py3: remove csv_reader and makedirs from utils/compat.py

dvc/cache.py

@@ -53,7 +53,7 @@ def getter(self):
 
         return Remote(self.repo, name=remote)
 
-    getter.__name__ = str(name)
+    getter.__name__ = name
     return cached_property(getter)
 
 

dvc/daemon.py

@@ -9,7 +9,6 @@
 from dvc.env import DVC_DAEMON
 from dvc.utils import fix_env
 from dvc.utils import is_binary
-from dvc.utils.compat import cast_bytes_py2
 
 
 logger = logging.getLogger(__name__)
@@ -98,9 +97,7 @@ def daemon(args):
 
     env = fix_env()
     file_path = os.path.abspath(inspect.stack()[0][1])
-    env[cast_bytes_py2("PYTHONPATH")] = cast_bytes_py2(
-        os.path.dirname(os.path.dirname(file_path))
-    )
-    env[cast_bytes_py2(DVC_DAEMON)] = cast_bytes_py2("1")
+    env["PYTHONPATH"] = os.path.dirname(os.path.dirname(file_path))
+    env[DVC_DAEMON] = "1"
 
     _spawn(cmd, env)

dvc/logger.py

@@ -7,7 +7,6 @@
 import colorama
 
 from dvc.progress import Tqdm
-from dvc.utils.compat import RecursionError
 
 
 FOOTER = (

dvc/path_info.py

@@ -51,7 +51,7 @@ def __str__(self):
         return relpath(path)
 
     def __repr__(self):
-        return str("{}: '{}'").format(type(self).__name__, self)
+        return "{}: '{}'".format(type(self).__name__, self)
 
     # This permits passing it to file utils directly in Python 3.6+
     # With Python 2.7, Python 3.5+ we are stuck with path_info.fspath for now

dvc/repo/metrics/show.py

@@ -10,7 +10,6 @@
 from dvc.exceptions import NoMetricsError
 from dvc.exceptions import OutputNotFoundError
 from dvc.repo import locked
-from dvc.utils.compat import csv_reader
 from dvc.utils.compat import open
 
 NO_METRICS_FILE_AT_REFERENCE_WARNING = (
@@ -101,7 +100,7 @@ def _format_csv(content, delimiter):
         "0.67528    0.289545        testing\n"
         "0.671502   0.297848        validation\n"
     """
-    reader = csv_reader(io.StringIO(content), delimiter=delimiter)
+    reader = csv.reader(io.StringIO(content), delimiter=delimiter)
     rows = [row for row in reader]
     max_widths = [max(map(len, column)) for column in zip(*rows)]
 

dvc/rwlock.py

@@ -9,12 +9,7 @@
 
 from .exceptions import DvcException
 from .lock import LockError
-from .utils.compat import (
-    convert_to_unicode,
-    FileNotFoundError,
-    JSONDecodeError,
-    str,
-)
+from .utils.compat import FileNotFoundError, JSONDecodeError, str
 from .utils.fs import relpath
 
 INFO_SCHEMA = {"pid": int, "cmd": str}
@@ -44,7 +39,7 @@ def _edit_rwlock(lock_dir):
     try:
         with open(path, "r") as fobj:
             lock = json.load(fobj)
-        lock = SCHEMA(convert_to_unicode(lock))
+        lock = SCHEMA(lock)
     except FileNotFoundError:
         lock = SCHEMA({})
     except JSONDecodeError as exc:

dvc/scm/git/__init__.py

@@ -17,7 +17,6 @@
 from dvc.utils import is_binary
 from dvc.utils import relpath
 from dvc.utils.fs import path_isin
-from dvc.utils.compat import cast_bytes_py2
 from dvc.utils.compat import open
 
 
@@ -76,7 +75,7 @@ def clone(url, to_path, rev=None):
             # LD_LIBRARY_PATH that has been set by PyInstaller.
             # See [1] for more info.
             # [1] https://github.com/gitpython-developers/GitPython/issues/924
-            env[cast_bytes_py2(ld_key)] = ""
+            env[ld_key] = ""
 
         try:
             tmp_repo = git.Repo.clone_from(

dvc/stage.py

@@ -421,10 +421,8 @@ def update(self):
 
     @staticmethod
     def validate(d, fname=None):
-        from dvc.utils.compat import convert_to_unicode
-
         try:
-            Stage.COMPILED_SCHEMA(convert_to_unicode(d))
+            Stage.COMPILED_SCHEMA(d)
         except MultipleInvalid as exc:
             raise StageFileFormatError(fname, exc)
 

dvc/utils/__init__.py

@@ -15,10 +15,8 @@
 from ruamel.yaml import YAML
 from shortuuid import uuid
 
-from dvc.utils.compat import cast_bytes_py2
 from dvc.utils.compat import fspath
 from dvc.utils.compat import fspath_py35
-from dvc.utils.compat import makedirs as _makedirs
 from dvc.utils.compat import open
 
 
@@ -143,15 +141,15 @@ def makedirs(path, exist_ok=False, mode=None):
     path = fspath_py35(path)
 
     if mode is None:
-        _makedirs(path, exist_ok=exist_ok)
+        os.makedirs(path, exist_ok=exist_ok)
         return
 
     # utilize umask to set proper permissions since Python 3.7 the `mode`
     # `makedirs` argument no longer affects the file permission bits of
     # newly-created intermediate-level directories.
     umask = os.umask(0o777 - mode)
     try:
-        _makedirs(path, exist_ok=exist_ok)
+        os.makedirs(path, exist_ok=exist_ok)
     finally:
         os.umask(umask)
 
@@ -205,8 +203,7 @@ def fix_env(env=None):
         lp_key = "LD_LIBRARY_PATH"
         lp_orig = env.get(lp_key + "_ORIG", None)
         if lp_orig is not None:
-            # NOTE: py2 doesn't like unicode strings in environ
-            env[cast_bytes_py2(lp_key)] = cast_bytes_py2(lp_orig)
+            env[lp_key] = lp_orig
         else:
             env.pop(lp_key, None)
 

dvc/utils/compat.py

@@ -1,7 +1,6 @@
 """Handle import compatibility between Python 2 and Python 3"""
 
 import errno
-import os
 import sys
 from contextlib import contextmanager
 
@@ -14,84 +13,6 @@
 #: Python 3.x?
 is_py3 = _ver[0] == 3
 
-# simplified version of ipython_genutils/encoding.py
-DEFAULT_ENCODING = sys.getdefaultencoding()
-
-if _ver[:2] < (3, 5):
-    RecursionError = RuntimeError
-else:
-    RecursionError = RecursionError
-
-
-def no_code(x, encoding=None):
-    return x
-
-
-def encode(u, encoding=None):
-    encoding = encoding or DEFAULT_ENCODING
-    return u.encode(encoding, "replace")
-
-
-def csv_reader(unicode_csv_data, dialect=None, **kwargs):
-    """csv.reader doesn't support Unicode input, so need to use some tricks
-    to work around this.
-
-    Source: https://docs.python.org/2/library/csv.html#csv-examples
-    """
-    import csv
-
-    dialect = dialect or csv.excel
-
-    if is_py3:
-        # Python3 supports encoding by default, so just return the object
-        for row in csv.reader(unicode_csv_data, dialect=dialect, **kwargs):
-            yield [cell for cell in row]
-
-    else:
-        # csv.py doesn't do Unicode; encode temporarily as UTF-8:
-        reader = csv.reader(
-            utf_8_encoder(unicode_csv_data), dialect=dialect, **kwargs
-        )
-        for row in reader:
-            # decode UTF-8 back to Unicode, cell by cell:
-            yield [unicode(cell, "utf-8") for cell in row]  # noqa: F821
-
-
-def utf_8_encoder(unicode_csv_data):
-    """Source: https://docs.python.org/2/library/csv.html#csv-examples"""
-    for line in unicode_csv_data:
-        yield line.encode("utf-8")
-
-
-def cast_bytes(s, encoding=None):
-    """Source: https://github.com/ipython/ipython_genutils"""
-    if not isinstance(s, bytes):
-        return encode(s, encoding)
-    return s
-
-
-def _makedirs(name, mode=0o777, exist_ok=False):
-    """Source: https://github.com/python/cpython/blob/
-        3ce3dea60646d8a5a1c952469a2eb65f937875b3/Lib/os.py#L196-L226
-    """
-    head, tail = os.path.split(name)
-    if not tail:
-        head, tail = os.path.split(head)
-    if head and tail and not os.path.exists(head):
-        try:
-            _makedirs(head, exist_ok=exist_ok)
-        except OSError as e:
-            if e.errno != errno.EEXIST:
-                raise
-        cdir = os.curdir
-        if tail == cdir:
-            return
-    try:
-        os.mkdir(name, mode)
-    except OSError:
-        if not exist_ok or not os.path.isdir(name):
-            raise
-
 
 @contextmanager
 def ignore_file_not_found():
@@ -118,8 +39,6 @@ def ignore_file_not_found():
     numeric_types = (int, long, float)  # noqa: F821
     integer_types = (int, long)  # noqa: F821
     input = raw_input  # noqa: F821
-    cast_bytes_py2 = cast_bytes
-    makedirs = _makedirs
     range = xrange  # noqa: F821
     FileNotFoundError = IOError
     JSONDecodeError = ValueError
@@ -141,19 +60,9 @@ def __enter__(self):
         def __exit__(self, *args):
             self.close()
 
-    def convert_to_unicode(data):
-        if isinstance(data, builtin_str):
-            return str(data)
-        if isinstance(data, dict):
-            return dict(map(convert_to_unicode, data.items()))
-        if isinstance(data, (list, tuple)):
-            return type(data)(map(convert_to_unicode, data))
-        return data
-
 
 elif is_py3:
     import pathlib  # noqa: F401
-    from os import makedirs  # noqa: F401
     from urllib.parse import (  # noqa: F401
         urlparse,  # noqa: F401
         urlunparse,  # noqa: F401
@@ -174,13 +83,9 @@ def convert_to_unicode(data):
     integer_types = (int,)  # noqa: F821
     input = input  # noqa: F821
     open = open  # noqa: F821
-    cast_bytes_py2 = no_code
     range = range  # noqa: F821
     FileNotFoundError = FileNotFoundError
 
-    def convert_to_unicode(data):
-        return data
-
 
 # Backport os.fspath() from Python 3.6
 try:

dvc/version.py

@@ -49,7 +49,7 @@ def _is_release(dir_path, base_version):
             ["git", "describe", "--tags", "--exact-match"],
             cwd=dir_path,
             stderr=subprocess.STDOUT,
-        ).decode("utf-8")
+        )
         tag = output.strip()
         return tag == base_version
     except subprocess.CalledProcessError:

tests/conftest.py

@@ -7,15 +7,14 @@
 
 from dvc.remote.ssh.connection import SSHConnection
 from dvc.repo import Repo as DvcRepo
-from dvc.utils.compat import cast_bytes_py2
 from .basic_env import TestDirFixture, TestDvcGitFixture, TestGitFixture
 from .dir_helpers import *  # noqa
 
 
 # Prevent updater and analytics from running their processes
-os.environ[cast_bytes_py2("DVC_TEST")] = cast_bytes_py2("true")
+os.environ["DVC_TEST"] = "true"
 # Ensure progress output even when not outputting to raw sys.stderr console
-os.environ[cast_bytes_py2("DVC_IGNORE_ISATTY")] = cast_bytes_py2("true")
+os.environ["DVC_IGNORE_ISATTY"] = "true"
 
 
 @pytest.fixture(autouse=True)

tests/unit/test_version.py

@@ -3,7 +3,6 @@
 import mock
 
 import dvc.version
-from dvc.utils.compat import cast_bytes
 
 
 def test_is_release():
@@ -13,10 +12,10 @@ def test_is_release():
         assert ret is False
 
         m.side_effect = None
-        m.return_value = cast_bytes(dvc.version._BASE_VERSION)
+        m.return_value = dvc.version._BASE_VERSION
         ret = dvc.version._is_release(None, dvc.version._BASE_VERSION)
         assert ret
 
-        m.return_value = cast_bytes("630d1741c2d5dd89a3176bd15b63121b905d35c9")
+        m.return_value = "630d1741c2d5dd89a3176bd15b63121b905d35c9"
         ret = dvc.version._is_release(None, dvc.version._BASE_VERSION)
         assert ret is False