Fix remote training and model validation pipeline

tests/test_time_series.py

@@ -304,6 +304,60 @@ def test_time_series_forecaster_attaches_quality_gate(make_settings, tmp_path) -
     assert forecast.quality_gate["status"] == "fail"
 
 
+def test_time_series_forecaster_uses_symbol_calibration(make_settings, tmp_path) -> None:
+    artifact_path = tmp_path / "lstm_forecaster.json"
+    _write_torch_gru_artifact(artifact_path, head_bias=0.2)
+    (tmp_path / "torch_threshold_calibration.json").write_text(
+        json.dumps(
+            {
+                "validation": {"status": "pass", "passed": True},
+                "recommended": {"edge": 0.08, "probability": 0.52, "confidence": 0.4},
+                "symbol_recommendations": {
+                    "BTCUSDT": {"edge": 0.03, "probability": 0.55, "confidence": 0.45}
+                },
+            }
+        ),
+        encoding="utf-8",
+    )
+    settings = make_settings(
+        tmp_path,
+        time_series_lstm_model_path=artifact_path,
+        time_series_forecast_horizon=1,
+    )
+
+    forecast = TimeSeriesForecaster(settings).forecast(
+        _candles_from_returns([0.0001] * 140), symbol="BTCUSDT"
+    )
+
+    assert forecast.calibrated_min_edge_percent == 0.03
+    assert forecast.calibrated_min_probability_up == 0.55
+    assert forecast.calibrated_min_confidence == 0.45
+
+
+def test_time_series_forecaster_averages_ensemble_members(make_settings, tmp_path) -> None:
+    artifact_path = tmp_path / "lstm_forecaster.json"
+    _write_torch_gru_artifact(artifact_path, head_bias=0.9)
+    artifact = json.loads(artifact_path.read_text(encoding="utf-8"))
+    entry = artifact["symbols"]["BTCUSDT"]
+    entry["ensemble_members"] = [
+        {"state_dict": entry["state_dict"], "head_weight": [0.0, 0.0], "head_bias": bias}
+        for bias in (0.1, 0.3)
+    ]
+    artifact_path.write_text(json.dumps(artifact), encoding="utf-8")
+    settings = make_settings(
+        tmp_path,
+        time_series_lstm_model_path=artifact_path,
+        time_series_forecast_horizon=1,
+    )
+
+    forecast = TimeSeriesForecaster(settings).forecast(
+        _candles_from_returns([0.0001] * 140), symbol="BTCUSDT"
+    )
+
+    assert forecast.usable is True
+    assert 0.015 <= forecast.expected_return_percent <= 0.025
+
+
 def test_time_series_forecaster_reads_multifeature_direct_horizon_artifact(make_settings, tmp_path) -> None:
     artifact_path = tmp_path / "lstm_forecaster.json"
     _write_multifeature_torch_gru_artifact(artifact_path, head_bias=0.2)