[tool.poetry]
name = "mwmbl"
version = "0.1.0"
description = ""
authors = ["Daoud Clarke <daoud.clarke@gmail.com>"]

[tool.poetry.dependencies]
python = "^3.10"
pandas = "^1.3.4"
zstandard = "^0.16.0"
mmh3 = "^3.0.0"
fastapi = "^0.70.1"
uvicorn = "^0.16.0"
numpy = "==1.21.1"
pyyaml = "==6.0"
# Optional dependencies do not get installed by default. Look under tool.poetry.extras section
# to see which extras to use.
botocore = {version= "==1.23.20", optional = true}
boto3 = {version= "==1.20.20", optional = true}
ujson = {version= "==4.3.0", optional = true}
warcio = {version= "==1.7.4", optional = true}
idna = {version= "==3.3", optional = true}
beautifulsoup4 = {version= "==4.10.0", optional = true}
lxml = {version= "==4.6.4", optional = true}
jusText = {version= "==3.0.0", optional = true}
langdetect = {version= "==1.0.9", optional = true}
pyarrow = {version= "==6.0.0", optional = true}
pyspark = {version= "==3.2.0", optional = true}
Levenshtein = {version= "==0.16.0", optional = true}
# en-core-web-sm requires a compatible version of spacy
spacy = {version= "==3.2.1", optional = true}
en-core-web-sm = {url = "https://github.com/explosion/spacy-models/releases/download/en_core_web_sm-3.2.0/en_core_web_sm-3.2.0.tar.gz", optional = true}


[tool.poetry.extras]
indexer = [
    "botocore",
    "boto3",
    "ujson",
    "warcio",
    "idna",
    "beautifulsoup4",
    "lxml",
    "jusText",
    "langdetect",
    "pyarrow",
    "pyspark",
    "Levenshtein",
    # en-core-web-sm requires a compatible version of spacy
    "spacy",
    "en-core-web-sm",
]

[tool.poetry.dev-dependencies]
# botocore = "^1.23.20"
# boto3 = "^1.20.20"
# ujson = "^4.3.0"
# warcio = "^1.7.4"
# idna = "^3.3"
# beautifulsoup4 = "^4.10.0"
# lxml = "^4.6.4"
# jusText = "^3.0.0"
# pyspark = "^3.2.0"
# langdetect = "^1.0.9"
# spacy = "^3.2.1"
# Levenshtein = "^0.16.0"
pytest = "^7.0.1"

[build-system]
requires = ["poetry-core>=1.0.0"]
build-backend = "poetry.core.masonry.api"

[tool.poetry.scripts]
mwmbl-tinysearchengine = "mwmbl.tinysearchengine.app:main"