fix typechecks for scripts/

Author: charleslien

evals/git-evals/gen-evals.ts

@@ -1,13 +1,13 @@
 import { execSync } from 'child_process'
 import fs from 'fs'
 import path from 'path'
-import { mapLimit } from 'async'
 
+import { disableLiveUserInputCheck } from '@codebuff/backend/live-user-inputs'
 import { promptAiSdk } from '@codebuff/backend/llm-apis/vercel-ai-sdk/ai-sdk'
 import { models } from '@codebuff/common/old-constants'
+import { mapLimit } from 'async'
 
 import { extractRepoNameFromUrl, setupTestRepo } from './setup-test-repo'
-import { disableLiveUserInputCheck } from '@codebuff/backend/live-user-inputs'
 
 import type { EvalData, EvalInput, FileState, EvalCommit } from './types'
 const SPEC_GENERATION_PROMPT = `Given a set of file changes and an optional description, write a clear specification describing WHAT needs to be implemented.
@@ -132,6 +132,7 @@ File Changes:\n${fileContext}`
       fingerprintId,
       userInputId,
       userId: undefined,
+      logger: console,
     })
 
     // Extract spec from <spec></spec> tags

evals/git-evals/judge-git-eval.ts

@@ -193,6 +193,7 @@ export async function judgeEvalRun(evalRun: EvalRunLog) {
       userInputId: generateCompactId(),
       userId: undefined,
       timeout: 10 * 60 * 1000, // 10 minute timeout
+      logger: console,
     }).catch((error) => {
       console.warn(`Judge ${index + 1} failed:`, error)
       return null

evals/git-evals/pick-commits.ts

@@ -3,14 +3,14 @@
 import { execFileSync } from 'child_process'
 import fs from 'fs'
 import path from 'path'
-import { mapLimit } from 'async'
 
+import { disableLiveUserInputCheck } from '@codebuff/backend/live-user-inputs'
 import { promptAiSdkStructured } from '@codebuff/backend/llm-apis/vercel-ai-sdk/ai-sdk'
 import { models } from '@codebuff/common/old-constants'
+import { mapLimit } from 'async'
 import { z } from 'zod/v4'
 
 import { extractRepoNameFromUrl, setupTestRepo } from './setup-test-repo'
-import { disableLiveUserInputCheck } from '@codebuff/backend/live-user-inputs'
 
 // Types for commit data
 export interface CommitDiff {
@@ -379,6 +379,7 @@ async function screenCommitsWithGpt5(
         fingerprintId,
         userInputId,
         userId: undefined,
+        logger: console,
       })
 
       // Handle empty or invalid response

evals/git-evals/post-eval-analysis.ts

@@ -188,5 +188,6 @@ export async function analyzeEvalResults(
     userInputId: generateCompactId(),
     userId: undefined,
     timeout: 10 * 60 * 1000, // 10 minute timeout
+    logger: console,
   })
 }

evals/git-evals/run-git-evals.ts

@@ -151,6 +151,7 @@ Explain your reasoning in detail. Do not ask Codebuff to git commit changes.`,
           userInputId: generateCompactId(),
           userId: undefined,
           timeout: 5 * 60_000, // 5 minute timeout
+          logger: console,
         })
       } catch (agentError) {
         throw new Error(

scripts/ft-file-selection/grade-traces.ts

@@ -46,7 +46,7 @@ async function gradeTraces({ logger }: { logger: Logger }) {
         batch.map(async (traceAndRelabels) => {
           try {
             console.log(`Grading trace ${traceAndRelabels.trace.id}`)
-            const result = await gradeRun(traceAndRelabels)
+            const result = await gradeRun({ ...traceAndRelabels, logger })
             return {
               traceId: traceAndRelabels.trace.id,
               status: 'success',

scripts/ft-file-selection/relabel-for-offline-scoring.ts

@@ -191,6 +191,7 @@ async function relabelTraceForModel(
     userInputId: 'relabel-offline-scoring',
     userId: TEST_USER_ID,
     maxOutputTokens: 1000,
+    logger: console,
   })
 
   const newRelabel: Relabel = {

scripts/ft-file-selection/relabel-traces.ts

@@ -67,22 +67,22 @@ async function runTraces() {
                   fingerprintId: 'relabel-trace-run',
                   userInputId: 'relabel-trace-run',
                   userId: TEST_USER_ID,
+                  logger: console,
                 })
               } else {
-                output = await promptFlashWithFallbacks(
-                  messagesWithSystem({
+                output = await promptFlashWithFallbacks({
+                  messages: messagesWithSystem({
                     messages: messages as Message[],
                     system: system as System,
                   }),
-                  {
-                    model:
-                      model as typeof models.openrouter_gemini2_5_pro_preview,
-                    clientSessionId: 'relabel-trace-run',
-                    fingerprintId: 'relabel-trace-run',
-                    userInputId: 'relabel-trace-run',
-                    userId: 'relabel-trace-run',
-                  },
-                )
+                  model:
+                    model as typeof models.openrouter_gemini2_5_pro_preview,
+                  clientSessionId: 'relabel-trace-run',
+                  fingerprintId: 'relabel-trace-run',
+                  userInputId: 'relabel-trace-run',
+                  userId: 'relabel-trace-run',
+                  logger: console,
+                })
               }
 
               // Create relabel record

scripts/get-changelog.ts

@@ -172,6 +172,7 @@ Start your response with a heading using ### (three hashes) and organize the con
       model: models.openrouter_claude_sonnet_4,
       userId: undefined,
       chargeUser: false,
+      logger: console,
     })
 
     // Clean up the AI response