Electric Barometer · Adapters (eb-adapters)

Adapter interfaces that normalize forecasting model APIs for consistent evaluation within the Electric Barometer ecosystem.

Overview

eb-adapters wraps heterogeneous forecasting libraries behind a common fit / predict contract for uniform evaluation. It isolates framework-specific behavior; metrics and evaluation live in sibling packages.

Installation

eb-adapters is distributed as a standard Python package.

pip install eb-adapters

The package supports Python 3.11 and later.

Core Concepts

• Interface normalization — Forecasting models from different libraries are wrapped behind a common training and prediction contract, enabling uniform downstream evaluation.
• Thin adaptation layer — Adapters aim to be minimal and non-invasive, preserving native model behavior while standardizing how models are invoked.
• Framework isolation — Library-specific configuration, defaults, and quirks are contained within adapters, preventing leakage into evaluation or orchestration layers.
• Explicit lifecycle boundaries — Model fitting, prediction, and state management are clearly separated to support reproducibility and controlled execution.
• Comparability over abstraction — Adapters do not attempt to hide meaningful differences between modeling approaches; they exist to make comparison feasible, not to enforce uniformity.

Minimal Example

The example below shows how a forecasting model is wrapped behind a standardized adapter interface so it can be trained and evaluated consistently alongside other models.

import numpy as np
from eb_adapters import ArimaAdapter, SarimaxAdapter

y_train = np.array([10.0, 12.0, 11.0, 13.0, 14.0, 15.0], dtype=float)
X_train = np.arange(len(y_train), dtype=float).reshape(-1, 1)
X_horizon = np.arange(7, dtype=float).reshape(-1, 1)

arima = ArimaAdapter(order=(1, 1, 1))
arima.fit(X_train, y_train)
y_pred = arima.predict(X_horizon)

sarimax = SarimaxAdapter(order=(1, 0, 0), seasonal_order=(0, 0, 0, 0))
sarimax.fit(X_train, y_train)
y_pred = sarimax.predict(X_horizon)

The same fit(X, y) / predict(X) contract is used with tree-based models:

import numpy as np
from eb_adapters import XGBoostRegressorAdapter

X_train = np.arange(20, dtype=float).reshape(-1, 1)
y_train = (2.0 * X_train[:, 0] + 1.0)
X_future = np.arange(20, 27, dtype=float).reshape(-1, 1)

adapter = XGBoostRegressorAdapter(n_estimators=50, max_depth=3)
adapter.fit(X_train, y_train)
y_pred = adapter.predict(X_future)

License

BSD 3-Clause License. © 2026 Kyle Corrie.