pyproject.toml

[tool.poetry]
name = "datalake"
version = "0.1.0"
description = "A data engineering programming exercise"
authors = ["praveer kumar <praveerkumar17@gmail.com>"]
license = "MIT"
readme = "README.md"
packages = [
    { include = "datalake", from = "." },
]

[tool.poetry.dependencies]
python = ">=3.11.9"
pyspark = "^3.5.1"
minio = "^7.2.7"
pydantic = "^2.8.2"
sqlalchemy = "^2.0.31"
pydantic-settings = "^2.3.4"
streamlit = "^1.36.0"
tqdm = "^4.66.4"
psycopg2-binary = "^2.9.9"
psycopg2 = { version = "^2.9.9", optional = true }
s3fs = "^2024.6.1"
httpx = {extras = ["http2"], version = "^0.27.0"}

[tool.poetry.group.dev.dependencies]
black = "^24.4.2"
pytest = "^8.2.2"

[tool.poetry.extras]
postgres = ["psycopg2"]

[tool.poetry.group.docs.dependencies]
sphinx = "^7.4.4"
sphinx-rtd-theme = "^2.0.0"
recommonmark = "^0.7.1"
furo = "^2024.5.6"
sphinxcontrib-plantuml = "^0.30"

[tool.poetry.scripts]
docs = "datalake.docs.__main__:main"
dashboard = "datalake.dashboard.__main__:cli"
download = "datalake.downloader.__main__:main"

[build-system]
requires = ["poetry-core"]
build-backend = "poetry.core.masonry.api"