Add v2 artifact metadata schema

src/research_signal_context_pipelines/__init__.py

@@ -4,7 +4,11 @@
 from .context_bundle import DEFAULT_UNIVERSE, build_context_bundle, build_context_from_source
 from .price_history import PriceExtractionSummary, write_filtered_price_history
 from .schema import SignalValidationError, validate_signal
-from .theme_momentum import build_theme_momentum_snapshot, write_theme_momentum_snapshot
+from .theme_momentum import (
+    build_theme_momentum_snapshot,
+    validate_theme_momentum_snapshot,
+    write_theme_momentum_snapshot,
+)
 from .theme_universe import build_theme_context, load_symbol_theme_exposure, load_theme_taxonomy
 
 __all__ = [
@@ -20,6 +24,7 @@
     "load_symbol_theme_exposure",
     "load_theme_taxonomy",
     "validate_signal",
+    "validate_theme_momentum_snapshot",
     "write_filtered_price_history",
     "write_theme_momentum_snapshot",
 ]

src/research_signal_context_pipelines/schema.py

@@ -28,6 +28,7 @@ class SignalValidationError(ValueError):
 ALLOWED_REGIMES = frozenset({"risk_on", "risk_off", "neutral", "mixed", "unknown"})
 ALLOWED_BIAS_VALUES = frozenset({"positive", "negative", "neutral", "watch", "avoid"})
 REQUIRED_SIGNAL_HORIZON = "1-3 years"
+SUPPORTED_SCHEMA_VERSIONS = frozenset({"1", "2"})
 
 
 def _require_mapping(value: Any, name: str) -> Mapping[str, Any]:
@@ -82,11 +83,15 @@ def validate_signal(payload: Mapping[str, Any]) -> None:
     if missing:
         raise SignalValidationError(f"missing required keys: {', '.join(missing)}")
 
-    if _require_string(payload["schema_version"], "schema_version") != "1":
-        raise SignalValidationError("schema_version must be '1'")
+    schema_version = _require_string(payload["schema_version"], "schema_version")
+    if schema_version not in SUPPORTED_SCHEMA_VERSIONS:
+        raise SignalValidationError("schema_version must be '1' or '2'")
     _require_iso_date(payload["as_of"], "as_of")
     _require_iso_datetime(payload["generated_at"], "generated_at")
     _require_iso_date(payload["expires_at"], "expires_at")
+    if schema_version == "2":
+        _require_string(payload.get("model_version"), "model_version")
+        _require_string(payload.get("scoring_version"), "scoring_version")
 
     if payload["mode"] != "shadow":
         raise SignalValidationError("mode must be 'shadow'")

src/research_signal_context_pipelines/theme_momentum.py

@@ -25,6 +25,9 @@
 THEME_MOMENTUM_HORIZON = "medium"
 THEME_MOMENTUM_HORIZON_WINDOW = "2-12 weeks"
 THEME_MOMENTUM_HORIZON_WINDOW_ZH = "2-12周"
+THEME_MOMENTUM_MODEL_VERSION = "theme-momentum-v1"
+THEME_MOMENTUM_SCORING_VERSION = "theme-momentum-rules-v1"
+THEME_MOMENTUM_EXPIRY_DAYS = 84
 
 
 def utc_now_iso() -> str:
@@ -218,9 +221,12 @@ def build_theme_momentum_snapshot(
 
     taxonomy_versions = sorted({theme.taxonomy_version for theme in themes.values() if theme.taxonomy_version})
     return {
-        "schema_version": "1",
+        "schema_version": "2",
         "as_of": snapshot_as_of,
         "generated_at": (generated_at or dt.datetime.now(dt.UTC)).isoformat().replace("+00:00", "Z"),
+        "expires_at": (parse_price_date(snapshot_as_of) + dt.timedelta(days=THEME_MOMENTUM_EXPIRY_DAYS)).isoformat(),
+        "model_version": THEME_MOMENTUM_MODEL_VERSION,
+        "scoring_version": THEME_MOMENTUM_SCORING_VERSION,
         "mode": "theme_momentum_snapshot",
         "artifact_type": THEME_MOMENTUM_ARTIFACT_TYPE,
         "horizon": THEME_MOMENTUM_HORIZON,
@@ -265,6 +271,29 @@ def build_theme_momentum_snapshot(
     }
 
 
+def validate_theme_momentum_snapshot(snapshot: Mapping[str, Any]) -> None:
+    """Validate the stable metadata and core shape of v1/v2 theme artifacts."""
+    required = ("schema_version", "as_of", "generated_at", "mode", "artifact_type", "theme_ranks", "data_quality", "policy")
+    missing = [key for key in required if key not in snapshot]
+    if missing:
+        raise ValueError(f"theme momentum snapshot missing required keys: {', '.join(missing)}")
+    schema_version = str(snapshot["schema_version"])
+    if schema_version not in {"1", "2"}:
+        raise ValueError("theme momentum snapshot schema_version must be '1' or '2'")
+    parse_price_date(snapshot["as_of"])
+    generated_at = snapshot["generated_at"]
+    if not isinstance(generated_at, str) or not generated_at.strip():
+        raise ValueError("theme momentum snapshot generated_at must be a non-empty string")
+    if schema_version == "2":
+        for key in ("expires_at", "model_version", "scoring_version"):
+            if not isinstance(snapshot.get(key), str) or not snapshot[key].strip():
+                raise ValueError(f"theme momentum snapshot {key} must be a non-empty string")
+        if parse_price_date(snapshot["expires_at"]) < parse_price_date(snapshot["as_of"]):
+            raise ValueError("theme momentum snapshot expires_at must not be before as_of")
+    if not isinstance(snapshot["theme_ranks"], list) or not isinstance(snapshot["data_quality"], Mapping):
+        raise ValueError("theme momentum snapshot core shape is invalid")
+
+
 def write_theme_momentum_snapshot(snapshot: Mapping[str, Any], path: str | Path) -> Path:
     output_path = Path(path)
     output_path.parent.mkdir(parents=True, exist_ok=True)

tests/test_signal_validation.py

@@ -21,6 +21,33 @@ def test_example_signal_is_valid() -> None:
     assert payload["horizon"] == "1-3 years"
 
 
+def test_v2_signal_requires_versioned_model_metadata() -> None:
+    payload = load_example()
+    payload.update(
+        {
+            "schema_version": "2",
+            "model_version": "shadow-v2",
+            "scoring_version": "rules-v2",
+        }
+    )
+
+    validate_signal(payload)
+
+
+def test_v1_signal_remains_readable_without_v2_metadata() -> None:
+    payload = load_example()
+
+    validate_signal(payload)
+
+
+def test_v2_signal_requires_model_and_scoring_versions() -> None:
+    payload = load_example()
+    payload["schema_version"] = "2"
+
+    with pytest.raises(SignalValidationError, match="model_version"):
+        validate_signal(payload)
+
+
 def test_signal_requires_long_horizon_contract() -> None:
     payload = load_example()
     payload["horizon"] = "1-3 months"

tests/test_theme_momentum.py

@@ -3,7 +3,10 @@
 import datetime as dt
 
 from research_signal_context_pipelines.price_history import PriceRow
-from research_signal_context_pipelines.theme_momentum import build_theme_momentum_snapshot
+from research_signal_context_pipelines.theme_momentum import (
+    build_theme_momentum_snapshot,
+    validate_theme_momentum_snapshot,
+)
 from research_signal_context_pipelines.theme_universe import SymbolThemeExposure, ThemeDefinition
 
 
@@ -58,6 +61,10 @@ def test_theme_momentum_ranks_strong_broad_theme_first() -> None:
 
     ranked = snapshot["theme_ranks"]
     assert snapshot["artifact_type"] == "medium_horizon_theme_context"
+    assert snapshot["schema_version"] == "2"
+    assert snapshot["expires_at"] == "2025-12-30"
+    assert snapshot["model_version"] == "theme-momentum-v1"
+    assert snapshot["scoring_version"] == "theme-momentum-rules-v1"
     assert snapshot["horizon"] == "medium"
     assert snapshot["horizon_window"] == "2-12 weeks"
     assert snapshot["horizon_window_label"] == "2-12周"
@@ -96,3 +103,31 @@ def test_theme_momentum_records_missing_price_coverage() -> None:
     assert snapshot["data_quality"]["coverage"]["price_coverage_ratio"] == 0.5
     assert snapshot["theme_ranks"][0]["component_count"] == 2
     assert snapshot["theme_ranks"][0]["priced_symbol_count"] == 1
+
+
+def test_theme_momentum_validator_reads_v1_and_v2_metadata() -> None:
+    themes = {
+        "hbm_memory": ThemeDefinition(
+            taxonomy_version="test-v1",
+            theme_id="hbm_memory",
+            theme_name="HBM and memory",
+            sector="technology",
+            horizon="6-24 months",
+            description="memory theme",
+            source_policy="primary evidence required",
+        )
+    }
+    exposures = {"MU": SymbolThemeExposure("MU", ("hbm_memory",), "high", "memory exposure")}
+    snapshot = build_theme_momentum_snapshot(
+        _trend_rows("MU", start_close=50, daily_step=0.22),
+        themes=themes,
+        exposures=exposures,
+        generated_at=dt.datetime(2026, 1, 1, tzinfo=dt.timezone.utc),
+    )
+
+    validate_theme_momentum_snapshot(snapshot)
+    legacy = dict(snapshot)
+    legacy["schema_version"] = "1"
+    for key in ("expires_at", "model_version", "scoring_version"):
+        legacy.pop(key)
+    validate_theme_momentum_snapshot(legacy)