functime 0.2.0

The easiest way to run and scale time-series machine learning in the Cloud.

Time-series machine learning and embeddings at scale

---

functime is a powerful Python library for production-ready AutoML forecasting and temporal embeddings.

functime also comes with time-series preprocessing (box-cox, differencing etc), cross-validation splitters (expanding and sliding window), and forecast metrics (MASE, SMAPE etc). All optimized as lazy Polars transforms.

Want to use functime for seamless time-series predictive analytics across your data team? Looking for production-grade time-series AutoML in a serverless Cloud deployment? Shoot Chris a message on LinkedIn to learn more about functime Enterprise.

Highlights

• Fast: Forecast 100,000 time series in seconds on your laptop
• Efficient: Embarrassingly parallel feature engineering for time-series using Polars
• Battle-tested: Machine learning algorithms that deliver real business impact and win competitions
• Exogenous features: supported by every forecaster
• Backtesting with expanding window and sliding window splitters
• AutoML: Automated lags and hyperparameter tuning using FLAML
• Censored model: for zero-inflated and thresholding forecasts

Getting Started

Install functime via the pip package manager.

pip install functime

Forecasting

import polars as pl
from functime.cross_validation import train_test_split
from functime.forecasting import lightgbm
from functime.metrics import mase

# Load example data
y = pl.read_parquet("https://github.com/descendant-ai/functime/raw/main/data/commodities.parquet")
entity_col, time_col = y.columns[:2]

# Time series split
y_train, y_test = y.pipe(train_test_split(test_size=3))

# Fit-predict
model = lightgbm(freq="1mo", lags=24, max_horizons=3, strategy="ensemble")
model.fit(y=y_train)
y_pred = model.predict(fh=3)

# functime ❤️ functional design
# fit-predict in a single line
y_pred = lightgbm(freq="1mo", lags=24)(y=y_train, fh=3)

# Score forecasts in parallel
scores = mase(y_true=y_test, y_pred=y_pred, y_train=y_train)

Classification

Only available on functime Enterprise.

import polars as pl
import functime
from sklearn.linear_model import RidgeClassifierCV
from sklearn.preprocessing import StandardScaler
from sklearn.metrics import accuracy_score
from sklearn.pipeline import make_pipeline

# Load GunPoint dataset (150 observations, 150 timestamps)
X_y_train = pl.read_parquet("https://github.com/descendant-ai/functime/raw/main/data/gunpoint_train.parquet")
X_y_test = pl.read_parquet("https://github.com/descendant-ai/functime/raw/main/data/gunpoint_test.parquet")

# Train-test split
X_train, y_train = X_y_train.select(pl.all().exclude("label")), X_y_train.select("label")
X_test, y_test = X_y_test.select(pl.all().exclude("label")), X_y_test.select("label")

X_train_embs = functime.embeddings.embed(X_train)

# Fit classifier on the embeddings
classifier = make_pipeline(
    StandardScaler(with_mean=False),
    RidgeClassifierCV(alphas=np.logspace(-3, 3, 10)),
)
classifier.fit(X_train_embs, y_train)

# Predict and
X_test_embs = embed(X_test)
labels = classifier.predict(X_test_embs)
accuracy = accuracy_score(predictions, y_test)

Clustering

Only available on functime Enterprise.

import functime
import polars as pl
from hdbscan import HDBSCAN
from umap import UMAP
from functime.preprocessing import roll

# Load S&P500 panel data from 2022-06-01 to 2023-06-01
# Columns: ticker, time, price
y = pl.read_parquet("https://github.com/descendant-ai/functime/raw/main/data/sp500.parquet")

# Create embeddings
embeddings = functime.embeddings.embed(y_ma_60)

# Reduce dimensionality with UMAP
reducer = UMAP(n_components=500, n_neighbors=10, metric="manhattan")
umap_embeddings = reducer.fit_transform(embeddings)

# Cluster with HDBSCAN
clusterer = HDBSCAN(metric="minkowski", p=1)
estimator.fit(X)

# Get predicted cluster labels
labels = estimator.predict(X)

Serverless Deployment

Only available on functime Enterprise.

Deploy and train forecasters the moment you call any .fit method. Run the functime list CLI command to list all deployed models. Finally, track data and forecasts usage using functime usage CLI command.

You can reuse a deployed model for predictions anywhere using the stub_id variable. Note: the .from_deployed model class must be the same as during .fit.

forecaster = LightGBM.from_deployed(stub_id)
y_pred = forecaster.predict(fh=3)

License

functime is distributed under AGPL-3.0-only. For Apache-2.0 exceptions, see LICENSING.md.