import json
import requests
import pandas as pd

# Create a tiny mock dataframe matching your query's feature layout
# Ensure the columns precisely match the ones your model trained on!
mock_batch = pd.DataFrame([{
    'person_id': 'usr_123456',
    'session_id': 'sess_normal_001',

    'session_start': '2026-06-08 19:18:42',
    'session_start_sin': -0.9659,
    'session_start_cos': 0.2588,
    'session_end': '2026-06-08 19:29:58',
    'session_end_sin': -0.8746,
    'session_end_cos': 0.4848,

    'event_count': 26,
    'unique_event_types': 9,
    'screens_visited': 11,
    'duration_seconds': 676,
    'events_per_minute': 5.15,

    'inter_event_time_mean_seconds': 11.7,
    'inter_event_time_median_seconds': 8.4,
    'inter_event_time_std_seconds': 10.6,
    'max_inter_event_gap_seconds': 74.2,

    'session_start_hour': 19,
    'session_day_of_week': 0,
    'session_day_of_week_sin': 0.0,
    'session_day_of_week_cos': 1.0,

    'distinct_hours_active': 1
}])

# Format into MLflow's split layout
payload = {"dataframe_split": mock_batch.to_dict(orient="split")}

# Hit the local model server
response = requests.post(
    "http://127.0.0.1:5001/invocations", 
    data=json.dumps(payload), 
    headers={"Content-Type": "application/json"}
)

print("Server Response Status:", response.status_code)
print("Predictions (1=Clean, -1=Anomaly):", response.json())