Files
Jyotisha/frontend/tests/rectification-server-focus.test.ts
T
Jesse_Chen 7de36e11d6
Independent Staging Quality Gate / validate (push) Successful in 27m23s
Independent Staging Quality Gate / publish (push) Successful in 52m48s
fix(rectification): align leftover prompt and focus tests with the shrunk agent contract
Independent Staging Quality Gate 2141 failed because four tests still matched the old Mastra prompt wall and treated an unrenderable focus as null.

Co-authored-by: Cursor <cursoragent@cursor.com>
2026-08-28 20:56:58 +08:00

435 lines
14 KiB
TypeScript

import assert from "node:assert/strict";
import test from "node:test";
import { buildInferenceState } from "../src/lib/rectification-agentic/core/build-state.ts";
import {
openQuestionFromPersistedFocus,
persistServerOwnedFocus,
shouldSkipDiscriminatorFollowup,
stableFollowupQuestionId,
} from "../src/lib/rectification-agentic/v9/server-focus.ts";
import { buildChoiceFrame, serverOwnedChoiceCopy } from "../src/lib/rectification-agentic/v9/choice-card.ts";
import type { MethodFollowup } from "../src/lib/rectification-agentic/v9/method-followup.ts";
import type { EventProbeStyleOption } from "../src/lib/rectification-agentic/v9/refinement-packet.ts";
import { fakeAccounting, CASE_ID, FOCUS_ID, USER_ID, activeFocusFixture } from "./rectification-v9-test-support.ts";
const DYNAMIC_STYLE_OPTIONS = [
{ label: "明确发生且时间吻合", answer_class: "yes" },
{ label: "发生过但程度较弱", answer_class: "weak_yes" },
{ label: "明确没有发生", answer_class: "no" },
{ label: "这段记不清楚", answer_class: "unsure" },
] as const;
function discriminatorFollowup(overrides: Partial<MethodFollowup> = {}): MethodFollowup {
const frame = buildChoiceFrame({
method_id: "dasha_events",
ask_theme: "dated_event",
domain: "education",
user_prompt_hint: "ask",
}, {
probes: [{
year: 2016,
year_label: "2016 年前后",
domain: "education",
event_family: "升学结果或学习环境出现明显变化",
source: "dasha_activation",
tracks: ["vimshottari", "narayana"],
tracks_agree: true,
unique_minute_claim: false,
user_meaning: "2016 年前后升学结果或学习环境出现明显变化",
role: "distinguish",
information_gain: 0.4,
semantic_key: "education:2016",
style_options: DYNAMIC_STYLE_OPTIONS,
}],
});
assert.ok(frame);
return {
method_id: "dasha_events",
intent: "distinguish_candidates",
ask_theme: "dated_event",
domain: "education",
kind_hint: null,
user_prompt_hint: "ask",
must_not_label: false,
choice_frame: frame,
source: "event_probe",
information_gain: 0.4,
semantic_key: "education:2016",
probe_year: 2016,
candidate_ids: ["05:00", "05:20"],
expected_outcomes: [
{ answer_class: "yes", supports: ["05:00"], conflicts: ["05:20"] },
{ answer_class: "no", supports: ["05:20"], conflicts: ["05:00"] },
],
...overrides,
};
}
function invalidStyleFollowup(
styleOptions: readonly EventProbeStyleOption[] | undefined,
eventFamily = "升学结果或学习环境出现明显变化",
choiceKind: "existence" | "varga_style" | "event_quality" = "existence",
): MethodFollowup {
const valid = discriminatorFollowup();
return {
...valid,
choice_kind: choiceKind,
choice_frame: buildChoiceFrame({
method_id: "dasha_events",
ask_theme: "dated_event",
domain: "education",
user_prompt_hint: "ask",
choice_kind: choiceKind,
}, {
probes: [{
year: 2016,
year_label: "2016 年前后",
domain: "education",
event_family: eventFamily,
source: "dasha_activation",
tracks: ["vimshottari", "narayana"],
tracks_agree: true,
unique_minute_claim: false,
user_meaning: "2016 年前后升学结果或学习环境出现明显变化",
role: "distinguish",
information_gain: 0.4,
semantic_key: "education:2016",
choice_kind: choiceKind,
style_options: styleOptions,
}],
}),
};
}
async function assertInvalidChoiceSkipped(followup: MethodFollowup) {
const accounting = fakeAccounting({}, {
fallback: () => {
throw new Error("invalid choice frame must not call the database");
},
});
const result = await persistServerOwnedFocus({
accounting: accounting.client,
userId: USER_ID,
caseId: CASE_ID,
activeFocus: null,
decisionReceipt: null,
followup,
});
assert.equal(result.status, "invalid_choice_schema");
assert.equal(accounting.calls.length, 0);
}
function persistedFocus(questionId = "probe:education:2016") {
const copy = serverOwnedChoiceCopy(discriminatorFollowup().choice_frame!);
assert.ok(copy);
return {
id: FOCUS_ID,
caseId: CASE_ID,
questionId,
intent: "distinguish_candidates",
targetEvidenceId: null,
targetDomain: "education",
targetKind: null,
expectedAnswerSchema: {
choice: copy,
semantic_key: "education:2016",
candidate_split_hash: "education:2016",
},
status: "active" as const,
askedAt: "2026-08-27T00:00:00.000Z",
resolvedAt: null,
};
}
test("only a successfully persisted focus can expose its discriminator question", () => {
const focus = persistedFocus();
for (const status of ["created", "already_open"] as const) {
assert.deepEqual(openQuestionFromPersistedFocus({
status,
focus,
questionId: focus.questionId,
prompt: "2016 年前后 · 升学结果或学习环境出现明显变化",
}), {
question_id: focus.questionId,
prompt: "2016 年前后 · 升学结果或学习环境出现明显变化",
status,
});
}
});
test("an unpersisted, mismatched, or incomplete focus cannot expose a discriminator question", () => {
const focus = persistedFocus();
assert.equal(openQuestionFromPersistedFocus({
status: "duplicate_focus",
focus,
questionId: focus.questionId,
prompt: "不能展示",
}), null);
assert.equal(openQuestionFromPersistedFocus({
status: "created",
focus: { ...focus, questionId: "another-question" },
questionId: focus.questionId,
prompt: "不能展示",
}), null);
const incomplete = openQuestionFromPersistedFocus({
status: "created",
focus: { ...focus, expectedAnswerSchema: { choice: { prompt: "不完整" } } },
questionId: focus.questionId,
prompt: "不能展示",
});
assert.equal(incomplete?.unrenderable, true);
assert.equal(incomplete?.prompt, null);
assert.equal(incomplete?.reason, "invalid_choice_schema");
});
test("complete dynamic event options persist a server-owned focus", async () => {
const followup = discriminatorFollowup({ source: "precision_stage", information_gain: 0.2 });
const accounting = fakeAccounting({
set_agentic_rectification_conversation_focus: (_fn, args) => ({
id: FOCUS_ID,
case_id: CASE_ID,
question_id: args.p_question_id,
intent: args.p_intent,
target_evidence_id: null,
target_domain: args.p_target_domain,
target_kind: args.p_target_kind,
expected_answer_schema: args.p_expected_answer_schema,
status: "active",
asked_at: "2026-08-27T00:00:00.000Z",
resolved_at: null,
idempotent: false,
}),
});
const result = await persistServerOwnedFocus({
accounting: accounting.client,
userId: USER_ID,
caseId: CASE_ID,
activeFocus: null,
decisionReceipt: null,
followup,
});
assert.equal(result.status, "created");
assert.equal(accounting.calls.length, 1);
assert.deepEqual(result.focus?.expectedAnswerSchema.choice, {
prompt: "2016 年前后,有没有升学结果或学习环境出现明显变化?",
option_a: DYNAMIC_STYLE_OPTIONS[0].label,
option_b: DYNAMIC_STYLE_OPTIONS[1].label,
option_c: DYNAMIC_STYLE_OPTIONS[2].label,
option_d: DYNAMIC_STYLE_OPTIONS[3].label,
options: [
{ key: "A", label: DYNAMIC_STYLE_OPTIONS[0].label, answer_class: "yes" },
{ key: "B", label: DYNAMIC_STYLE_OPTIONS[1].label, answer_class: "weak_yes" },
{ key: "C", label: DYNAMIC_STYLE_OPTIONS[2].label, answer_class: "no" },
{ key: "D", label: DYNAMIC_STYLE_OPTIONS[3].label, answer_class: "unsure" },
],
});
});
test("non-renderable varga styles and empty event family skip focus persistence", async () => {
await assertInvalidChoiceSkipped(invalidStyleFollowup(
[{ label: "巨蟹相处主动热情", answer_class: "yes" }],
"升学结果或学习环境出现明显变化",
"varga_style",
));
await assertInvalidChoiceSkipped(invalidStyleFollowup(DYNAMIC_STYLE_OPTIONS, " "));
});
test("does not ask an already-answered discriminator probe again", async () => {
const followup = discriminatorFollowup();
const active = activeFocusFixture({
expectedAnswerSchema: { probe_id: "education:2016", choice: serverOwnedChoiceCopy(followup.choice_frame!) },
});
const accounting = fakeAccounting({
set_agentic_rectification_conversation_focus: () => {
throw new Error("should not create a second focus");
},
});
const result = await persistServerOwnedFocus({
accounting: accounting.client,
userId: USER_ID,
caseId: CASE_ID,
activeFocus: {
...active,
id: FOCUS_ID,
caseId: CASE_ID,
questionId: stableFollowupQuestionId(followup),
intent: "distinguish_candidates",
targetEvidenceId: null,
targetDomain: "education",
targetKind: null,
expectedAnswerSchema: { probe_id: "education:2016" },
status: "active",
askedAt: "2026-08-24T00:00:00.000Z",
resolvedAt: null,
},
decisionReceipt: {
inference_state: {
answered_probes: [{ probe_id: "education:2016", semantic_key: "education:2016", answer_class: "yes" }],
probes: [],
},
},
followup,
});
assert.equal(result.status, "probe_already_answered");
assert.equal(accounting.calls.length, 0);
});
test("zero information gain does not open a discriminator", () => {
assert.equal(
shouldSkipDiscriminatorFollowup(discriminatorFollowup({ information_gain: 0 })),
"zero_information_gain",
);
});
test("recorded event quality cannot open a zero-information scoring card", () => {
assert.equal(
shouldSkipDiscriminatorFollowup(discriminatorFollowup({
choice_kind: "event_quality",
information_gain: 0,
})),
"zero_information_gain",
);
});
test("an unmatched scoring identity is skipped instead of being synthesized", async () => {
const followup = discriminatorFollowup({
semantic_key: "education.synthetic",
candidate_split_hash: "education:synthetic",
});
const accounting = fakeAccounting({
set_agentic_rectification_conversation_focus: () => {
throw new Error("must not persist an unmatched scoring focus");
},
});
const result = await persistServerOwnedFocus({
accounting: accounting.client,
userId: USER_ID,
caseId: CASE_ID,
activeFocus: null,
decisionReceipt: {
inference_state: {
algorithm_version: "birth-time-event-scoring-v1",
candidates: [
{ id: "05:00", birth_time: "05:00", posterior: 0.5 },
{ id: "05:20", birth_time: "05:20", posterior: 0.5 },
],
answered_probes: [],
probes: [{
id: "p-live",
semantic_key: "education.live",
candidate_split_hash: "education:live",
domain: "education",
year: 2016,
question: "live",
candidate_ids: ["05:00", "05:20"],
expected_outcomes: [
{ answer_class: "yes", supports: ["05:00"], conflicts: ["05:20"] },
{ answer_class: "no", supports: ["05:20"], conflicts: ["05:00"] },
],
information_gain: 0.2,
source: "dasha_activation",
}],
},
},
followup,
});
assert.equal(result.status, "invalid_choice_schema");
assert.equal(accounting.calls.length, 0);
});
test("duplicate focus conflict does not throw", async () => {
const followup = discriminatorFollowup({ source: "precision_stage", information_gain: 0.2 });
const accounting = fakeAccounting({
set_agentic_rectification_conversation_focus: () => {
throw new Error("agentic_rectification_focus_idempotency_conflict");
},
});
const result = await persistServerOwnedFocus({
accounting: accounting.client,
userId: USER_ID,
caseId: CASE_ID,
activeFocus: null,
decisionReceipt: null,
followup,
});
assert.equal(result.status, "duplicate_focus");
});
test("contrast probe is not replaced by an already-answered education quality probe", async () => {
const followup = discriminatorFollowup({
domain: "education",
ask_theme: "education_style",
information_gain: 0.16,
semantic_key: "varga.d24.05:00/05:06|05:07",
candidate_split_hash: "varga.d24.05:00/05:06|05:07",
probe_year: undefined,
});
const accounting = fakeAccounting({
set_agentic_rectification_conversation_focus: (_fn, args) => ({
id: FOCUS_ID,
case_id: CASE_ID,
question_id: args.p_question_id,
intent: args.p_intent,
target_evidence_id: null,
target_domain: args.p_target_domain,
target_kind: args.p_target_kind,
expected_answer_schema: args.p_expected_answer_schema,
status: "active",
asked_at: "2026-08-25T00:00:00.000Z",
resolved_at: null,
idempotent: false,
}),
});
const result = await persistServerOwnedFocus({
accounting: accounting.client,
userId: USER_ID,
caseId: CASE_ID,
activeFocus: null,
decisionReceipt: {
inference_state: buildInferenceState({
range_start: "05:00",
range_end: "05:20",
candidates: [
{ id: "05:00", time: "05:00", relative_support: 10 },
{ id: "05:20", time: "05:20", relative_support: 10 },
],
events: [],
probes: [{
id: "probe:education.2016",
semantic_key: "education.2016",
candidate_split_hash: "education:2016",
domain: "education",
year: 2016,
question: "发挥失常",
candidate_ids: [],
expected_outcomes: [],
information_gain: 0,
source: "known_event_quality",
}, {
id: "contrast:varga.d24.05:00/05:06|05:07",
semantic_key: "varga.d24.05:00/05:06|05:07",
candidate_split_hash: "varga.d24.05:00/05:06|05:07",
domain: "education",
year: 0,
question: "学业盘候选差异",
candidate_ids: ["05:00", "05:06", "05:07"],
expected_outcomes: [
{ answer_class: "yes", supports: ["05:00"], conflicts: ["05:06", "05:07"] },
{ answer_class: "no", supports: ["05:06", "05:07"], conflicts: ["05:00"] },
],
information_gain: 0.16,
source: "varga_contrast",
}],
}),
},
followup,
});
assert.equal(result.status, "created");
assert.ok(result.focus);
const schema = result.focus.expectedAnswerSchema;
assert.equal(schema.semantic_key, "varga.d24.05:00/05:06|05:07");
assert.notEqual(schema.probe_id, "probe:education.2016");
assert.match(String(schema.probe_id), /varga\.d24/);
});