Flowise/packages/components/nodes/chatmodels/ChatGoogleGenerativeAI/ChatGoogleGenerativeAI.ts

195 lines
7.6 KiB
TypeScript

import { ICommonObject, INode, INodeData, INodeParams } from '../../../src/Interface'
import { convertMultiOptionsToStringArray, getBaseClasses, getCredentialData, getCredentialParam } from '../../../src/utils'
import { BaseCache } from 'langchain/schema'
import { ChatGoogleGenerativeAI, GoogleGenerativeAIChatInput } from '@langchain/google-genai'
import { HarmBlockThreshold, HarmCategory } from '@google/generative-ai'
import type { SafetySetting } from '@google/generative-ai'
class GoogleGenerativeAI_ChatModels implements INode {
label: string
name: string
version: number
type: string
icon: string
category: string
description: string
baseClasses: string[]
credential: INodeParams
inputs: INodeParams[]
constructor() {
this.label = 'ChatGoogleGenerativeAI'
this.name = 'chatGoogleGenerativeAI'
this.version = 1.0
this.type = 'ChatGoogleGenerativeAI'
this.icon = 'GoogleGemini.svg'
this.category = 'Chat Models'
this.description = 'Wrapper around Google Gemini large language models that use the Chat endpoint'
this.baseClasses = [this.type, ...getBaseClasses(ChatGoogleGenerativeAI)]
this.credential = {
label: 'Connect Credential',
name: 'credential',
type: 'credential',
credentialNames: ['googleGenerativeAI'],
optional: false,
description: 'Google Generative AI credential.'
}
this.inputs = [
{
label: 'Cache',
name: 'cache',
type: 'BaseCache',
optional: true
},
{
label: 'Model Name',
name: 'modelName',
type: 'options',
options: [
{
label: 'gemini-pro',
name: 'gemini-pro'
}
],
default: 'gemini-pro'
},
{
label: 'Temperature',
name: 'temperature',
type: 'number',
step: 0.1,
default: 0.9,
optional: true
},
{
label: 'Max Output Tokens',
name: 'maxOutputTokens',
type: 'number',
step: 1,
optional: true,
additionalParams: true
},
{
label: 'Top Probability',
name: 'topP',
type: 'number',
step: 0.1,
optional: true,
additionalParams: true
},
{
label: 'Top Next Highest Probability Tokens',
name: 'topK',
type: 'number',
description: `Decode using top-k sampling: consider the set of top_k most probable tokens. Must be positive`,
step: 1,
optional: true,
additionalParams: true
},
{
label: 'Harm Category',
name: 'harmCategory',
type: 'multiOptions',
description:
'Refer to <a target="_blank" href="https://cloud.google.com/vertex-ai/docs/generative-ai/multimodal/configure-safety-attributes#safety_attribute_definitions">official guide</a> on how to use Harm Category',
options: [
{
label: 'Dangerous',
name: HarmCategory.HARM_CATEGORY_DANGEROUS_CONTENT
},
{
label: 'Harassment',
name: HarmCategory.HARM_CATEGORY_HARASSMENT
},
{
label: 'Hate Speech',
name: HarmCategory.HARM_CATEGORY_HATE_SPEECH
},
{
label: 'Sexually Explicit',
name: HarmCategory.HARM_CATEGORY_SEXUALLY_EXPLICIT
}
],
optional: true,
additionalParams: true
},
{
label: 'Harm Block Threshold',
name: 'harmBlockThreshold',
type: 'multiOptions',
description:
'Refer to <a target="_blank" href="https://cloud.google.com/vertex-ai/docs/generative-ai/multimodal/configure-safety-attributes#safety_setting_thresholds">official guide</a> on how to use Harm Block Threshold',
options: [
{
label: 'Low and Above',
name: HarmBlockThreshold.BLOCK_LOW_AND_ABOVE
},
{
label: 'Medium and Above',
name: HarmBlockThreshold.BLOCK_MEDIUM_AND_ABOVE
},
{
label: 'None',
name: HarmBlockThreshold.BLOCK_NONE
},
{
label: 'Only High',
name: HarmBlockThreshold.BLOCK_ONLY_HIGH
},
{
label: 'Threshold Unspecified',
name: HarmBlockThreshold.HARM_BLOCK_THRESHOLD_UNSPECIFIED
}
],
optional: true,
additionalParams: true
}
]
}
async init(nodeData: INodeData, _: string, options: ICommonObject): Promise<any> {
const credentialData = await getCredentialData(nodeData.credential ?? '', options)
const apiKey = getCredentialParam('googleGenerativeAPIKey', credentialData, nodeData)
const temperature = nodeData.inputs?.temperature as string
const modelName = nodeData.inputs?.modelName as string
const maxOutputTokens = nodeData.inputs?.maxOutputTokens as string
const topP = nodeData.inputs?.topP as string
const topK = nodeData.inputs?.topK as string
const harmCategory = nodeData.inputs?.harmCategory as string
const harmBlockThreshold = nodeData.inputs?.harmBlockThreshold as string
const cache = nodeData.inputs?.cache as BaseCache
const obj: Partial<GoogleGenerativeAIChatInput> = {
apiKey: apiKey,
modelName: modelName,
maxOutputTokens: 2048
}
if (maxOutputTokens) obj.maxOutputTokens = parseInt(maxOutputTokens, 10)
const model = new ChatGoogleGenerativeAI(obj)
if (topP) model.topP = parseFloat(topP)
if (topK) model.topK = parseFloat(topK)
if (cache) model.cache = cache
if (temperature) model.temperature = parseFloat(temperature)
// Safety Settings
let harmCategories: string[] = convertMultiOptionsToStringArray(harmCategory)
let harmBlockThresholds: string[] = convertMultiOptionsToStringArray(harmBlockThreshold)
if (harmCategories.length != harmBlockThresholds.length)
throw new Error(`Harm Category & Harm Block Threshold are not the same length`)
const safetySettings: SafetySetting[] = harmCategories.map((harmCategory, index) => {
return {
category: harmCategory as HarmCategory,
threshold: harmBlockThresholds[index] as HarmBlockThreshold
}
})
if (safetySettings.length > 0) model.safetySettings = safetySettings
return model
}
}
module.exports = { nodeClass: GoogleGenerativeAI_ChatModels }