
if(modal.type === 'image'){
const dataurl = modal.base64
const base64 = dataurl.split(',')[1]
const mediaType = dataurl.split(';')[0].split(':')[1]
formatedChat.content.unshift({
type: 'image',
source: {
type: 'base64',
media_type: mediaType,
data: base64
}
})
}
}
}
claudeChat.push(formatedChat)
}
}
for(const chat of formated){
switch(chat.role){
case 'user':{
addClaudeChat({
role: 'user',
content: chat.content
}, chat.multimodals)
break
}
case 'assistant':{
addClaudeChat({
role: 'assistant',
content: chat.content
}, chat.multimodals)
break
}
case 'system':{
if(claudeChat.length === 0){
systemPrompt += '\n\n' + chat.content
}
else{
addClaudeChat({
role: 'user',
content: "System: " + chat.content
})
}
break
}
case 'function':{
//ignore function for now
break
}
}
}
if(claudeChat.length === 0 && systemPrompt === ''){
return {
type: 'fail',
result: 'No input'
}
}
if(claudeChat.length === 0 && systemPrompt !== ''){
claudeChat.push({
role: 'user',
content: [{
type: 'text',
text: 'Start'
}]
})
systemPrompt = ''
}
if(claudeChat[0].role !== 'user'){
claudeChat.unshift({
role: 'user',
content: [{
type: 'text',
text: 'Start'
}]
})
}
if(db.claudeCachingExperimental){
for(let i = 0;i<4;i++){
const ind = claudeChat.findLastIndex((v) => {
if(v.role !== 'user'){
return false
}
if(v.content.length === 0){
return false
}
if(v.content[0].cache_control){ // if it already has cache control, skip
return false
}
return true
})
console.log(ind)
if(ind === -1){
break
}
claudeChat[ind].content[0].cache_control = {
type: 'ephemeral'
}
}
}
let finalChat:Claude3ExtendedChat[] = claudeChat
if(aiModel === 'reverse_proxy'){
finalChat = claudeChat.map((v) => {
if(v.content.length > 0 && v.content[0].type === 'text'){
return {
role: v.role,
content: v.content[0].text
}
}
})
}
console.log(arg.modelInfo.parameters)
let body = applyParameters({
model: arg.modelInfo.internalID,
messages: finalChat,
system: systemPrompt.trim(),
max_tokens: maxTokens,
stream: useStreaming ?? false
}, arg.modelInfo.parameters, {
'thinking_tokens': 'thinking.budget_tokens'
}, arg.mode)
if(body?.thinking?.budget_tokens === 0){
delete body.thinking
}
else if(body?.thinking?.budget_tokens && body?.thinking?.budget_tokens > 0){
body.thinking.type = 'enabled'
}
if(systemPrompt === ''){
delete body.system
}
const bedrock = arg.modelInfo.format === LLMFormat.AWSBedrockClaude
if(bedrock && aiModel !== 'reverse_proxy'){
function getCredentialParts(key:string) {
const [accessKeyId, secretAccessKey, region] = key.split(":");
if (!accessKeyId || !secretAccessKey || !region) {
throw new Error("The key assigned to this request is invalid.");
}
return { accessKeyId, secretAccessKey, region };
}
const { accessKeyId, secretAccessKey, region } = getCredentialParts(apiKey);
const AMZ_HOST = "bedrock-runtime.%REGION%.amazonaws.com";
const host = AMZ_HOST.replace("%REGION%", region);
const stream = false; // todo?
const awsModel = "us." + arg.modelInfo.internalID;
const url = `https://${host}/model/${awsModel}/invoke${stream ? "-with-response-stream" : ""}`
const params = {
messages : claudeChat,
system: systemPrompt.trim(),
max_tokens: maxTokens,
// stop_sequences: null,
temperature: arg.temperature,
top_p: db.top_p,
top_k: db.top_k,
anthropic_version: "bedrock-2023-05-31",
}
const rq = new HttpRequest({
method: "POST",
protocol: "https:",
hostname: host,
path: `/model/${awsModel}/invoke${stream ? "-with-response-stream" : ""}`,
headers: {
["Host"]: host,
["Content-Type"]: "application/json",
["accept"]: "application/json",
},
body: JSON.stringify(params),
});
const signer = new SignatureV4({
sha256: Sha256,
credentials: { accessKeyId, secretAccessKey },
region,
service: "bedrock",
});
const signed = await signer.sign(rq);
const res = await globalFetch(url, {
method: "POST",
body: params,
headers: signed.headers,
plainFetchForce: true,
chatId: arg.chatId
})
if(!res.ok){
return {
type: 'fail',
result: JSON.stringify(res.data)
}
}
if(res.data.error){
return {
type: 'fail',
result: JSON.stringify(res.data.error)
}
}
return {
type: 'success',
result: res.data.content[0].text
}
}
let headers:{
[key:string]:string
} = {
"Content-Type": "application/json",
"x-api-key": apiKey,
"anthropic-version": "2023-06-01",
"accept": "application/json",
}
let betas:string[] = []
if(db.claudeCachingExperimental){
betas.push('prompt-caching-2024-07-31')
}
if(body.max_tokens > 8192){
betas.push('output-128k-2025-02-19')
}
if(betas.length > 0){
headers['anthropic-beta'] = betas.join(',')
}
if(db.usePlainFetch){
headers['anthropic-dangerous-direct-browser-access'] = 'true'
}
if(useStreaming){
const res = await fetchNative(replacerURL, {
body: JSON.stringify(body),
headers: headers,
method: "POST",
chatId: arg.chatId
})
if(res.status !== 200){
return {
type: 'fail',
result: await textifyReadableStream(res.body)
}
}
let breakError = ''
let thinking = false
const stream = new ReadableStream<StreamResponseChunk>({
async start(controller){
let text = ''
let reader = res.body.getReader()
let parserData = ''
const decoder = new TextDecoder()
const parseEvent = (async (e:string) => {
try {
const parsedData = JSON.parse(e)
if(parsedData?.type === 'content_block_delta'){
if(parsedData?.delta?.type === 'text' || parsedData.delta?.type === 'text_delta'){
if(thinking){
text += "</Thoughts>\n\n"
thinking = false
}
text += parsedData.delta?.text ?? ''
}
if(parsedData?.delta?.type === 'thinking' || parsedData.delta?.type === 'thinking_delta'){
if(!thinking){
text += "<Thoughts>\n"
thinking = true
}
text += parsedData.delta?.thinking ?? ''
}
if(parsedData?.delta?.type === 'redacted_thinking'){
if(!thinking){
text += "<Thoughts>\n"
thinking = true
}
text += '\n{{redacted_thinking}}\n'
}
}
if(parsedData?.type === 'error'){
const errormsg:string = parsedData?.error?.message
if(errormsg && errormsg.toLocaleLowerCase().includes('overload') && db.antiServerOverloads){
// console.log('Overload detected, retrying...')
controller.enqueue({
"0": "Overload detected, retrying..."
})
return 'overload'
}
text += "Error:" + parsedData?.error?.message
}
}
catch (error) {
}
})
let breakWhile = false
let i = 0;
let prevText = ''
while(true){
try {
if(arg?.abortSignal?.aborted || breakWhile){
break
}
const {done, value} = await reader.read()
if(done){
break
}
parserData += (decoder.decode(value))
let parts = parserData.split('\n')
for(;i<parts.length-1;i++){
prevText = text
if(parts?.[i]?.startsWith('data: ')){
const d = await parseEvent(parts[i].slice(6))
if(d === 'overload'){
parserData = ''
prevText = ''
text = ''
reader.cancel()
const res = await fetchNative(replacerURL, {
body: JSON.stringify(body),
headers: headers,
method: "POST",
chatId: arg.chatId
})
if(res.status !== 200){
controller.enqueue({
"0": await textifyReadableStream(res.body)
})
breakWhile = true
break
}
reader = res.body.getReader()
break
}
}
}
i--;
text = prevText
controller.enqueue({
"0": text
})
} catch (error) {
await sleep(1)
}
}
controller.close()
},
cancel(){
}
})
return {
type: 'streaming',
result: stream
}
}
const res = await globalFetch(replacerURL, {
body: body,
headers: headers,
method: "POST",
chatId: arg.chatId
})
if(!res.ok){
const stringlified = JSON.stringify(res.data)
return {
type: 'fail',
result: stringlified,
failByServerError: stringlified?.toLocaleLowerCase()?.includes('overload')
}
}
if(res.data.error){
const stringlified = JSON.stringify(res.data.error)
return {
type: 'fail',
result: stringlified,
failByServerError: stringlified?.toLocaleLowerCase()?.includes('overload')
}
}
const contents = res?.data?.content
if(!contents || contents.length === 0){
return {
type: 'fail',
result: JSON.stringify(res.data)
}
}
let resText = ''
let thinking = false
for(const content of contents){
if(content.type === 'text'){
if(thinking){
resText += "</Thoughts>\n\n"
thinking = false
}
resText += content.text
}
if(content.type === 'thinking'){
if(!thinking){
resText += "<Thoughts>\n"
thinking = true
}
resText += content.thinking ?? ''
}
if(content.type === 'redacted_thinking'){
if(!thinking){
resText += "<Thoughts>\n"
thinking = true
}
resText += '\n{{redacted_thinking}}\n'
}
}
if(arg.extractJson && db.jsonSchemaEnabled){
return {
type: 'success',
result: extractJSON(resText, db.jsonSchema)
}
}
return {
type: 'success',
result: resText
}
}
async function requestHorde(arg:RequestDataArgumentExtended):Promise<requestDataResponse> {
const formated = arg.formated
const db = getDatabase()
const aiModel = arg.aiModel
const currentChar = getCurrentCharacter()
const abortSignal = arg.abortSignal
const prompt = applyChatTemplate(formated)
const realModel = aiModel.split(":::")[1]
const argument = {
"prompt": prompt,
"params": {
"n": 1,
"max_context_length": db.maxContext + 100,
"max_length": db.maxResponse,
"singleline": false,
"temperature": db.temperature / 100,
"top_k": db.top_k,
"top_p": db.top_p,
},
"trusted_workers": false,
"workerslow_workers": true,
"_blacklist": false,
"dry_run": false,
"models": [realModel, realModel.trim(), ' ' + realModel, realModel + ' ']
}
if(realModel === 'auto'){
delete argument.models
}
let apiKey = '0000000000'
if(db.hordeConfig.apiKey.length > 2){
apiKey = db.hordeConfig.apiKey
}
const da = await fetch("https://stablehorde.net/api/v2/generate/text/async", {
body: JSON.stringify(argument),
method: "POST",
headers: {
"content-type": "application/json",
"apikey": apiKey
},
signal: abortSignal
})
if(da.status !== 202){
return {
type: "fail",
result: await da.text()
}
}
const json:{
id:string,
kudos:number,
message:string
} = await da.json()
let warnMessage = ""
if(json.message){
warnMessage = "with " + json.message
}
while(true){
await sleep(2000)
const data = await (await fetch("https://stablehorde.net/api/v2/generate/text/status/" + json.id)).json()
if(!data.is_possible){
fetch("https://stablehorde.net/api/v2/generate/text/status/" + json.id, {
method: "DELETE"
})
return {
type: 'fail',
result: "Response not possible" + warnMessage,
noRetry: true
}
}
if(data.done && Array.isArray(data.generations) && data.generations.length > 0){
const generations:{text:string}[] = data.generations
if(generations && generations.length > 0){
return {
type: "success",
result: unstringlizeChat(generations[0].text, formated, currentChar?.name ?? '')
}
}
return {
type: 'fail',
result: "No Generations when done",
noRetry: true
}
}
}
}
async function requestWebLLM(arg:RequestDataArgumentExtended):Promise<requestDataResponse> {
const formated = arg.formated
const db = getDatabase()
const aiModel = arg.aiModel
const currentChar = getCurrentCharacter()
const maxTokens = arg.maxTokens
const temperature = arg.temperature
const realModel = aiModel.split(":::")[1]
const prompt = applyChatTemplate(formated)
const v = await runTransformers(prompt, realModel, {
temperature: temperature,
max_new_tokens: maxTokens,
top_k: db.ooba.top_k,
top_p: db.ooba.top_p,
repetition_penalty: db.ooba.repetition_penalty,
typical_p: db.ooba.typical_p,
} as any)
return {
type: 'success',
result: unstringlizeChat(v.generated_text as string, formated, currentChar?.name ?? '')
}
}
export interface KoboldSamplerSettingsSchema {
rep_pen?: number;
rep_pen_range?: number;
rep_pen_slope?: number;
top_k?: number;
top_a?: number;
top_p?: number;
tfs?: number;
typical?: number;
temperature?: number;
}
export interface KoboldGenerationInputSchema extends KoboldSamplerSettingsSchema {
prompt: string;
use_memory?: boolean;
use_story?: boolean;
use_authors_note?: boolean;
use_world_info?: boolean;
use_userscripts?: boolean;
soft_prompt?: string;
max_length?: number;
max_context_length?: number;
n: number;
disable_output_formatting?: boolean;
frmttriminc?: boolean;
frmtrmblln?: boolean;
frmtrmspch?: boolean;
singleline?: boolean;
disable_input_formatting?: boolean;
frmtadsnsp?: boolean;
quiet?: boolean;
sampler_order?: number[];
sampler_seed?: number;
sampler_full_determinism?: boolean;
}