Add incremental material consumption integrator

scripts/validate_k7_material_consumption.py

@@ -0,0 +1,65 @@
+#!/usr/bin/env python3
+"""Replay common timestamps from ignored K7 run 00842 captures, without network access."""
+
+import argparse
+import json
+from datetime import UTC, datetime
+from pathlib import Path
+
+from production_analytics.calculations import (
+    MaterialConsumptionIntegrator,
+    MaterialIntegratorConfig,
+    MaterialSample,
+)
+
+RATE_UUID = "c9d06af5-f6d6-4ede-b6c4-5a98bac77129"  # Stundenleistung Anlage [kg/h]
+GATE_UUID = "823867bb-f5d2-40eb-b875-657155addfd0"  # Geschwindigkeit Gesamtanlage [m/min]
+CAPTURE_DIRECTORY = Path("data/raw/enlyze")
+
+
+def read_capture(path: Path, variable_uuid: str) -> dict[datetime, float]:
+    """Read unique, ordered observations; refuse ambiguous duplicate source records."""
+    data = json.loads(path.read_text(encoding="utf-8"))["response"]["body"]["data"]
+    time_index = data["columns"].index("time")
+    value_index = data["columns"].index(variable_uuid)
+    values = {}
+    previous = None
+    for record in data["records"]:
+        timestamp = datetime.fromisoformat(record[time_index].replace("Z", "+00:00"))
+        if timestamp.utcoffset() is None:
+            raise ValueError(f"Naive timestamp in {path}")
+        timestamp = timestamp.astimezone(UTC)
+        if previous is not None and timestamp <= previous:
+            raise ValueError(f"Capture timestamps must be strictly increasing: {path}")
+        previous = timestamp
+        if record[value_index] is not None:
+            values[timestamp] = float(record[value_index])
+    return values
+
+
+def main() -> int:
+    parser = argparse.ArgumentParser(description=__doc__)
+    parser.add_argument("--max-sample-gap-seconds", type=float, default=20.0)
+    args = parser.parse_args()
+    integrator = MaterialConsumptionIntegrator(
+        MaterialIntegratorConfig(0.5, args.max_sample_gap_seconds)
+    )
+    try:
+        rates = read_capture(CAPTURE_DIRECTORY / "k7-00842-throughput.raw.json", RATE_UUID)
+        gates = read_capture(CAPTURE_DIRECTORY / "k7-00842-speed.raw.json", GATE_UUID)
+    except (OSError, ValueError, KeyError) as error:
+        parser.exit(1, f"Cannot replay K7 captures: {error}\n")
+    # Dict insertion order preserves validated capture order. No resampling or filling.
+    samples = [MaterialSample(t, rate, gates[t]) for t, rate in rates.items() if t in gates]
+    if len(samples) < 2:
+        parser.exit(1, "Need at least two common non-null samples\n")
+    state = integrator.process_many(samples)
+    print(f"Common samples: {len(samples)}")
+    print(f"Maximum sample gap: {integrator.config.max_sample_gap_seconds:g} s")
+    print(f"Integrated running time: {state.integrated_running_seconds / 3600:.9f} h")
+    print(f"Cumulative material consumption: {state.cumulative_consumption_kg:.9f} kg")
+    return 0
+
+
+if __name__ == "__main__":
+    raise SystemExit(main())