Add remove-em-dash hook

.pre-commit-hooks.yaml

@@ -190,6 +190,14 @@
     language: python
     pass_filenames: false
     always_run: true
+-   id: remove-em-dash
+    name: remove em-dash
+    description: replaces em-dashes with a plain hyphen.
+    entry: remove-em-dash
+    language: python
+    types: [text]
+    stages: [pre-commit, pre-push, manual]
+    minimum_pre_commit_version: 3.2.0
 -   id: requirements-txt-fixer
     name: fix requirements.txt
     description: sorts entries in requirements.txt.

README.md

@@ -175,6 +175,11 @@ the following commandline options:
   - `--no-sort-keys` - when autofixing, retain the original key ordering (instead of sorting the keys)
   - `--top-keys comma,separated,keys` - Keys to keep at the top of mappings.
 
+#### `remove-em-dash`
+Replaces em-dashes (Unicode `U+2014`) with a plain hyphen (`-`).
+  - Only the UTF-8 encoding of the em-dash is replaced; files using other
+    encodings are left untouched.
+
 #### `requirements-txt-fixer`
 Sorts entries in requirements.txt and constraints.txt and removes incorrect entry for `pkg-resources==0.0.0`
 

pre_commit_hooks/remove_em_dash.py

@@ -0,0 +1,34 @@
+from __future__ import annotations
+
+import argparse
+from collections.abc import Sequence
+
+EM_DASH = '\N{EM DASH}'.encode()
+
+
+def _fix_file(filename: str) -> bool:
+    with open(filename, 'rb') as f:
+        contents = f.read()
+    new_contents = contents.replace(EM_DASH, b'-')
+    if new_contents == contents:
+        return False
+    with open(filename, 'wb') as f:
+        f.write(new_contents)
+    return True
+
+
+def main(argv: Sequence[str] | None = None) -> int:
+    parser = argparse.ArgumentParser()
+    parser.add_argument('filenames', nargs='*', help='Filenames to fix')
+    args = parser.parse_args(argv)
+
+    retv = 0
+    for filename in args.filenames:
+        if _fix_file(filename):
+            print(f'Fixing {filename}')
+            retv = 1
+    return retv
+
+
+if __name__ == '__main__':
+    raise SystemExit(main())

setup.cfg

@@ -57,6 +57,7 @@ console_scripts =
     no-commit-to-branch = pre_commit_hooks.no_commit_to_branch:main
     pre-commit-hooks-removed = pre_commit_hooks.removed:main
     pretty-format-json = pre_commit_hooks.pretty_format_json:main
+    remove-em-dash = pre_commit_hooks.remove_em_dash:main
     requirements-txt-fixer = pre_commit_hooks.requirements_txt_fixer:main
     sort-simple-yaml = pre_commit_hooks.sort_simple_yaml:main
     trailing-whitespace-fixer = pre_commit_hooks.trailing_whitespace_fixer:main

tests/remove_em_dash_test.py

@@ -0,0 +1,36 @@
+from __future__ import annotations
+
+import pytest
+
+from pre_commit_hooks.remove_em_dash import main
+
+
+@pytest.mark.parametrize(
+    ('text', 'expected'),
+    (
+        ('foo\N{EM DASH}bar\n', b'foo-bar\n'),
+        ('foo \N{EM DASH} bar\n', b'foo - bar\n'),
+        ('a\N{EM DASH}b\N{EM DASH}c\n', b'a-b-c\n'),
+        ('x\N{EM DASH}y\r\nz\r\n', b'x-y\r\nz\r\n'),
+    ),
+)
+def test_fixes_em_dash(text, expected, tmpdir):
+    path = tmpdir.join('file.txt')
+    path.write_binary(text.encode())
+    assert main((str(path),)) == 1
+    assert path.read_binary() == expected
+
+
+@pytest.mark.parametrize(
+    'contents',
+    (
+        pytest.param(b'foo-bar\n', id='plain-hyphen'),
+        pytest.param(b'no em dashes here\n', id='no-dash'),
+        pytest.param(b'<a>\x97</a>\n', id='windows-1252-em-dash'),
+    ),
+)
+def test_noop_without_utf8_em_dash(contents, tmpdir):
+    path = tmpdir.join('file.txt')
+    path.write_binary(contents)
+    assert main((str(path),)) == 0
+    assert path.read_binary() == contents