feat: compare Vimshottari case timing states

frontend/src/mastra/index.ts

@@ -113,6 +113,7 @@ When reference_transparency is present:
 - Present candidate_windows and exact_triggers when relevant, but describe exact_triggers as technical trigger points, never guaranteed events.
 - Share a public case only when similar_public_cases.status is high_similarity_public_references_available. State the listed matching factors, dissimilar factors, event source URL, and that the case is reference-only.
 - If a shared case has reference_status public_context_only, state that it has not been replayed for calibration and cannot increase timing confidence.
+- Treat similarity.timing_state as authoritative: only status matched permits saying Vimshottari Mahadasha matches; different or not_compared must be described as such, and never imply Antardasha, Narayana, or transits also match.
 - When similar_public_cases.coverage.requested_uncovered_domains is non-empty, say the current public-case catalog does not yet cover those themes; do not infer that no comparable real-world case exists.
 - When method_variants applies, present parallel methods and their source paths rather than silently picking one result as the only truth.
 - If should_lead_with_limitations is false, do not lead with limitations. If a limitation is relevant, put it in one short sentence at the end.

frontend/tests/consultation-context.test.ts

@@ -9,5 +9,6 @@ test("passes transparent public-case references into the agent context", () => {
   assert.match(source, /high_similarity_public_references_available/);
   assert.match(source, /requested_uncovered_domains/);
   assert.match(source, /public_context_only/);
+  assert.match(source, /timing_state/);
   assert.match(source, /exact_triggers as technical trigger points/);
 });

scripts/jyotish_api_server.py

@@ -720,6 +720,7 @@ def execute_consultation_workflow(
         chart,
         themes,
         timing=body.get('timing'),
+        reference_date=_consultation_reference_date(body).date().isoformat(),
     )
     return result
 

scripts/reference_transparency_contract.py

@@ -5,6 +5,7 @@ from __future__ import annotations
 
 import json
 import sys
+from datetime import datetime
 from functools import lru_cache
 from pathlib import Path
 from typing import Any
@@ -14,11 +15,11 @@ if str(SCRIPT_DIR) not in sys.path:
     sys.path.insert(0, str(SCRIPT_DIR))
 
 try:
-    from scripts.domain_calculation_service import compute_chart
+    from scripts.domain_calculation_service import compute_chart, compute_vimshottari_timeline
     from scripts.timing_precision_contract import build_timing_precision_contract
     from scripts.varga import SIGN_LORDS, calc_varga
 except ModuleNotFoundError:  # pragma: no cover - direct script execution
-    from domain_calculation_service import compute_chart
+    from domain_calculation_service import compute_chart, compute_vimshottari_timeline
     from timing_precision_contract import build_timing_precision_contract
     from varga import SIGN_LORDS, calc_varga
 
@@ -41,6 +42,7 @@ FEATURE_WEIGHTS = {
     "d9_venus": 0.10,
     "d10_ascendant": 0.10,
     "d10_sun": 0.10,
+    "vimshottari_mahadasha": 0.15,
 }
 HIGH_SIMILARITY_THRESHOLD = 0.75
 
@@ -85,6 +87,43 @@ def _varga_sign(chart: dict[str, Any], key: str, division: int) -> str | None:
     return calc_varga(longitude, division)["sign"] if longitude is not None else None
 
 
+def _birth_datetime(chart: dict[str, Any]) -> datetime | None:
+    birth = chart.get("birth_info") if isinstance(chart, dict) else None
+    if not isinstance(birth, dict):
+        birth = chart.get("birth") if isinstance(chart, dict) else None
+    if not isinstance(birth, dict) or not isinstance(birth.get("date"), str):
+        return None
+    raw_time = birth.get("time")
+    if not isinstance(raw_time, str):
+        raw_time = f"{int(birth.get('hour', 0)):02d}:{int(birth.get('minute', 0)):02d}:{int(birth.get('second', 0)):02d}"
+    try:
+        return datetime.fromisoformat(f"{birth['date']}T{raw_time}")
+    except ValueError:
+        return None
+
+
+def _vimshottari_mahadasha(chart: dict[str, Any], reference_date: str | None) -> str | None:
+    if not isinstance(reference_date, str):
+        return None
+    try:
+        target = datetime.fromisoformat(reference_date[:10])
+    except ValueError:
+        return None
+    birth_dt, moon_lon = _birth_datetime(chart), _longitude(chart, "Moon")
+    if birth_dt is None or moon_lon is None:
+        return None
+    try:
+        periods = compute_vimshottari_timeline(birth_dt=birth_dt, moon_lon=moon_lon)["periods"]
+        for period in periods:
+            start = datetime.fromisoformat(period["start"])
+            end = datetime.fromisoformat(period["end"])
+            if start <= target <= end:
+                return period["lord"]
+    except (KeyError, TypeError, ValueError):
+        return None
+    return None
+
+
 def _features(chart: dict[str, Any], domain: str) -> dict[str, Any]:
     ascendant = chart.get("ascendant") if isinstance(chart, dict) else None
     rahu, ketu = _sign(chart, "Rahu"), _sign(chart, "Ketu")
@@ -136,9 +175,26 @@ def _chart_for_case(case: dict[str, Any]) -> dict[str, Any] | None:
     )
 
 
-def _similarity(user_chart: dict[str, Any], case_chart: dict[str, Any], domain: str) -> dict[str, Any]:
+def _similarity(
+    user_chart: dict[str, Any], case_chart: dict[str, Any], domain: str,
+    *, reference_date: str | None = None, case_event_date: str | None = None,
+) -> dict[str, Any]:
     user = _features(user_chart, domain)
     candidate = _features(case_chart, domain)
+    user_md = _vimshottari_mahadasha(user_chart, reference_date)
+    case_md = _vimshottari_mahadasha(case_chart, case_event_date)
+    if user_md is not None and case_md is not None:
+        user["vimshottari_mahadasha"] = user_md
+        candidate["vimshottari_mahadasha"] = case_md
+        timing_state = {
+            "status": "matched" if user_md == case_md else "different",
+            "user_vimshottari_mahadasha": user_md,
+            "case_event_vimshottari_mahadasha": case_md,
+            "reference_date": reference_date,
+            "case_event_date": case_event_date,
+        }
+    else:
+        timing_state = {"status": "not_compared"}
     matching, dissimilar, total = [], [], 0.0
     for name, weight in FEATURE_WEIGHTS.items():
         if name not in user or name not in candidate:
@@ -156,7 +212,9 @@ def _similarity(user_chart: dict[str, Any], case_chart: dict[str, Any], domain:
         compared_vargas.append("D9")
     if domain == "career" and all(user.get(name) is not None and candidate.get(name) is not None for name in ("d10_ascendant", "d10_sun")):
         compared_vargas.append("D10")
-    uncompared = ["dasha_event_state", "transit_event_state"]
+    uncompared = ["vimshottari_antardasha", "narayana_dasha", "transit_event_state"]
+    if timing_state["status"] == "not_compared":
+        uncompared.insert(0, "vimshottari_mahadasha")
     if domain == "marriage" and "D9" not in compared_vargas:
         uncompared.insert(0, "D9")
     if domain == "career" and "D10" not in compared_vargas:
@@ -165,8 +223,9 @@ def _similarity(user_chart: dict[str, Any], case_chart: dict[str, Any], domain:
         "score": score,
         "matching_factors": matching,
         "dissimilar_factors": dissimilar,
-        "feature_scope": "D1 ascendant, Moon, theme-house lord, Rahu/Ketu axis" + (f", {'/'.join(compared_vargas)}" if compared_vargas else ""),
+        "feature_scope": "D1 ascendant, Moon, theme-house lord, Rahu/Ketu axis" + (f", {'/'.join(compared_vargas)}" if compared_vargas else "") + (", Vimshottari MD" if timing_state["status"] != "not_compared" else ""),
         "uncompared_layers": uncompared,
+        "timing_state": timing_state,
     }
 
 
@@ -207,6 +266,7 @@ def select_similar_public_cases(
     themes: list[str],
     *,
     cases: list[dict[str, Any]] | None = None,
+    reference_date: str | None = None,
     threshold: float = HIGH_SIMILARITY_THRESHOLD,
     max_cases: int = 3,
 ) -> dict[str, Any]:
@@ -231,7 +291,13 @@ def select_similar_public_cases(
         for event in case.get("event_outcomes", []):
             if not isinstance(event, dict) or event.get("domain") not in themes:
                 continue
-            similarity = _similarity(user_chart, case_chart, event["domain"])
+            similarity = _similarity(
+                user_chart,
+                case_chart,
+                event["domain"],
+                reference_date=reference_date,
+                case_event_date=event.get("event_date"),
+            )
             if similarity["score"] < threshold:
                 continue
             source = case.get("source") if isinstance(case.get("source"), dict) else {}
@@ -270,7 +336,7 @@ def select_similar_public_cases(
 
 def build_reference_transparency_contract(
     chart: dict[str, Any], themes: list[str], *, timing: dict[str, Any] | None = None,
-    cases: list[dict[str, Any]] | None = None,
+    cases: list[dict[str, Any]] | None = None, reference_date: str | None = None,
 ) -> dict[str, Any]:
     timing_contract = build_timing_precision_contract(timing)
     return {
@@ -296,5 +362,10 @@ def build_reference_transparency_contract(
             "source": "references/oracle/xalen_formula_unit_attribution_2026_07_17.json",
             "boundary": "流派/公式差异并列展示；不以单一引擎多数投票决定真值。",
         },
-        "similar_public_cases": select_similar_public_cases(chart, themes, cases=cases),
+        "similar_public_cases": select_similar_public_cases(
+            chart,
+            themes,
+            cases=cases,
+            reference_date=reference_date,
+        ),
     }

tests/test_reference_transparency_contract.py

@@ -106,6 +106,27 @@ def test_default_public_manifest_can_surface_a_matching_replayed_case() -> None:
     assert selected["cases"][0]["reference_only"] is True
 
 
+def test_same_event_date_compares_vimshottari_mahadasha() -> None:
+    from scripts.domain_calculation_service import compute_chart
+
+    jobs_chart = compute_chart({
+        "year": 1955, "month": 2, "day": 24, "hour": 19, "minute": 15,
+        "lat": 37.7833, "lon": -122.4167, "tz": -8.0,
+        "ayanamsa": "lahiri", "node_mode": "mean",
+    })
+
+    selected = select_similar_public_cases(
+        jobs_chart,
+        ["career"],
+        reference_date="2007-01-09",
+    )
+
+    similarity = selected["cases"][0]["similarity"]
+    assert "vimshottari_mahadasha" in similarity["matching_factors"]
+    assert similarity["timing_state"]["status"] == "matched"
+    assert "vimshottari_antardasha" in similarity["uncompared_layers"]
+
+
 def test_pending_health_case_is_context_only_not_calibration() -> None:
     user_chart = _chart("Leo", "Pisces", "Libra")
     cases = [{
@@ -203,3 +224,4 @@ def test_default_manifest_exposes_kahlo_health_as_context_only() -> None:
 def test_consultation_api_exposes_reference_transparency_contract() -> None:
     source = (ROOT / "scripts" / "jyotish_api_server.py").read_text(encoding="utf-8")
     assert "result['reference_transparency'] = build_reference_transparency_contract(" in source
+    assert "reference_date=_consultation_reference_date(body).date().isoformat()" in source