move metadata to pyproject.toml

CHANGELOG.md

@@ -10,6 +10,15 @@ The format is based on [Keep a Changelog](http://keepachangelog.com/) and this p
 
 [Unreleased]: https://github.com/LoicGrobol/zeldarose/compare/v0.4.0...HEAD
 
+### Added
+
+- `lint` extra that install linting tools and plugins
+- Config for [flakeheaven](https://github.com/flakeheaven/flakeheaven)
+
+### Changed
+
+- Move packaging config to `pyproject.toml` and require `setuptools>=61".
+
 ## [0.4.0] — 2022-03-18
 
 [0.4.0]: https://github.com/LoicGrobol/zeldarose/compare/v0.3.4...v0.4.0

pyproject.toml

@@ -1,10 +1,78 @@
+[build-system]
+requires = ["setuptools>=61", "wheel"]
+build-backend = "setuptools.build_meta"
+
 [project]
 name = "zeldarose"
+version = "0.4.0"
+description = "Train transformer-based models"
+license = {text = "MIT"}
+authors = [{name = "Loïc Grobol", email = "loic.grobol@gmail.com"}]
+classifiers = [
+    "License :: OSI Approved :: MIT License",
+    "Programming Language :: Python :: 3.8",
+    "Programming Language :: Python :: 3.9",
+    "Programming Language :: Python :: 3.10",
+    "Environment :: Console",
+]
+keywords = [
+    "nlp",
+    "transformers",
+    "language-model",
+]
 requires-python = ">=3.8"
+dependencies = [
+    "click",
+    "click_pathlib",
+    "datasets",
+    "filelock",
+    "loguru",
+    "pydantic",
+    "pytorch-lightning >= 1.5.6, < 1.6.0",
+    "rich",
+    "torch > 1.8, < 1.12",
+    "torchmetrics",
+    "tokenizers ~= 0.10",
+    "toml",
+    "tqdm",
+    "transformers >= 4.0.0, < 5.0.0",
+]
 
-[build-system]
-requires = ["setuptools>=54", "wheel"]
-build-backend = "setuptools.build_meta" 
+[project.readme]
+file = "README.md"
+content-type = "text/markdown"
+
+[project.optional-dependencies]
+lint = [
+    "flakeheaven",
+    "flake8-bugbear",
+    "flake8-pytest-style",
+    "mypy",
+]
+tests = [
+    "pytest",
+    "pytest-console-scripts",
+    "pytest-lazy-fixture",
+]
+
+
+[project.scripts]
+zeldarose-tokenizer = "zeldarose.train_tokenizer:main"
+zeldarose-transformer = "zeldarose.train_transformer:main"
+
+[tool.flakeheaven]
+max_line_length = 100
+
+[tool.flakeheaven.plugins]
+bugbear = ["+*"]
+pycodestyle = ["+*", "-W503", "-E501"]
+pyflakes = ["+*", "-E203", "-E266"]
+pytest-style = ["+*"]
+
+[tool.mypy]
+allow_redefinition = true
+warn_unreachable = true
+plugins = ["pydantic.mypy"]
 
 [tool.black]
-line-length = 88
+line-length = 100

setup.cfg

@@ -1,61 +1,7 @@
-[metadata]
-name = zeldarose
-version = 0.4.0
-description = Train transformer-based models
-license = MIT
-author = Loïc Grobol
-author_email = loic.grobol@gmail.com
-long_description = file: README.md
-long_description_content_type = text/markdown
-classifiers =
-    License :: OSI Approved :: MIT License
-    Programming Language :: Python :: 3.8
-    Programming Language :: Python :: 3.9
-    Programming Language :: Python :: 3.10
-    Environment :: Console
-keywords =
-    nlp
-    transformers
-    language-model
-
-
 [options]
 packages = find:
-include_package_data = True
-python_requires = >=3.8
-install_requires =
-    click
-    click_pathlib
-    datasets
-    filelock
-    loguru
-    pydantic
-    pytorch-lightning >= 1.5.6, < 1.6.0
-    rich
-    torch > 1.8, < 1.12
-    torchmetrics
-    tokenizers ~= 0.10
-    toml
-    tqdm
-    transformers >= 4.0.0, < 5.0.0
-
-[options.extras_require]
-tests =
-    pytest
-    pytest-console-scripts
-    pytest-lazy-fixture
-
-[options.entry_points]
-console_scripts =
-    zeldarose-tokenizer = zeldarose.train_tokenizer:main
-    zeldarose-transformer = zeldarose.train_transformer:main
 
 [flake8]
 max-line-length = 100
 ignore = E203,E266,E501,W503
 max-complexity = 18
-
-[mypy]
-allow_redefinition = True
-warn_unreachable = True
-plugins = pydantic.mypy

tests/conftest.py

@@ -1,11 +1,7 @@
 import pathlib
-from typing import Union
 
 import pytest
 
-from pytest_lazyfixture import lazy_fixture
-
-
 fixtures_dir = pathlib.Path(__file__).parent / "fixtures"
 
 

tests/test_smoke.py

@@ -166,4 +166,4 @@ def test_train_mlm_with_remote_dataset(
         "--max-epochs",
         "2",
     )
-    assert ret.success
+    assert ret.success

zeldarose/utils.py

@@ -155,7 +155,7 @@ def __init__(
                 "Sharing embeddings between different model types might not work well:"
                 f" {type(self.follower_transformer.embeddings)} vs {type(self.leader_transformer.embeddings)}"
             )
-        
+
         self.replaced_layer: Dict[str, CombiningLayer] = dict()
 
     def on_train_start(self, trainer, pl_module):