pyproject.toml

[tool.poetry]
name = "my-itmo-ru-to-ical"
version = "1.0.0"
description = "A simple app exporting the scheudle from my.itmo.ru in iCalendar format."
authors = ["iburakov <speedwatson@gmail.com>"]
license = "MIT"

[tool.poetry.dependencies]
python = "^3.11"
ics = "^0.7.2"
Flask = {extras = ["async"], version = "^2.3.3"}
sentry-sdk = {extras = ["flask"], version = "^1.4.3"}
aiohttp = "^3.8.5"
async-cache = "^1.1.1"

[tool.poetry.group.dev.dependencies]
ruff = "^0.0.291"
black = "^23.9.1"
mypy = "^1.5.1"
vulture = "^2.9.1"
types-requests = "^2.31.0.3"
types-python-dateutil = "^2.8.19.14"


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


[tool.black]
line-length = 120


[tool.mypy]
files = ["src"]

[[tool.mypy.overrides]]
# https://mypy.readthedocs.io/en/stable/running_mypy.html#missing-imports
# it's better to disable untyped modules on a per-module basis
module = ["ics.*", "cache.*"]
ignore_missing_imports = true


[tool.ruff]
line-length = 120
select = [
    # A set of chosen linter rules is specified here.
    # See https://beta.ruff.rs/docs/rules/ for more info.
    # pyflakes (simple obvious errors)
    "F",
    # pycodestyle (style linter, pep8, black-compatible)
    "E",
    "W",
    # pep8-naming (style linter, pep8 naming conventions)
    "N",
    # isort (imports sorting)
    "I",
    # mccabe (cyclomatic complexity analyzer to prevent overcomplicated functions)
    "C90",
    # pyupgrade (possible syntax upgrades checker)
    "UP",
    # eradicate (commented-out code finder for it to be removed)
    "ERA",
    # flake8-2020 (errors related to sys.version or sys.version_info, just in case)
    "YTT",
    # flake8-async (async-related mistakes/errors)
    "ASYNC",
    # flake8-builtins (checks builtin names shadowing, it's better not to)
    "A",
    # flake8-commas (to enforce trailing commas)
    "COM",
    # flake8-future-annotations (to ensure enabling modern (3.7+) postponed evaluation of type hints)
    "FA",
    # flake8-import-conventions (to enforce standartized import aliases like "import pandas as pd")
    "ICN",
    # flake8-no-pep420 (to enforce presence of __init__.py in packages)
    "INP",
    # flake8-print (to disallow print statements)
    "T20",
    # flake8-pytest-style (to improve pytest-related style consistency)
    "PT",
    # flake8-quotes (to enforce quotes style)
    "Q",
    # flake8-return (checks mistakes related to return values)
    "RET",
    # flake8-use-pathlib (to enforce pathlib usage instead of os.path)
    "PTH",
    # tryceratops (try/except-related advices)
    "TRY",
    # ruff-specific advices
    "RUF",
]
ignore = [
    "E501",   # allow lines longer than 80 symbols (up to 120)
    "RET504", # allow "unnecessary" assignments before return statements as it's often useful for chaning data transformations
    "TRY003", # allow something like raise ValueError(f"Formatting {this} error message on the spot here"). too strict!
]
unfixable = [
    "ERA001", # prohibit to "auto-fix" (=remove) commented-out code
]
src = ["src"]

[tool.ruff.mccabe]
max-complexity = 7


[tool.vulture]
# Vulture is a tool that statically finds unreachable code or unused attributes in Python programs.
# Doing it without running the code in a language as dynamic as Python is non-trivial.
# Hence, Vulture has some notion of "confidence", but imho it's pretty weird, since obvious unused global functions are 
# given 60% confidence anyway just because they're functions (I guess?). Confidence depends on object's type, among other things.
min_confidence = 60
# 60% confidence threshold yields a lot of false positives, but 70% threshold yields a lot of false negatives.
# The only way to get a good result is to use a 60% with a whitelist, adding false positives to it as they appear:
#   1) Initialize a whitelist (once per project): 
#           $ vulture --make-whitelist > vulture_whitelist.txt
#   2) During development, run vulture and review the results. This config in pyproject.toml is picked up, so 
#      vulture_whitelist.txt is picked up too (it's specified in the `paths` below) and false positives are ignored:
#           $ vulture
#   3) When all the new true positives are fixed, add remaining false positives to the whitelist: 
#           $ vulture --make-whitelist >> vulture_whitelist.txt
#   4) Keep in mind that if some false positives tend to appear again (like FastAPI's `@app`-decorated functions),
#      they better be added to the ignore list below, not to the whitelist.
paths = ["src", "vulture_whitelist.txt"]
ignore_decorators = [
    # FastAPI, Flask
    "@app.route",
    "@app.exception_handler",
    "@app.get",
    "@app.post",
    "@app.put",
    "@app.delete",
    "@app.patch",
]
ignore_names = []
sort_by_size = true