setup.py

#
# Copyright (c) 2023 salesforce.com, inc.
# All rights reserved.
# SPDX-License-Identifier: BSD-3-Clause
# For full license text, see the LICENSE file in the repo root or https://opensource.org/licenses/BSD-3-Clause
#
from setuptools import setup, find_namespace_packages

MERLION_JARS = [
    "resources/gson-2.8.9.jar",
    "resources/randomcutforest-core-1.0.jar",
    "resources/randomcutforest-serialization-json-1.0.jar",
]

MERLION_DASHBOARD_ASSETS = [
    "dashboard/assets/fonts/SalesforceSans-Bold.woff",
    "dashboard/assets/fonts/SalesforceSans-BoldItalic.woff",
    "dashboard/assets/fonts/SalesforceSans-Italic.woff",
    "dashboard/assets/fonts/SalesforceSans-Light.woff",
    "dashboard/assets/fonts/SalesforceSans-LightItalic.woff",
    "dashboard/assets/fonts/SalesforceSans-Regular.woff",
    "dashboard/assets/fonts/SalesforceSans-Thin.woff",
    "dashboard/assets/fonts/SalesforceSans-ThinItalic.woff",
    "dashboard/assets/Acumin-BdPro.otf",
    "dashboard/assets/base.css",
    "dashboard/assets/merlion.css",
    "dashboard/assets/merlion_small.svg",
    "dashboard/assets/modal.css",
    "dashboard/assets/resizing.js",
    "dashboard/assets/styles.css",
    "dashboard/assets/upload.svg",
]

# optional dependencies
extra_require = {
    "dashboard": ["dash[diskcache]>=2.4", "dash_bootstrap_components>=1.0", "diskcache"],
    "deep-learning": ["torch>=1.9.0", "einops>=0.4.0"],
    "spark": ["pyspark[sql]>=3"],
}
extra_require["all"] = sum(extra_require.values(), [])


def read_file(fname):
    with open(fname, "r", encoding="utf-8") as f:
        return f.read()


setup(
    name="salesforce-merlion",
    version="2.0.2",
    author=", ".join(read_file("AUTHORS.md").split("\n")),
    author_email="abhatnagar@salesforce.com",
    description="Merlion: A Machine Learning Framework for Time Series Intelligence",
    long_description=read_file("README.md"),
    long_description_content_type="text/markdown",
    keywords="time series, forecasting, anomaly detection, machine learning, autoML, "
    "ensemble learning, benchmarking, Python, scientific toolkit",
    url="https://github.com/salesforce/Merlion",
    license="3-Clause BSD",
    packages=find_namespace_packages(include="merlion.*"),
    package_dir={"merlion": "merlion"},
    package_data={"merlion": MERLION_JARS + MERLION_DASHBOARD_ASSETS},
    install_requires=[
        "cython",
        "dill",
        "GitPython",
        "py4j",
        "matplotlib",
        "plotly>=4.13",
        "numpy>=1.21,<2.0",  # 1.21 remediates a security risk
        "packaging",
        "pandas>=1.1.0",  # >=1.1.0 for origin kwarg to df.resample()
        "prophet>=1.1",  # 1.1 removes dependency on pystan
        "scikit-learn>=0.22",  # >=0.22 for changes to isolation forest algorithm
        "scipy>=1.6.0",  # 1.6.0 adds multivariate_t density to scipy.stats
        "statsmodels>=0.12.2",
        "lightgbm",  # if running at MacOS, need OpenMP: "brew install libomp"
        "tqdm",
    ],
    extras_require=extra_require,
    python_requires=">=3.7.0",
    zip_safe=False,
)