Merge pull request #25 from akx/bytestrings

Don't require f-strings when the formatee is a bytestring

flake8_use_fstring/base.py

@@ -1,5 +1,5 @@
-import token as _token
 import typing as _typing
+
 from tokenize import (
     TokenInfo as _TokenInfo,
 )
@@ -8,6 +8,10 @@
     OptionManager as _OptionManager,
 )
 
+from flake8_use_fstring.utils import (
+    is_text_string_token as _is_text_string_token,
+)
+
 Flake8Output = _typing.Tuple[_typing.Tuple[int, int], str]
 
 
@@ -38,7 +42,7 @@ def __iter__(self) -> _typing.Iterator[Flake8Output]:
         met_string = False
 
         for i in range(len(self.tokens)):
-            if self.tokens[i].exact_type == _token.STRING:
+            if _is_text_string_token(self.tokens[i]):
                 met_string = True
 
             if not self[i]:
@@ -48,7 +52,7 @@ def __iter__(self) -> _typing.Iterator[Flake8Output]:
                 # only if last token is string
                 if i == 0:  # cannot use IndexError because -1 is a valid index
                     continue  # pragma: no cover (syntax error)
-                if self.tokens[i - 1].exact_type != _token.STRING:
+                if not _is_text_string_token(self.tokens[i - 1]):
                     continue
 
             elif self.greedy == self.GREEDY_MET_STRING:

flake8_use_fstring/prefix.py

@@ -9,6 +9,10 @@
     BaseLogicalLineChecker as _Base,
 )
 
+from .utils import (
+    is_text_string_token as _is_text_string_token,
+)
+
 from . import __version__
 
 FSTRING_REGEX = _re.compile(r'^([a-zA-Z]*?[fF][a-zA-Z]*?){1}["\']')
@@ -26,7 +30,7 @@ class MissingPrefixDetector(_Base):
 
     def __getitem__(self, i: int) -> bool:
         token = self.tokens[i]
-        if token.exact_type != _token.STRING:
+        if not _is_text_string_token(token):
             return False
 
         if FSTRING_REGEX.search(token.string):  # already is an f-string

flake8_use_fstring/utils.py

@@ -0,0 +1,15 @@
+import re as _re
+import token as _token
+from tokenize import TokenInfo as _TokenInfo
+
+PREFIX_RE = _re.compile(r"^[^'\"]*")
+
+
+def is_text_string_token(token: _TokenInfo) -> bool:
+    if not token.type == _token.STRING:
+        return False  # Not a string at all? Ignore.
+    # Get the prefix of the string (anything before the first quote)
+    prefix = PREFIX_RE.match(token.string).group(0).lower()
+    if 'b' in prefix:
+        return False  # Smells like a bytestring. Ignore it.
+    return True

tests/example.py

@@ -50,5 +50,13 @@ def format(self):  # noqa: A003
      '{{m}} {n}'
      r'[a-z]{1,3}')  # Should not be matched.
 
+# should not be matched on any greedy level < 2
+# (bytestrings can't be f-strings)
+p = b'%d' % f
+
+# should not be matched for missing prefix
+# (bytestrings can't be f-strings)
+q = b'{n}'
+
 # no errors below; coverage
 ''.strip()

tests/test_00.py

@@ -34,6 +34,7 @@ def test_greedy_2(test_flake8_cmd):
 tests/example.py:25:6: FS002 '.format' used
 tests/example.py:34:13: FS002 '.format' used
 tests/example.py:37:8: FS002 '.format' used
+tests/example.py:55:11: FS001 '%' operator used
 """
     test_flake8_cmd.test()
 
@@ -47,6 +48,7 @@ def test_greedy_different(test_flake8_cmd):
 tests/example.py:12:16: FS001 '%' operator used
 tests/example.py:15:7: FS001 '%' operator used
 tests/example.py:18:12: FS002 '.format' used
+tests/example.py:55:11: FS001 '%' operator used
 """
     test_flake8_cmd.test()