py2: add unicode literals

Fix #1512

dvc/__init__.py

@@ -3,6 +3,9 @@ DVC
 ----
 Make your data science projects reproducible and shareable.
 """
+
+from __future__ import unicode_literals
+
 import os
 import warnings
 

dvc/__main__.py

@@ -1,5 +1,7 @@
 """Main entry point for dvc command line tool."""
 
+from __future__ import unicode_literals
+
 import sys
 
 from dvc.main import main

dvc/analytics.py

@@ -1,5 +1,9 @@
 """Collect and send usage analytics"""
 
+from __future__ import unicode_literals
+
+from dvc.utils.compat import str
+
 import os
 import json
 import errno

dvc/cache.py

@@ -1,5 +1,7 @@
 """Manages cache of a dvc project."""
 
+from __future__ import unicode_literals
+
 import os
 
 from dvc.config import Config

dvc/cli.py

@@ -1,6 +1,7 @@
 """Command line interface for dvc."""
 
 from __future__ import print_function
+from __future__ import unicode_literals
 
 import sys
 import argparse

dvc/command/add.py

@@ -1,3 +1,5 @@
+from __future__ import unicode_literals
+
 import dvc.logger as logger
 from dvc.exceptions import DvcException
 from dvc.command.base import CmdBase

dvc/command/base.py

@@ -1,4 +1,4 @@
-import sys
+from __future__ import unicode_literals
 
 import dvc.logger as logger
 
@@ -11,7 +11,9 @@ def fix_subparsers(subparsers):
         Args:
             subparsers: subparsers to fix.
     """
-    if sys.version_info[0] == 3:  # pragma: no cover
+    from dvc.utils.compat import is_py3
+
+    if is_py3:  # pragma: no cover
         subparsers.required = True
         subparsers.dest = 'cmd'
 

dvc/command/cache.py

@@ -1,3 +1,5 @@
+from __future__ import unicode_literals
+
 import argparse
 
 from dvc.command.base import fix_subparsers

dvc/command/checkout.py

@@ -1,3 +1,5 @@
+from __future__ import unicode_literals
+
 from dvc.command.base import CmdBase
 
 

dvc/command/config.py

@@ -1,3 +1,5 @@
+from __future__ import unicode_literals
+
 import os
 import argparse
 

dvc/command/daemon.py

@@ -1,3 +1,5 @@
+from __future__ import unicode_literals
+
 from dvc.command.base import CmdBase, fix_subparsers
 
 

dvc/command/data_sync.py

@@ -1,3 +1,5 @@
+from __future__ import unicode_literals
+
 import argparse
 
 import dvc.logger as logger

dvc/command/destroy.py

@@ -1,3 +1,5 @@
+from __future__ import unicode_literals
+
 import dvc.prompt as prompt
 import dvc.logger as logger
 from dvc.command.base import CmdBase

dvc/command/gc.py

@@ -1,3 +1,5 @@
+from __future__ import unicode_literals
+
 import os
 import dvc.prompt as prompt
 import dvc.logger as logger

dvc/command/imp.py

@@ -1,3 +1,5 @@
+from __future__ import unicode_literals
+
 import os
 
 import dvc.logger as logger

dvc/command/init.py

@@ -1,3 +1,5 @@
+from __future__ import unicode_literals
+
 import dvc.logger as logger
 
 

dvc/command/install.py

@@ -1,3 +1,5 @@
+from __future__ import unicode_literals
+
 import dvc.logger as logger
 from dvc.command.base import CmdBase
 

dvc/command/lock.py

@@ -1,3 +1,5 @@
+from __future__ import unicode_literals
+
 import dvc.logger as logger
 from dvc.exceptions import DvcException
 from dvc.command.base import CmdBase

dvc/command/metrics.py

@@ -1,3 +1,5 @@
+from __future__ import unicode_literals
+
 import dvc.logger as logger
 from dvc.exceptions import DvcException
 from dvc.command.base import CmdBase, fix_subparsers

dvc/command/move.py

@@ -1,3 +1,5 @@
+from __future__ import unicode_literals
+
 import dvc.logger as logger
 from dvc.exceptions import DvcException
 from dvc.command.base import CmdBase

dvc/command/pipeline.py

@@ -1,3 +1,7 @@
+from __future__ import unicode_literals
+
+from dvc.utils.compat import str
+
 import os
 
 import dvc.logger as logger

dvc/command/remote.py

@@ -1,3 +1,5 @@
+from __future__ import unicode_literals
+
 import os
 import re
 

dvc/command/remove.py

@@ -1,3 +1,5 @@
+from __future__ import unicode_literals
+
 import dvc.prompt as prompt
 import dvc.logger as logger
 from dvc.exceptions import DvcException
@@ -13,7 +15,7 @@ class CmdRemove(CmdBase):
             return False
 
         msg = (
-            u'Are you sure you want to remove {} with its outputs?'
+            'Are you sure you want to remove {} with its outputs?'
             .format(target)
         )
 
@@ -21,8 +23,8 @@ class CmdRemove(CmdBase):
             return False
 
         raise DvcException(
-            u'Cannot purge without a confirmation from the user.'
-            u" Use '-f' to force."
+            'Cannot purge without a confirmation from the user.'
+            " Use '-f' to force."
         )
 
     def run(self):

dvc/command/repro.py

@@ -1,3 +1,5 @@
+from __future__ import unicode_literals
+
 import os
 
 import dvc.logger as logger

dvc/command/root.py

@@ -1,3 +1,5 @@
+from __future__ import unicode_literals
+
 import os
 
 import dvc.logger as logger

dvc/command/run.py

@@ -1,3 +1,5 @@
+from __future__ import unicode_literals
+
 import os
 import argparse
 

dvc/command/status.py

@@ -1,3 +1,5 @@
+from __future__ import unicode_literals
+
 import dvc.logger as logger
 from dvc.command.data_sync import CmdDataBase
 

dvc/command/unprotect.py

@@ -1,3 +1,5 @@
+from __future__ import unicode_literals
+
 import dvc.logger as logger
 from dvc.exceptions import DvcException
 from dvc.command.base import CmdBase

dvc/config.py

@@ -1,6 +1,9 @@
-"""
-DVC config objects.
-"""
+"""DVC config objects."""
+
+from __future__ import unicode_literals
+
+from dvc.utils.compat import str
+
 import os
 import errno
 import configobj
@@ -18,8 +21,9 @@ class ConfigError(DvcException):
         ex (Exception): optional exception that has caused this error.
     """
     def __init__(self, msg, ex=None):
-        super(ConfigError, self).__init__('config file error: {}'.format(msg),
-                                          ex)
+        super(ConfigError, self).__init__(
+            'config file error: {}'.format(msg), ex
+        )
 
 
 def supported_cache_type(types):
@@ -531,6 +535,6 @@ class Config(object):  # pylint: disable=too-many-instance-attributes
         for s_key, s_value in config.items():
             new_s = {}
             for key, value in s_value.items():
-                new_s[key.lower()] = value
+                new_s[key.lower()] = str(value)
             new_config[s_key.lower()] = new_s
         return new_config

dvc/daemon.py

@@ -1,5 +1,7 @@
 """Launch `dvc daemon` command in a separate detached process."""
 
+from __future__ import unicode_literals
+
 import os
 import sys
 import inspect

dvc/dagascii.py

@@ -1,4 +1,8 @@
 """Draws DAG in ASCII."""
+
+from __future__ import unicode_literals
+from __future__ import print_function
+
 import sys
 import math
 

dvc/data_cloud.py

@@ -1,5 +1,7 @@
 """Manages dvc remotes that user can use with push/pull/status comamnds."""
 
+from __future__ import unicode_literals
+
 import dvc.logger as logger
 from dvc.config import Config, ConfigError
 

dvc/dependency/__init__.py

@@ -1,3 +1,5 @@
+from __future__ import unicode_literals
+
 import schema
 
 from dvc.config import Config

dvc/dependency/base.py

@@ -1,3 +1,5 @@
+from __future__ import unicode_literals
+
 from dvc.exceptions import DvcException
 
 

dvc/dependency/gs.py

@@ -1,3 +1,5 @@
+from __future__ import unicode_literals
+
 from dvc.output.gs import OutputGS
 from dvc.dependency.base import DependencyBase
 

dvc/dependency/hdfs.py

@@ -1,3 +1,5 @@
+from __future__ import unicode_literals
+
 from dvc.output.hdfs import OutputHDFS
 from dvc.dependency.base import DependencyBase
 

dvc/dependency/http.py

@@ -1,3 +1,5 @@
+from __future__ import unicode_literals
+
 from dvc.utils.compat import urlparse, urljoin
 from dvc.output.base import OutputBase
 from dvc.remote.http import RemoteHTTP

dvc/dependency/local.py

@@ -1,3 +1,5 @@
+from __future__ import unicode_literals
+
 from dvc.dependency.base import DependencyBase
 from dvc.output.local import OutputLOCAL
 

dvc/dependency/s3.py

@@ -1,3 +1,5 @@
+from __future__ import unicode_literals
+
 from dvc.output.s3 import OutputS3
 from dvc.dependency.base import DependencyBase
 

dvc/dependency/ssh.py

@@ -1,3 +1,5 @@
+from __future__ import unicode_literals
+
 from dvc.output.ssh import OutputSSH
 from dvc.dependency.base import DependencyBase
 

dvc/exceptions.py

@@ -1,5 +1,9 @@
 """Exceptions raised by the dvc."""
 
+from __future__ import unicode_literals
+
+from dvc.utils.compat import str, builtin_str
+
 import traceback
 
 
@@ -32,9 +36,12 @@ class OutputDuplicationError(DvcException):
         stages (list): list of paths to stages.
     """
     def __init__(self, output, stages):
-        assert isinstance(output, str)
+        assert isinstance(output, str) or isinstance(output, builtin_str)
         assert isinstance(stages, list)
-        assert all(isinstance(stage, str) for stage in stages)
+        assert all(
+            isinstance(stage, str) or isinstance(stage, builtin_str)
+            for stage in stages
+        )
         msg = ("file/directory '{}' is specified as an output in more than one"
                "stage: {}").format(output, "\n    ".join(stages))
         super(OutputDuplicationError, self).__init__(msg)
@@ -50,8 +57,8 @@ class WorkingDirectoryAsOutputError(DvcException):
             output.
     """
     def __init__(self, cwd, fname):
-        assert isinstance(cwd, str)
-        assert isinstance(fname, str)
+        assert isinstance(cwd, str) or isinstance(cwd, builtin_str)
+        assert isinstance(fname, str) or isinstance(fname, builtin_str)
         msg = (
             "current working directory '{cwd}' is specified as an output in"
             " '{fname}'. Use another CWD to prevent any data removal."
@@ -68,7 +75,9 @@ class CircularDependencyError(DvcException):
         dependency (str): path to the dependency.
     """
     def __init__(self, dependency):
-        assert isinstance(dependency, str)
+        assert (isinstance(dependency, str) or
+                isinstance(dependency, builtin_str))
+
         msg = ("file/directory '{}' is specified as an output and as a "
                "dependency.")
         super(CircularDependencyError, self).__init__(msg.format(dependency))
@@ -82,7 +91,7 @@ class ArgumentDuplicationError(DvcException):
         path (str): path to the file/directory.
     """
     def __init__(self, path):
-        assert isinstance(path, str)
+        assert isinstance(path, str) or isinstance(path, builtin_str)
         msg = "file '{}' is specified more than once."
         super(ArgumentDuplicationError, self).__init__(msg.format(path))
 

dvc/istextfile.py

@@ -1,18 +1,19 @@
 """Use heuristics to guess if it is a text file or a binary file."""
 
+from __future__ import unicode_literals
+
+from dvc.utils.compat import is_py3
+
 # Based on https://eli.thegreenplace.net/2011/10/19/
 # perls-guess-if-file-is-text-or-binary-implemented-in-python
 
-import sys
-PY3 = sys.version_info[0] == 3
-
 
 # A function that takes an integer in the 8-bit range and returns
 # a single-character byte object in py3 / a single-character string
 # in py2.
 #
 def _int2byte(i):
-    if PY3:
+    if is_py3:
         return bytes((i,))
     return chr(i)
 

dvc/lock.py

@@ -1,5 +1,7 @@
 """Manages dvc lock file."""
 
+from __future__ import unicode_literals
+
 import os
 import time
 import zc.lockfile

dvc/logger.py

@@ -1,5 +1,9 @@
 """Manages logger for dvc project."""
 
+from __future__ import unicode_literals
+
+from dvc.utils.compat import str
+
 import re
 import sys
 import logging
@@ -184,7 +188,7 @@ def colorize(message, color=None):
         'red': colorama.Fore.RED,
     }
 
-    return u'{color}{message}{nc}'.format(
+    return '{color}{message}{nc}'.format(
         color=colors.get(color, ''),
         message=message,
         nc=colorama.Fore.RESET,

dvc/main.py

@@ -1,5 +1,7 @@
 """Main entry point for dvc CLI."""
 
+from __future__ import unicode_literals
+
 import dvc.logger as logger
 from dvc.cli import parse_args
 from dvc.command.base import CmdBase

dvc/output/__init__.py

@@ -1,7 +1,9 @@
+from __future__ import unicode_literals
+
 import schema
 
 from dvc.config import Config
-from dvc.utils.compat import urlparse
+from dvc.utils.compat import urlparse, str
 
 from dvc.output.base import OutputBase
 from dvc.output.s3 import OutputS3

dvc/output/base.py

@@ -1,3 +1,6 @@
+from __future__ import unicode_literals
+from dvc.utils.compat import str
+
 import re
 from schema import Or, Optional
 

dvc/output/gs.py

@@ -1,3 +1,5 @@
+from __future__ import unicode_literals
+
 from dvc.remote.gs import RemoteGS
 from dvc.output.s3 import OutputS3
 

dvc/output/hdfs.py

@@ -1,3 +1,5 @@
+from __future__ import unicode_literals
+
 import posixpath
 
 from dvc.utils.compat import urlparse

dvc/output/local.py

@@ -1,3 +1,5 @@
+from __future__ import unicode_literals
+
 import os
 
 import dvc.logger as logger

dvc/output/s3.py

@@ -1,3 +1,5 @@
+from __future__ import unicode_literals
+
 import posixpath
 
 from dvc.remote.s3 import RemoteS3

dvc/output/ssh.py

@@ -1,3 +1,5 @@
+from __future__ import unicode_literals
+
 import getpass
 import posixpath
 

dvc/progress.py

@@ -1,6 +1,10 @@
 """Manages progress bars for dvc project."""
 
 from __future__ import print_function
+from __future__ import unicode_literals
+
+from dvc.utils.compat import str
+
 import sys
 import threading
 import dvc.logger as logger

dvc/project.py

@@ -1,3 +1,7 @@
+from __future__ import unicode_literals
+
+from dvc.utils.compat import str, builtin_str
+
 import collections
 import os
 import dvc.prompt as prompt
@@ -1001,7 +1005,7 @@ class Project(object):
 
         col, row = hxsv_path.split(',')
         row = int(row)
-        reader = list(csv.DictReader(fd, delimiter=delimiter))
+        reader = list(csv.DictReader(fd, delimiter=builtin_str(delimiter)))
         return self._do_read_metric_xsv(reader, row, col)
 
     def _read_metric_xsv(self, fd, xsv_path, delimiter):
@@ -1010,7 +1014,7 @@ class Project(object):
         col, row = xsv_path.split(',')
         row = int(row)
         col = int(col)
-        reader = list(csv.reader(fd, delimiter=delimiter))
+        reader = list(csv.reader(fd, delimiter=builtin_str(delimiter)))
         return self._do_read_metric_xsv(reader, row, col)
 
     def _read_metric(self, path, typ=None, xpath=None):

dvc/prompt.py

@@ -1,13 +1,12 @@
 """Manages user prompts."""
 
+from __future__ import unicode_literals
+from __future__ import print_function
+
 import sys
 from getpass import getpass
 
-try:
-    # NOTE: in Python3 raw_input() was renamed to input()
-    input = raw_input  # pylint: disable=redefined-builtin, invalid-name
-except NameError:
-    pass
+from dvc.utils.compat import input
 
 
 def _ask(prompt, limited_to=None):

dvc/remote/__init__.py

@@ -1,3 +1,5 @@
+from __future__ import unicode_literals
+
 from dvc.remote.azure import RemoteAzure
 from dvc.remote.gs import RemoteGS
 from dvc.remote.hdfs import RemoteHDFS

dvc/remote/azure.py

@@ -1,4 +1,6 @@
 from __future__ import absolute_import
+from __future__ import unicode_literals
+
 import os
 import re
 

dvc/remote/base.py

@@ -1,3 +1,7 @@
+from __future__ import unicode_literals
+
+from dvc.utils.compat import str
+
 import os
 import re
 import errno

dvc/remote/gs.py

@@ -1,3 +1,5 @@
+from __future__ import unicode_literals
+
 import os
 
 try:

dvc/remote/hdfs.py

@@ -1,3 +1,5 @@
+from __future__ import unicode_literals
+
 import os
 import re
 import getpass

dvc/remote/http.py

@@ -1,3 +1,5 @@
+from __future__ import unicode_literals
+
 import os
 import threading
 import requests

dvc/remote/local.py

@@ -1,3 +1,7 @@
+from __future__ import unicode_literals
+
+from dvc.utils.compat import str
+
 import os
 import stat
 import uuid
@@ -239,12 +243,12 @@ class RemoteLOCAL(RemoteBase):
             with open(path, 'r') as fd:
                 d = json.load(fd)
         except Exception:
-            msg = u"Failed to load dir cache '{}'"
+            msg = "Failed to load dir cache '{}'"
             logger.error(msg.format(os.path.relpath(path)))
             return []
 
         if not isinstance(d, list):
-            msg = u"dir cache file format error '{}' [skipping the file]"
+            msg = "dir cache file format error '{}' [skipping the file]"
             logger.error(msg.format(os.path.relpath(path)))
             return []
 

dvc/remote/s3.py

@@ -1,3 +1,5 @@
+from __future__ import unicode_literals
+
 import os
 import threading
 

dvc/remote/ssh.py

@@ -1,3 +1,5 @@
+from __future__ import unicode_literals
+
 import os
 import getpass
 import posixpath

dvc/scm/__init__.py

@@ -1,5 +1,7 @@
 """Manages source control systems(e.g. Git)."""
 
+from __future__ import unicode_literals
+
 from dvc.scm.base import Base
 from dvc.scm.git import Git
 

dvc/scm/base.py

@@ -1,5 +1,7 @@
 """Manages source control systems(e.g. Git) in dvc."""
 
+from __future__ import unicode_literals
+
 import os
 
 from dvc.exceptions import DvcException

dvc/scm/git.py

@@ -1,9 +1,13 @@
 """Manages Git."""
 
+from __future__ import unicode_literals
+
 # NOTE: need this one because python 2 would re-import this module
 # (called git.py) instead of importing gitpython's git module.
 from __future__ import absolute_import
 
+from dvc.utils.compat import str
+
 import os
 
 import dvc.logger as logger

dvc/stage.py

@@ -1,3 +1,7 @@
+from __future__ import unicode_literals
+
+from dvc.utils.compat import str
+
 import os
 import yaml
 import posixpath
@@ -14,7 +18,7 @@ from dvc.utils import dict_md5, fix_env
 
 class StageCmdFailedError(DvcException):
     def __init__(self, stage):
-        msg = u"stage '{}' cmd {} failed".format(stage.relpath, stage.cmd)
+        msg = "stage '{}' cmd {} failed".format(stage.relpath, stage.cmd)
         super(StageCmdFailedError, self).__init__(msg)
 
 
@@ -72,7 +76,7 @@ class MissingDep(DvcException):
         else:
             dep = 'dependency'
 
-        msg = u'missing {}: {}'.format(dep, ', '.join(map(str, deps)))
+        msg = 'missing {}: {}'.format(dep, ', '.join(map(str, deps)))
         super(MissingDep, self).__init__(msg)
 
 
@@ -84,7 +88,7 @@ class MissingDataSource(DvcException):
         if len(missing_files) > 1:
             source += 's'
 
-        msg = u'missing data {}: {}'.format(source, ', '.join(missing_files))
+        msg = 'missing data {}: {}'.format(source, ', '.join(missing_files))
         super(MissingDataSource, self).__init__(msg)
 
 
@@ -251,7 +255,7 @@ class Stage(object):
         if interactive and not prompt.confirm(msg):
             raise DvcException('reproduction aborted by the user')
 
-        logger.info(u"Reproducing '{stage}'".format(stage=self.relpath))
+        logger.info("Reproducing '{stage}'".format(stage=self.relpath))
 
         self.run(dry=dry)
 
@@ -261,8 +265,10 @@ class Stage(object):
 
     @staticmethod
     def validate(d, fname=None):
+        from dvc.utils import convert_to_unicode
+
         try:
-            Schema(Stage.SCHEMA).validate(d)
+            Schema(Stage.SCHEMA).validate(convert_to_unicode(d))
         except SchemaError as exc:
             raise StageFileFormatError(fname, exc)
 
@@ -556,7 +562,7 @@ class Stage(object):
                     self.deps[0].download(self.outs[0].path_info)
 
         elif self.is_data_source:
-            msg = u"Verifying data sources in '{}'".format(self.relpath)
+            msg = "Verifying data sources in '{}'".format(self.relpath)
             logger.info(msg)
             if not dry:
                 self.check_missing_outputs()

dvc/state.py

@@ -1,5 +1,9 @@
 """Manages state database used for checksum caching."""
 
+from __future__ import unicode_literals
+
+from dvc.utils.compat import str
+
 import os
 import time
 import sqlite3

dvc/system.py

@@ -1,3 +1,7 @@
+from __future__ import unicode_literals
+
+from dvc.utils.compat import str
+
 import os
 
 

dvc/updater.py

@@ -1,3 +1,5 @@
+from __future__ import unicode_literals
+
 import sys
 import os
 import time

dvc/utils/__init__.py

@@ -1,6 +1,8 @@
-"""
-Helpers for other modules.
-"""
+"""Helpers for other modules."""
+
+from __future__ import unicode_literals
+from dvc.utils.compat import str, builtin_str
+
 import os
 import sys
 import stat
@@ -9,11 +11,6 @@ import json
 import shutil
 import hashlib
 
-import dvc.logger as logger
-from dvc.progress import progress
-from dvc.istextfile import istextfile
-
-
 LOCAL_CHUNK_SIZE = 1024*1024
 LARGE_FILE_SIZE = 1024*1024*1024
 LARGE_DIR_SIZE = 100
@@ -25,6 +22,10 @@ def dos2unix(data):
 
 def file_md5(fname):
     """ get the (md5 hexdigest, md5 digest) of a file """
+    import dvc.logger as logger
+    from dvc.progress import progress
+    from dvc.istextfile import istextfile
+
     if os.path.exists(fname):
         hash_md5 = hashlib.md5()
         binary = not istextfile(fname)
@@ -81,6 +82,9 @@ def dict_filter(d, exclude=[]):
     elif isinstance(d, dict):
         ret = {}
         for k, v in d.items():
+            if isinstance(k, builtin_str):
+                k = str(k)
+
             assert isinstance(k, str)
             if k in exclude:
                 continue
@@ -97,7 +101,9 @@ def dict_md5(d, exclude=[]):
 
 
 def copyfile(src, dest, no_progress_bar=False, name=None):
-    '''Copy file with progress bar'''
+    """Copy file with progress bar"""
+    from dvc.progress import progress
+
     copied = 0
     name = name if name else os.path.basename(dest)
     total = os.stat(src).st_size
@@ -140,10 +146,12 @@ def move(src, dst):
 
 
 def remove(path):
+    import dvc.logger as logger
+
     if not os.path.exists(path):
         return
 
-    logger.debug(u"Removing '{}'".format(os.path.relpath(path)))
+    logger.debug("Removing '{}'".format(os.path.relpath(path)))
 
     def _chmod(func, p, excinfo):
         perm = os.stat(p).st_mode
@@ -192,3 +200,14 @@ def fix_env(env=None):
             env.pop(lp_key, None)
 
     return env
+
+
+def convert_to_unicode(data):
+    if isinstance(data, builtin_str):
+        return str(data)
+    elif isinstance(data, dict):
+        return dict(map(convert_to_unicode, data.items()))
+    elif isinstance(data, list) or isinstance(data, tuple):
+        return type(data)(map(convert_to_unicode, data))
+    else:
+        return data

dvc/utils/compat.py

@@ -1,6 +1,45 @@
 """Handle import compatibility between Python 2 and Python 3"""
 
-try:
-    from urlparse import urlparse, urljoin  # noqa: F401
-except ImportError:
+import sys
+
+
+# Syntax sugar.
+_ver = sys.version_info
+
+#: Python 2.x?
+is_py2 = (_ver[0] == 2)
+
+#: Python 3.x?
+is_py3 = (_ver[0] == 3)
+
+if is_py2:
+    from urlparse import urlparse, urljoin                 # noqa: F401
+    from StringIO import StringIO                          # noqa: F401
+    from BaseHTTPServer import HTTPServer                  # noqa: F401
+    from SimpleHTTPServer import SimpleHTTPRequestHandler  # noqa: F401
+    import ConfigParser                                    # noqa: F401
+
+    builtin_str = str                   # noqa: F821
+    bytes = str                         # noqa: F821
+    str = unicode                       # noqa: F821
+    basestring = basestring             # noqa: F821
+    numeric_types = (int, long, float)  # noqa: F821
+    integer_types = (int, long)         # noqa: F821
+    input = raw_input                   # noqa: F821
+
+elif is_py3:
     from urllib.parse import urlparse, urljoin  # noqa: F401
+    from io import StringIO                     # noqa: F401
+    from http.server import (                   # noqa: F401
+        HTTPServer,                             # noqa: F401
+        SimpleHTTPRequestHandler                # noqa: F401
+    )                                           # noqa: F401
+    import configparser as ConfigParser         # noqa: F401
+
+    builtin_str = str             # noqa: F821
+    str = str                     # noqa: F821
+    bytes = bytes                 # noqa: F821
+    basestring = (str, bytes)     # noqa: F821
+    numeric_types = (int, float)  # noqa: F821
+    integer_types = (int,)        # noqa: F821
+    input = input                 # noqa: F821

scripts/innosetup/config_gen.py

@@ -1,10 +1,6 @@
 # This script generates config.ini for setup.iss script
 from dvc import VERSION
-
-try:
-    import configparser as ConfigParser
-except ImportError:
-    import ConfigParser
+from dvc.utils.compat import ConfigParser
 
 config = ConfigParser.ConfigParser()
 config.add_section('Version')

tests/test_add.py

@@ -1,3 +1,5 @@
+from __future__ import unicode_literals
+
 import os
 import yaml
 import time

tests/test_data_cloud.py

@@ -11,6 +11,7 @@ from dvc.state import State
 from mock import patch
 
 import dvc.logger as logger
+from dvc.utils.compat import str
 from dvc.main import main
 from dvc.config import Config
 from dvc.data_cloud import (DataCloud, RemoteS3, RemoteGS, RemoteAzure,
@@ -21,10 +22,7 @@ from dvc.utils import file_md5
 from tests.basic_env import TestDvc
 from tests.utils import spy
 
-try:
-    from StringIO import StringIO
-except ImportError:
-    from io import StringIO
+from dvc.utils.compat import StringIO
 
 TEST_REMOTE = 'upstream'
 TEST_SECTION = 'remote "{}"'.format(TEST_REMOTE)

tests/test_import.py

@@ -1,3 +1,5 @@
+from dvc.utils.compat import str
+
 import os
 from uuid import uuid4
 
@@ -7,10 +9,7 @@ from dvc.main import main
 from mock import patch
 from tests.basic_env import TestDvc
 
-try:
-    from StringIO import StringIO
-except ImportError:
-    from io import StringIO
+from dvc.utils.compat import StringIO
 
 
 class TestCmdImport(TestDvc):

tests/test_logger.py

@@ -2,15 +2,11 @@ import traceback
 from mock import patch
 from unittest import TestCase
 
-try:
-    from StringIO import StringIO
-except ImportError:
-    from io import StringIO
-
 import logging
 import dvc.logger as logger
 from dvc.command.base import CmdBase
 from dvc.exceptions import DvcException
+from dvc.utils.compat import StringIO
 
 
 class TestLogger(TestCase):

tests/test_repro.py

@@ -1,3 +1,5 @@
+from dvc.utils.compat import str
+
 import os
 import yaml
 import shutil

tests/test_state.py

@@ -1,5 +1,7 @@
 import os
 
+from dvc.utils.compat import str
+
 from dvc.state import State
 from dvc.utils import file_md5
 from dvc.main import main

tests/unit/state.py

@@ -1,3 +1,5 @@
+from dvc.utils.compat import str
+
 import os
 
 from dvc.state import State

tests/utils/httpd.py

@@ -2,11 +2,7 @@ import hashlib
 import os
 import threading
 
-try:
-    from http.server import HTTPServer, SimpleHTTPRequestHandler
-except ImportError:
-    from BaseHTTPServer import HTTPServer
-    from SimpleHTTPServer import SimpleHTTPRequestHandler
+from dvc.utils.compat import HTTPServer, SimpleHTTPRequestHandler
 
 
 class ETagHandler(SimpleHTTPRequestHandler):