[Feature] Repair JSON from LLM in StructuredOutputParserAdvanced (#3723)

Repair JSON from LLM in StructuredOutputParserAdvanced

Co-authored-by: Henry <hzj94@hotmail.com>

packages/components/nodes/outputparsers/StructuredOutputParser/StructuredOutputParser.ts

@@ -3,6 +3,7 @@ import { BaseOutputParser } from '@langchain/core/output_parsers'
 import { StructuredOutputParser as LangchainStructuredOutputParser } from 'langchain/output_parsers'
 import { CATEGORY } from '../OutputParserHelpers'
 import { convertSchemaToZod, getBaseClasses, INode, INodeData, INodeParams } from '../../../src'
+import { jsonrepair } from 'jsonrepair'
 
 class StructuredOutputParser implements INode {
     label: string
@@ -74,6 +75,14 @@ class StructuredOutputParser implements INode {
             const zodSchema = z.object(convertSchemaToZod(jsonStructure)) as any
             const structuredOutputParser = LangchainStructuredOutputParser.fromZodSchema(zodSchema)
 
+            const baseParse = structuredOutputParser.parse
+
+            // Fix broken JSON from LLM
+            structuredOutputParser.parse = (text) => {
+                const jsonString = text.includes('```') ? text.trim().split(/```(?:json)?/)[1] : text.trim()
+                return baseParse.call(structuredOutputParser, jsonrepair(jsonString))
+            }
+
             // NOTE: When we change Flowise to return a json response, the following has to be changed to: JsonStructuredOutputParser
             Object.defineProperty(structuredOutputParser, 'autoFix', {
                 enumerable: true,

packages/components/nodes/outputparsers/StructuredOutputParserAdvanced/StructuredOutputParserAdvanced.ts

@@ -3,6 +3,7 @@ import { BaseOutputParser } from '@langchain/core/output_parsers'
 import { StructuredOutputParser as LangchainStructuredOutputParser } from 'langchain/output_parsers'
 import { CATEGORY } from '../OutputParserHelpers'
 import { z } from 'zod'
+import { jsonrepair } from 'jsonrepair'
 
 class AdvancedStructuredOutputParser implements INode {
     label: string
@@ -62,6 +63,14 @@ class AdvancedStructuredOutputParser implements INode {
         try {
             const structuredOutputParser = LangchainStructuredOutputParser.fromZodSchema(zodSchema)
 
+            const baseParse = structuredOutputParser.parse
+
+            // Fix broken JSON from LLM
+            structuredOutputParser.parse = (text) => {
+                const jsonString = text.includes('```') ? text.trim().split(/```(?:json)?/)[1] : text.trim()
+                return baseParse.call(structuredOutputParser, jsonrepair(jsonString))
+            }
+
             // NOTE: When we change Flowise to return a json response, the following has to be changed to: JsonStructuredOutputParser
             Object.defineProperty(structuredOutputParser, 'autoFix', {
                 enumerable: true,

packages/components/package.json

@@ -92,6 +92,7 @@
         "ioredis": "^5.3.2",
         "jsdom": "^22.1.0",
         "jsonpointer": "^5.0.1",
+        "jsonrepair": "^3.11.1",
         "langchain": "^0.3.5",
         "langfuse": "3.3.4",
         "langfuse-langchain": "^3.3.4",

pnpm-lock.yaml

@@ -319,6 +319,9 @@ importers:
             jsonpointer:
                 specifier: ^5.0.1
                 version: 5.0.1
+            jsonrepair:
+                specifier: ^3.11.1
+                version: 3.11.2
             langchain:
                 specifier: ^0.3.5
                 version: 0.3.5(@langchain/anthropic@0.3.7(@langchain/core@0.3.18(openai@4.57.3(encoding@0.1.13)(zod@3.22.4)))(encoding@0.1.13))(@langchain/aws@0.1.2(@aws-sdk/client-sso-oidc@3.624.0(@aws-sdk/client-sts@3.624.0))(@aws-sdk/client-sts@3.624.0)(@langchain/core@0.3.18(openai@4.57.3(encoding@0.1.13)(zod@3.22.4))))(@langchain/cohere@0.0.7(encoding@0.1.13)(openai@4.57.3(encoding@0.1.13)(zod@3.22.4)))(@langchain/core@0.3.18(openai@4.57.3(encoding@0.1.13)(zod@3.22.4)))(@langchain/google-genai@0.1.3(@langchain/core@0.3.18(openai@4.57.3(encoding@0.1.13)(zod@3.22.4)))(zod@3.22.4))(@langchain/google-vertexai@0.1.2(@langchain/core@0.3.18(openai@4.57.3(encoding@0.1.13)(zod@3.22.4)))(encoding@0.1.13)(zod@3.22.4))(@langchain/groq@0.1.2(@langchain/core@0.3.18(openai@4.57.3(encoding@0.1.13)(zod@3.22.4)))(encoding@0.1.13))(@langchain/mistralai@0.0.26(encoding@0.1.13)(openai@4.57.3(encoding@0.1.13)(zod@3.22.4)))(@langchain/ollama@0.1.2(@langchain/core@0.3.18(openai@4.57.3(encoding@0.1.13)(zod@3.22.4))))(axios@1.6.2)(cheerio@1.0.0-rc.12)(encoding@0.1.13)(openai@4.57.3(encoding@0.1.13)(zod@3.22.4))(typeorm@0.3.20(ioredis@5.3.2)(mongodb@6.3.0(gcp-metadata@6.1.0(encoding@0.1.13))(socks@2.8.1))(mysql2@3.11.4)(pg@8.11.3)(redis@4.6.13)(sqlite3@5.1.7)(ts-node@10.9.2(@swc/core@1.4.6)(@types/node@20.12.12)(typescript@5.5.2)))
@@ -10807,6 +10810,10 @@ packages:
         resolution: { integrity: sha512-p/nXbhSEcu3pZRdkW1OfJhpsVtW1gd4Wa1fnQc9YLiTfAjn0312eMKimbdIQzuZl9aa9xUGaRlP9T/CJE/ditQ== }
         engines: { node: '>=0.10.0' }
 
+    jsonrepair@3.11.2:
+        resolution: { integrity: sha512-ejydGcTq0qKk1r0NUBwjtvswbPFhs19+QEfwSeGwB8KJZ59W7/AOFmQh04c68mkJ+2hGk+OkOmkr2bKG4tGlLQ== }
+        hasBin: true
+
     jsonwebtoken@9.0.2:
         resolution: { integrity: sha512-PRp66vJ865SSqOlgqS8hujT5U4AOgMfhrwYIuIhfKaoSCZcirrmASQr8CX7cUg+RMih+hgznrjp99o+W4pJLHQ== }
         engines: { node: '>=12', npm: '>=6' }
@@ -29873,6 +29880,8 @@ snapshots:
 
     jsonpointer@5.0.1: {}
 
+    jsonrepair@3.11.2: {}
+
     jsonwebtoken@9.0.2:
         dependencies:
             jws: 3.2.2