scripts/verify_finales.py

import csv
ids={r["ID"] for r in csv.DictReader(open("/root/Test_phase2.csv",encoding="utf-8"))}
for f in ["/root/sub_finale1_mms1b.csv","/root/sub_finale2_multi.csv"]:
    d={r["ID"]:r["Target"] for r in csv.DictReader(open(f,encoding="utf-8"))}
    emp=sum(1 for v in d.values() if not v.strip())
    name=f.split("/")[-1]
    print(f"{name}: {len(d)} IDs | couvre={ids==set(d)} | manquants={len(ids-set(d))} | vides={emp}")
d1={r["ID"]:r["Target"] for r in csv.DictReader(open("/root/sub_finale1_mms1b.csv",encoding="utf-8"))}
d2={r["ID"]:r["Target"] for r in csv.DictReader(open("/root/sub_finale2_multi.csv",encoding="utf-8"))}
diff=sum(1 for k in d1 if d1[k]!=d2.get(k))
print(f"lignes differentes F1(routage) vs F2(multi): {diff}/{len(d1)}")
print("=== 4 exemples Finale 2 (multilingue, sans routage) ===")
for t in list(d2)[:4]: print("  ",t,":",repr(d2[t][:75]))