From 99e378286865c1eb7e3105641b8f91ab33729d42 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Germ=C3=A1n=20Larra=C3=ADn?=
 <glarrain@users.noreply.github.com>
Date: Wed, 24 Apr 2019 18:53:01 -0400
Subject: [PATCH] libs: add module `encoding_utils`

---
 cl_sii/libs/encoding_utils.py     | 55 +++++++++++++++++++++++++++++++
 tests/test_libs_encoding_utils.py | 18 ++++++++++
 2 files changed, 73 insertions(+)
 create mode 100644 cl_sii/libs/encoding_utils.py
 create mode 100644 tests/test_libs_encoding_utils.py

diff --git a/cl_sii/libs/encoding_utils.py b/cl_sii/libs/encoding_utils.py
new file mode 100644
index 00000000..9a6adb14
--- /dev/null
+++ b/cl_sii/libs/encoding_utils.py
@@ -0,0 +1,55 @@
+import base64
+import binascii
+from typing import Union
+
+
+def clean_base64(value: Union[str, bytes]) -> bytes:
+    """
+    Force bytes and remove line breaks and spaces.
+
+    Does not validate base64 format.
+
+    :raises ValueError:
+    :raises TypeError:
+
+    """
+    if isinstance(value, bytes):
+        value_base64_bytes = value
+    elif isinstance(value, str):
+        try:
+            value_base64_bytes = value.strip().encode(encoding='ascii', errors='strict')
+        except UnicodeEncodeError as exc:
+            raise ValueError("Only ASCII characters are accepted.", str(exc)) from exc
+    else:
+        raise TypeError("Value must be str or bytes.")
+
+    # remove line breaks and spaces
+    value_base64_bytes_cleaned = value_base64_bytes.replace(b'\n', b'').replace(b' ', b'')
+
+    return value_base64_bytes_cleaned
+
+
+def decode_base64_strict(value: Union[str, bytes]) -> bytes:
+    """
+    Strict conversion for str/bytes, tolerating only line breaks and spaces.
+
+    :raises ValueError: non-base64 input or non-ASCII characters included
+
+    """
+    value_base64_bytes_cleaned = clean_base64(value)
+    try:
+        value_bytes = base64.b64decode(value_base64_bytes_cleaned, validate=True)
+    except binascii.Error as exc:
+        raise ValueError("Input is not a valid base64 value.", str(exc)) from exc
+    return value_bytes
+
+
+def validate_base64(value: Union[str, bytes]) -> None:
+    """
+    Validate that ``value`` is base64-encoded data.
+
+    :raises ValueError:
+    :raises TypeError:
+
+    """
+    decode_base64_strict(value)
diff --git a/tests/test_libs_encoding_utils.py b/tests/test_libs_encoding_utils.py
new file mode 100644
index 00000000..0065d6f6
--- /dev/null
+++ b/tests/test_libs_encoding_utils.py
@@ -0,0 +1,18 @@
+import unittest
+
+from cl_sii.libs.encoding_utils import clean_base64, decode_base64_strict, validate_base64  # noqa: F401,E501
+
+
+class FunctionsTest(unittest.TestCase):
+
+    def test_clean_base64(self):
+        # TODO: implement for function 'clean_base64'.
+        pass
+
+    def test_decode_base64_strict(self):
+        # TODO: implement for function 'decode_base64_strict'.
+        pass
+
+    def test_validate_base64(self):
+        # TODO: implement for function 'validate_base64'.
+        pass