"""Reproducible check for sensor-quality-demo.csv only; not engineering limits."""
import csv, json
from pathlib import Path
from collections import defaultdict
from datetime import datetime, timedelta
path = Path(__file__).with_name("sensor-quality-demo.csv")
rows = list(csv.DictReader(path.open(newline="")))
groups = defaultdict(list)
for row in rows:
    groups[(row["asset_id"],row["sensor_id"],row["timestamp"])].append(row)
exact_duplicates = sum(len(items)-1 for items in groups.values() if all(x==items[0] for x in items))
conflicts = [key for key,items in groups.items() if any(x!=items[0] for x in items)]
series = {key[2]: items[0] for key,items in groups.items() if key[:2]==("A","T-1") and key not in conflicts}
start=datetime.fromisoformat("2026-10-20T09:00:00+03:00")
expected=[(start+timedelta(minutes=i)).isoformat() for i in range(6)]
missing=[t for t in expected if t not in series]
bad=[t for t,row in series.items() if row["quality"]!="OK"]
normalized=[]
for t,row in series.items():
    if row["quality"]!="OK": continue
    v=float(row["value"])
    if row["unit"]=="F": v=(v-32)*5/9
    elif row["unit"]!="C": raise ValueError("Unknown unit")
    normalized.append({"timestamp":t,"value_c":v,"source_value":row["value"],"source_unit":row["unit"]})
result={"source_rows":len(rows),"exact_duplicates":exact_duplicates,"conflicts":conflicts,"missing":missing,"bad_quality":bad,"usable":len(normalized),"expected":len(expected),"usable_pct":round(100*len(normalized)/len(expected),1),"normalized":normalized}
assert len(rows)==7 and exact_duplicates==1 and len(missing)==1 and len(bad)==1 and len(normalized)==4
assert normalized[-1]["value_c"]==70
print(json.dumps(result,ensure_ascii=False,indent=2))