--- a
+++ b/setup.py
@@ -0,0 +1,108 @@
+#!/usr/bin/env python
+# -*- coding: utf-8 -*-
+
+from setuptools import setup, find_packages
+
+requirements = [
+    "argh",
+    "attr",
+    "related",
+    "cloudpickle>=1.0.0",
+
+    "concise==0.6.7",
+    "deepexplain",
+
+    # ml
+    "gin-config",
+    "keras==2.2.4",
+    "scikit-learn",
+    # "tensorflow",
+
+    # numerics
+    "h5py",
+    "numpy",
+    "pandas",
+    "scipy",
+    "statsmodels",
+
+    # Plotting
+    "matplotlib>=3.0.2",
+    "plotnine",
+    "seaborn",
+
+    # genomics
+    "pybigwig",
+    "pybedtools",  # remove?
+    "modisco==0.5.3.0",
+    # "pyranges",
+
+    "joblib",
+    "cloudpickle>=1.0.0",  # - remove?
+    "kipoi>=0.6.8",
+    "kipoi-utils>=0.3.0",
+    "kipoiseq>=0.2.2",
+
+    "papermill",
+    "jupyter_client>=6.1.2",
+    "ipykernel",
+    "nbconvert>=5.5.0",
+    "vdom>=0.6",
+
+    # utils
+    "ipython",
+    "tqdm",
+    # "pprint",
+
+
+    # Remove
+    "genomelake",
+    "pysam",  # replace with pyfaidx
+]
+
+optional = [
+    "comet_ml",
+    "wandb",
+    "fastparquet",
+    "python-snappy",
+    "ipywidgets",  # for motif simulation
+]
+
+test_requirements = [
+    "pytest>=3.3.1",
+    "pytest-cov>=2.6.1",
+    # "pytest-xdist",
+    "gdown",   # download files from google drive
+    "virtualenv",
+]
+
+dependency_links = [
+    "deepexplain @ git+https://github.com/kundajelab/DeepExplain.git@#egg=deepexplain-0.1"
+]
+
+
+setup(
+    name="bpnet",
+    version='0.0.23',
+    description=("BPNet: toolkit to learn motif synthax from high-resolution functional genomics data"
+                 " using convolutional neural networks"),
+    author="Ziga Avsec",
+    author_email="avsec@in.tum.de",
+    url="https://github.com/kundajelab/bpnet",
+    packages=find_packages(),
+    install_requires=requirements,
+    extras_require={
+        "dev": test_requirements,
+        "extras": optional,
+    },
+    license="MIT license",
+    entry_points={'console_scripts': ['bpnet = bpnet.__main__:main']},
+    zip_safe=False,
+    keywords=["deep learning",
+              "computational biology",
+              "bioinformatics",
+              "genomics"],
+    test_suite="tests",
+    package_data={'bpnet': ['logging.conf']},
+    include_package_data=True,
+    tests_require=test_requirements
+)