fix(triage): normalize OpenAI response schemas

Author: duncanleo

packages/triage/src/llm/common/jsonSchema.test.ts

@@ -0,0 +1,34 @@
+import { describe, expect, it } from 'vitest';
+import { z } from 'zod';
+import { toOpenAiJsonSchema } from './jsonSchema.js';
+
+function findSchemaKey(value: unknown, key: string): boolean {
+  if (Array.isArray(value)) {
+    return value.some((item) => findSchemaKey(item, key));
+  }
+
+  if (value == null || typeof value !== 'object') {
+    return false;
+  }
+
+  return Object.entries(value).some(
+    ([entryKey, entryValue]) =>
+      entryKey === key || findSchemaKey(entryValue, key),
+  );
+}
+
+describe('toOpenAiJsonSchema', () => {
+  it('rewrites Zod discriminated-union oneOf entries to anyOf', () => {
+    const schema = z.object({
+      result: z.discriminatedUnion('kind', [
+        z.object({ kind: z.literal('existing'), id: z.string() }),
+        z.object({ kind: z.literal('new'), type: z.string() }),
+      ]),
+    });
+
+    const jsonSchema = toOpenAiJsonSchema(schema);
+
+    expect(findSchemaKey(jsonSchema, 'oneOf')).toBe(false);
+    expect(findSchemaKey(jsonSchema, 'anyOf')).toBe(true);
+  });
+});

packages/triage/src/llm/common/jsonSchema.ts

@@ -0,0 +1,23 @@
+import { z } from 'zod';
+
+type JsonObject = { [key: string]: unknown };
+
+function replaceOneOfWithAnyOf(value: unknown): unknown {
+  if (Array.isArray(value)) {
+    return value.map(replaceOneOfWithAnyOf);
+  }
+
+  if (value == null || typeof value !== 'object') {
+    return value;
+  }
+
+  const output: JsonObject = {};
+  for (const [key, child] of Object.entries(value)) {
+    output[key === 'oneOf' ? 'anyOf' : key] = replaceOneOfWithAnyOf(child);
+  }
+  return output;
+}
+
+export function toOpenAiJsonSchema(schema: z.ZodType): JsonObject {
+  return replaceOneOfWithAnyOf(z.toJSONSchema(schema)) as JsonObject;
+}

packages/triage/src/llm/functions/extractClaimsFromNewEvidence/index.ts

@@ -12,6 +12,7 @@ import {
 } from '../../../helpers/estimateOpenAICost.js';
 import { assert } from '../../../util/assert.js';
 import { getOpenAiClient } from '../../client.js';
+import { toOpenAiJsonSchema } from '../../common/jsonSchema.js';
 import type { ToolRegistry } from '../../common/tool.js';
 import { normalizeClaimsForEvidence } from './normalizeClaimsForEvidence.js';
 import { buildSystemPrompt } from './prompt.js';
@@ -90,7 +91,7 @@ Timestamp: ${evidenceTs.toISO({ includeOffset: true })}
           type: 'json_schema',
           name: 'Response',
           strict: true,
-          schema: z.toJSONSchema(ResponseSchema),
+          schema: toOpenAiJsonSchema(ResponseSchema),
         },
       },
       tools: Object.values(toolRegistry).map((tool) => {

packages/triage/src/llm/functions/generateIssueTitleAndSlug/index.ts

@@ -2,6 +2,7 @@ import type { ResponseInputItem } from 'openai/resources/responses/responses.mjs
 import { z } from 'zod';
 import { assert } from '../../../util/assert.js';
 import { getOpenAiClient } from '../../client.js';
+import { toOpenAiJsonSchema } from '../../common/jsonSchema.js';
 import { buildSystemPrompt } from './prompt.js';
 
 export const ResponseSchema = z.object({
@@ -46,7 +47,7 @@ Text: ${params.text}
         type: 'json_schema',
         name: 'Response',
         strict: true,
-        schema: z.toJSONSchema(ResponseSchema),
+        schema: toOpenAiJsonSchema(ResponseSchema),
       },
     },
   });

packages/triage/src/llm/functions/translate/index.ts

@@ -1,12 +1,12 @@
 import { TranslationsSchema } from '@mrtdown/core';
 import type { ResponseInputItem } from 'openai/resources/responses/responses.js';
-import { z } from 'zod';
 import {
   estimateOpenAICostFromUsage,
   normalizeOpenAIResponsesUsage,
 } from '../../../helpers/estimateOpenAICost.js';
 import { assert } from '../../../util/assert.js';
 import { getOpenAiClient } from '../../client.js';
+import { toOpenAiJsonSchema } from '../../common/jsonSchema.js';
 
 export async function translate(text: string) {
   const model = 'gpt-5-nano';
@@ -31,7 +31,7 @@ Keep the names in English as much as possible, do not provide any translations f
         type: 'json_schema',
         name: 'Translation',
         strict: true,
-        schema: z.toJSONSchema(TranslationsSchema),
+        schema: toOpenAiJsonSchema(TranslationsSchema),
       },
     },
     reasoning: {

packages/triage/src/llm/functions/triageNewEvidence/index.ts

@@ -8,6 +8,7 @@ import type {
 import z from 'zod';
 import { assert } from '../../../util/assert.js';
 import { getOpenAiClient } from '../../client.js';
+import { toOpenAiJsonSchema } from '../../common/jsonSchema.js';
 import type { ToolRegistry } from '../../common/tool.js';
 import { buildSystemPrompt } from './prompt.js';
 import { FindIssuesByDateRangeTool } from './tools/FindIssuesByDateRangeTool.js';
@@ -82,7 +83,7 @@ Timestamp: ${evidenceTs.toISO({ includeOffset: true })}
           type: 'json_schema',
           name: 'Response',
           strict: true,
-          schema: z.toJSONSchema(ResponseSchema),
+          schema: toOpenAiJsonSchema(ResponseSchema),
         },
       },
       tools: Object.values(toolRegistry).map((tool) => {

packages/triage/src/util/ingestContent/index.ts

@@ -36,12 +36,12 @@ function createWriter(dataDir: string): MRTDownWriter {
 async function runLlmStep<T>(
   label: string,
   operation: () => Promise<T>,
-): Promise<T | null> {
+): Promise<T> {
   try {
     return await operation();
   } catch (error) {
     console.error(`[ingestContent] ${label} failed:`, error);
-    return null;
+    throw error;
   }
 }