Using pre-commit built-in text files detection feature (#72)

Author: Lucas-C

.pre-commit-config.yaml

@@ -43,21 +43,10 @@ repos:
         args:
           - --py37-plus
         exclude: ^tests/resources/.*
-  - repo: https://github.com/pre-commit/mirrors-pylint
-    rev: v3.0.0a5
-    hooks:
-      - id: pylint
-        args:
-          - --rcfile=.pylintrc
-          - --reports=no
-          - --py-version=3.7
-        additional_dependencies: [pytest, rapidfuzz]
-        exclude: ^tests/resources/.*(init_with_license\.py|todo).*$
   - repo: https://github.com/psf/black
     rev: 23.3.0
     hooks:
       - id: black
-        args: [--quiet]
         exclude: ^tests/resources/
   - repo: https://github.com/Lucas-C/pre-commit-hooks-bandit
     rev: v1.0.6
@@ -77,8 +66,16 @@ repos:
           - --show-error-context
   - repo: local
     hooks:
-      - id: py.test
-        name: py.test
+      - id: pylint
+        name: pylint
+        # 3x faster than the official pylint hook, and has no issue with imports
+        # (tested with: time pre-commit run pylint --all-files)
+        language: system
+        entry: pylint
+        files: \.py$
+        exclude: ^tests/resources/.*(init_with_license|todo)
+      - id: pytest
+        name: pytest
         language: python
         additional_dependencies: [pytest, pytest-cov, coverage, rapidfuzz]
         entry: pytest -sv

.pre-commit-hooks.yaml

@@ -3,27 +3,27 @@
     description: "Forbid files containing CRLF end-lines to be committed"
     entry: forbid_crlf
     language: python
-    files: ''
+    types: [text]
 -   id: remove-crlf
     name: CRLF end-lines remover
     description: "Replace CRLF end-lines by LF ones before committing"
     entry: remove_crlf
     language: python
-    files: ''
+    types: [text]
 -   id: forbid-tabs
     name: No-tabs checker
     description: "Forbid files containing tabs to be committed"
     entry: forbid_tabs
     language: python
-    files: ''
+    types: [text]
     exclude: (Makefile|debian/rules|.gitmodules)(\.in)?$
 -   id: remove-tabs
     name: Tabs remover
     description: "Replace tabs by whitespaces before committing"
     entry: remove_tabs
     language: python
     args: [ --whitespaces-count, '4' ]
-    files: ''
+    types: [text]
     exclude: (Makefile|debian/rules|.gitmodules)(\.in)?$
 -   id: chmod
     name: Set file permissions
@@ -34,4 +34,4 @@
     description: "Insert a short license disclaimer as a header comment in source files"
     entry: insert_license
     language: python
-    files: '.*/.*'
+    types: [text]

pre_commit_hooks/forbid_crlf.py

@@ -1,5 +1,4 @@
 import argparse, sys
-from .utils import is_textfile
 
 
 def contains_crlf(filename):
@@ -14,8 +13,7 @@ def main(argv=None):
     parser = argparse.ArgumentParser()
     parser.add_argument("filenames", nargs="*", help="filenames to check")
     args = parser.parse_args(argv)
-    text_files = [f for f in args.filenames if is_textfile(f)]
-    files_with_crlf = [f for f in text_files if contains_crlf(f)]
+    files_with_crlf = [f for f in args.filenames if contains_crlf(f)]
     return_code = 0
     for file_with_crlf in files_with_crlf:
         print(f"CRLF end-lines detected in file: {file_with_crlf}")

pre_commit_hooks/forbid_tabs.py

@@ -1,5 +1,4 @@
 import argparse, sys
-from .utils import is_textfile
 
 
 def contains_tabs(filename):
@@ -11,8 +10,7 @@ def main(argv=None):
     parser = argparse.ArgumentParser()
     parser.add_argument("filenames", nargs="*", help="filenames to check")
     args = parser.parse_args(argv)
-    text_files = [f for f in args.filenames if is_textfile(f)]
-    files_with_tabs = [f for f in text_files if contains_tabs(f)]
+    files_with_tabs = [f for f in args.filenames if contains_tabs(f)]
     return_code = 0
     for file_with_tabs in files_with_tabs:
         print(f"Tabs detected in file: {file_with_tabs}")

pre_commit_hooks/remove_crlf.py

@@ -1,5 +1,4 @@
 import argparse, sys
-from .utils import is_textfile
 
 
 def contains_crlf(filename):
@@ -23,8 +22,7 @@ def main(argv=None):
     parser = argparse.ArgumentParser()
     parser.add_argument("filenames", nargs="*", help="filenames to check")
     args = parser.parse_args(argv)
-    text_files = [f for f in args.filenames if is_textfile(f)]
-    files_with_crlf = [f for f in text_files if contains_crlf(f)]
+    files_with_crlf = [f for f in args.filenames if contains_crlf(f)]
     for file_with_crlf in files_with_crlf:
         print(f"Removing CRLF end-lines in: {file_with_crlf}")
         removes_crlf_in_file(file_with_crlf)

pre_commit_hooks/remove_tabs.py

@@ -1,5 +1,4 @@
 import argparse, sys
-from .utils import is_textfile
 
 
 def contains_tabs(filename):
@@ -26,8 +25,7 @@ def main(argv=None):
     )
     parser.add_argument("filenames", nargs="*", help="filenames to check")
     args = parser.parse_args(argv)
-    text_files = [f for f in args.filenames if is_textfile(f)]
-    files_with_tabs = [f for f in text_files if contains_tabs(f)]
+    files_with_tabs = [f for f in args.filenames if contains_tabs(f)]
     for file_with_tabs in files_with_tabs:
         print(
             f"Substituting tabs in: {file_with_tabs} by {args.whitespaces_count} whitespaces"

pre_commit_hooks/utils.py

@@ -2,6 +2,7 @@
 -e .
 # Install development specific dependencies
 pre-commit
+pylint
 pytest
 pytest-cov
 coverage

requirements-dev.txt

@@ -19,20 +19,6 @@ def test_remove_crlf(input_s, expected, tmpdir):
     assert input_file.read_bytes() == bytes(expected, "UTF-8")
 
 
-@pytest.mark.parametrize(
-    ("input_s", "expected"),
-    (
-        ("foo\r\nbar", "foo\r\nbar"),
-        ("bar\nbaz\r\n", "bar\nbaz\r\n"),
-    ),
-)
-def test_noremove_crlf(input_s, expected, tmpdir):
-    input_file = Path(tmpdir.join("file.pdf"))
-    input_file.write_bytes(bytes(input_s, "UTF-8"))
-    assert remove_crlf([str(input_file)]) == 0
-    assert input_file.read_bytes() == bytes(expected, "UTF-8")
-
-
 @pytest.mark.parametrize(("arg"), ("", "a.b", "a/b"))
 def test_badopt(arg):
     with pytest.raises(