osc-lib/osc_lib/api/api.py
Doug Hellmann 5b80552b11 turn on warning-is-error for sphinx
Update a docstring that had an rst formatting error and then turn on the
warning-is-error flag to prevent future issues with documentation.

Change-Id: I29fc90a2eb1599c9da88489415a9e080d88cc707
Signed-off-by: Doug Hellmann <doug@doughellmann.com>
2017-06-22 14:38:17 -04:00

418 lines
12 KiB
Python

# 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.
#
"""Base API Library"""
import simplejson as json
import six
from keystoneauth1 import exceptions as ksa_exceptions
from keystoneauth1 import session as ksa_session
from osc_lib import exceptions
from osc_lib.i18n import _
class BaseAPI(object):
"""Base API wrapper for keystoneauth1.session.Session
Encapsulate the translation between keystoneauth1.session.Session
and requests.Session in a single layer:
* Restore some requests.session.Session compatibility;
keystoneauth1.session.Session.request() has the method and url
arguments swapped from the rest of the requests-using world.
* Provide basic endpoint handling when a Service Catalog is not
available.
"""
# Which service are we? Set in API-specific subclasses
SERVICE_TYPE = ""
# The common OpenStack microversion header
HEADER_NAME = "OpenStack-API-Version"
def __init__(
self,
session=None,
service_type=None,
endpoint=None,
**kwargs
):
"""Base object that contains some common API objects and methods
:param keystoneauth1.session.Session session:
The session to be used for making the HTTP API calls. If None,
a default keystoneauth1.session.Session will be created.
:param string service_type:
API name, i.e. ``identity`` or ``compute``
:param string endpoint:
An optional URL to be used as the base for API requests on
this API.
:param kwargs:
Keyword arguments passed to keystoneauth1.session.Session().
"""
super(BaseAPI, self).__init__()
# Create a keystoneauth1.session.Session if one is not supplied
if not session:
self.session = ksa_session.Session(**kwargs)
else:
self.session = session
self.service_type = service_type
self.endpoint = self._munge_endpoint(endpoint)
def _munge_endpoint(self, endpoint):
"""Hook to allow subclasses to massage the passed-in endpoint
Hook to massage passed-in endpoints from arbitrary sources,
including direct user input. By default just remove trailing
'/' as all of our path info strings start with '/' and not all
services can handle '//' in their URLs.
Some subclasses will override this to do additional work, most
likely with regard to API versions.
:param string endpoint: The service endpoint, generally direct
from the service catalog.
:return: The modified endpoint
"""
if isinstance(endpoint, six.string_types):
return endpoint.rstrip('/')
else:
return endpoint
def _request(self, method, url, session=None, **kwargs):
"""Perform call into session
All API calls are funneled through this method to provide a common
place to finalize the passed URL and other things.
:param string method:
The HTTP method name, i.e. ``GET``, ``PUT``, etc
:param string url:
The API-specific portion of the URL path, or a full URL if
``endpoint`` was not supplied at initialization.
:param keystoneauth1.session.Session session:
An initialized session to override the one created at
initialization.
:param kwargs:
Keyword arguments passed to requests.request().
:return: the requests.Response object
"""
# If session arg is supplied, use it this time, but don't save it
if not session:
session = self.session
# Do the auto-endpoint magic
if self.endpoint:
if url:
url = '/'.join([self.endpoint.rstrip('/'), url.lstrip('/')])
else:
# NOTE(dtroyer): This is left here after _munge_endpoint() is
# added because endpoint is public and there is
# no accounting for what may happen.
url = self.endpoint.rstrip('/')
else:
# Pass on the lack of URL unmolested to maintain the same error
# handling from keystoneauth: raise EndpointNotFound
pass
# Hack out empty headers 'cause KSA can't stomach it
if 'headers' in kwargs and kwargs['headers'] is None:
kwargs.pop('headers')
# Why is ksc session backwards???
return session.request(url, method, **kwargs)
# The basic action methods all take a Session and return dict/lists
def create(
self,
url,
session=None,
method=None,
**params
):
"""Create a new resource
:param string url:
The API-specific portion of the URL path
:param Session session:
HTTP client session
:param string method:
HTTP method (default POST)
"""
if not method:
method = 'POST'
ret = self._request(method, url, session=session, **params)
# Should this move into _requests()?
try:
return ret.json()
except json.JSONDecodeError:
return ret
def delete(
self,
url,
session=None,
**params
):
"""Delete a resource
:param string url:
The API-specific portion of the URL path
:param Session session:
HTTP client session
"""
return self._request('DELETE', url, **params)
def list(
self,
path,
session=None,
body=None,
detailed=False,
headers=None,
**params
):
"""Return a list of resources
GET ${ENDPOINT}/${PATH}?${PARAMS}
path is often the object's plural resource type
:param string path:
The API-specific portion of the URL path
:param Session session:
HTTP client session
:param body: data that will be encoded as JSON and passed in POST
request (GET will be sent by default)
:param bool detailed:
Adds '/details' to path for some APIs to return extended attributes
:param dict headers:
Headers dictionary to pass to requests
:returns:
JSON-decoded response, could be a list or a dict-wrapped-list
"""
if detailed:
path = '/'.join([path.rstrip('/'), 'details'])
if body:
ret = self._request(
'POST',
path,
# service=self.service_type,
json=body,
params=params,
headers=headers,
)
else:
ret = self._request(
'GET',
path,
# service=self.service_type,
params=params,
headers=headers,
)
try:
return ret.json()
except json.JSONDecodeError:
return ret
# Layered actions built on top of the basic action methods do not
# explicitly take a Session but one may still be passed in kwargs
def find_attr(
self,
path,
value=None,
attr=None,
resource=None,
):
"""Find a resource via attribute or ID
Most APIs return a list wrapped by a dict with the resource
name as key. Some APIs (Identity) return a dict when a query
string is present and there is one return value. Take steps to
unwrap these bodies and return a single dict without any resource
wrappers.
:param string path:
The API-specific portion of the URL path
:param string value:
value to search for
:param string attr:
attribute to use for resource search
:param string resource:
plural of the object resource name; defaults to path
For example:
n = find(netclient, 'network', 'networks', 'matrix')
"""
# Default attr is 'name'
if attr is None:
attr = 'name'
# Default resource is path - in many APIs they are the same
if resource is None:
resource = path
def getlist(kw):
"""Do list call, unwrap resource dict if present"""
ret = self.list(path, **kw)
if isinstance(ret, dict) and resource in ret:
ret = ret[resource]
return ret
# Search by attribute
kwargs = {attr: value}
data = getlist(kwargs)
if isinstance(data, dict):
return data
if len(data) == 1:
return data[0]
if len(data) > 1:
msg = _("Multiple %(resource)s exist with %(attr)s='%(value)s'")
raise exceptions.CommandError(
msg % {'resource': resource,
'attr': attr,
'value': value}
)
# Search by id
kwargs = {'id': value}
data = getlist(kwargs)
if len(data) == 1:
return data[0]
msg = _("No %(resource)s with a %(attr)s or ID of '%(value)s' found")
raise exceptions.CommandError(
msg % {'resource': resource,
'attr': attr,
'value': value}
)
def find_bulk(
self,
path,
headers=None,
**kwargs
):
"""Bulk load and filter locally
:param string path:
The API-specific portion of the URL path
:param kwargs:
A dict of AVPs to match - logical AND
:param dict headers:
Headers dictionary to pass to requests
:returns: list of resource dicts
"""
print("keys: %s" % kwargs.keys())
items = self.list(path)
if isinstance(items, dict):
# strip off the enclosing dict
key = list(items.keys())[0]
items = items[key]
ret = []
for o in items:
try:
if all(o[attr] == kwargs[attr] for attr in kwargs.keys()):
ret.append(o)
except KeyError:
continue
return ret
def find_one(
self,
path,
**kwargs
):
"""Find a resource by name or ID
:param string path:
The API-specific portion of the URL path
:returns:
resource dict
"""
bulk_list = self.find_bulk(path, **kwargs)
num_bulk = len(bulk_list)
if num_bulk == 0:
msg = _("none found")
raise exceptions.NotFound(msg)
elif num_bulk > 1:
msg = _("many found")
raise RuntimeError(msg)
return bulk_list[0]
def find(
self,
path,
value=None,
attr=None,
headers=None,
):
"""Find a single resource by name or ID
:param string path:
The API-specific portion of the URL path
:param string value:
search expression (required, really)
:param string attr:
name of attribute for secondary search
:param dict headers:
Headers dictionary to pass to requests
"""
try:
ret = self._request(
'GET', "/%s/%s" % (path, value),
headers=headers,
).json()
if isinstance(ret, dict):
# strip off the enclosing dict
key = list(ret.keys())[0]
ret = ret[key]
except (
ksa_exceptions.NotFound,
ksa_exceptions.BadRequest,
):
kwargs = {attr: value}
try:
ret = self.find_one(
path,
headers=headers,
**kwargs
)
except (
exceptions.NotFound,
ksa_exceptions.NotFound,
):
msg = _("%s not found") % value
raise exceptions.NotFound(msg)
return ret