refactor: add parseLlmJson utility for improved JSON handling across LLM queries

Author: skwowet

services/apps/organizations_enrichment_worker/src/activities/llm.ts

@@ -1,3 +1,4 @@
+import { parseLlmJson } from '@crowd/common'
 import { LlmService } from '@crowd/common_services'
 import {
   OrganizationField,
@@ -132,7 +133,7 @@ export async function selectMostRelevantDomainWithLLM(
       organizationId,
     )
     if (!response) throw new Error('LLM returned no response')
-    return JSON.parse(response.answer) as LlmDomainSelection
+    return parseLlmJson<LlmDomainSelection>(response.answer)
   }
 
   const MAX_RETRIES = 1

services/apps/profiles_worker/src/workflows/member/processMemberBotAnalysisWithLLM.ts

@@ -1,5 +1,6 @@
-import { ApplicationFailure, proxyActivities } from '@temporalio/workflow'
+import { proxyActivities } from '@temporalio/workflow'
 
+import { parseLlmJson } from '@crowd/common'
 import { LlmQueryType } from '@crowd/types'
 
 import * as activities from '../../activities'
@@ -72,18 +73,7 @@ export async function processMemberBotAnalysisWithLLM(
 
   const llm = await getLLMResult(LlmQueryType.MEMBER_BOT_VALIDATION, PROMPT, memberId)
 
-  const start = llm.answer.indexOf('{')
-  const end = llm.answer.lastIndexOf('}')
-
-  if (start === -1 || end === -1) {
-    throw ApplicationFailure.retryable(
-      `LLM returned no valid JSON object for member ${memberId}: ${llm.answer.substring(0, 200)}`,
-    )
-  }
-
-  const { isBot, signals } = JSON.parse(
-    llm.answer.substring(start, end + 1),
-  ) as MemberBotSuggestionResult
+  const { isBot, signals } = parseLlmJson<MemberBotSuggestionResult>(llm.answer)
 
   if (!isBot) {
     await createMemberNoBot(memberId)

services/libs/common/src/index.ts

@@ -36,6 +36,7 @@ export * from './displayName'
 export * from './jira'
 export * from './email'
 export * from './bot'
+export * from './llm'
 
 export * from './i18n'
 export * from './member'

services/libs/common/src/llm.ts

@@ -0,0 +1,59 @@
+export function parseLlmJson<T>(answer: string): T {
+  const raw = answer.trim()
+
+  try {
+    return JSON.parse(raw) as T
+  } catch {
+    // continue with normalization
+  }
+
+  const fenced = raw.match(/^```(?:json)?\s*([\s\S]*?)\s*```$/i)
+  if (fenced?.[1]) {
+    return JSON.parse(fenced[1].trim()) as T
+  }
+
+  const starts = [
+    { idx: raw.indexOf('{'), open: '{', close: '}' },
+    { idx: raw.indexOf('['), open: '[', close: ']' },
+  ]
+    .filter((s) => s.idx >= 0)
+    .sort((a, b) => a.idx - b.idx)
+
+  for (const candidate of starts) {
+    let depth = 0
+    let inString = false
+    let escaped = false
+
+    for (let i = candidate.idx; i < raw.length; i++) {
+      const char = raw[i]
+
+      if (escaped) {
+        escaped = false
+        continue
+      }
+
+      if (char === '\\') {
+        escaped = true
+        continue
+      }
+
+      if (char === '"') {
+        inString = !inString
+        continue
+      }
+
+      if (!inString) {
+        if (char === candidate.open) {
+          depth += 1
+        } else if (char === candidate.close) {
+          depth -= 1
+          if (depth === 0) {
+            return JSON.parse(raw.slice(candidate.idx, i + 1)) as T
+          }
+        }
+      }
+    }
+  }
+
+  throw new SyntaxError('LLM response does not contain valid JSON content')
+}

services/libs/common_services/src/services/llm.service.ts

@@ -5,7 +5,7 @@ import {
 } from '@aws-sdk/client-bedrock-runtime'
 import { performance } from 'perf_hooks'
 
-import { IS_LLM_ENABLED } from '@crowd/common'
+import { IS_LLM_ENABLED, parseLlmJson } from '@crowd/common'
 import { insertPromptHistoryEntry } from '@crowd/data-access-layer'
 import { QueryExecutor } from '@crowd/data-access-layer'
 import { Logger, LoggerBase } from '@crowd/logging'
@@ -177,7 +177,7 @@ export class LlmService extends LoggerBase {
       } as ILlmResult<LlmMemberEnrichmentResult>
     }
 
-    const result = JSON.parse(response.answer)
+    const result = parseLlmJson<LlmMemberEnrichmentResult>(response.answer)
 
     return {
       result,
@@ -200,7 +200,7 @@ export class LlmService extends LoggerBase {
       } as ILlmResult<{ profileIndex: number }>
     }
 
-    const result = JSON.parse(response.answer)
+    const result = parseLlmJson<{ profileIndex: number }>(response.answer)
 
     return {
       result,
@@ -224,7 +224,7 @@ export class LlmService extends LoggerBase {
       } as ILlmResult<T>
     }
 
-    const result = JSON.parse(response.answer)
+    const result = parseLlmJson<T>(response.answer)
 
     return {
       result,
@@ -248,7 +248,7 @@ export class LlmService extends LoggerBase {
       } as ILlmResult<T>
     }
 
-    const result = JSON.parse(response.answer)
+    const result = parseLlmJson<T>(response.answer)
 
     return {
       result,
@@ -268,7 +268,7 @@ export class LlmService extends LoggerBase {
       } as ILlmResult<T>
     }
 
-    const result = JSON.parse(response.answer)
+    const result = parseLlmJson<T>(response.answer)
 
     return {
       result,
@@ -285,7 +285,7 @@ export class LlmService extends LoggerBase {
       } as ILlmResult<T>
     }
 
-    const result = JSON.parse(response.answer)
+    const result = parseLlmJson<T>(response.answer)
 
     return {
       result,
@@ -302,7 +302,7 @@ export class LlmService extends LoggerBase {
       } as ILlmResult<T>
     }
 
-    const result = JSON.parse(response.answer)
+    const result = parseLlmJson<T>(response.answer)
 
     return {
       result,