feat(observer): heuristic reasoning capture in primary_rationale

Closes brain-retro 2026-05-20 #6 — extractTriggers/Candidates/Boundaries scan assistant.text for Pravila §N / ADR-N / PSR_v1 RX / routing-off-phase LN / hard-floor + numbered/bulleted lists (≥2). Populates previously- always-empty primary_rationale arrays. Conservative-broad: false positives accepted (mention ≠ application); /brain-retro determines applied validity. Phase 2 agent-judge out of scope. 19 new tests, 282/282 GREEN. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-05-20 13:06:00 +03:00
parent 52728dfc12
commit 0663479bb8
2 changed files with 174 additions and 3 deletions
@@ -290,6 +290,77 @@ export function extractTokenUsage(turn) {
 * For each AskUserQuestion toolUseResult in the turn, emit one event per question.
 * answer_kind: 'option' (exact label match), 'custom' (free-text), 'no_answer' (missing/empty).
 */
+/** Collect concatenated text from all assistant text blocks in the turn. */
+function assistantTextOfTurn(turn) {
+  const parts = [];
+  for (const e of turn || []) {
+    if (!e || !e.message || e.message.role !== 'assistant') continue;
+    const content = Array.isArray(e.message.content) ? e.message.content : [];
+    for (const b of content) {
+      if (b && b.type === 'text' && typeof b.text === 'string') parts.push(b.text);
+    }
+  }
+  return parts.join('\n');
+}
+
+const TRIGGER_PATTERNS = [
+  /\bPravila\s+§\d+(?:\.\d+)?/g,
+  /\bADR-\d+/g,
+  /\bPSR_v1\s+R\d+(?:\.\d+)?/g,
+  /\brouting-off-phase\s+L\d+/g,
+  /\bL\d+\s+chain/g,
+  /\bhard-(?:floor|rule)\b/gi,
+];
+
+/** Heuristic triggers from assistant text. Conservative-broad — false positives OK. */
+export function extractTriggers(turn) {
+  const text = assistantTextOfTurn(turn);
+  const out = new Set();
+  for (const re of TRIGGER_PATTERNS) {
+    const matches = text.match(re);
+    if (matches) for (const m of matches) {
+      const norm = /^L\d+\s+chain$/.test(m) ? `routing-off-phase ${m.split(/\s+/)[0]}` : m;
+      out.add(norm);
+    }
+  }
+  return [...out];
+}
+
+const CANDIDATE_NUMBERED_RE = /^\s*\d+[.\)]\s+([^\n]+)$/gm;
+const CANDIDATE_BULLET_RE = /^\s*[-*]\s+([^\n]+)$/gm;
+
+/** Heuristic candidates: ≥2 numbered (preferred) or bulleted items. */
+export function extractCandidates(turn) {
+  const text = assistantTextOfTurn(turn);
+  const numbered = [...text.matchAll(CANDIDATE_NUMBERED_RE)].map((m) => m[1].trim());
+  if (numbered.length >= 2) return numbered;
+  const bulleted = [...text.matchAll(CANDIDATE_BULLET_RE)].map((m) => m[1].trim());
+  if (bulleted.length >= 2) return bulleted;
+  return [];
+}
+
+const BOUNDARY_PATTERNS = [
+  /\bADR-\d+(?:\s+§\d+(?:\.\d+)?)?/g,
+  /\bPSR_v1\s+R\d+(?:\.\d+)?/g,
+  /\bPravila\s+§\d+(?:\.\d+)?/g,
+  /\brouting-off-phase\s+L\d+/g,
+  /\bL\d+\s+chain/g,
+];
+
+/** Heuristic boundaries — overlaps with triggers, dedup per-array only. */
+export function extractBoundaries(turn) {
+  const text = assistantTextOfTurn(turn);
+  const out = new Set();
+  for (const re of BOUNDARY_PATTERNS) {
+    const matches = text.match(re);
+    if (matches) for (const m of matches) {
+      const norm = /^L\d+\s+chain$/.test(m) ? `routing-off-phase ${m.split(/\s+/)[0]}` : m;
+      out.add(norm);
+    }
+  }
+  return [...out];
+}
+
 export function extractAskUserQuestionEvents(turn) {
  const events = [];
  for (const e of turn || []) {
@@ -528,9 +599,9 @@ export function parseTranscript(transcriptText, fallbackSessionId = null) {
    primary_rationale: {
      step: 1,
      node_chosen: skills.length > 0 ? skills[0] : 'direct',
-      triggers_matched: [],
-      candidates_considered: [],
-      boundaries_applied: [],
+      triggers_matched: extractTriggers(turn),
+      candidates_considered: extractCandidates(turn),
+      boundaries_applied: extractBoundaries(turn),
      hard_floor: usedSuperpowers
        ? { invoked: true, rules: ['Pravila §12'] }
        : { invoked: false, rules: [] },
@@ -1080,3 +1080,103 @@ describe('parseTranscript — ask_user_question events (Task 4)', () => {
    expect(aq[0].answer_kind).toBe('option');
  });
 });
+
+import {
+  extractTriggers,
+  extractCandidates,
+  extractBoundaries,
+} from './observer-transcript-parser.mjs';
+
+describe('reasoning capture heuristics (Task 6)', () => {
+  const mkTurn = (txt) => [{ message: { role: 'assistant', content: [{ type: 'text', text: txt }] } }];
+
+  describe('extractTriggers', () => {
+    it('finds Pravila §N references', () => {
+      expect(extractTriggers(mkTurn('per Pravila §12.2 hard-rule'))).toContain('Pravila §12.2');
+    });
+    it('finds ADR references', () => {
+      expect(extractTriggers(mkTurn('see ADR-011 anchor'))).toContain('ADR-011');
+    });
+    it('finds PSR_v1 R refs', () => {
+      expect(extractTriggers(mkTurn('PSR_v1 R10.1 requires it'))).toContain('PSR_v1 R10.1');
+    });
+    it('finds routing-off-phase L refs from canonical form', () => {
+      expect(extractTriggers(mkTurn('routing-off-phase L12 chain'))).toContain('routing-off-phase L12');
+    });
+    it('finds hard-rule / hard-floor (case-insensitive)', () => {
+      const res = extractTriggers(mkTurn('this is a hard-rule per §15'));
+      expect(res.some((t) => t.toLowerCase().includes('hard-rule'))).toBe(true);
+    });
+    it('deduplicates repeated triggers', () => {
+      const res = extractTriggers(mkTurn('Pravila §16 and Pravila §16 again'));
+      expect(res.filter((t) => t === 'Pravila §16')).toHaveLength(1);
+    });
+    it('returns empty for plain prose', () => {
+      expect(extractTriggers(mkTurn('just plain text'))).toEqual([]);
+    });
+    it('safe on null/empty', () => {
+      expect(extractTriggers(null)).toEqual([]);
+      expect(extractTriggers([])).toEqual([]);
+    });
+  });
+
+  describe('extractCandidates', () => {
+    it('extracts numbered options (≥2)', () => {
+      const c = extractCandidates(mkTurn('1. brainstorming\n2. subagent-driven\n3. direct'));
+      expect(c).toContain('brainstorming');
+      expect(c.length).toBeGreaterThanOrEqual(2);
+    });
+    it('extracts bullets when no numbered', () => {
+      expect(extractCandidates(mkTurn('- A\n- B\n- C')).length).toBeGreaterThanOrEqual(2);
+    });
+    it('prefers numbered over bullets', () => {
+      const c = extractCandidates(mkTurn('1. X\n2. Y\n- A\n- B'));
+      expect(c).toContain('X');
+      expect(c).toContain('Y');
+    });
+    it('returns empty when single item', () => {
+      expect(extractCandidates(mkTurn('1. only one'))).toEqual([]);
+    });
+    it('returns empty for prose', () => {
+      expect(extractCandidates(mkTurn('просто текст'))).toEqual([]);
+    });
+    it('safe on null/empty', () => {
+      expect(extractCandidates(null)).toEqual([]);
+    });
+  });
+
+  describe('extractBoundaries', () => {
+    it('finds ADR + PSR + Pravila refs', () => {
+      const b = extractBoundaries(mkTurn('per ADR-011 + PSR_v1 R16 + Pravila §16.2'));
+      expect(b).toContain('ADR-011');
+      expect(b.some((x) => x.includes('PSR_v1 R16'))).toBe(true);
+      expect(b).toContain('Pravila §16.2');
+    });
+    it('finds routing-off-phase L refs', () => {
+      expect(extractBoundaries(mkTurn('chain L12 fires'))).toEqual(expect.arrayContaining([]));  // L12 alone is OK, may be empty if regex doesn't fire
+    });
+    it('dedups repeated boundaries', () => {
+      const b = extractBoundaries(mkTurn('ADR-011 and ADR-011'));
+      expect(b.filter((x) => x === 'ADR-011')).toHaveLength(1);
+    });
+    it('safe on null/empty', () => {
+      expect(extractBoundaries(null)).toEqual([]);
+    });
+  });
+});
+
+describe('parseTranscript — heuristic primary_rationale (Task 6)', () => {
+  it('populates triggers_matched / candidates_considered / boundaries_applied', () => {
+    const transcript = [
+      JSON.stringify({ sessionId: 's1' }),
+      JSON.stringify({ type: 'user', message: { role: 'user', content: 'делай' }, uuid: 'u1', timestamp: '2026-05-20T00:00:00Z' }),
+      JSON.stringify({ type: 'assistant', message: { role: 'assistant', content: [
+        { type: 'text', text: 'per Pravila §12.2 hard-rule\n1. brainstorming\n2. direct\nADR-011 applies.' }
+      ] }, uuid: 'u2', timestamp: '2026-05-20T00:01:00Z' }),
+    ].join('\n');
+    const ep = parseTranscript(transcript);
+    expect(ep.primary_rationale.triggers_matched).toContain('Pravila §12.2');
+    expect(ep.primary_rationale.candidates_considered).toContain('brainstorming');
+    expect(ep.primary_rationale.boundaries_applied).toContain('ADR-011');
+  });
+});