File size: 2,499 Bytes
5c0af90 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 | """
inference.py — SARIMAX Groundwater Level Forecasting
=====================================================
Usage
-----
from inference import load_model, forecast
model = load_model()
predictions, conf_int = forecast(model, steps=12, exog_future=X_future)
"""
import json
import joblib
import numpy as np
import pandas as pd
from pathlib import Path
MODEL_PATH = Path(__file__).parent / "sarimax_model.pkl"
CONFIG_PATH = Path(__file__).parent / "model_config.json"
def load_model():
"""Load the fitted SARIMAX model."""
return joblib.load(MODEL_PATH)
def load_config():
"""Load model configuration and metrics."""
with open(CONFIG_PATH) as f:
return json.load(f)
def forecast(model, steps: int, exog_future: pd.DataFrame):
"""
Generate a forecast.
Parameters
----------
model : fitted SARIMAX result (from load_model)
steps : number of months ahead to forecast
exog_future : DataFrame with columns [temperature, precipitation, wind_speed]
Must have a DatetimeIndex with freq='MS' and exactly `steps` rows.
Returns
-------
predictions : pd.Series — point forecasts
conf_int : pd.DataFrame — 95% confidence interval
"""
cfg = load_config()
required_cols = cfg['exog_cols']
missing = [c for c in required_cols if c not in exog_future.columns]
if missing:
raise ValueError(f"exog_future is missing columns: {missing}")
if len(exog_future) != steps:
raise ValueError(f"exog_future must have {steps} rows, got {len(exog_future)}")
fc = model.get_forecast(steps=steps, exog=exog_future[required_cols])
predictions = fc.predicted_mean
conf_int = fc.conf_int()
return predictions, conf_int
if __name__ == "__main__":
model = load_model()
cfg = load_config()
print(f"Model loaded: SARIMAX{cfg['order']}x{cfg['seasonal_order']}")
print(f"Test RMSE: {cfg['test_metrics']['RMSE']}")
# Dummy forecast — replace with real future met data
idx = pd.date_range(start="2024-01-01", periods=12, freq="MS")
X_fut = pd.DataFrame({
"temperature" : np.random.uniform(3, 15, 12),
"precipitation": np.random.uniform(20, 120, 12),
"wind_speed" : np.random.uniform(10, 25, 12),
}, index=idx)
preds, ci = forecast(model, steps=12, exog_future=X_fut)
print(pd.DataFrame({"predicted": preds, "lower": ci.iloc[:,0], "upper": ci.iloc[:,1]})) |