Flavio Percoco ddd41adab1 Don't stream by default
Current `_Iterator` streams data from the server by default. This
behavior, although useful, shouldn't be considered a sane default. A
good example is when using `limit` on listing operations. If the default
is to stream, the limit will be ignored because the server client will
get everything from the server until there's anything left to consume.
This behavior can also be harmful if not used carefully.

This patch adds a `stream` method to the iterator and makes using this
behavior an explicit operation. The user has to opt-in for data
streaming.

Change-Id: Ib1af24960dff97cb956990d4caf88705f2f7a0d5
2014-04-22 13:20:58 +02:00

110 lines
3.6 KiB
Python

# Copyright (c) 2014 Rackspace
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
# implied.
# See the License for the specific language governing permissions and
# limitations under the License.
class _Iterator(object):
"""Base Iterator
This iterator is not meant to be used outside
the scope of this package. The iterator gets
a dictionary as returned by a listing endpoint.
Subclasses of this base class determine the key
to iterate over, as well as the means of creating
the objects contained within.
If there are no objects left to return, the iterator
will try to load more by following the `next` rel link
type.
The iterator raises a StopIteration exception if the server
doesn't return more objects after a `next-page` call.
:param client: The client instance used by the queue
:type client: `v1.Client`
:param listing_response: Response returned by the listing call
:type listing_response: Dict
"""
def __init__(self, client, listing_response, iter_key, create_function):
self._client = client
self._iter_key = iter_key
self._create_function = create_function
self._links = []
self._stream = False
self._listing_response = listing_response
# NOTE(flaper87): Simple hack to
# re-use the iterator for get_many_messages
# and message listing.
if isinstance(listing_response, dict):
self._links = listing_response['links']
self._listing_response = listing_response[self._iter_key]
def __iter__(self):
return self
def get_iterables(self, iterables):
self._links = iterables['links']
self._listing_response = iterables[self._iter_key]
def stream(self, enabled=True):
"""Make this `_Iterator` a stream iterator.
Since `_Iterator`'s default is to *not* stream,
this method's default value is to *stream* data
from the server. That is, unless explicitly specified
this method will enable make this iterator a stream
iterator.
:param enabled: Whether streaming should be
enabled or not.
:type enabled: bool
"""
self._stream = enabled
return self
def _next_page(self):
for link in self._links:
if link['rel'] == 'next':
# NOTE(flaper87): We already have the
# ref for the next set of messages, lets
# just follow it.
iterables = self._client.follow(link['href'])
# NOTE(flaper87): Since we're using
# `.follow`, the empty result will
# be None. Consider making the API
# return an empty dict for consistency.
if iterables:
self.get_iterables(iterables)
return
raise StopIteration
def __next__(self):
try:
args = self._listing_response.pop(0)
except IndexError:
if not self._stream:
raise StopIteration
self._next_page()
return self.next()
return self._create_function(args)
# NOTE(flaper87): Py2K support
next = __next__