PYTHON-674 python 2/3 single-source for the pymongo module

bson/json_util.py

@@ -152,7 +152,8 @@ def _json_convert(obj):
     """
     if hasattr(obj, 'iteritems') or hasattr(obj, 'items'):  # PY3 support
         return SON(((k, _json_convert(v)) for k, v in iteritems(obj)))
-    elif hasattr(obj, '__iter__') and not isinstance(obj, string_types):
+    elif hasattr(obj, '__iter__') and not isinstance(obj, (text_type,
+                                                           binary_type)):
         return list((_json_convert(v) for v in obj))
     try:
         return default(obj)

bson/py3compat.py

@@ -40,17 +40,26 @@ def bytes_from_hex(h):
         return bytes.fromhex(h)
 
     def iteritems(d):
-        return d.items()
+        return iter(d.items())
+
+    def itervalues(d):
+        return iter(d.values())
 
     def reraise(exctype, value, trace=None):
         raise exctype(str(value)).with_traceback(trace)
 
+    def _unicode(s):
+        return s
+
     binary_type = bytes
     text_type = str
     string_type = str
     integer_types = int
     next_item = "__next__"
 
+    # TODO: remove when gridfs module is made single-source
+    next_item = '__next__'
+    string_types = (bytes, text_type)
 else:
     try:
         from cStringIO import StringIO
@@ -73,15 +82,21 @@ def bytes_from_hex(h):
     def iteritems(d):
         return d.iteritems()
 
+    def itervalues(d):
+        return d.itervalues()
+
     # "raise x, y, z" raises SyntaxError in Python 3
     exec("""def reraise(exctype, value, trace=None):
     raise exctype, str(value), trace
 """)
 
+    _unicode = unicode
+
     binary_type = str
     string_type = basestring
     text_type = unicode
     integer_types = (int, long)
-    next_item = "next"
 
-string_types = (binary_type, text_type)
+    # TODO: remove when gridfs module is made single-source
+    next_item = 'next'
+    string_types = (bytes, text_type)

pymongo/__init__.py

@@ -14,7 +14,6 @@
 
 """Python driver for MongoDB."""
 
-
 ASCENDING = 1
 """Ascending sort order."""
 DESCENDING = -1
@@ -70,7 +69,7 @@
 version_tuple = (3, 0, ".dev0")
 
 def get_version_string():
-    if isinstance(version_tuple[-1], basestring):
+    if isinstance(version_tuple[-1], str):
         return '.'.join(map(str, version_tuple[:-1])) + version_tuple[-1]
     return '.'.join(map(str, version_tuple))
 

pymongo/auth.py

@@ -14,24 +14,20 @@
 
 """Authentication helpers."""
 
+from __future__ import unicode_literals
+
 import hmac
-try:
-    import hashlib
-    _MD5 = hashlib.md5
-    _DMOD = _MD5
-except ImportError:  # for Python < 2.5
-    import md5
-    _MD5 = md5.new
-    _DMOD = md5
 
 HAVE_KERBEROS = True
 try:
     import kerberos
 except ImportError:
     HAVE_KERBEROS = False
 
+from hashlib import md5
+
 from bson.binary import Binary
-from bson.py3compat import b
+from bson.py3compat import b, string_type, _unicode
 from bson.son import SON
 from pymongo.errors import ConfigurationError, OperationFailure
 
@@ -55,29 +51,29 @@ def _build_credentials_tuple(mech, source, user, passwd, extra):
 def _password_digest(username, password):
     """Get a password digest to use for authentication.
     """
-    if not isinstance(password, basestring):
-        raise TypeError("password must be an instance "
-                        "of %s" % (basestring.__name__,))
+    if not isinstance(password, string_type):
+        raise TypeError("password must be an "
+                        "instance of %s" % (string_type.__name__,))
     if len(password) == 0:
         raise ValueError("password can't be empty")
-    if not isinstance(username, basestring):
-        raise TypeError("username must be an instance "
-                        "of %s" % (basestring.__name__,))
+    if not isinstance(username, string_type):
+        raise TypeError("password must be an "
+                        "instance of  %s" % (string_type.__name__,))
 
-    md5hash = _MD5()
+    md5hash = md5()
     data = "%s:mongo:%s" % (username, password)
     md5hash.update(data.encode('utf-8'))
-    return unicode(md5hash.hexdigest())
+    return _unicode(md5hash.hexdigest())
 
 
 def _auth_key(nonce, username, password):
     """Get an auth key to use for authentication.
     """
     digest = _password_digest(username, password)
-    md5hash = _MD5()
-    data = "%s%s%s" % (nonce, unicode(username), digest)
+    md5hash = md5()
+    data = "%s%s%s" % (nonce, _unicode(username), digest)
     md5hash.update(data.encode('utf-8'))
-    return unicode(md5hash.hexdigest())
+    return _unicode(md5hash.hexdigest())
 
 
 def _authenticate_gssapi(credentials, sock_info, cmd_func):
@@ -114,7 +110,7 @@ def _authenticate_gssapi(credentials, sock_info, cmd_func):
             response, _ = cmd_func(sock_info, '$external', cmd)
 
             # Limit how many times we loop to catch protocol / library issues
-            for _ in xrange(10):
+            for _ in range(10):
                 result = kerberos.authGSSClientStep(ctx,
                                                     str(response['payload']))
                 if result == -1:
@@ -156,7 +152,7 @@ def _authenticate_gssapi(credentials, sock_info, cmd_func):
         finally:
             kerberos.authGSSClientClean(ctx)
 
-    except kerberos.KrbError, exc:
+    except kerberos.KrbError as exc:
         raise OperationFailure(str(exc))
 
 
@@ -181,14 +177,14 @@ def _authenticate_cram_md5(credentials, sock_info, cmd_func):
     passwd = _password_digest(username, password)
     cmd = SON([('saslStart', 1),
                ('mechanism', 'CRAM-MD5'),
-               ('payload', Binary(b(''))),
+               ('payload', Binary(b'')),
                ('autoAuthorize', 1)])
     response, _ = cmd_func(sock_info, source, cmd)
     # MD5 as implicit default digest for digestmod is deprecated
     # in python 3.4
-    mac = hmac.HMAC(key=passwd.encode('utf-8'), digestmod=_DMOD)
+    mac = hmac.HMAC(key=passwd.encode('utf-8'), digestmod=md5)
     mac.update(response['payload'])
-    challenge = username.encode('utf-8') + b(' ') + b(mac.hexdigest())
+    challenge = username.encode('utf-8') + b' ' + b(mac.hexdigest())
     cmd = SON([('saslContinue', 1),
                ('conversationId', response['conversationId']),
                ('payload', Binary(challenge))])

pymongo/bulk.py

@@ -17,6 +17,8 @@
 .. versionadded:: 2.7
 """
 
+from __future__ import unicode_literals
+
 from bson.objectid import ObjectId
 from bson.son import SON
 from pymongo.errors import (BulkWriteError,
@@ -72,10 +74,10 @@ def _make_error(index, code, errmsg, operation):
     """Create and return an error document.
     """
     return {
-        u"index": index,
-        u"code": code,
-        u"errmsg": errmsg,
-        u"op": operation
+        "index": index,
+        "code": code,
+        "errmsg": errmsg,
+        "op": operation
     }
 
 
@@ -113,7 +115,7 @@ def _merge_legacy(run, full_result, result, index):
         full_result['nInserted'] += 1
     elif run.op_type == _UPDATE:
         if "upserted" in result:
-            doc = {u"index": run.index(index), u"_id": result["upserted"]}
+            doc = {"index": run.index(index), "_id": result["upserted"]}
             full_result["upserted"].append(doc)
             full_result['nUpserted'] += affected
         else:
@@ -146,7 +148,7 @@ def _merge_command(run, full_result, results):
                 else:
                     n_upserted = 1
                     index = run.index(offset)
-                    doc = {u"index": index, u"_id": upserted}
+                    doc = {"index": index, "_id": upserted}
                     full_result["upserted"].append(doc)
                 full_result["nUpserted"] += n_upserted
                 full_result["nMatched"] += (affected - n_upserted)
@@ -168,7 +170,7 @@ def _merge_command(run, full_result, results):
                 idx = doc["index"] + offset
                 doc["index"] = run.index(idx)
                 # Add the failed operation to the error document.
-                doc[u"op"] = run.ops[idx]
+                doc["op"] = run.ops[idx]
             full_result["writeErrors"].extend(write_errors)
 
         wc_error = result.get("writeConcernError")
@@ -207,7 +209,7 @@ def add_update(self, selector, update, multi=False, upsert=False):
         # Update can not be {}
         if not update:
             raise ValueError('update only works with $ operators')
-        first = iter(update).next()
+        first = next(iter(update))
         if not first.startswith('$'):
             raise ValueError('update only works with $ operators')
         cmd = SON([('q', selector), ('u', update),
@@ -221,7 +223,7 @@ def add_replace(self, selector, replacement, upsert=False):
             raise TypeError('replacement must be an instance of dict')
         # Replacement can be {}
         if replacement:
-            first = iter(replacement).next()
+            first = next(iter(replacement))
             if first.startswith('$'):
                 raise ValueError('replacement can not include $ operators')
         cmd = SON([('q', selector), ('u', replacement),
@@ -378,15 +380,15 @@ def execute_legacy(self, generator, write_concern):
                                              multi=(not operation['limit']),
                                              **write_concern)
                     _merge_legacy(run, full_result, result, idx)
-                except DocumentTooLarge, exc:
+                except DocumentTooLarge as exc:
                     # MongoDB 2.6 uses error code 2 for "too large".
                     error = _make_error(
                         run.index(idx), _BAD_VALUE, str(exc), operation)
                     full_result['writeErrors'].append(error)
                     if self.ordered:
                         stop = True
                         break
-                except OperationFailure, exc:
+                except OperationFailure as exc:
                     if not exc.details:
                         # Some error not related to the write operation
                         # (e.g. kerberos failure). Re-raise immediately.