2021-12-05 20:31:49 +00:00
|
|
|
[tool.poetry]
|
2021-12-28 11:02:48 +00:00
|
|
|
name = "mwmbl"
|
2021-12-05 20:31:49 +00:00
|
|
|
version = "0.1.0"
|
|
|
|
|
description = ""
|
|
|
|
|
authors = ["Daoud Clarke <daoud.clarke@gmail.com>"]
|
|
|
|
|
|
|
|
|
|
[tool.poetry.dependencies]
|
2022-03-20 22:01:45 +00:00
|
|
|
python = ">=3.10,<3.11"
|
|
|
|
|
pandas = "^1.3.5"
|
|
|
|
|
scipy = "^1.8.0"
|
|
|
|
|
scikit-learn = "^1.0.2"
|
2021-12-26 08:47:33 +00:00
|
|
|
zstandard = "^0.16.0"
|
|
|
|
|
mmh3 = "^3.0.0"
|
|
|
|
|
fastapi = "^0.70.1"
|
|
|
|
|
uvicorn = "^0.16.0"
|
2021-12-29 14:18:02 +00:00
|
|
|
pyyaml = "==6.0"
|
2022-06-16 21:49:41 +00:00
|
|
|
boto3 = "^1.20.37"
|
2022-03-20 22:01:45 +00:00
|
|
|
|
2021-12-27 14:46:24 +00:00
|
|
|
# Optional dependencies do not get installed by default. Look under tool.poetry.extras section
|
|
|
|
|
# to see which extras to use.
|
|
|
|
|
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}
|
2022-06-05 08:15:04 +00:00
|
|
|
requests = "^2.27.1"
|
2022-06-16 21:49:41 +00:00
|
|
|
psycopg2-binary = "^2.9.3"
|
2021-12-26 08:47:33 +00:00
|
|
|
|
2021-12-27 14:46:24 +00:00
|
|
|
|
|
|
|
|
[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",
|
|
|
|
|
]
|
2021-12-26 08:47:33 +00:00
|
|
|
|
|
|
|
|
[tool.poetry.dev-dependencies]
|
2021-12-22 23:21:23 +00:00
|
|
|
# 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"
|
2022-02-18 22:12:22 +00:00
|
|
|
pytest = "^7.0.1"
|
2021-12-05 20:31:49 +00:00
|
|
|
|
|
|
|
|
[build-system]
|
|
|
|
|
requires = ["poetry-core>=1.0.0"]
|
|
|
|
|
build-backend = "poetry.core.masonry.api"
|
2021-12-29 14:18:02 +00:00
|
|
|
|
|
|
|
|
[tool.poetry.scripts]
|
2022-06-17 20:26:21 +00:00
|
|
|
mwmbl-tinysearchengine = "mwmbl.main:run"
|
