CM-60540: remove binaryornot dep

Author: gotbadger

cycode/cli/utils/binary_utils.py

@@ -0,0 +1,42 @@
+import logging
+
+logger = logging.getLogger(__name__)
+
+_CONTROL_CHARS = b'\n\r\t\f\b'
+_PRINTABLE_ASCII = _CONTROL_CHARS + bytes(range(32, 127))
+_PRINTABLE_HIGH_ASCII = bytes(range(127, 256))
+
+
+def is_binary_string(bytes_to_check: bytes) -> bool:
+    """Check if a chunk of bytes appears to be binary content.
+
+    Uses a simplified version of the Perl detection algorithm.
+    """
+    if not bytes_to_check:
+        return False
+
+    # Null bytes are a strong binary indicator
+    if b'\x00' in bytes_to_check:
+        return True
+
+    # Binary if control chars are > 30% of the string
+    low_chars = bytes_to_check.translate(None, _PRINTABLE_ASCII)
+    nontext_ratio1 = len(low_chars) / len(bytes_to_check)
+
+    # Binary if high ASCII chars are < 5% of the string
+    high_chars = bytes_to_check.translate(None, _PRINTABLE_HIGH_ASCII)
+    nontext_ratio2 = len(high_chars) / len(bytes_to_check)
+
+    is_likely_binary = (nontext_ratio1 > 0.3 and nontext_ratio2 < 0.05) or (
+        nontext_ratio1 > 0.8 and nontext_ratio2 > 0.8
+    )
+
+    if not is_likely_binary:
+        return False
+
+    # Try UTF-8 decode to rescue files that look binary by ratios but are valid text
+    try:
+        bytes_to_check.decode('utf-8')
+        return False
+    except UnicodeDecodeError:
+        return True

cycode/cli/utils/path_utils.py

@@ -4,7 +4,7 @@
 from typing import TYPE_CHECKING, AnyStr, Optional, Union
 
 import typer
-from binaryornot.helpers import is_binary_string
+from cycode.cli.utils.binary_utils import is_binary_string
 
 from cycode.cli.logger import logger
 

cycode/cli/utils/string_utils.py

@@ -5,7 +5,7 @@
 import string
 from sys import getsizeof
 
-from binaryornot.check import is_binary_string
+from cycode.cli.utils.binary_utils import is_binary_string
 
 from cycode.cli.consts import SCA_SHORTCUT_DEPENDENCY_PATHS
 

cycode/logger.py

@@ -31,8 +31,6 @@ def _set_io_encodings() -> None:
 logging.getLogger('werkzeug').setLevel(logging.WARNING)
 logging.getLogger('schedule').setLevel(logging.WARNING)
 logging.getLogger('kubernetes').setLevel(logging.WARNING)
-logging.getLogger('binaryornot').setLevel(logging.WARNING)
-logging.getLogger('chardet').setLevel(logging.WARNING)
 logging.getLogger('git.cmd').setLevel(logging.WARNING)
 logging.getLogger('git.util').setLevel(logging.WARNING)
 

poetry.lock

@@ -39,7 +39,6 @@ pyyaml = ">=6.0,<7.0"
 marshmallow = ">=3.15.0,<4.0.0"
 gitpython = ">=3.1.30,<3.2.0"
 arrow = ">=1.0.0,<1.4.0"
-binaryornot = ">=0.4.4,<0.5.0"
 requests = ">=2.32.4,<3.0"
 urllib3 = ">=2.4.0,<3.0.0"
 pyjwt = ">=2.8.0,<3.0"