root/Jyotisha
1
0
Fork 0
Code Issues Pull Requests 1 Actions 1 Packages Projects Releases Wiki Activity

fix(consult): treat pinched answers as failed and restore reply actions

Browse Source 
Incomplete Flash generations were billed as completed consultations. Fail
those runs, keep the partial text, and reuse the rectification like/copy/rerun
bar on ordinary chat replies.

Co-authored-by: Cursor <cursoragent@cursor.com>
This commit is contained in:
Jesse_Chen
2026-08-19 19:37:02 +08:00
parent c38f11dbd3
commit 7d667fecbf
17 changed files with 487 additions and 89 deletions
Download Patch File Download Diff File Expand all files Collapse all files
frontend/src/app/api/consult/route.ts
+2
View File
@@ -37,6 +37,7 @@ import type { AgentExecutionReceipt, WorkflowReceipt } from "@/lib/consultation-
import {
AGENT_MAX_STEPS,
AGENT_TIMEOUT_MS,
consultationGenerationSettings,
createConsultationAgentContext,
consultationModelStepTelemetry,
consultationStepBudgetReceipt,
@@ -653,6 +654,7 @@ export async function POST(request: Request) {
maxSteps: AGENT_MAX_STEPS,
abortSignal: agentAbortSignal,
hooks,
...consultationGenerationSettings(selectedModel.model),
};
const workflowReceipt: WorkflowReceipt = consultationMode === "general_no_birth_time"
? {
frontend/src/app/globals.css
+8 -8
View File
@@ -679,14 +679,14 @@ button:disabled { cursor: default; opacity: .45; }
.message-bubble { overflow: hidden; border: 0; padding: var(--space-3) var(--space-4); border-radius: var(--radius-lg); background: var(--color-canvas-muted); }
.message-assistant .message-bubble { border-radius: 0; background: transparent; padding: var(--space-3) 0; }
.rectification-message-entry { min-width: 0; }
.rectification-message-actions {
.message-actions {
display: flex;
align-items: center;
gap: 1px;
margin: -4px 0 8px 42px;
color: var(--color-ink-tertiary);
}
.rectification-message-actions button {
.message-actions button {
display: inline-flex;
width: 26px;
height: 26px;
@@ -702,17 +702,17 @@ button:disabled { cursor: default; opacity: .45; }
opacity: .7;
transition: background-color 120ms ease-out, color 120ms ease-out, opacity 120ms ease-out, transform 120ms ease-out;
}
.rectification-message-actions button:hover:not(:disabled) {
.message-actions button:hover:not(:disabled) {
background: color-mix(in srgb, var(--color-canvas-muted) 72%, transparent);
color: var(--color-ink-secondary);
opacity: 1;
transform: translateY(-1px);
}
.rectification-message-actions button:active:not(:disabled) { transform: translateY(0); }
.rectification-message-actions button.is-active { color: var(--color-action); opacity: 1; }
.rectification-message-actions button:focus-visible { outline: 2px solid color-mix(in srgb, var(--color-focus) 52%, transparent); outline-offset: 1px; }
.rectification-message-actions button:disabled { cursor: default; opacity: 0.32; }
.rectification-message-actions svg { width: 13px; height: 13px; stroke-width: 1.65; }
.message-actions button:active:not(:disabled) { transform: translateY(0); }
.message-actions button.is-active { color: var(--color-action); opacity: 1; }
.message-actions button:focus-visible { outline: 2px solid color-mix(in srgb, var(--color-focus) 52%, transparent); outline-offset: 1px; }
.message-actions button:disabled { cursor: default; opacity: 0.32; }
.message-actions svg { width: 13px; height: 13px; stroke-width: 1.65; }
.rectification-analysis {
width: min(620px, calc(100% - 42px));
margin: -2px 0 var(--space-1) 42px;
frontend/src/app/page.tsx
+139 -11
View File
@@ -26,6 +26,11 @@ import {
import { ConversationalBirthTimeRectification, type PersistedRectificationTurn } from "@/components/conversational-birth-time-rectification";
import { ChatMessageContent } from "@/components/chat-message-content";
import { AgentAvatar, ChatMessageRow } from "@/components/chat-message-row";
import {
ChatMessageActions,
toggleChatMessageFeedback,
type ChatMessageFeedback,
} from "@/components/chat-message-actions";
import { ModelSelector } from "@/components/model-selector";
import { OnboardingRedeemPaywall } from "@/components/onboarding-redeem-paywall";
import { BirthPlacePicker } from "@/components/birth-place-picker";
@@ -1059,6 +1064,8 @@ export default function Home() {
const [pendingSessionId, setPendingSessionId] = useState<string | null>(null);
const [pendingRequestId, setPendingRequestId] = useState<string | null>(null);
const [streamingReply, setStreamingReply] = useState<StreamingReply | null>(null);
const [messageFeedback, setMessageFeedback] = useState<Record<string, ChatMessageFeedback>>({});
const [copiedMessageKey, setCopiedMessageKey] = useState<string | null>(null);
const [replyOutcome, setReplyOutcome] = useState<ReplyOutcome | null>(null);
const [requestError, setRequestError] = useState<RequestError | null>(null);
const [birthTimeConsultationConsent, setBirthTimeConsultationConsent] = useState<BirthTimeConsultationConsentState>(
@@ -2907,16 +2914,22 @@ export default function Home() {
entrypoint: ConsultationEntrypoint | null = null,
consentGrantedForRequest: ConsultationBirthTimeMode | null = null,
targetSessionId: string | null = null,
options: { resumeRequestId?: string } = {},
options: {
resumeRequestId?: string;
sessionOverride?: ChatSession;
restoreOnFailure?: ChatSession;
} = {},
): Promise<boolean> {
const originalQuestion = text;
const question = text.trim();
const resumeRequestId = options.resumeRequestId;
const resuming = Boolean(resumeRequestId);
const currentSession = targetSessionId
const liveSession = targetSessionId
? sessions.find((session) => session.id === targetSessionId)
: activeSession;
const currentSession = options.sessionOverride ?? liveSession;
if (!question || !currentSession || !modelCatalog || !account) return false;
const rollbackSession = options.restoreOnFailure ?? liveSession ?? currentSession;
if (!resuming && (pendingSessionId || cancellationInFlight.current || pendingConsultation.current)) return false;
if (resuming) {
const pending = pendingConsultation.current;
@@ -3003,7 +3016,7 @@ export default function Home() {
question: originalQuestion,
entrypoint,
theme,
previousSession: currentSession,
previousSession: rollbackSession,
optimisticSession: userSession,
previousOnboardingState,
controller,
@@ -3077,9 +3090,9 @@ export default function Home() {
await persistSession(userSession);
} catch (caught) {
if (controller.signal.aborted) return false;
updateSession(sessionId, () => currentSession);
updateSession(sessionId, () => rollbackSession);
setOnboardingJustCompleted(previousOnboardingState);
if (activeSessionIdRef.current === sessionId) {
if (!options.restoreOnFailure && activeSessionIdRef.current === sessionId) {
setDraft(originalQuestion);
setDraftTheme(theme);
setDraftEntrypoint(entrypoint);
@@ -3088,7 +3101,9 @@ export default function Home() {
sessionId,
message: `${caught instanceof Error ? caught.message : "问题保存失败，请稍后重试。"} 问题已放回输入框。`,
});
setComposerNotice("问题保存失败，未开始生成；问题已放回输入框。");
setComposerNotice(options.restoreOnFailure
? "问题保存失败，未开始生成；已恢复原来的回答。"
: "问题保存失败，未开始生成；问题已放回输入框。");
completeConsultationInterface(requestId);
window.requestAnimationFrame(() => composerInput.current?.focus());
return false;
@@ -3164,6 +3179,7 @@ export default function Home() {
};
let agentExecutionReceipt: AgentExecutionReceipt | undefined;
let runCompleted = false;
let truncatedFailure: Extract<ConsultationAgentPublicEvent, { type: "run.failed" }> | undefined;
const reader = response.body.getReader();
const decoder = new TextDecoder();
let answer = "";
@@ -3200,7 +3216,18 @@ export default function Home() {
workflowReceipt = event.receipt.workflow;
techniqueTruth = event.receipt.techniqueTruth ?? "unknown";
}
if (event.type === "run.failed") throw new ConsultationResponseError(502, event.message);
if (event.type === "run.failed") {
if (event.code === "answer_truncated") {
truncatedFailure = event;
if (event.receipt) {
agentExecutionReceipt = event.receipt;
workflowReceipt = event.receipt.workflow;
techniqueTruth = event.receipt.techniqueTruth ?? techniqueTruth;
}
return;
}
throw new ConsultationResponseError(502, event.message);
}
updateActivity(event);
});
while (true) {
@@ -3209,7 +3236,38 @@ export default function Home() {
parser.push(decoder.decode(value, { stream: true }));
}
parser.finish(decoder.decode());
if (!runCompleted) throw new ConsultationResponseError(502, "Agent 回答未完成，本次不会保存为成功咨询。");
if (truncatedFailure) {
const reply = parseAgentReply(answer);
if (!reply.text) throw new ConsultationResponseError(502, truncatedFailure.message);
const truncatedSession: ChatSession = {
...userSession,
title: userSession.title,
messages: [...userSession.messages, {
role: "assistant",
text: reply.text,
techniqueTruth,
workflowReceipt,
agentExecutionReceipt,
}],
updatedAt: timestamp(),
};
updateSession(sessionId, () => truncatedSession);
setStreamingReply(null);
setReplyOutcome({ sessionId, phase: "failed", replyOrdinal: 0 });
setComposerNotice(truncatedFailure.message);
try {
await persistSession(truncatedSession);
} catch (error) {
setRequestError({
sessionId,
message: error instanceof Error ? error.message : "未完成的回答暂时无法同步。",
});
}
completeConsultationInterface(requestId);
void refreshAccount();
return true;
}
if (!runCompleted && !truncatedFailure) throw new ConsultationResponseError(502, "Agent 回答未完成，本次不会保存为成功咨询。");
} else {
while (true) {
const { done, value } = await reader.read();
@@ -3264,6 +3322,11 @@ export default function Home() {
setRequestError({ sessionId, message: caught.message });
setReplyOutcome({ sessionId, phase: "failed", replyOrdinal: 0 });
setComposerNotice(caught.message);
const restore = options.restoreOnFailure;
if (restore) {
updateSession(sessionId, () => restore);
void persistSession(restore).catch(() => {});
}
completeConsultationInterface(requestId);
return false;
}
@@ -3299,6 +3362,42 @@ export default function Home() {
}
}
async function copyAssistantMessage(renderKey: string, text: string) {
try {
await navigator.clipboard.writeText(text);
setCopiedMessageKey(renderKey);
window.setTimeout(() => setCopiedMessageKey((current) => (
current === renderKey ? null : current
)), 1_500);
} catch {
// Clipboard permission failures must not interrupt the conversation.
}
}
function regenerateLatestAnswer(renderKey: string) {
const session = activeSession;
if (!session || isLoading || cancellationPending || pendingConsultation.current) return;
const last = session.messages.at(-1);
if (last?.role !== "assistant" || last.text.trim() === "") return;
const previous = session.messages.at(-2);
if (previous?.role !== "user" || previous.text.trim() === "") return;
if (`message-${session.messages.length - 1}` !== renderKey) return;
const sessionOverride: ChatSession = {
...session,
messages: session.messages.slice(0, -1),
updatedAt: timestamp(),
};
setMessageFeedback((current) => {
const next = { ...current };
delete next[`${session.id}:${renderKey}`];
return next;
});
void send(previous.text, session.theme, null, null, session.id, {
sessionOverride,
restoreOnFailure: session,
});
}
consultationReplay.current = () => {
const pending = pendingConsultation.current;
if (!pending || pending.cancelled || pending.phase !== "recovering" || !pending.question.trim()) return;
@@ -3636,9 +3735,38 @@ export default function Home() {
</div>
) : (
<div className="message-list" aria-busy={isLoading}>
{chatMessageViews(activeSession.messages, isLoading, activeStreamingText, activeStreamingActivity).map((message) => (
<ChatMessageRow key={message.renderKey} message={message} />
))}
{chatMessageViews(activeSession.messages, isLoading, activeStreamingText, activeStreamingActivity).map((message, _, views) => {
const showActions = message.role === "assistant"
&& message.state === "settled"
&& Boolean(message.text);
const feedbackKey = `${activeSession.id}:${message.renderKey}`;
const latestRegeneratableKey = !isLoading && !cancellationPending
? [...views].reverse().find((item) => (
item.role === "assistant" && item.state === "settled" && Boolean(item.text)
))?.renderKey
: undefined;
return (
<div className="message-entry" key={message.renderKey}>
<ChatMessageRow message={message} />
{showActions && (
<ChatMessageActions
feedback={messageFeedback[feedbackKey]}
copied={copiedMessageKey === feedbackKey}
canRegenerate={message.renderKey === latestRegeneratableKey}
onFeedback={(requested) => setMessageFeedback((current) => {
const next = { ...current };
const value = toggleChatMessageFeedback(current[feedbackKey], requested);
if (value) next[feedbackKey] = value;
else delete next[feedbackKey];
return next;
})}
onCopy={() => void copyAssistantMessage(feedbackKey, message.text)}
onRegenerate={() => regenerateLatestAnswer(message.renderKey)}
/>
)}
</div>
);
})}
{activeError && <p className="error-message" role="alert">{activeError}</p>}
</div>
)}
frontend/src/components/chat-message-actions.tsx
+65
View File
@@ -0,0 +1,65 @@
"use client";
import { Check, Copy, RotateCcw, ThumbsDown, ThumbsUp } from "lucide-react";
export type ChatMessageFeedback = "up" | "down";
export function toggleChatMessageFeedback(
current: ChatMessageFeedback | undefined,
requested: ChatMessageFeedback,
): ChatMessageFeedback | undefined {
return current === requested ? undefined : requested;
}
export function ChatMessageActions({
feedback,
copied = false,
canRegenerate,
onFeedback,
onCopy,
onRegenerate,
}: Readonly<{
feedback?: ChatMessageFeedback;
copied?: boolean;
canRegenerate: boolean;
onFeedback: (value: ChatMessageFeedback) => void;
onCopy: () => void;
onRegenerate: () => void;
}>) {
return (
<div className="message-actions" aria-label="Agent 回答操作">
<button
aria-label="赞"
aria-pressed={feedback === "up"}
className={feedback === "up" ? "is-active" : ""}
title="赞"
type="button"
onClick={() => onFeedback("up")}
>
<ThumbsUp aria-hidden="true" />
</button>
<button
aria-label="踩"
aria-pressed={feedback === "down"}
className={feedback === "down" ? "is-active" : ""}
title="踩"
type="button"
onClick={() => onFeedback("down")}
>
<ThumbsDown aria-hidden="true" />
</button>
<button aria-label="复制回答" title="复制" type="button" onClick={onCopy}>
{copied ? <Check aria-hidden="true" /> : <Copy aria-hidden="true" />}
</button>
<button
aria-label="重新生成回答"
disabled={!canRegenerate}
title={canRegenerate ? "重新生成" : "只能重新生成最近一条回答"}
type="button"
onClick={onRegenerate}
>
<RotateCcw aria-hidden="true" />
</button>
</div>
);
}
frontend/src/components/rectification-agentic-chat.tsx
+17 -44
View File
@@ -1,6 +1,6 @@
"use client";
import { ArrowUp, Check, Copy, RotateCcw, ThumbsDown, ThumbsUp } from "lucide-react";
import { ArrowUp } from "lucide-react";
import { useCallback, useEffect, useRef, useState } from "react";
import { parseAgentReply } from "@/lib/agent-reply";
import type { ChatMessage, ChatMessageView } from "@/lib/chat-message-view";
@@ -23,6 +23,10 @@ import {
} from "@/lib/rectification-agentic/v9/public-receipt";
import type { PublicLanguageModel } from "@/lib/public-models";
import { ChatMessageRow } from "./chat-message-row";
import {
ChatMessageActions,
toggleChatMessageFeedback,
} from "./chat-message-actions";
import { CompletedActivityReceipt } from "./completed-activity-receipt";
import { ModelSelector } from "./model-selector";
import { Button } from "./ui/button";
@@ -125,7 +129,7 @@ export function toggleRectificationFeedback(
current: "up" | "down" | undefined,
requested: "up" | "down",
): "up" | "down" | undefined {
return current === requested ? undefined : requested;
return toggleChatMessageFeedback(current, requested);
}
function activityPhase(tool: PublicRectificationTool): NonNullable<ChatMessageView["activity"]>["phase"] {
@@ -584,48 +588,17 @@ export function RectificationAgenticChat(props: RectificationAgenticChatProps) {
/>
)}
{showActions && !regenerating && (
<div className="rectification-message-actions" aria-label="Agent 回答操作">
<button
aria-label="赞"
aria-pressed={feedback[message.renderKey] === "up"}
className={feedback[message.renderKey] === "up" ? "is-active" : ""}
title="赞"
type="button"
onClick={() => setFeedback((current) => ({
...current,
[message.renderKey]: toggleRectificationFeedback(current[message.renderKey], "up"),
}))}
>
<ThumbsUp aria-hidden="true" />
</button>
<button
aria-label="踩"
aria-pressed={feedback[message.renderKey] === "down"}
className={feedback[message.renderKey] === "down" ? "is-active" : ""}
title="踩"
type="button"
onClick={() => setFeedback((current) => ({
...current,
[message.renderKey]: toggleRectificationFeedback(current[message.renderKey], "down"),
}))}
>
<ThumbsDown aria-hidden="true" />
</button>
<button aria-label="复制回答" title="复制" type="button" onClick={() => void copyMessage(message)}>
{copiedMessageKey === message.renderKey
? <Check aria-hidden="true" />
: <Copy aria-hidden="true" />}
</button>
<button
aria-label="重新生成回答"
disabled={!canRegenerate}
title={canRegenerate ? "重新生成" : "只能重新生成最近一条回答"}
type="button"
onClick={() => void regenerateMessage(message)}
>
<RotateCcw aria-hidden="true" />
</button>
</div>
<ChatMessageActions
feedback={feedback[message.renderKey]}
copied={copiedMessageKey === message.renderKey}
canRegenerate={canRegenerate}
onFeedback={(requested) => setFeedback((current) => ({
...current,
[message.renderKey]: toggleRectificationFeedback(current[message.renderKey], requested),
}))}
onCopy={() => void copyMessage(message)}
onRegenerate={() => void regenerateMessage(message)}
/>
)}
</div>
);
frontend/src/lib/agent-observability.ts
+1
View File
@@ -180,6 +180,7 @@ export const logAgentObservability = createAgentObservabilityLogger();
const knownErrorCodes = new Set([
"runtime_contract_incomplete",
"empty_answer",
"answer_truncated",
"calculation_failed",
"timeout",
"cancelled",
frontend/src/lib/consultation-agent-events.ts
+1 -1
View File
@@ -100,7 +100,7 @@ const runCompletedSchema = z.object({ type: z.literal("run.completed"), receipt:
// and losing the whole failure event would be worse than losing its receipt.
const runFailedSchema = z.object({
type: z.literal("run.failed"),
code: z.enum(["runtime_contract_incomplete", "calculation_failed", "empty_answer", "cancelled"]),
code: z.enum(["runtime_contract_incomplete", "calculation_failed", "empty_answer", "answer_truncated", "cancelled"]),
message: z.string().max(200),
receipt: agentExecutionReceiptSchema.optional(),
}).strict();
frontend/src/lib/stream-agent-response.ts
+45 -21
View File
@@ -77,6 +77,27 @@ function safeToolError(error: unknown) {
return "calculation_failed" as const;
}
function isTimeoutOrAbort(error: unknown) {
return error instanceof Error && (error.name === "TimeoutError" || error.name === "AbortError");
}
type RunFailedCode = "runtime_contract_incomplete" | "empty_answer" | "answer_truncated" | "calculation_failed";
function runFailedCode(error: unknown, emitted: boolean): RunFailedCode {
if (error instanceof Error && error.message === "runtime_contract_incomplete") return "runtime_contract_incomplete";
if (error instanceof Error && error.message === "empty_answer") return "empty_answer";
if (error instanceof Error && error.message === "answer_truncated") return "answer_truncated";
if (emitted && isTimeoutOrAbort(error)) return "answer_truncated";
return "calculation_failed";
}
function runFailedMessage(code: RunFailedCode) {
if (code === "runtime_contract_incomplete") return "Agent 未完成必要的方法与计算步骤，本次不会扣点。";
if (code === "empty_answer") return "计算已完成，但这次没有生成回答，本次不会扣点。请再发送一次。";
if (code === "answer_truncated") return "回答未完成，已保留现有内容；本次不会扣点。";
return "咨询暂时无法完成，本次不会扣点。";
}
/**
* Whether a tool result is really an input rejection Mastra resolved with.
*
@@ -287,19 +308,26 @@ export function streamAgentResponse(options: StreamAgentResponseOptions) {
// A retry runs a second model loop under the same step budget, so the run
// total accumulates while the finish reason describes the latest attempt.
const stepCountBeforeAttempt = options.state.modelStepCount;
for await (const chunk of readChunks(stream)) {
for (const event of mapChunk(chunk, options, startedAt, toolErrors)) send(controller, event);
if (chunk.type === "step-finish") options.state.modelStepCount += 1;
if (chunk.type === "finish") {
const finish = finishTelemetry(chunk);
options.state.modelFinishReason = finish.reason;
if (finish.stepCount !== null) options.state.modelStepCount = stepCountBeforeAttempt + finish.stepCount;
}
if (chunk.type === "text-delta" && typeof chunk.payload?.text === "string") {
await outputText(visible.push(chunk.payload.text));
try {
for await (const chunk of readChunks(stream)) {
for (const event of mapChunk(chunk, options, startedAt, toolErrors)) send(controller, event);
if (chunk.type === "step-finish") options.state.modelStepCount += 1;
if (chunk.type === "finish") {
const finish = finishTelemetry(chunk);
options.state.modelFinishReason = finish.reason;
if (finish.stepCount !== null) options.state.modelStepCount = stepCountBeforeAttempt + finish.stepCount;
}
if (chunk.type === "text-delta" && typeof chunk.payload?.text === "string") {
await outputText(visible.push(chunk.payload.text));
}
}
await outputText(visible.finish(""));
} catch (error) {
try {
await outputText(visible.finish(""));
} catch {}
throw error;
}
await outputText(visible.finish(""));
}
const body = new ReadableStream<Uint8Array>({
@@ -330,6 +358,10 @@ export function streamAgentResponse(options: StreamAgentResponseOptions) {
// Still nothing to show. Failing is the honest outcome and it is the
// one that does not charge for the run.
if (!/\S/.test(fullOutput)) throw new Error("empty_answer");
// A spoken answer that stopped because the token budget ran out is
// not a completed consultation. The heading may already be on screen,
// so keep it and refuse to bill.
if (options.state.modelFinishReason === "length") throw new Error("answer_truncated");
settling = true;
const receipt = agentExecutionReceiptSchema.parse(options.receipt());
await options.onComplete?.(fullOutput, receipt);
@@ -344,11 +376,7 @@ export function streamAgentResponse(options: StreamAgentResponseOptions) {
try {
await options.onError?.(error, emitted, fullOutput);
} catch {}
const code = error instanceof Error && error.message === "runtime_contract_incomplete"
? "runtime_contract_incomplete" as const
: error instanceof Error && error.message === "empty_answer"
? "empty_answer" as const
: "calculation_failed" as const;
const code = runFailedCode(error, emitted);
// Step durations, the step budget and the workflow route are the only
// evidence the caller has for why a run failed. Building the receipt
// must not be able to replace the failure event with a silent close.
@@ -359,11 +387,7 @@ export function streamAgentResponse(options: StreamAgentResponseOptions) {
send(controller, {
type: "run.failed",
code,
message: code === "runtime_contract_incomplete"
? "Agent 未完成必要的方法与计算步骤，本次不会扣点。"
: code === "empty_answer"
? "计算已完成，但这次没有生成回答，本次不会扣点。请再发送一次。"
: "咨询暂时无法完成，本次不会扣点。",
message: runFailedMessage(code),
...(failureReceipt ? { receipt: failureReceipt } : {}),
});
if (!disconnected) controller.close();
frontend/src/mastra/consultation-tools.ts
+28
View File
@@ -42,6 +42,7 @@ import {
// calculation inside the 110s budget—so it is measured, not guessed.
export const AGENT_MAX_STEPS = 8;
export const AGENT_TIMEOUT_MS = 110_000;
export const CONSULTATION_MAX_OUTPUT_TOKENS = 8192;
const CONSULTATION_DOMAIN_DURATION_MS = 21_000;
const CONSULTATION_ANSWER_RESERVE_MS = 45_000;
export const CONSULTATION_DOMAIN_WALL_CLOCK_MS = AGENT_TIMEOUT_MS - CONSULTATION_ANSWER_RESERVE_MS;
@@ -50,6 +51,33 @@ export const MAX_CONSULTATION_DOMAINS = Math.max(
Math.floor(CONSULTATION_DOMAIN_WALL_CLOCK_MS / CONSULTATION_DOMAIN_DURATION_MS),
);
const thinkingDisabled = { thinking: { type: "disabled" as const } };
/**
* Visible-answer generation settings for one consult stream.
*
* DeepSeek V4 Flash thinks by default, and those hidden tokens share
* `max_tokens` with the spoken answer. Without an explicit visible budget and
* thinking turned off, a finished-looking stream can stop mid-heading with
* `finish_reason=length`. The provider id is repeated under `openai` because
* OpenAI-compatible adapters often look there first.
*/
export function consultationGenerationSettings(model?: unknown) {
const providerId = typeof model === "string"
? model
: model && typeof model === "object" && "providerId" in model && typeof model.providerId === "string"
? model.providerId
: undefined;
const providerOptions: Record<string, typeof thinkingDisabled> = {
openai: thinkingDisabled,
};
if (providerId) providerOptions[providerId] = thinkingDisabled;
return {
modelSettings: { maxOutputTokens: CONSULTATION_MAX_OUTPUT_TOKENS },
providerOptions,
};
}
// The raw plan bound stays at the registry default so a duplicate-heavy list
// canonicalizes instead of failing outright. The executable cap is enforced
// after canonicalization, where it can degrade and disclose rather than throw.