Chore/Remove Redundant Check for Image Processing (#5980)

HenryHengZJ · web-flow · commit 378d3ec3bd2d · 2026-03-16T22:35:57.000+08:00
get rid of redundant check for model during image processing
diff --git a/packages/components/nodes/agents/ConversationalAgent/ConversationalAgent.ts b/packages/components/nodes/agents/ConversationalAgent/ConversationalAgent.ts
@@ -10,16 +10,7 @@ import { RunnableSequence } from '@langchain/core/runnables'
 import { ChatConversationalAgent } from '@langchain/classic/agents'
 import { getBaseClasses, transformBracesWithColon } from '../../../src/utils'
 import { ConsoleCallbackHandler, CustomChainHandler, additionalCallbacks } from '../../../src/handler'
-import {
-    IVisionChatModal,
-    FlowiseMemory,
-    ICommonObject,
-    INode,
-    INodeData,
-    INodeParams,
-    IUsedTool,
-    IServerSideEventStreamer
-} from '../../../src/Interface'
+import { FlowiseMemory, ICommonObject, INode, INodeData, INodeParams, IUsedTool, IServerSideEventStreamer } from '../../../src/Interface'
 import { AgentExecutor } from '../../../src/agents'
 import { addImagesToMessages, llmSupportsVision } from '../../../src/multiModalUtils'
 import { checkInputs, Moderation, streamResponse } from '../../moderation/Moderation'
@@ -241,12 +232,9 @@ const prepareAgent = async (
     })
 
     if (llmSupportsVision(model)) {
-        const visionChatModel = model as IVisionChatModal
         const messageContent = await addImagesToMessages(nodeData, options, model.multiModalOption)
 
         if (messageContent?.length) {
-            visionChatModel.setVisionModel()
-
             // Pop the `agent_scratchpad` MessagePlaceHolder
             let messagePlaceholder = prompt.promptMessages.pop() as MessagesPlaceholder
             if (prompt.promptMessages.at(-1) instanceof HumanMessagePromptTemplate) {
@@ -264,8 +252,6 @@ const prepareAgent = async (
 
             // Add the `agent_scratchpad` MessagePlaceHolder back
             prompt.promptMessages.push(messagePlaceholder)
-        } else {
-            visionChatModel.revertToOriginalModel()
         }
     }
 
diff --git a/packages/components/nodes/agents/ConversationalRetrievalToolAgent/ConversationalRetrievalToolAgent.ts b/packages/components/nodes/agents/ConversationalRetrievalToolAgent/ConversationalRetrievalToolAgent.ts
@@ -13,16 +13,7 @@ import {
     createTextOnlyOutputParser
 } from '../../../src/utils'
 import { type ToolsAgentStep } from '@langchain/classic/agents/openai/output_parser'
-import {
-    FlowiseMemory,
-    ICommonObject,
-    INode,
-    INodeData,
-    INodeParams,
-    IServerSideEventStreamer,
-    IUsedTool,
-    IVisionChatModal
-} from '../../../src/Interface'
+import { FlowiseMemory, ICommonObject, INode, INodeData, INodeParams, IServerSideEventStreamer, IUsedTool } from '../../../src/Interface'
 import { ConsoleCallbackHandler, CustomChainHandler, additionalCallbacks } from '../../../src/handler'
 import { AgentExecutor, ToolCallingAgentOutputParser } from '../../../src/agents'
 import { Moderation, checkInputs, streamResponse } from '../../moderation/Moderation'
@@ -279,12 +270,9 @@ const prepareAgent = async (
     ])
 
     if (llmSupportsVision(model)) {
-        const visionChatModel = model as IVisionChatModal
         const messageContent = await addImagesToMessages(nodeData, options, model.multiModalOption)
 
         if (messageContent?.length) {
-            visionChatModel.setVisionModel()
-
             // Pop the `agent_scratchpad` MessagePlaceHolder
             let messagePlaceholder = prompt.promptMessages.pop() as MessagesPlaceholder
             if (prompt.promptMessages.at(-1) instanceof HumanMessagePromptTemplate) {
@@ -302,8 +290,6 @@ const prepareAgent = async (
 
             // Add the `agent_scratchpad` MessagePlaceHolder back
             prompt.promptMessages.push(messagePlaceholder)
-        } else {
-            visionChatModel.revertToOriginalModel()
         }
     }
 
diff --git a/packages/components/nodes/agents/ReActAgentChat/ReActAgentChat.ts b/packages/components/nodes/agents/ReActAgentChat/ReActAgentChat.ts
@@ -6,7 +6,7 @@ import type { PromptTemplate } from '@langchain/core/prompts'
 import { BaseChatModel } from '@langchain/core/language_models/chat_models'
 import { pull } from 'langchain/hub'
 import { additionalCallbacks } from '../../../src/handler'
-import { IVisionChatModal, FlowiseMemory, ICommonObject, IMessage, INode, INodeData, INodeParams } from '../../../src/Interface'
+import { FlowiseMemory, ICommonObject, IMessage, INode, INodeData, INodeParams } from '../../../src/Interface'
 import { getBaseClasses } from '../../../src/utils'
 import { createReactAgent } from '../../../src/agents'
 import { addImagesToMessages, llmSupportsVision } from '../../../src/multiModalUtils'
@@ -105,14 +105,10 @@ class ReActAgentChat_Agents implements INode {
         let chatPromptTemplate = undefined
 
         if (llmSupportsVision(model)) {
-            const visionChatModel = model as IVisionChatModal
             const messageContent = await addImagesToMessages(nodeData, options, model.multiModalOption)
 
             if (messageContent?.length) {
-                // Change model to vision supported
-                visionChatModel.setVisionModel()
                 const oldTemplate = prompt.template as string
-
                 const msg = HumanMessagePromptTemplate.fromTemplate([
                     ...messageContent,
                     {
@@ -121,9 +117,6 @@ class ReActAgentChat_Agents implements INode {
                 ])
                 msg.inputVariables = prompt.inputVariables
                 chatPromptTemplate = ChatPromptTemplate.fromMessages([msg])
-            } else {
-                // revert to previous values if image upload is empty
-                visionChatModel.revertToOriginalModel()
             }
         }
 
diff --git a/packages/components/nodes/agents/ToolAgent/ToolAgent.ts b/packages/components/nodes/agents/ToolAgent/ToolAgent.ts
@@ -14,16 +14,7 @@ import {
     removeInvalidImageMarkdown,
     transformBracesWithColon
 } from '../../../src/utils'
-import {
-    FlowiseMemory,
-    ICommonObject,
-    INode,
-    INodeData,
-    INodeParams,
-    IServerSideEventStreamer,
-    IUsedTool,
-    IVisionChatModal
-} from '../../../src/Interface'
+import { FlowiseMemory, ICommonObject, INode, INodeData, INodeParams, IServerSideEventStreamer, IUsedTool } from '../../../src/Interface'
 import { ConsoleCallbackHandler, CustomChainHandler, CustomStreamingHandler, additionalCallbacks } from '../../../src/handler'
 import { AgentExecutor, ToolCallingAgentOutputParser } from '../../../src/agents'
 import { Moderation, checkInputs, streamResponse } from '../../moderation/Moderation'
@@ -315,12 +306,9 @@ const prepareAgent = async (
     }
 
     if (llmSupportsVision(model)) {
-        const visionChatModel = model as IVisionChatModal
         const messageContent = await addImagesToMessages(nodeData, options, model.multiModalOption)
 
         if (messageContent?.length) {
-            visionChatModel.setVisionModel()
-
             // Pop the `agent_scratchpad` MessagePlaceHolder
             let messagePlaceholder = prompt.promptMessages.pop() as MessagesPlaceholder
             if (prompt.promptMessages.at(-1) instanceof HumanMessagePromptTemplate) {
@@ -338,8 +326,6 @@ const prepareAgent = async (
 
             // Add the `agent_scratchpad` MessagePlaceHolder back
             prompt.promptMessages.push(messagePlaceholder)
-        } else {
-            visionChatModel.revertToOriginalModel()
         }
     }
 
diff --git a/packages/components/nodes/chains/ConversationChain/ConversationChain.ts b/packages/components/nodes/chains/ConversationChain/ConversationChain.ts
@@ -16,7 +16,6 @@ import { formatResponse } from '../../outputparsers/OutputParserHelpers'
 import { addImagesToMessages, llmSupportsVision } from '../../../src/multiModalUtils'
 import { ChatOpenAI } from '../../chatmodels/ChatOpenAI/FlowiseChatOpenAI'
 import {
-    IVisionChatModal,
     FlowiseMemory,
     ICommonObject,
     INode,
@@ -232,13 +231,6 @@ const prepareChain = async (nodeData: INodeData, options: ICommonObject, session
     let messageContent: MessageContentImageUrl[] = []
     if (llmSupportsVision(model)) {
         messageContent = await addImagesToMessages(nodeData, options, model.multiModalOption)
-        const visionChatModel = model as IVisionChatModal
-        if (messageContent?.length) {
-            visionChatModel.setVisionModel()
-        } else {
-            // revert to previous values if image upload is empty
-            visionChatModel.revertToOriginalModel()
-        }
     }
 
     const chatPrompt = prepareChatPrompt(nodeData, messageContent)
diff --git a/packages/components/nodes/chains/LLMChain/LLMChain.ts b/packages/components/nodes/chains/LLMChain/LLMChain.ts
@@ -201,8 +201,6 @@ const runPrediction = async (
         const visionChatModel = chain.llm as IVisionChatModal
         const messageContent = await addImagesToMessages(nodeData, options, visionChatModel.multiModalOption)
         if (messageContent?.length) {
-            // Change model to gpt-4-vision && max token to higher when using gpt-4-vision
-            visionChatModel.setVisionModel()
             // Add image to the message
             if (chain.prompt instanceof PromptTemplate) {
                 const existingPromptTemplate = chain.prompt.template as string
@@ -238,9 +236,6 @@ const runPrediction = async (
                 // @ts-ignore
                 chain.prompt.examplePrompt = newFewShotPromptTemplate
             }
-        } else {
-            // revert to previous values if image upload is empty
-            visionChatModel.revertToOriginalModel()
         }
     }
 
diff --git a/packages/components/nodes/chatmodels/AWSBedrock/FlowiseAWSChatBedrock.ts b/packages/components/nodes/chatmodels/AWSBedrock/FlowiseAWSChatBedrock.ts
@@ -1,9 +1,6 @@
 import { IVisionChatModal, IMultiModalOption } from '../../../src'
 import { ChatBedrockConverse as LCBedrockChat, ChatBedrockConverseInput } from '@langchain/aws'
 
-const DEFAULT_IMAGE_MODEL = 'anthropic.claude-3-haiku-20240307-v1:0'
-const DEFAULT_IMAGE_MAX_TOKEN = 1024
-
 export class BedrockChat extends LCBedrockChat implements IVisionChatModal {
     configuredModel: string
     configuredMaxToken?: number
@@ -17,19 +14,7 @@ export class BedrockChat extends LCBedrockChat implements IVisionChatModal {
         this.configuredMaxToken = fields?.maxTokens
     }
 
-    revertToOriginalModel(): void {
-        this.model = this.configuredModel
-        this.maxTokens = this.configuredMaxToken
-    }
-
     setMultiModalOption(multiModalOption: IMultiModalOption): void {
         this.multiModalOption = multiModalOption
     }
-
-    setVisionModel(): void {
-        if (!this.model.includes('claude-3')) {
-            this.model = DEFAULT_IMAGE_MODEL
-            this.maxTokens = this.configuredMaxToken ? this.configuredMaxToken : DEFAULT_IMAGE_MAX_TOKEN
-        }
-    }
 }
diff --git a/packages/components/nodes/chatmodels/AzureChatOpenAI/FlowiseAzureChatOpenAI.ts b/packages/components/nodes/chatmodels/AzureChatOpenAI/FlowiseAzureChatOpenAI.ts
@@ -17,19 +17,10 @@ export class AzureChatOpenAI extends LangchainAzureChatOpenAI implements IVision
         this.configuredMaxToken = fields?.maxTokens
     }
 
-    revertToOriginalModel(): void {
-        this.model = this.configuredModel
-        this.maxTokens = this.configuredMaxToken
-    }
-
     setMultiModalOption(multiModalOption: IMultiModalOption): void {
         this.multiModalOption = multiModalOption
     }
 
-    setVisionModel(): void {
-        // pass
-    }
-
     addBuiltInTools(builtInTool: Record<string, any>): void {
         this.builtInTools.push(builtInTool)
     }
diff --git a/packages/components/nodes/chatmodels/ChatAnthropic/FlowiseChatAnthropic.ts b/packages/components/nodes/chatmodels/ChatAnthropic/FlowiseChatAnthropic.ts
@@ -2,9 +2,6 @@ import { AnthropicInput, ChatAnthropic as LangchainChatAnthropic } from '@langch
 import { type BaseChatModelParams } from '@langchain/core/language_models/chat_models'
 import { IVisionChatModal, IMultiModalOption } from '../../../src'
 
-const DEFAULT_IMAGE_MODEL = 'claude-3-5-haiku-latest'
-const DEFAULT_IMAGE_MAX_TOKEN = 2048
-
 export class ChatAnthropic extends LangchainChatAnthropic implements IVisionChatModal {
     configuredModel: string
     configuredMaxToken: number
@@ -19,19 +16,7 @@ export class ChatAnthropic extends LangchainChatAnthropic implements IVisionChat
         this.configuredMaxToken = fields?.maxTokens ?? 2048
     }
 
-    revertToOriginalModel(): void {
-        this.modelName = this.configuredModel
-        this.maxTokens = this.configuredMaxToken
-    }
-
     setMultiModalOption(multiModalOption: IMultiModalOption): void {
         this.multiModalOption = multiModalOption
     }
-
-    setVisionModel(): void {
-        if (!this.modelName.startsWith('claude-3')) {
-            this.modelName = DEFAULT_IMAGE_MODEL
-            this.maxTokens = this.configuredMaxToken ? this.configuredMaxToken : DEFAULT_IMAGE_MAX_TOKEN
-        }
-    }
 }
diff --git a/packages/components/nodes/chatmodels/ChatGoogleGenerativeAI/FlowiseChatGoogleGenerativeAI.ts b/packages/components/nodes/chatmodels/ChatGoogleGenerativeAI/FlowiseChatGoogleGenerativeAI.ts
@@ -930,16 +930,7 @@ export class ChatGoogleGenerativeAI extends LangchainChatGoogleGenerativeAI impl
         }
     }
 
-    revertToOriginalModel(): void {
-        this.model = this.configuredModel
-        this.maxOutputTokens = this.configuredMaxToken
-    }
-
     setMultiModalOption(multiModalOption: IMultiModalOption): void {
         this.multiModalOption = multiModalOption
     }
-
-    setVisionModel(): void {
-        // pass
-    }
 }
diff --git a/packages/components/nodes/chatmodels/ChatGoogleVertexAI/ChatGoogleVertexAI.ts b/packages/components/nodes/chatmodels/ChatGoogleVertexAI/ChatGoogleVertexAI.ts
@@ -14,9 +14,6 @@ import {
 import { getModels, getRegions, MODEL_TYPE } from '../../../src/modelLoader'
 import { getBaseClasses } from '../../../src/utils'
 
-const DEFAULT_IMAGE_MAX_TOKEN = 8192
-const DEFAULT_IMAGE_MODEL = 'gemini-1.5-flash-latest'
-
 class ChatVertexAI extends LcChatVertexAI implements IVisionChatModal {
     configuredModel: string
     configuredMaxToken: number
@@ -35,21 +32,9 @@ class ChatVertexAI extends LcChatVertexAI implements IVisionChatModal {
         this.configuredMaxToken = fields?.maxOutputTokens ?? 2048
     }
 
-    revertToOriginalModel(): void {
-        this.modelName = this.configuredModel
-        this.maxOutputTokens = this.configuredMaxToken
-    }
-
     setMultiModalOption(multiModalOption: IMultiModalOption): void {
         this.multiModalOption = multiModalOption
     }
-
-    setVisionModel(): void {
-        if (!this.modelName.startsWith('claude-3')) {
-            this.modelName = DEFAULT_IMAGE_MODEL
-            this.maxOutputTokens = this.configuredMaxToken ? this.configuredMaxToken : DEFAULT_IMAGE_MAX_TOKEN
-        }
-    }
 }
 
 class GoogleVertexAI_ChatModels implements INode {
diff --git a/packages/components/nodes/chatmodels/ChatOllama/FlowiseChatOllama.ts b/packages/components/nodes/chatmodels/ChatOllama/FlowiseChatOllama.ts
@@ -13,15 +13,7 @@ export class ChatOllama extends LCChatOllama implements IVisionChatModal {
         this.configuredModel = fields?.model ?? ''
     }
 
-    revertToOriginalModel(): void {
-        this.model = this.configuredModel
-    }
-
     setMultiModalOption(multiModalOption: IMultiModalOption): void {
         this.multiModalOption = multiModalOption
     }
-
-    setVisionModel(): void {
-        // pass
-    }
 }
diff --git a/packages/components/nodes/chatmodels/ChatOpenAI/FlowiseChatOpenAI.ts b/packages/components/nodes/chatmodels/ChatOpenAI/FlowiseChatOpenAI.ts
@@ -15,19 +15,10 @@ export class ChatOpenAI extends LangchainChatOpenAI implements IVisionChatModal
         this.configuredMaxToken = fields?.maxTokens
     }
 
-    revertToOriginalModel(): void {
-        this.model = this.configuredModel
-        this.maxTokens = this.configuredMaxToken
-    }
-
     setMultiModalOption(multiModalOption: IMultiModalOption): void {
         this.multiModalOption = multiModalOption
     }
 
-    setVisionModel(): void {
-        // pass
-    }
-
     addBuiltInTools(builtInTool: Record<string, any>): void {
         this.builtInTools.push(builtInTool)
     }
diff --git a/packages/components/nodes/chatmodels/ChatOpenRouter/FlowiseChatOpenRouter.ts b/packages/components/nodes/chatmodels/ChatOpenRouter/FlowiseChatOpenRouter.ts
@@ -14,16 +14,7 @@ export class ChatOpenRouter extends LangchainChatOpenAI implements IVisionChatMo
         this.configuredMaxToken = fields?.maxTokens
     }
 
-    revertToOriginalModel(): void {
-        this.model = this.configuredModel
-        this.maxTokens = this.configuredMaxToken
-    }
-
     setMultiModalOption(multiModalOption: IMultiModalOption): void {
         this.multiModalOption = multiModalOption
     }
-
-    setVisionModel(): void {
-        // pass - OpenRouter models don't need model switching
-    }
 }
diff --git a/packages/components/nodes/chatmodels/ChatPerplexity/FlowiseChatPerplexity.ts b/packages/components/nodes/chatmodels/ChatPerplexity/FlowiseChatPerplexity.ts
@@ -15,18 +15,8 @@ export class ChatPerplexity extends LangchainChatPerplexity implements IVisionCh
         this.configuredMaxToken = fields?.maxTokens
     }
 
-    // Method to revert to the original model configuration
-    revertToOriginalModel(): void {
-        this.model = this.configuredModel
-        this.maxTokens = this.configuredMaxToken
-    }
-
     // Method to set multimodal options
     setMultiModalOption(multiModalOption: IMultiModalOption): void {
         this.multiModalOption = multiModalOption
     }
-
-    setVisionModel(): void {
-        // pass
-    }
 }
diff --git a/packages/components/nodes/chatmodels/ChatXAI/FlowiseChatXAI.ts b/packages/components/nodes/chatmodels/ChatXAI/FlowiseChatXAI.ts
diff --git a/packages/components/nodes/multiagents/Supervisor/Supervisor.ts b/packages/components/nodes/multiagents/Supervisor/Supervisor.ts
diff --git a/packages/components/nodes/sequentialagents/commonUtils.ts b/packages/components/nodes/sequentialagents/commonUtils.ts
diff --git a/packages/components/src/Interface.ts b/packages/components/src/Interface.ts
diff --git a/packages/components/src/multiModalUtils.ts b/packages/components/src/multiModalUtils.ts

Original file line number	Diff line number	Diff line change
`@@ -201,8 +201,6 @@ const runPrediction = async (`
`201`	`201`	`const visionChatModel = chain.llm as IVisionChatModal`
`202`	`202`	`const messageContent = await addImagesToMessages(nodeData, options, visionChatModel.multiModalOption)`
`203`	`203`	`if (messageContent?.length) {`
`204`		`- // Change model to gpt-4-vision && max token to higher when using gpt-4-vision`
`205`		`- visionChatModel.setVisionModel()`
`206`	`204`	`// Add image to the message`
`207`	`205`	`if (chain.prompt instanceof PromptTemplate) {`
`208`	`206`	`const existingPromptTemplate = chain.prompt.template as string`
`@@ -238,9 +236,6 @@ const runPrediction = async (`
`238`	`236`	`// @ts-ignore`
`239`	`237`	`chain.prompt.examplePrompt = newFewShotPromptTemplate`
`240`	`238`	`}`
`241`		`- } else {`
`242`		`- // revert to previous values if image upload is empty`
`243`		`- visionChatModel.revertToOriginalModel()`
`244`	`239`	`}`
`245`	`240`	`}`
`246`	`241`
Original file line number	Diff line number	Diff line change
`@@ -17,19 +17,10 @@ export class AzureChatOpenAI extends LangchainAzureChatOpenAI implements IVision`
`17`	`17`	`this.configuredMaxToken = fields?.maxTokens`
`18`	`18`	`}`
`19`	`19`
`20`		`- revertToOriginalModel(): void {`
`21`		`- this.model = this.configuredModel`
`22`		`- this.maxTokens = this.configuredMaxToken`
`23`		`- }`
`24`		`-`
`25`	`20`	`setMultiModalOption(multiModalOption: IMultiModalOption): void {`
`26`	`21`	`this.multiModalOption = multiModalOption`
`27`	`22`	`}`
`28`	`23`
`29`		`- setVisionModel(): void {`
`30`		`- // pass`
`31`		`- }`
`32`		`-`
`33`	`24`	`addBuiltInTools(builtInTool: Record<string, any>): void {`
`34`	`25`	`this.builtInTools.push(builtInTool)`
`35`	`26`	`}`
Original file line number	Diff line number	Diff line change
`@@ -930,16 +930,7 @@ export class ChatGoogleGenerativeAI extends LangchainChatGoogleGenerativeAI impl`
`930`	`930`	`}`
`931`	`931`	`}`
`932`	`932`
`933`		`- revertToOriginalModel(): void {`
`934`		`- this.model = this.configuredModel`
`935`		`- this.maxOutputTokens = this.configuredMaxToken`
`936`		`- }`
`937`		`-`
`938`	`933`	`setMultiModalOption(multiModalOption: IMultiModalOption): void {`
`939`	`934`	`this.multiModalOption = multiModalOption`
`940`	`935`	`}`
`941`		`-`
`942`		`- setVisionModel(): void {`
`943`		`- // pass`
`944`		`- }`
`945`	`936`	`}`