ported freezing as csv to python 3.4

dataset/freeze/format/common.py

@@ -3,10 +3,7 @@ import re
 import sys
 import locale
 
-try:
-    str = unicode
-except NameError:
-    pass
+from six import binary_type, text_type
 
 from dataset.util import FreezeException
 from slugify import slugify
@@ -16,7 +13,7 @@ TMPL_KEY = re.compile("{{([^}]*)}}")
 
 OPERATIONS = {
         'identity': lambda x: x,
-        'lower': lambda x: str(x).lower(),
+        'lower': lambda x: text_type(x).lower(),
         'slug': slugify
         }
 
@@ -24,6 +21,7 @@ OPERATIONS = {
 class Serializer(object):
 
     def __init__(self, export, query):
+        self._encoding = locale.getpreferredencoding()
         self.export = export
         self.query = query
         self._paths = []
@@ -35,10 +33,14 @@ class Serializer(object):
 
     def _get_basepath(self):
         prefix = self.export.get('prefix', '')
+        if isinstance(prefix, binary_type):
+            prefix = text_type(prefix, encoding=self._encoding)
         prefix = os.path.abspath(prefix)
         prefix = os.path.realpath(prefix)
         self._prefix = prefix
         filename = self.export.get('filename')
+        if isinstance(filename, binary_type):
+            filename = text_type(filename, encoding=self._encoding)
         if filename is None:
             raise FreezeException("No 'filename' is specified")
         self._basepath = os.path.join(prefix, filename)
@@ -50,8 +52,7 @@ class Serializer(object):
                 op, key = key.split(':', 1)
             return str(OPERATIONS.get(op)(data.get(key, '')))
         path = TMPL_KEY.sub(repl, self._basepath)
-        enc = locale.getpreferredencoding()
-        return os.path.realpath(path.encode(enc, 'replace'))
+        return os.path.realpath(path)
 
     def file_name(self, row):
         # signal that there is a fileobj available:

dataset/freeze/format/fcsv.py

@@ -1,13 +1,23 @@
+from __future__ import unicode_literals
 import csv
-from datetime import datetime
+from datetime import datetime, date
+
+from six import PY3, binary_type, text_type
+from sqlalchemy import DateTime
 
 from dataset.freeze.format.common import Serializer
 
 
 def value_to_str(value):
-    if isinstance(value, datetime):
-        return value.isoformat()
-    if hasattr(value, 'encode'):
+    if isinstance(value, (datetime, date)):
+        #
+        # FIXME: this check does not work for values returned from a db query!
+        # As a workaround, we make sure, the isoformat call returns the regular
+        # str representation.
+        #
+        sep = ' ' if PY3 else str(' ')
+        return text_type(value.isoformat(sep=sep))
+    if not PY3 and hasattr(value, 'encode'):
         return value.encode('utf-8')
     if value is None:
         return ''
@@ -25,12 +35,18 @@ class CSVSerializer(Serializer):
 
             # handle fileobj that has been passed in:
             if path is not None:
-                fh = open(path, 'wb')
+                if PY3:
+                    fh = open(path, 'wt', encoding='utf8', newline='')
+                else:
+                    fh = open(path, 'wb')
             else:
                 fh = self.fileobj
 
             writer = csv.writer(fh)
-            writer.writerow([k.encode('utf-8') for k in keys])
+            if PY3:
+                writer.writerow(keys)
+            else:
+                writer.writerow([k.encode('utf-8') for k in keys])
             self.handles[path] = (writer, fh)
         writer, fh = self.handles[path]
         values = [value_to_str(result.get(k)) for k in keys]