Support text open(), though not by default (python#32)

Closes python#31

Author: warsaw

importlib_resources/_py2.py

@@ -2,10 +2,10 @@
 import tempfile
 
 from ._compat import FileNotFoundError
-from __builtin__ import open as builtin_open
+from ._util import _wrap_file
 from contextlib import contextmanager
 from importlib import import_module
-from io import BytesIO
+from io import BytesIO, open as io_open
 from pathlib2 import Path
 
 
@@ -32,17 +32,21 @@ def _normalize_path(path):
         return file_name
 
 
-def open(package, file_name):
-    """Return a file-like object opened for binary-reading of the resource."""
+def open(package, file_name, encoding=None, errors=None):
+    """Return a file-like object opened for reading of the resource."""
     file_name = _normalize_path(file_name)
     package = _get_package(package)
     # Using pathlib doesn't work well here due to the lack of 'strict' argument
     # for pathlib.Path.resolve() prior to Python 3.6.
     package_path = os.path.dirname(package.__file__)
     relative_path = os.path.join(package_path, file_name)
     full_path = os.path.abspath(relative_path)
+    if encoding is None:
+        args = dict(mode='rb')
+    else:
+        args = dict(mode='r', encoding=encoding, errors=errors)
     try:
-        return builtin_open(full_path, 'rb')
+        return io_open(full_path, **args)
     except IOError:
         # This might be a package in a zip file.  zipimport provides a loader
         # with a functioning get_data() method, however we have to strip the
@@ -60,7 +64,7 @@ def open(package, file_name):
                 file_name, package_name)
             raise FileNotFoundError(message)
         else:
-            return BytesIO(data)
+            return _wrap_file(BytesIO(data), encoding, errors)
 
 
 def read(package, file_name, encoding='utf-8', errors='strict'):

importlib_resources/_py3.py

@@ -3,6 +3,7 @@
 import tempfile
 
 from . import abc as resources_abc
+from ._util import _wrap_file
 from builtins import open as builtins_open
 from contextlib import contextmanager
 from importlib import import_module
@@ -12,7 +13,7 @@
 from types import ModuleType
 from typing import Iterator, Union
 from typing import cast
-from typing.io import BinaryIO
+from typing.io import IO
 
 
 Package = Union[ModuleType, str]
@@ -46,21 +47,28 @@ def _normalize_path(path) -> str:
         return file_name
 
 
-def open(package: Package, file_name: FileName) -> BinaryIO:
-    """Return a file-like object opened for binary-reading of the resource."""
+def open(package: Package,
+         file_name: FileName,
+         encoding: str = None,
+         errors: str = None) -> IO:
+    """Return a file-like object opened for reading of the resource."""
     file_name = _normalize_path(file_name)
     package = _get_package(package)
     if hasattr(package.__spec__.loader, 'open_resource'):
         reader = cast(resources_abc.ResourceReader, package.__spec__.loader)
-        return reader.open_resource(file_name)
+        return _wrap_file(reader.open_resource(file_name), encoding, errors)
     else:
         # Using pathlib doesn't work well here due to the lack of 'strict'
         # argument for pathlib.Path.resolve() prior to Python 3.6.
         absolute_package_path = os.path.abspath(package.__spec__.origin)
         package_path = os.path.dirname(absolute_package_path)
         full_path = os.path.join(package_path, file_name)
+        if encoding is None:
+            args = dict(mode='rb')
+        else:
+            args = dict(mode='r', encoding=encoding, errors=errors)
         try:
-            return builtins_open(full_path, 'rb')
+            return builtins_open(full_path, **args)   # type: ignore
         except IOError:
             # Just assume the loader is a resource loader; all the relevant
             # importlib.machinery loaders are and an AttributeError for
@@ -74,7 +82,7 @@ def open(package: Package, file_name: FileName) -> BinaryIO:
                     file_name, package_name)
                 raise FileNotFoundError(message)
             else:
-                return BytesIO(data)
+                return _wrap_file(BytesIO(data), encoding, errors)
 
 
 def read(package: Package,

importlib_resources/_util.py

@@ -0,0 +1,7 @@
+from io import TextIOWrapper
+
+
+def _wrap_file(resource, encoding, errors):
+    if encoding is None:
+        return resource
+    return TextIOWrapper(resource, encoding=encoding, errors=errors)

importlib_resources/tests/test_open.py

@@ -17,22 +17,27 @@ class OpenTests:
 
     # Subclasses are expected to set the 'data' attribute.
 
-    def test_opened_for_reading(self):
-        # The file-like object is ready for reading.
-        with resources.open(self.data, 'utf-8.file') as file:
-            self.assertEqual(b"Hello, UTF-8 world!\n", file.read())
+    def test_open_for_binary(self):
+        # By default, the resource is opened for binary reads.
+        with resources.open(self.data, 'utf-8.file') as fp:
+            self.assertEqual(b'Hello, UTF-8 world!\n', fp.read())
 
     def test_wrap_for_text(self):
         # The file-like object can be wrapped for text reading.
-        with resources.open(self.data, 'utf-8.file') as file:
-            text = file.read().decode(encoding='utf-8')
+        with resources.open(self.data, 'utf-8.file') as fp:
+            text = fp.read().decode(encoding='utf-8')
             self.assertEqual('Hello, UTF-8 world!\n', text)
 
     def test_FileNotFoundError(self):
         with self.assertRaises(FileNotFoundError):
             with resources.open(self.data, 'does-not-exist'):
                 pass
 
+    def test_open_for_text(self):
+        # open() takes an optional encoding and errors parameter.
+        with resources.open(self.data, 'utf-8.file', 'utf-8', 'strict') as fp:
+            self.assertEqual('Hello, UTF-8 world!\n', fp.read())
+
 
 class OpenDiskTests(OpenTests, unittest.TestCase):