Add trace_agg/src/trace_agg/core.py

trace_agg/src/trace_agg/core.py

@@ -0,0 +1,56 @@
+from __future__ import annotations
+import statistics
+from typing import List, Dict, Any
+
+
+class InvalidTraceDataError(ValueError):
+    """Raised when trace data validation fails."""
+    pass
+
+
+def _validate_record(record: Dict[str, Any]) -> None:
+    required_fields = {
+        'run_id': str,
+        'mischfenster_p50': (int, float),
+        'mischfenster_p95': (int, float),
+        'mischfenster_max': (int, float),
+        'step_order_stability': (int, float),
+        'read_between_steps': int
+    }
+    for field, ftype in required_fields.items():
+        if field not in record:
+            raise InvalidTraceDataError(f"Missing field '{field}' in trace record")
+        if not isinstance(record[field], ftype):
+            raise InvalidTraceDataError(
+                f"Field '{field}' has invalid type: {type(record[field])}, expected {ftype}"
+            )
+
+
+def generate_summary(trace_data: List[Dict[str, Any]]) -> Dict[str, Any]:
+    """Aggregiert Trace-Daten und erzeugt eine zusammengefasste JSON-Statistik pro Run."""
+    assert isinstance(trace_data, list), "trace_data muss eine Liste sein."
+    if not trace_data:
+        raise InvalidTraceDataError("Eingabeliste trace_data darf nicht leer sein.")
+
+    for rec in trace_data:
+        _validate_record(rec)
+
+    p50_values = [float(r['mischfenster_p50']) for r in trace_data]
+    p95_values = [float(r['mischfenster_p95']) for r in trace_data]
+    max_values = [float(r['mischfenster_max']) for r in trace_data]
+    stability_values = [float(r['step_order_stability']) for r in trace_data]
+    read_values = [int(r['read_between_steps']) for r in trace_data]
+
+    result = {
+        'num_records': len(trace_data),
+        'mischfenster_p50_mean': statistics.mean(p50_values),
+        'mischfenster_p95_mean': statistics.mean(p95_values),
+        'mischfenster_max_mean': statistics.mean(max_values),
+        'mischfenster_p50_median': statistics.median(p50_values),
+        'mischfenster_p95_median': statistics.median(p95_values),
+        'mischfenster_max_median': statistics.median(max_values),
+        'step_order_stability_avg': statistics.mean(stability_values),
+        'read_between_steps_total': sum(read_values),
+    }
+
+    return result