- update LangchainJS version

- add types & interfaces - small ui bug fix - update marketplace files
digesto · Apr 13, 2023 · 9281b57 · 9281b57
1 parent 02d8284
commit 9281b57
Show file tree

Hide file tree

Showing 47 changed files with 1,043 additions and 1,006 deletions.
diff --git a/.github/workflows/main.yml b/.github/workflows/main.yml
@@ -17,7 +17,7 @@ jobs:
         strategy:
             matrix:
                 platform: [ubuntu-latest]
-                node-version: [14.x, 16.x]
+                node-version: [18.15.0]
         runs-on: ${{ matrix.platform }}
 
         steps:

diff --git a/packages/components/nodes/agents/ConversationalAgent/ConversationalAgent.ts b/packages/components/nodes/agents/ConversationalAgent/ConversationalAgent.ts
@@ -1,4 +1,9 @@
 import { INode, INodeData, INodeParams } from '../../../src/Interface'
+import { initializeAgentExecutor, AgentExecutor } from 'langchain/agents'
+import { Tool } from 'langchain/tools'
+import { BaseChatModel } from 'langchain/chat_models/base'
+import { BaseChatMemory } from 'langchain/memory'
+import { getBaseClasses } from '../../../src/utils'
 
 class ConversationalAgent_Agents implements INode {
     label: string
@@ -17,6 +22,7 @@ class ConversationalAgent_Agents implements INode {
         this.category = 'Agents'
         this.icon = 'agent.svg'
         this.description = 'Conversational agent for a chat model. It will utilize chat specific prompts'
+        this.baseClasses = [this.type, ...getBaseClasses(AgentExecutor)]
         this.inputs = [
             {
                 label: 'Allowed Tools',
@@ -37,24 +43,18 @@ class ConversationalAgent_Agents implements INode {
         ]
     }
 
-    async getBaseClasses(): Promise<string[]> {
-        return ['AgentExecutor']
-    }
-
     async init(nodeData: INodeData): Promise<any> {
-        const { initializeAgentExecutor } = await import('langchain/agents')
-
-        const model = nodeData.inputs?.model
-        const tools = nodeData.inputs?.tools
-        const memory = nodeData.inputs?.memory
+        const model = nodeData.inputs?.model as BaseChatModel
+        const tools = nodeData.inputs?.tools as Tool[]
+        const memory = nodeData.inputs?.memory as BaseChatMemory
 
         const executor = await initializeAgentExecutor(tools, model, 'chat-conversational-react-description', true)
         executor.memory = memory
         return executor
     }
 
     async run(nodeData: INodeData, input: string): Promise<string> {
-        const executor = nodeData.instance
+        const executor = nodeData.instance as AgentExecutor
         const result = await executor.call({ input })
 
         return result?.output

diff --git a/...des/agents/MRLKAgentChat/MRLKAgentChat.ts → ...des/agents/MRKLAgentChat/MRKLAgentChat.ts b/...des/agents/MRLKAgentChat/MRLKAgentChat.ts → ...des/agents/MRKLAgentChat/MRKLAgentChat.ts
@@ -1,6 +1,10 @@
 import { INode, INodeData, INodeParams } from '../../../src/Interface'
+import { initializeAgentExecutor, AgentExecutor } from 'langchain/agents'
+import { Tool } from 'langchain/tools'
+import { BaseChatModel } from 'langchain/chat_models/base'
+import { getBaseClasses } from '../../../src/utils'
 
-class MRLKAgentChat_Agents implements INode {
+class MRKLAgentChat_Agents implements INode {
     label: string
     name: string
     description: string
@@ -11,12 +15,13 @@ class MRLKAgentChat_Agents implements INode {
     inputs: INodeParams[]
 
     constructor() {
-        this.label = 'MRLK Agent for Chat Models'
-        this.name = 'mrlkAgentChat'
+        this.label = 'MRKL Agent for Chat Models'
+        this.name = 'mrklAgentChat'
         this.type = 'AgentExecutor'
         this.category = 'Agents'
         this.icon = 'agent.svg'
         this.description = 'Agent that uses the ReAct Framework to decide what action to take, optimized to be used with Chat Models'
+        this.baseClasses = [this.type, ...getBaseClasses(AgentExecutor)]
         this.inputs = [
             {
                 label: 'Allowed Tools',
@@ -32,27 +37,20 @@ class MRLKAgentChat_Agents implements INode {
         ]
     }
 
-    async getBaseClasses(): Promise<string[]> {
-        return ['AgentExecutor']
-    }
-
     async init(nodeData: INodeData): Promise<any> {
-        const { initializeAgentExecutor } = await import('langchain/agents')
-
-        const model = nodeData.inputs?.model
-        const tools = nodeData.inputs?.tools
+        const model = nodeData.inputs?.model as BaseChatModel
+        const tools = nodeData.inputs?.tools as Tool[]
 
         const executor = await initializeAgentExecutor(tools, model, 'chat-zero-shot-react-description', true)
-
         return executor
     }
 
     async run(nodeData: INodeData, input: string): Promise<string> {
-        const executor = nodeData.instance
+        const executor = nodeData.instance as AgentExecutor
         const result = await executor.call({ input })
 
         return result?.output
     }
 }
 
-module.exports = { nodeClass: MRLKAgentChat_Agents }
+module.exports = { nodeClass: MRKLAgentChat_Agents }
diff --git a/...ents/nodes/agents/MRLKAgentChat/agent.svg → ...ents/nodes/agents/MRKLAgentChat/agent.svg b/...ents/nodes/agents/MRLKAgentChat/agent.svg → ...ents/nodes/agents/MRKLAgentChat/agent.svg
diff --git a/...nodes/agents/MRLKAgentLLM/MRLKAgentLLM.ts → ...nodes/agents/MRKLAgentLLM/MRKLAgentLLM.ts b/...nodes/agents/MRLKAgentLLM/MRLKAgentLLM.ts → ...nodes/agents/MRKLAgentLLM/MRKLAgentLLM.ts
@@ -1,6 +1,10 @@
 import { INode, INodeData, INodeParams } from '../../../src/Interface'
+import { initializeAgentExecutor, AgentExecutor } from 'langchain/agents'
+import { Tool } from 'langchain/tools'
+import { BaseLLM } from 'langchain/llms/base'
+import { getBaseClasses } from '../../../src/utils'
 
-class MRLKAgentLLM_Agents implements INode {
+class MRKLAgentLLM_Agents implements INode {
     label: string
     name: string
     description: string
@@ -11,12 +15,13 @@ class MRLKAgentLLM_Agents implements INode {
     inputs: INodeParams[]
 
     constructor() {
-        this.label = 'MRLK Agent for LLMs'
-        this.name = 'mrlkAgentLLM'
+        this.label = 'MRKL Agent for LLMs'
+        this.name = 'mrklAgentLLM'
         this.type = 'AgentExecutor'
         this.category = 'Agents'
         this.icon = 'agent.svg'
         this.description = 'Agent that uses the ReAct Framework to decide what action to take, optimized to be used with LLMs'
+        this.baseClasses = [this.type, ...getBaseClasses(AgentExecutor)]
         this.inputs = [
             {
                 label: 'Allowed Tools',
@@ -27,32 +32,25 @@ class MRLKAgentLLM_Agents implements INode {
             {
                 label: 'LLM Model',
                 name: 'model',
-                type: 'BaseLanguageModel'
+                type: 'BaseLLM'
             }
         ]
     }
 
-    async getBaseClasses(): Promise<string[]> {
-        return ['AgentExecutor']
-    }
-
     async init(nodeData: INodeData): Promise<any> {
-        const { initializeAgentExecutor } = await import('langchain/agents')
-
-        const model = nodeData.inputs?.model
-        const tools = nodeData.inputs?.tools
+        const model = nodeData.inputs?.model as BaseLLM
+        const tools = nodeData.inputs?.tools as Tool[]
 
         const executor = await initializeAgentExecutor(tools, model, 'zero-shot-react-description', true)
-
         return executor
     }
 
     async run(nodeData: INodeData, input: string): Promise<string> {
-        const executor = nodeData.instance
+        const executor = nodeData.instance as AgentExecutor
         const result = await executor.call({ input })
 
         return result?.output
     }
 }
 
-module.exports = { nodeClass: MRLKAgentLLM_Agents }
+module.exports = { nodeClass: MRKLAgentLLM_Agents }
diff --git a/...nents/nodes/agents/MRLKAgentLLM/agent.svg → ...nents/nodes/agents/MRKLAgentLLM/agent.svg b/...nents/nodes/agents/MRLKAgentLLM/agent.svg → ...nents/nodes/agents/MRKLAgentLLM/agent.svg
diff --git a/.../components/nodes/chains/ConversationalRetrievalQAChain/ConversationalRetrievalQAChain.ts b/.../components/nodes/chains/ConversationalRetrievalQAChain/ConversationalRetrievalQAChain.ts
@@ -1,5 +1,8 @@
 import { ICommonObject, IMessage, INode, INodeData, INodeParams } from '../../../src/Interface'
 import { getBaseClasses } from '../../../src/utils'
+import { ConversationalRetrievalQAChain } from 'langchain/chains'
+import { BaseLLM } from 'langchain/llms/base'
+import { BaseRetriever } from 'langchain/schema'
 
 class ConversationalRetrievalQAChain_Chains implements INode {
     label: string
@@ -18,11 +21,12 @@ class ConversationalRetrievalQAChain_Chains implements INode {
         this.icon = 'chain.svg'
         this.category = 'Chains'
         this.description = 'Document QA - built on RetrievalQAChain to provide a chat history component'
+        this.baseClasses = [this.type, ...getBaseClasses(ConversationalRetrievalQAChain)]
         this.inputs = [
             {
                 label: 'LLM',
                 name: 'llm',
-                type: 'BaseLanguageModel'
+                type: 'BaseLLM'
             },
             {
                 label: 'Vector Store Retriever',
@@ -32,23 +36,16 @@ class ConversationalRetrievalQAChain_Chains implements INode {
         ]
     }
 
-    async getBaseClasses(): Promise<string[]> {
-        const { ConversationalRetrievalQAChain } = await import('langchain/chains')
-        return getBaseClasses(ConversationalRetrievalQAChain)
-    }
-
     async init(nodeData: INodeData): Promise<any> {
-        const { ConversationalRetrievalQAChain } = await import('langchain/chains')
-
-        const llm = nodeData.inputs?.llm
-        const vectorStoreRetriever = nodeData.inputs?.vectorStoreRetriever
+        const llm = nodeData.inputs?.llm as BaseLLM
+        const vectorStoreRetriever = nodeData.inputs?.vectorStoreRetriever as BaseRetriever
 
         const chain = ConversationalRetrievalQAChain.fromLLM(llm, vectorStoreRetriever)
         return chain
     }
 
     async run(nodeData: INodeData, input: string, options: ICommonObject): Promise<string> {
-        const chain = nodeData.instance
+        const chain = nodeData.instance as ConversationalRetrievalQAChain
         let chatHistory = ''
 
         if (options && options.chatHistory) {

diff --git a/packages/components/nodes/chains/LLMChain/LLMChain.ts b/packages/components/nodes/chains/LLMChain/LLMChain.ts
@@ -1,5 +1,8 @@
 import { INode, INodeData, INodeParams } from '../../../src/Interface'
 import { getBaseClasses } from '../../../src/utils'
+import { LLMChain } from 'langchain/chains'
+import { BaseLanguageModel } from 'langchain/base_language'
+import { BasePromptTemplate } from 'langchain/prompts'
 
 class LLMChain_Chains implements INode {
     label: string
@@ -18,10 +21,11 @@ class LLMChain_Chains implements INode {
         this.icon = 'chain.svg'
         this.category = 'Chains'
         this.description = 'Chain to run queries against LLMs'
+        this.baseClasses = [this.type, ...getBaseClasses(LLMChain)]
         this.inputs = [
             {
-                label: 'LLM',
-                name: 'llm',
+                label: 'Language Model',
+                name: 'model',
                 type: 'BaseLanguageModel'
             },
             {
@@ -43,24 +47,17 @@ class LLMChain_Chains implements INode {
         ]
     }
 
-    async getBaseClasses(): Promise<string[]> {
-        const { LLMChain } = await import('langchain/chains')
-        return getBaseClasses(LLMChain)
-    }
-
     async init(nodeData: INodeData): Promise<any> {
-        const { LLMChain } = await import('langchain/chains')
+        const model = nodeData.inputs?.model as BaseLanguageModel
+        const prompt = nodeData.inputs?.prompt as BasePromptTemplate
 
-        const llm = nodeData.inputs?.llm
-        const prompt = nodeData.inputs?.prompt
-
-        const chain = new LLMChain({ llm, prompt })
+        const chain = new LLMChain({ llm: model, prompt })
         return chain
     }
 
     async run(nodeData: INodeData, input: string): Promise<string> {
-        const inputVariables = nodeData.instance.prompt.inputVariables // ["product"]
-        const chain = nodeData.instance
+        const inputVariables = nodeData.instance.prompt.inputVariables as string[] // ["product"]
+        const chain = nodeData.instance as LLMChain
 
         if (inputVariables.length === 1) {
             const res = await chain.run(input)
@@ -71,7 +68,7 @@ class LLMChain_Chains implements INode {
 
             const promptValues = JSON.parse(promptValuesStr.replace(/\s/g, ''))
 
-            let seen = []
+            let seen: string[] = []
 
             for (const variable of inputVariables) {
                 seen.push(variable)
@@ -81,13 +78,17 @@ class LLMChain_Chains implements INode {
             }
 
             if (seen.length === 1) {
+                const lastValue = seen.pop()
+                if (!lastValue) throw new Error('Please provide Prompt Values')
                 const options = {
                     ...promptValues,
-                    [seen.pop()]: input
+                    [lastValue]: input
                 }
                 const res = await chain.call(options)
                 return res?.text
-            } else throw new Error('Please provide Prompt Values')
+            } else {
+                throw new Error('Please provide Prompt Values')
+            }
         } else {
             const res = await chain.run(input)
             return res

diff --git a/packages/components/nodes/chains/RetrievalQAChain/RetrievalQAChain.ts b/packages/components/nodes/chains/RetrievalQAChain/RetrievalQAChain.ts
@@ -1,4 +1,8 @@
 import { INode, INodeData, INodeParams } from '../../../src/Interface'
+import { RetrievalQAChain } from 'langchain/chains'
+import { BaseLLM } from 'langchain/llms/base'
+import { BaseRetriever } from 'langchain/schema'
+import { getBaseClasses } from '../../../src/utils'
 
 class RetrievalQAChain_Chains implements INode {
     label: string
@@ -17,11 +21,12 @@ class RetrievalQAChain_Chains implements INode {
         this.icon = 'chain.svg'
         this.category = 'Chains'
         this.description = 'QA chain to answer a question based on the retrieved documents'
+        this.baseClasses = [this.type, ...getBaseClasses(RetrievalQAChain)]
         this.inputs = [
             {
                 label: 'LLM',
                 name: 'llm',
-                type: 'BaseLanguageModel'
+                type: 'BaseLLM'
             },
             {
                 label: 'Vector Store Retriever',
@@ -31,21 +36,16 @@ class RetrievalQAChain_Chains implements INode {
         ]
     }
 
-    async getBaseClasses(): Promise<string[]> {
-        return ['BaseChain']
-    }
-
     async init(nodeData: INodeData): Promise<any> {
-        const { RetrievalQAChain } = await import('langchain/chains')
-        const llm = nodeData.inputs?.llm
-        const vectorStoreRetriever = nodeData.inputs?.vectorStoreRetriever
+        const llm = nodeData.inputs?.llm as BaseLLM
+        const vectorStoreRetriever = nodeData.inputs?.vectorStoreRetriever as BaseRetriever
 
         const chain = RetrievalQAChain.fromLLM(llm, vectorStoreRetriever)
         return chain
     }
 
     async run(nodeData: INodeData, input: string): Promise<string> {
-        const chain = nodeData.instance
+        const chain = nodeData.instance as RetrievalQAChain
         const obj = {
             query: input
         }