Optimize performance of character class predicates

Note that because letters are contiguous blocks in ASCII,
we can compare with lower and upper boundaries,
which is more efficient than checking containment in a set or string.

Author: Cito

.coveragerc

@@ -4,9 +4,10 @@ source = src
 omit =
     */conftest.py
     */test_*_fuzz.py
+    */assert_valid_name.py
     */cached_property.py
+    */character_classes.py
     */is_iterable.py
-    */assert_valid_name.py
 
 [report]
 exclude_lines =

README.md

@@ -14,7 +14,7 @@ a query language for APIs created by Facebook.
 
 The current version 3.1.6 of GraphQL-core is up-to-date with GraphQL.js version 15.5.1.
 
-An extensive test suite with nearly 2300 unit tests and 100% coverage comprises a
+An extensive test suite with over 2300 unit tests and 100% coverage comprises a
 replication of the complete test suite of GraphQL.js, making sure this port is
 reliable and compatible with GraphQL.js.
 

src/graphql/language/character_classes.py

@@ -1,33 +1,68 @@
 __all__ = ["is_digit", "is_letter", "is_name_start", "is_name_continue"]
 
+try:
+    "string".isascii()
+except AttributeError:  # Python < 3.7
 
-def is_digit(char: str) -> bool:
-    """Check whether char is a digit
+    def is_digit(char: str) -> bool:
+        """Check whether char is a digit
 
-    For internal use by the lexer only.
-    """
-    return "0" <= char <= "9"
+        For internal use by the lexer only.
+        """
+        return "0" <= char <= "9"
 
+    def is_letter(char: str) -> bool:
+        """Check whether char is a plain ASCII letter
 
-def is_letter(char: str) -> bool:
-    """Check whether char is a plain ASCII letter
+        For internal use by the lexer only.
+        """
+        return "a" <= char <= "z" or "A" <= char <= "Z"
 
-    For internal use by the lexer only.
-    """
-    return "A" <= char <= "Z" or "a" <= char <= "z"
+    def is_name_start(char: str) -> bool:
+        """Check whether char is allowed at the beginning of a GraphQL name
 
+        For internal use by the lexer only.
+        """
+        return "a" <= char <= "z" or "A" <= char <= "Z" or char == "_"
 
-def is_name_start(char: str) -> bool:
-    """Check whether char is allowed at the beginning of a GraphQL name
+    def is_name_continue(char: str) -> bool:
+        """Check whether char is allowed in the continuation of a GraphQL name
 
-    For internal use by the lexer only.
-    """
-    return is_letter(char) or char == "_"
+        For internal use by the lexer only.
+        """
+        return (
+            "a" <= char <= "z"
+            or "A" <= char <= "Z"
+            or "0" <= char <= "9"
+            or char == "_"
+        )
 
+else:
 
-def is_name_continue(char: str) -> bool:
-    """Check whether char is allowed in the continuation of a GraphQL name
+    def is_digit(char: str) -> bool:
+        """Check whether char is a digit
 
-    For internal use by the lexer only.
-    """
-    return is_letter(char) or is_digit(char) or char == "_"
+        For internal use by the lexer only.
+        """
+        return char.isascii() and char.isdigit()
+
+    def is_letter(char: str) -> bool:
+        """Check whether char is a plain ASCII letter
+
+        For internal use by the lexer only.
+        """
+        return char.isascii() and char.isalpha()
+
+    def is_name_start(char: str) -> bool:
+        """Check whether char is allowed at the beginning of a GraphQL name
+
+        For internal use by the lexer only.
+        """
+        return char.isascii() and (char.isalpha() or char == "_")
+
+    def is_name_continue(char: str) -> bool:
+        """Check whether char is allowed in the continuation of a GraphQL name
+
+        For internal use by the lexer only.
+        """
+        return char.isascii() and (char.isalnum() or char == "_")

src/graphql/language/lexer.py

@@ -524,7 +524,7 @@ def read_hex_digit(char: str) -> int:
         return ord(char) - 48
     elif "A" <= char <= "F":
         return ord(char) - 55
-    elif "a" <= char <= "f":  # a-f
+    elif "a" <= char <= "f":
         return ord(char) - 87
     return -1
 

tests/language/test_character_classes.py

@@ -0,0 +1,90 @@
+from string import ascii_letters as letters, digits, punctuation
+
+from graphql.language.character_classes import (
+    is_digit,
+    is_letter,
+    is_name_start,
+    is_name_continue,
+)
+
+non_ascii = "¯＿±¹²³½£ºµÄäÖöØø×〇᧐〸αΑωΩ"
+
+
+def describe_digit():
+    def accepts_digits():
+        assert all(is_digit(char) for char in digits)
+
+    def rejects_letters():
+        assert not any(is_digit(char) for char in letters)
+
+    def rejects_underscore():
+        assert not is_digit("_")
+
+    def rejects_punctuation():
+        assert not any(is_digit(char) for char in punctuation)
+
+    def rejects_non_ascii():
+        assert not any(is_digit(char) for char in non_ascii)
+
+    def rejects_empty_string():
+        assert not is_digit("")
+
+
+def describe_letter():
+    def accepts_letters():
+        assert all(is_letter(char) for char in letters)
+
+    def rejects_digits():
+        assert not any(is_letter(char) for char in digits)
+
+    def rejects_underscore():
+        assert not is_letter("_")
+
+    def rejects_punctuation():
+        assert not any(is_letter(char) for char in punctuation)
+
+    def rejects_non_ascii():
+        assert not any(is_letter(char) for char in non_ascii)
+
+    def rejects_empty_string():
+        assert not is_letter("")
+
+
+def describe_name_start():
+    def accepts_letters():
+        assert all(is_name_start(char) for char in letters)
+
+    def accepts_underscore():
+        assert is_name_start("_")
+
+    def rejects_digits():
+        assert not any(is_name_start(char) for char in digits)
+
+    def rejects_punctuation():
+        assert not any(is_name_start(char) for char in punctuation if char != "_")
+
+    def rejects_non_ascii():
+        assert not any(is_name_start(char) for char in non_ascii)
+
+    def rejects_empty_string():
+        assert not is_name_start("")
+
+
+def describe_name_continue():
+    def accepts_letters():
+        assert all(is_name_continue(char) for char in letters)
+
+    def accepts_digits():
+        assert all(is_name_continue(char) for char in digits)
+
+    def accepts_underscore():
+        assert is_name_continue("_")
+
+    def rejects_punctuation():
+        assert not any(is_name_continue(char) for char in punctuation if char != "_")
+
+    def rejects_non_ascii():
+        assert not any(is_name_continue(char) for char in non_ascii)
+
+    def rejects_empty_string():
+        assert not is_name_continue("")