asterisk/rest-api-templates/swagger_model.py

756 lines
25 KiB
Python
Raw Normal View History

# Asterisk -- An open source telephony toolkit.
#
# Copyright (C) 2013, Digium, Inc.
#
# David M. Lee, II <dlee@digium.com>
#
# See http://www.asterisk.org for more information about
# the Asterisk project. Please do not directly contact
# any of the maintainers of this project for assistance;
# the project provides a web site, mailing lists and IRC
# channels for your use.
#
# This program is free software, distributed under the terms of
# the GNU General Public License Version 2. See the LICENSE file
# at the top of the source tree.
#
"""Swagger data model objects.
These objects should map directly to the Swagger api-docs, without a lot of
additional fields. In the process of translation, it should also validate the
model for consistency against the Swagger spec (i.e., fail if fields are
missing, or have incorrect values).
See https://github.com/wordnik/swagger-core/wiki/API-Declaration for the spec.
"""
import json
import os.path
import pprint
Update events to use Swagger 1.3 subtyping, and related aftermath This patch started with the simple idea of changing the /events data model to be more sane. The original model would send out events like: { "stasis_start": { "args": [], "channel": { ... } } } The event discriminator was the field name instead of being a value in the object, due to limitations in how Swagger 1.1 could model objects. While technically sufficient in communicating event information, it was really difficult to deal with in terms of client side JSON handling. This patch takes advantage of a proposed extension[1] to Swagger which allows type variance through the use of a discriminator field. This had a domino effect that made this a surprisingly large patch. [1]: https://groups.google.com/d/msg/wordnik-api/EC3rGajE0os/ey_5dBI_jWcJ In changing the models, I also had to change the swagger_model.py processor so it can handle the type discriminator and subtyping. I took that a big step forward, and using that information to generate an ari_model module, which can validate a JSON object against the Swagger model. The REST and WebSocket generators were changed to take advantage of the validators. If compiled with AST_DEVMODE enabled, JSON objects that don't match their corresponding models will not be sent out. For REST API calls, a 500 Internal Server response is sent. For WebSockets, the invalid JSON message is replaced with an error message. Since this took over about half of the job of the existing JSON generators, and the .to_json virtual function on messages took over the other half, I reluctantly removed the generators. The validators turned up all sorts of errors and inconsistencies in our data models, and the code. These were cleaned up, with checks in the code generator avoid some of the consistency problems in the future. * The model for a channel snapshot was trimmed down to match the information sent via AMI. Many of the field being sent were not useful in the general case. * The model for a bridge snapshot was updated to be more consistent with the other ARI models. Another impact of introducing subtyping was that the swagger-codegen documentation generator was insufficient (at least until it catches up with Swagger 1.2). I wanted it to be easier to generate docs for the API anyways, so I ported the wiki pages to use the Asterisk Swagger generator. In the process, I was able to clean up many of the model links, which would occasionally give inconsistent results on the wiki. I also added error responses to the wiki docs, making the wiki documentation more complete. Finally, since Stasis-HTTP will now be named Asterisk REST Interface (ARI), any new functions and files I created carry the ari_ prefix. I changed a few stasis_http references to ari where it was non-intrusive and made sense. (closes issue ASTERISK-21885) Review: https://reviewboard.asterisk.org/r/2639/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@393529 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-07-03 16:32:41 +00:00
import re
import sys
import traceback
# We don't fully support Swagger 1.2, but we need it for subtyping
SWAGGER_VERSIONS = ["1.1", "1.2"]
Update events to use Swagger 1.3 subtyping, and related aftermath This patch started with the simple idea of changing the /events data model to be more sane. The original model would send out events like: { "stasis_start": { "args": [], "channel": { ... } } } The event discriminator was the field name instead of being a value in the object, due to limitations in how Swagger 1.1 could model objects. While technically sufficient in communicating event information, it was really difficult to deal with in terms of client side JSON handling. This patch takes advantage of a proposed extension[1] to Swagger which allows type variance through the use of a discriminator field. This had a domino effect that made this a surprisingly large patch. [1]: https://groups.google.com/d/msg/wordnik-api/EC3rGajE0os/ey_5dBI_jWcJ In changing the models, I also had to change the swagger_model.py processor so it can handle the type discriminator and subtyping. I took that a big step forward, and using that information to generate an ari_model module, which can validate a JSON object against the Swagger model. The REST and WebSocket generators were changed to take advantage of the validators. If compiled with AST_DEVMODE enabled, JSON objects that don't match their corresponding models will not be sent out. For REST API calls, a 500 Internal Server response is sent. For WebSockets, the invalid JSON message is replaced with an error message. Since this took over about half of the job of the existing JSON generators, and the .to_json virtual function on messages took over the other half, I reluctantly removed the generators. The validators turned up all sorts of errors and inconsistencies in our data models, and the code. These were cleaned up, with checks in the code generator avoid some of the consistency problems in the future. * The model for a channel snapshot was trimmed down to match the information sent via AMI. Many of the field being sent were not useful in the general case. * The model for a bridge snapshot was updated to be more consistent with the other ARI models. Another impact of introducing subtyping was that the swagger-codegen documentation generator was insufficient (at least until it catches up with Swagger 1.2). I wanted it to be easier to generate docs for the API anyways, so I ported the wiki pages to use the Asterisk Swagger generator. In the process, I was able to clean up many of the model links, which would occasionally give inconsistent results on the wiki. I also added error responses to the wiki docs, making the wiki documentation more complete. Finally, since Stasis-HTTP will now be named Asterisk REST Interface (ARI), any new functions and files I created carry the ari_ prefix. I changed a few stasis_http references to ari where it was non-intrusive and made sense. (closes issue ASTERISK-21885) Review: https://reviewboard.asterisk.org/r/2639/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@393529 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-07-03 16:32:41 +00:00
SWAGGER_PRIMITIVES = [
'void',
'string',
'boolean',
'number',
'int',
'long',
'double',
'float',
'Date',
]
Update events to use Swagger 1.3 subtyping, and related aftermath This patch started with the simple idea of changing the /events data model to be more sane. The original model would send out events like: { "stasis_start": { "args": [], "channel": { ... } } } The event discriminator was the field name instead of being a value in the object, due to limitations in how Swagger 1.1 could model objects. While technically sufficient in communicating event information, it was really difficult to deal with in terms of client side JSON handling. This patch takes advantage of a proposed extension[1] to Swagger which allows type variance through the use of a discriminator field. This had a domino effect that made this a surprisingly large patch. [1]: https://groups.google.com/d/msg/wordnik-api/EC3rGajE0os/ey_5dBI_jWcJ In changing the models, I also had to change the swagger_model.py processor so it can handle the type discriminator and subtyping. I took that a big step forward, and using that information to generate an ari_model module, which can validate a JSON object against the Swagger model. The REST and WebSocket generators were changed to take advantage of the validators. If compiled with AST_DEVMODE enabled, JSON objects that don't match their corresponding models will not be sent out. For REST API calls, a 500 Internal Server response is sent. For WebSockets, the invalid JSON message is replaced with an error message. Since this took over about half of the job of the existing JSON generators, and the .to_json virtual function on messages took over the other half, I reluctantly removed the generators. The validators turned up all sorts of errors and inconsistencies in our data models, and the code. These were cleaned up, with checks in the code generator avoid some of the consistency problems in the future. * The model for a channel snapshot was trimmed down to match the information sent via AMI. Many of the field being sent were not useful in the general case. * The model for a bridge snapshot was updated to be more consistent with the other ARI models. Another impact of introducing subtyping was that the swagger-codegen documentation generator was insufficient (at least until it catches up with Swagger 1.2). I wanted it to be easier to generate docs for the API anyways, so I ported the wiki pages to use the Asterisk Swagger generator. In the process, I was able to clean up many of the model links, which would occasionally give inconsistent results on the wiki. I also added error responses to the wiki docs, making the wiki documentation more complete. Finally, since Stasis-HTTP will now be named Asterisk REST Interface (ARI), any new functions and files I created carry the ari_ prefix. I changed a few stasis_http references to ari where it was non-intrusive and made sense. (closes issue ASTERISK-21885) Review: https://reviewboard.asterisk.org/r/2639/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@393529 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-07-03 16:32:41 +00:00
class Stringify(object):
"""Simple mix-in to make the repr of the model classes more meaningful.
"""
def __repr__(self):
return "%s(%s)" % (self.__class__, pprint.saferepr(self.__dict__))
def compare_versions(lhs, rhs):
'''Performs a lexicographical comparison between two version numbers.
This properly handles simple major.minor.whatever.sure.why.not version
numbers, but fails miserably if there's any letters in there.
For reference:
1.0 == 1.0
1.0 < 1.0.1
1.2 < 1.10
@param lhs Left hand side of the comparison
@param rhs Right hand side of the comparison
@return < 0 if lhs < rhs
@return == 0 if lhs == rhs
@return > 0 if lhs > rhs
'''
lhs = [int(v) for v in lhs.split('.')]
rhs = [int(v) for v in rhs.split('.')]
return cmp(lhs, rhs)
class ParsingContext(object):
"""Context information for parsing.
This object is immutable. To change contexts (like adding an item to the
stack), use the next() and next_stack() functions to build a new one.
"""
def __init__(self, swagger_version, stack):
self.__swagger_version = swagger_version
self.__stack = stack
def __repr__(self):
return "ParsingContext(swagger_version=%s, stack=%s)" % (
self.swagger_version, self.stack)
def get_swagger_version(self):
return self.__swagger_version
def get_stack(self):
return self.__stack
swagger_version = property(get_swagger_version)
stack = property(get_stack)
def version_less_than(self, ver):
return compare_versions(self.swagger_version, ver) < 0
def next_stack(self, json, id_field):
"""Returns a new item pushed to the stack.
@param json: Current JSON object.
@param id_field: Field identifying this object.
@return New context with additional item in the stack.
"""
if not id_field in json:
raise SwaggerError("Missing id_field: %s" % id_field, self)
new_stack = self.stack + ['%s=%s' % (id_field, str(json[id_field]))]
return ParsingContext(self.swagger_version, new_stack)
def next(self, version=None, stack=None):
if version is None:
version = self.version
if stack is None:
stack = self.stack
return ParsingContext(version, stack)
class SwaggerError(Exception):
"""Raised when an error is encountered mapping the JSON objects into the
model.
"""
def __init__(self, msg, context, cause=None):
"""Ctor.
@param msg: String message for the error.
Update events to use Swagger 1.3 subtyping, and related aftermath This patch started with the simple idea of changing the /events data model to be more sane. The original model would send out events like: { "stasis_start": { "args": [], "channel": { ... } } } The event discriminator was the field name instead of being a value in the object, due to limitations in how Swagger 1.1 could model objects. While technically sufficient in communicating event information, it was really difficult to deal with in terms of client side JSON handling. This patch takes advantage of a proposed extension[1] to Swagger which allows type variance through the use of a discriminator field. This had a domino effect that made this a surprisingly large patch. [1]: https://groups.google.com/d/msg/wordnik-api/EC3rGajE0os/ey_5dBI_jWcJ In changing the models, I also had to change the swagger_model.py processor so it can handle the type discriminator and subtyping. I took that a big step forward, and using that information to generate an ari_model module, which can validate a JSON object against the Swagger model. The REST and WebSocket generators were changed to take advantage of the validators. If compiled with AST_DEVMODE enabled, JSON objects that don't match their corresponding models will not be sent out. For REST API calls, a 500 Internal Server response is sent. For WebSockets, the invalid JSON message is replaced with an error message. Since this took over about half of the job of the existing JSON generators, and the .to_json virtual function on messages took over the other half, I reluctantly removed the generators. The validators turned up all sorts of errors and inconsistencies in our data models, and the code. These were cleaned up, with checks in the code generator avoid some of the consistency problems in the future. * The model for a channel snapshot was trimmed down to match the information sent via AMI. Many of the field being sent were not useful in the general case. * The model for a bridge snapshot was updated to be more consistent with the other ARI models. Another impact of introducing subtyping was that the swagger-codegen documentation generator was insufficient (at least until it catches up with Swagger 1.2). I wanted it to be easier to generate docs for the API anyways, so I ported the wiki pages to use the Asterisk Swagger generator. In the process, I was able to clean up many of the model links, which would occasionally give inconsistent results on the wiki. I also added error responses to the wiki docs, making the wiki documentation more complete. Finally, since Stasis-HTTP will now be named Asterisk REST Interface (ARI), any new functions and files I created carry the ari_ prefix. I changed a few stasis_http references to ari where it was non-intrusive and made sense. (closes issue ASTERISK-21885) Review: https://reviewboard.asterisk.org/r/2639/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@393529 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-07-03 16:32:41 +00:00
@param context: ParsingContext object
@param cause: Optional exception that caused this one.
"""
super(Exception, self).__init__(msg, context, cause)
class SwaggerPostProcessor(object):
"""Post processing interface for model objects. This processor can add
fields to model objects for additional information to use in the
templates.
"""
Update events to use Swagger 1.3 subtyping, and related aftermath This patch started with the simple idea of changing the /events data model to be more sane. The original model would send out events like: { "stasis_start": { "args": [], "channel": { ... } } } The event discriminator was the field name instead of being a value in the object, due to limitations in how Swagger 1.1 could model objects. While technically sufficient in communicating event information, it was really difficult to deal with in terms of client side JSON handling. This patch takes advantage of a proposed extension[1] to Swagger which allows type variance through the use of a discriminator field. This had a domino effect that made this a surprisingly large patch. [1]: https://groups.google.com/d/msg/wordnik-api/EC3rGajE0os/ey_5dBI_jWcJ In changing the models, I also had to change the swagger_model.py processor so it can handle the type discriminator and subtyping. I took that a big step forward, and using that information to generate an ari_model module, which can validate a JSON object against the Swagger model. The REST and WebSocket generators were changed to take advantage of the validators. If compiled with AST_DEVMODE enabled, JSON objects that don't match their corresponding models will not be sent out. For REST API calls, a 500 Internal Server response is sent. For WebSockets, the invalid JSON message is replaced with an error message. Since this took over about half of the job of the existing JSON generators, and the .to_json virtual function on messages took over the other half, I reluctantly removed the generators. The validators turned up all sorts of errors and inconsistencies in our data models, and the code. These were cleaned up, with checks in the code generator avoid some of the consistency problems in the future. * The model for a channel snapshot was trimmed down to match the information sent via AMI. Many of the field being sent were not useful in the general case. * The model for a bridge snapshot was updated to be more consistent with the other ARI models. Another impact of introducing subtyping was that the swagger-codegen documentation generator was insufficient (at least until it catches up with Swagger 1.2). I wanted it to be easier to generate docs for the API anyways, so I ported the wiki pages to use the Asterisk Swagger generator. In the process, I was able to clean up many of the model links, which would occasionally give inconsistent results on the wiki. I also added error responses to the wiki docs, making the wiki documentation more complete. Finally, since Stasis-HTTP will now be named Asterisk REST Interface (ARI), any new functions and files I created carry the ari_ prefix. I changed a few stasis_http references to ari where it was non-intrusive and made sense. (closes issue ASTERISK-21885) Review: https://reviewboard.asterisk.org/r/2639/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@393529 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-07-03 16:32:41 +00:00
def process_resource_api(self, resource_api, context):
"""Post process a ResourceApi object.
@param resource_api: ResourceApi object.
@param context: Current context in the API.
"""
pass
Update events to use Swagger 1.3 subtyping, and related aftermath This patch started with the simple idea of changing the /events data model to be more sane. The original model would send out events like: { "stasis_start": { "args": [], "channel": { ... } } } The event discriminator was the field name instead of being a value in the object, due to limitations in how Swagger 1.1 could model objects. While technically sufficient in communicating event information, it was really difficult to deal with in terms of client side JSON handling. This patch takes advantage of a proposed extension[1] to Swagger which allows type variance through the use of a discriminator field. This had a domino effect that made this a surprisingly large patch. [1]: https://groups.google.com/d/msg/wordnik-api/EC3rGajE0os/ey_5dBI_jWcJ In changing the models, I also had to change the swagger_model.py processor so it can handle the type discriminator and subtyping. I took that a big step forward, and using that information to generate an ari_model module, which can validate a JSON object against the Swagger model. The REST and WebSocket generators were changed to take advantage of the validators. If compiled with AST_DEVMODE enabled, JSON objects that don't match their corresponding models will not be sent out. For REST API calls, a 500 Internal Server response is sent. For WebSockets, the invalid JSON message is replaced with an error message. Since this took over about half of the job of the existing JSON generators, and the .to_json virtual function on messages took over the other half, I reluctantly removed the generators. The validators turned up all sorts of errors and inconsistencies in our data models, and the code. These were cleaned up, with checks in the code generator avoid some of the consistency problems in the future. * The model for a channel snapshot was trimmed down to match the information sent via AMI. Many of the field being sent were not useful in the general case. * The model for a bridge snapshot was updated to be more consistent with the other ARI models. Another impact of introducing subtyping was that the swagger-codegen documentation generator was insufficient (at least until it catches up with Swagger 1.2). I wanted it to be easier to generate docs for the API anyways, so I ported the wiki pages to use the Asterisk Swagger generator. In the process, I was able to clean up many of the model links, which would occasionally give inconsistent results on the wiki. I also added error responses to the wiki docs, making the wiki documentation more complete. Finally, since Stasis-HTTP will now be named Asterisk REST Interface (ARI), any new functions and files I created carry the ari_ prefix. I changed a few stasis_http references to ari where it was non-intrusive and made sense. (closes issue ASTERISK-21885) Review: https://reviewboard.asterisk.org/r/2639/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@393529 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-07-03 16:32:41 +00:00
def process_api(self, api, context):
"""Post process an Api object.
@param api: Api object.
@param context: Current context in the API.
"""
pass
def process_operation(self, operation, context):
"""Post process a Operation object.
@param operation: Operation object.
@param context: Current context in the API.
"""
pass
def process_parameter(self, parameter, context):
"""Post process a Parameter object.
@param parameter: Parameter object.
@param context: Current context in the API.
"""
pass
Update events to use Swagger 1.3 subtyping, and related aftermath This patch started with the simple idea of changing the /events data model to be more sane. The original model would send out events like: { "stasis_start": { "args": [], "channel": { ... } } } The event discriminator was the field name instead of being a value in the object, due to limitations in how Swagger 1.1 could model objects. While technically sufficient in communicating event information, it was really difficult to deal with in terms of client side JSON handling. This patch takes advantage of a proposed extension[1] to Swagger which allows type variance through the use of a discriminator field. This had a domino effect that made this a surprisingly large patch. [1]: https://groups.google.com/d/msg/wordnik-api/EC3rGajE0os/ey_5dBI_jWcJ In changing the models, I also had to change the swagger_model.py processor so it can handle the type discriminator and subtyping. I took that a big step forward, and using that information to generate an ari_model module, which can validate a JSON object against the Swagger model. The REST and WebSocket generators were changed to take advantage of the validators. If compiled with AST_DEVMODE enabled, JSON objects that don't match their corresponding models will not be sent out. For REST API calls, a 500 Internal Server response is sent. For WebSockets, the invalid JSON message is replaced with an error message. Since this took over about half of the job of the existing JSON generators, and the .to_json virtual function on messages took over the other half, I reluctantly removed the generators. The validators turned up all sorts of errors and inconsistencies in our data models, and the code. These were cleaned up, with checks in the code generator avoid some of the consistency problems in the future. * The model for a channel snapshot was trimmed down to match the information sent via AMI. Many of the field being sent were not useful in the general case. * The model for a bridge snapshot was updated to be more consistent with the other ARI models. Another impact of introducing subtyping was that the swagger-codegen documentation generator was insufficient (at least until it catches up with Swagger 1.2). I wanted it to be easier to generate docs for the API anyways, so I ported the wiki pages to use the Asterisk Swagger generator. In the process, I was able to clean up many of the model links, which would occasionally give inconsistent results on the wiki. I also added error responses to the wiki docs, making the wiki documentation more complete. Finally, since Stasis-HTTP will now be named Asterisk REST Interface (ARI), any new functions and files I created carry the ari_ prefix. I changed a few stasis_http references to ari where it was non-intrusive and made sense. (closes issue ASTERISK-21885) Review: https://reviewboard.asterisk.org/r/2639/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@393529 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-07-03 16:32:41 +00:00
def process_model(self, model, context):
"""Post process a Model object.
Update events to use Swagger 1.3 subtyping, and related aftermath This patch started with the simple idea of changing the /events data model to be more sane. The original model would send out events like: { "stasis_start": { "args": [], "channel": { ... } } } The event discriminator was the field name instead of being a value in the object, due to limitations in how Swagger 1.1 could model objects. While technically sufficient in communicating event information, it was really difficult to deal with in terms of client side JSON handling. This patch takes advantage of a proposed extension[1] to Swagger which allows type variance through the use of a discriminator field. This had a domino effect that made this a surprisingly large patch. [1]: https://groups.google.com/d/msg/wordnik-api/EC3rGajE0os/ey_5dBI_jWcJ In changing the models, I also had to change the swagger_model.py processor so it can handle the type discriminator and subtyping. I took that a big step forward, and using that information to generate an ari_model module, which can validate a JSON object against the Swagger model. The REST and WebSocket generators were changed to take advantage of the validators. If compiled with AST_DEVMODE enabled, JSON objects that don't match their corresponding models will not be sent out. For REST API calls, a 500 Internal Server response is sent. For WebSockets, the invalid JSON message is replaced with an error message. Since this took over about half of the job of the existing JSON generators, and the .to_json virtual function on messages took over the other half, I reluctantly removed the generators. The validators turned up all sorts of errors and inconsistencies in our data models, and the code. These were cleaned up, with checks in the code generator avoid some of the consistency problems in the future. * The model for a channel snapshot was trimmed down to match the information sent via AMI. Many of the field being sent were not useful in the general case. * The model for a bridge snapshot was updated to be more consistent with the other ARI models. Another impact of introducing subtyping was that the swagger-codegen documentation generator was insufficient (at least until it catches up with Swagger 1.2). I wanted it to be easier to generate docs for the API anyways, so I ported the wiki pages to use the Asterisk Swagger generator. In the process, I was able to clean up many of the model links, which would occasionally give inconsistent results on the wiki. I also added error responses to the wiki docs, making the wiki documentation more complete. Finally, since Stasis-HTTP will now be named Asterisk REST Interface (ARI), any new functions and files I created carry the ari_ prefix. I changed a few stasis_http references to ari where it was non-intrusive and made sense. (closes issue ASTERISK-21885) Review: https://reviewboard.asterisk.org/r/2639/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@393529 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-07-03 16:32:41 +00:00
@param model: Model object.
@param context: Current context in the API.
"""
pass
def process_property(self, property, context):
"""Post process a Property object.
@param property: Property object.
@param context: Current context in the API.
"""
pass
def process_type(self, swagger_type, context):
"""Post process a SwaggerType object.
@param swagger_type: ResourceListing object.
@param context: Current context in the API.
"""
pass
def process_resource_listing(self, resource_listing, context):
"""Post process the overall ResourceListing object.
@param resource_listing: ResourceListing object.
@param context: Current context in the API.
"""
pass
class AllowableRange(Stringify):
"""Model of a allowableValues of type RANGE
See https://github.com/wordnik/swagger-core/wiki/datatypes#complex-types
"""
def __init__(self, min_value, max_value):
self.min_value = min_value
self.max_value = max_value
class AllowableList(Stringify):
"""Model of a allowableValues of type LIST
See https://github.com/wordnik/swagger-core/wiki/datatypes#complex-types
"""
def __init__(self, values):
self.values = values
def load_allowable_values(json, context):
"""Parse a JSON allowableValues object.
This returns None, AllowableList or AllowableRange, depending on the
valueType in the JSON. If the valueType is not recognized, a SwaggerError
is raised.
"""
if not json:
return None
if not 'valueType' in json:
raise SwaggerError("Missing valueType field", context)
value_type = json['valueType']
if value_type == 'RANGE':
if not 'min' in json and not 'max' in json:
raise SwaggerError("Missing fields min/max", context)
return AllowableRange(json.get('min'), json.get('max'))
if value_type == 'LIST':
if not 'values' in json:
raise SwaggerError("Missing field values", context)
return AllowableList(json['values'])
raise SwaggerError("Unkown valueType %s" % value_type, context)
class Parameter(Stringify):
"""Model of an operation's parameter.
See https://github.com/wordnik/swagger-core/wiki/parameters
"""
required_fields = ['name', 'paramType', 'dataType']
def __init__(self):
self.param_type = None
self.name = None
self.description = None
self.data_type = None
self.required = None
self.allowable_values = None
self.allow_multiple = None
def load(self, parameter_json, processor, context):
Update events to use Swagger 1.3 subtyping, and related aftermath This patch started with the simple idea of changing the /events data model to be more sane. The original model would send out events like: { "stasis_start": { "args": [], "channel": { ... } } } The event discriminator was the field name instead of being a value in the object, due to limitations in how Swagger 1.1 could model objects. While technically sufficient in communicating event information, it was really difficult to deal with in terms of client side JSON handling. This patch takes advantage of a proposed extension[1] to Swagger which allows type variance through the use of a discriminator field. This had a domino effect that made this a surprisingly large patch. [1]: https://groups.google.com/d/msg/wordnik-api/EC3rGajE0os/ey_5dBI_jWcJ In changing the models, I also had to change the swagger_model.py processor so it can handle the type discriminator and subtyping. I took that a big step forward, and using that information to generate an ari_model module, which can validate a JSON object against the Swagger model. The REST and WebSocket generators were changed to take advantage of the validators. If compiled with AST_DEVMODE enabled, JSON objects that don't match their corresponding models will not be sent out. For REST API calls, a 500 Internal Server response is sent. For WebSockets, the invalid JSON message is replaced with an error message. Since this took over about half of the job of the existing JSON generators, and the .to_json virtual function on messages took over the other half, I reluctantly removed the generators. The validators turned up all sorts of errors and inconsistencies in our data models, and the code. These were cleaned up, with checks in the code generator avoid some of the consistency problems in the future. * The model for a channel snapshot was trimmed down to match the information sent via AMI. Many of the field being sent were not useful in the general case. * The model for a bridge snapshot was updated to be more consistent with the other ARI models. Another impact of introducing subtyping was that the swagger-codegen documentation generator was insufficient (at least until it catches up with Swagger 1.2). I wanted it to be easier to generate docs for the API anyways, so I ported the wiki pages to use the Asterisk Swagger generator. In the process, I was able to clean up many of the model links, which would occasionally give inconsistent results on the wiki. I also added error responses to the wiki docs, making the wiki documentation more complete. Finally, since Stasis-HTTP will now be named Asterisk REST Interface (ARI), any new functions and files I created carry the ari_ prefix. I changed a few stasis_http references to ari where it was non-intrusive and made sense. (closes issue ASTERISK-21885) Review: https://reviewboard.asterisk.org/r/2639/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@393529 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-07-03 16:32:41 +00:00
context = context.next_stack(parameter_json, 'name')
validate_required_fields(parameter_json, self.required_fields, context)
self.name = parameter_json.get('name')
self.param_type = parameter_json.get('paramType')
self.description = parameter_json.get('description') or ''
self.data_type = parameter_json.get('dataType')
self.required = parameter_json.get('required') or False
Update events to use Swagger 1.3 subtyping, and related aftermath This patch started with the simple idea of changing the /events data model to be more sane. The original model would send out events like: { "stasis_start": { "args": [], "channel": { ... } } } The event discriminator was the field name instead of being a value in the object, due to limitations in how Swagger 1.1 could model objects. While technically sufficient in communicating event information, it was really difficult to deal with in terms of client side JSON handling. This patch takes advantage of a proposed extension[1] to Swagger which allows type variance through the use of a discriminator field. This had a domino effect that made this a surprisingly large patch. [1]: https://groups.google.com/d/msg/wordnik-api/EC3rGajE0os/ey_5dBI_jWcJ In changing the models, I also had to change the swagger_model.py processor so it can handle the type discriminator and subtyping. I took that a big step forward, and using that information to generate an ari_model module, which can validate a JSON object against the Swagger model. The REST and WebSocket generators were changed to take advantage of the validators. If compiled with AST_DEVMODE enabled, JSON objects that don't match their corresponding models will not be sent out. For REST API calls, a 500 Internal Server response is sent. For WebSockets, the invalid JSON message is replaced with an error message. Since this took over about half of the job of the existing JSON generators, and the .to_json virtual function on messages took over the other half, I reluctantly removed the generators. The validators turned up all sorts of errors and inconsistencies in our data models, and the code. These were cleaned up, with checks in the code generator avoid some of the consistency problems in the future. * The model for a channel snapshot was trimmed down to match the information sent via AMI. Many of the field being sent were not useful in the general case. * The model for a bridge snapshot was updated to be more consistent with the other ARI models. Another impact of introducing subtyping was that the swagger-codegen documentation generator was insufficient (at least until it catches up with Swagger 1.2). I wanted it to be easier to generate docs for the API anyways, so I ported the wiki pages to use the Asterisk Swagger generator. In the process, I was able to clean up many of the model links, which would occasionally give inconsistent results on the wiki. I also added error responses to the wiki docs, making the wiki documentation more complete. Finally, since Stasis-HTTP will now be named Asterisk REST Interface (ARI), any new functions and files I created carry the ari_ prefix. I changed a few stasis_http references to ari where it was non-intrusive and made sense. (closes issue ASTERISK-21885) Review: https://reviewboard.asterisk.org/r/2639/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@393529 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-07-03 16:32:41 +00:00
self.default_value = parameter_json.get('defaultValue')
self.allowable_values = load_allowable_values(
parameter_json.get('allowableValues'), context)
self.allow_multiple = parameter_json.get('allowMultiple') or False
processor.process_parameter(self, context)
Update events to use Swagger 1.3 subtyping, and related aftermath This patch started with the simple idea of changing the /events data model to be more sane. The original model would send out events like: { "stasis_start": { "args": [], "channel": { ... } } } The event discriminator was the field name instead of being a value in the object, due to limitations in how Swagger 1.1 could model objects. While technically sufficient in communicating event information, it was really difficult to deal with in terms of client side JSON handling. This patch takes advantage of a proposed extension[1] to Swagger which allows type variance through the use of a discriminator field. This had a domino effect that made this a surprisingly large patch. [1]: https://groups.google.com/d/msg/wordnik-api/EC3rGajE0os/ey_5dBI_jWcJ In changing the models, I also had to change the swagger_model.py processor so it can handle the type discriminator and subtyping. I took that a big step forward, and using that information to generate an ari_model module, which can validate a JSON object against the Swagger model. The REST and WebSocket generators were changed to take advantage of the validators. If compiled with AST_DEVMODE enabled, JSON objects that don't match their corresponding models will not be sent out. For REST API calls, a 500 Internal Server response is sent. For WebSockets, the invalid JSON message is replaced with an error message. Since this took over about half of the job of the existing JSON generators, and the .to_json virtual function on messages took over the other half, I reluctantly removed the generators. The validators turned up all sorts of errors and inconsistencies in our data models, and the code. These were cleaned up, with checks in the code generator avoid some of the consistency problems in the future. * The model for a channel snapshot was trimmed down to match the information sent via AMI. Many of the field being sent were not useful in the general case. * The model for a bridge snapshot was updated to be more consistent with the other ARI models. Another impact of introducing subtyping was that the swagger-codegen documentation generator was insufficient (at least until it catches up with Swagger 1.2). I wanted it to be easier to generate docs for the API anyways, so I ported the wiki pages to use the Asterisk Swagger generator. In the process, I was able to clean up many of the model links, which would occasionally give inconsistent results on the wiki. I also added error responses to the wiki docs, making the wiki documentation more complete. Finally, since Stasis-HTTP will now be named Asterisk REST Interface (ARI), any new functions and files I created carry the ari_ prefix. I changed a few stasis_http references to ari where it was non-intrusive and made sense. (closes issue ASTERISK-21885) Review: https://reviewboard.asterisk.org/r/2639/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@393529 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-07-03 16:32:41 +00:00
if parameter_json.get('allowedValues'):
raise SwaggerError(
"Field 'allowedValues' invalid; use 'allowableValues'",
context)
return self
def is_type(self, other_type):
return self.param_type == other_type
class ErrorResponse(Stringify):
"""Model of an error response.
See https://github.com/wordnik/swagger-core/wiki/errors
"""
required_fields = ['code', 'reason']
def __init__(self):
self.code = None
self.reason = None
def load(self, err_json, processor, context):
Update events to use Swagger 1.3 subtyping, and related aftermath This patch started with the simple idea of changing the /events data model to be more sane. The original model would send out events like: { "stasis_start": { "args": [], "channel": { ... } } } The event discriminator was the field name instead of being a value in the object, due to limitations in how Swagger 1.1 could model objects. While technically sufficient in communicating event information, it was really difficult to deal with in terms of client side JSON handling. This patch takes advantage of a proposed extension[1] to Swagger which allows type variance through the use of a discriminator field. This had a domino effect that made this a surprisingly large patch. [1]: https://groups.google.com/d/msg/wordnik-api/EC3rGajE0os/ey_5dBI_jWcJ In changing the models, I also had to change the swagger_model.py processor so it can handle the type discriminator and subtyping. I took that a big step forward, and using that information to generate an ari_model module, which can validate a JSON object against the Swagger model. The REST and WebSocket generators were changed to take advantage of the validators. If compiled with AST_DEVMODE enabled, JSON objects that don't match their corresponding models will not be sent out. For REST API calls, a 500 Internal Server response is sent. For WebSockets, the invalid JSON message is replaced with an error message. Since this took over about half of the job of the existing JSON generators, and the .to_json virtual function on messages took over the other half, I reluctantly removed the generators. The validators turned up all sorts of errors and inconsistencies in our data models, and the code. These were cleaned up, with checks in the code generator avoid some of the consistency problems in the future. * The model for a channel snapshot was trimmed down to match the information sent via AMI. Many of the field being sent were not useful in the general case. * The model for a bridge snapshot was updated to be more consistent with the other ARI models. Another impact of introducing subtyping was that the swagger-codegen documentation generator was insufficient (at least until it catches up with Swagger 1.2). I wanted it to be easier to generate docs for the API anyways, so I ported the wiki pages to use the Asterisk Swagger generator. In the process, I was able to clean up many of the model links, which would occasionally give inconsistent results on the wiki. I also added error responses to the wiki docs, making the wiki documentation more complete. Finally, since Stasis-HTTP will now be named Asterisk REST Interface (ARI), any new functions and files I created carry the ari_ prefix. I changed a few stasis_http references to ari where it was non-intrusive and made sense. (closes issue ASTERISK-21885) Review: https://reviewboard.asterisk.org/r/2639/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@393529 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-07-03 16:32:41 +00:00
context = context.next_stack(err_json, 'code')
validate_required_fields(err_json, self.required_fields, context)
self.code = err_json.get('code')
self.reason = err_json.get('reason')
return self
Update events to use Swagger 1.3 subtyping, and related aftermath This patch started with the simple idea of changing the /events data model to be more sane. The original model would send out events like: { "stasis_start": { "args": [], "channel": { ... } } } The event discriminator was the field name instead of being a value in the object, due to limitations in how Swagger 1.1 could model objects. While technically sufficient in communicating event information, it was really difficult to deal with in terms of client side JSON handling. This patch takes advantage of a proposed extension[1] to Swagger which allows type variance through the use of a discriminator field. This had a domino effect that made this a surprisingly large patch. [1]: https://groups.google.com/d/msg/wordnik-api/EC3rGajE0os/ey_5dBI_jWcJ In changing the models, I also had to change the swagger_model.py processor so it can handle the type discriminator and subtyping. I took that a big step forward, and using that information to generate an ari_model module, which can validate a JSON object against the Swagger model. The REST and WebSocket generators were changed to take advantage of the validators. If compiled with AST_DEVMODE enabled, JSON objects that don't match their corresponding models will not be sent out. For REST API calls, a 500 Internal Server response is sent. For WebSockets, the invalid JSON message is replaced with an error message. Since this took over about half of the job of the existing JSON generators, and the .to_json virtual function on messages took over the other half, I reluctantly removed the generators. The validators turned up all sorts of errors and inconsistencies in our data models, and the code. These were cleaned up, with checks in the code generator avoid some of the consistency problems in the future. * The model for a channel snapshot was trimmed down to match the information sent via AMI. Many of the field being sent were not useful in the general case. * The model for a bridge snapshot was updated to be more consistent with the other ARI models. Another impact of introducing subtyping was that the swagger-codegen documentation generator was insufficient (at least until it catches up with Swagger 1.2). I wanted it to be easier to generate docs for the API anyways, so I ported the wiki pages to use the Asterisk Swagger generator. In the process, I was able to clean up many of the model links, which would occasionally give inconsistent results on the wiki. I also added error responses to the wiki docs, making the wiki documentation more complete. Finally, since Stasis-HTTP will now be named Asterisk REST Interface (ARI), any new functions and files I created carry the ari_ prefix. I changed a few stasis_http references to ari where it was non-intrusive and made sense. (closes issue ASTERISK-21885) Review: https://reviewboard.asterisk.org/r/2639/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@393529 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-07-03 16:32:41 +00:00
class SwaggerType(Stringify):
"""Model of a data type.
"""
def __init__(self):
self.name = None
self.is_discriminator = None
self.is_list = None
self.singular_name = None
self.is_primitive = None
def load(self, type_name, processor, context):
# Some common errors
if type_name == 'integer':
raise SwaggerError("The type for integer should be 'int'", context)
self.name = type_name
type_param = get_list_parameter_type(self.name)
self.is_list = type_param is not None
if self.is_list:
self.singular_name = type_param
else:
self.singular_name = self.name
self.is_primitive = self.singular_name in SWAGGER_PRIMITIVES
processor.process_type(self, context)
return self
class Operation(Stringify):
"""Model of an operation on an API
See https://github.com/wordnik/swagger-core/wiki/API-Declaration#apis
"""
required_fields = ['httpMethod', 'nickname', 'responseClass', 'summary']
def __init__(self):
self.http_method = None
self.nickname = None
self.response_class = None
self.parameters = []
self.summary = None
self.notes = None
self.error_responses = []
def load(self, op_json, processor, context):
Update events to use Swagger 1.3 subtyping, and related aftermath This patch started with the simple idea of changing the /events data model to be more sane. The original model would send out events like: { "stasis_start": { "args": [], "channel": { ... } } } The event discriminator was the field name instead of being a value in the object, due to limitations in how Swagger 1.1 could model objects. While technically sufficient in communicating event information, it was really difficult to deal with in terms of client side JSON handling. This patch takes advantage of a proposed extension[1] to Swagger which allows type variance through the use of a discriminator field. This had a domino effect that made this a surprisingly large patch. [1]: https://groups.google.com/d/msg/wordnik-api/EC3rGajE0os/ey_5dBI_jWcJ In changing the models, I also had to change the swagger_model.py processor so it can handle the type discriminator and subtyping. I took that a big step forward, and using that information to generate an ari_model module, which can validate a JSON object against the Swagger model. The REST and WebSocket generators were changed to take advantage of the validators. If compiled with AST_DEVMODE enabled, JSON objects that don't match their corresponding models will not be sent out. For REST API calls, a 500 Internal Server response is sent. For WebSockets, the invalid JSON message is replaced with an error message. Since this took over about half of the job of the existing JSON generators, and the .to_json virtual function on messages took over the other half, I reluctantly removed the generators. The validators turned up all sorts of errors and inconsistencies in our data models, and the code. These were cleaned up, with checks in the code generator avoid some of the consistency problems in the future. * The model for a channel snapshot was trimmed down to match the information sent via AMI. Many of the field being sent were not useful in the general case. * The model for a bridge snapshot was updated to be more consistent with the other ARI models. Another impact of introducing subtyping was that the swagger-codegen documentation generator was insufficient (at least until it catches up with Swagger 1.2). I wanted it to be easier to generate docs for the API anyways, so I ported the wiki pages to use the Asterisk Swagger generator. In the process, I was able to clean up many of the model links, which would occasionally give inconsistent results on the wiki. I also added error responses to the wiki docs, making the wiki documentation more complete. Finally, since Stasis-HTTP will now be named Asterisk REST Interface (ARI), any new functions and files I created carry the ari_ prefix. I changed a few stasis_http references to ari where it was non-intrusive and made sense. (closes issue ASTERISK-21885) Review: https://reviewboard.asterisk.org/r/2639/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@393529 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-07-03 16:32:41 +00:00
context = context.next_stack(op_json, 'nickname')
validate_required_fields(op_json, self.required_fields, context)
self.http_method = op_json.get('httpMethod')
self.nickname = op_json.get('nickname')
Update events to use Swagger 1.3 subtyping, and related aftermath This patch started with the simple idea of changing the /events data model to be more sane. The original model would send out events like: { "stasis_start": { "args": [], "channel": { ... } } } The event discriminator was the field name instead of being a value in the object, due to limitations in how Swagger 1.1 could model objects. While technically sufficient in communicating event information, it was really difficult to deal with in terms of client side JSON handling. This patch takes advantage of a proposed extension[1] to Swagger which allows type variance through the use of a discriminator field. This had a domino effect that made this a surprisingly large patch. [1]: https://groups.google.com/d/msg/wordnik-api/EC3rGajE0os/ey_5dBI_jWcJ In changing the models, I also had to change the swagger_model.py processor so it can handle the type discriminator and subtyping. I took that a big step forward, and using that information to generate an ari_model module, which can validate a JSON object against the Swagger model. The REST and WebSocket generators were changed to take advantage of the validators. If compiled with AST_DEVMODE enabled, JSON objects that don't match their corresponding models will not be sent out. For REST API calls, a 500 Internal Server response is sent. For WebSockets, the invalid JSON message is replaced with an error message. Since this took over about half of the job of the existing JSON generators, and the .to_json virtual function on messages took over the other half, I reluctantly removed the generators. The validators turned up all sorts of errors and inconsistencies in our data models, and the code. These were cleaned up, with checks in the code generator avoid some of the consistency problems in the future. * The model for a channel snapshot was trimmed down to match the information sent via AMI. Many of the field being sent were not useful in the general case. * The model for a bridge snapshot was updated to be more consistent with the other ARI models. Another impact of introducing subtyping was that the swagger-codegen documentation generator was insufficient (at least until it catches up with Swagger 1.2). I wanted it to be easier to generate docs for the API anyways, so I ported the wiki pages to use the Asterisk Swagger generator. In the process, I was able to clean up many of the model links, which would occasionally give inconsistent results on the wiki. I also added error responses to the wiki docs, making the wiki documentation more complete. Finally, since Stasis-HTTP will now be named Asterisk REST Interface (ARI), any new functions and files I created carry the ari_ prefix. I changed a few stasis_http references to ari where it was non-intrusive and made sense. (closes issue ASTERISK-21885) Review: https://reviewboard.asterisk.org/r/2639/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@393529 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-07-03 16:32:41 +00:00
response_class = op_json.get('responseClass')
self.response_class = response_class and SwaggerType().load(
response_class, processor, context)
# Specifying WebSocket URL's is our own extension
self.is_websocket = op_json.get('upgrade') == 'websocket'
self.is_req = not self.is_websocket
if self.is_websocket:
self.websocket_protocol = op_json.get('websocketProtocol')
if self.http_method != 'GET':
raise SwaggerError(
"upgrade: websocket is only valid on GET operations",
context)
params_json = op_json.get('parameters') or []
self.parameters = [
Parameter().load(j, processor, context) for j in params_json]
self.query_parameters = [
p for p in self.parameters if p.is_type('query')]
self.has_query_parameters = self.query_parameters and True
self.path_parameters = [
p for p in self.parameters if p.is_type('path')]
self.has_path_parameters = self.path_parameters and True
self.header_parameters = [
p for p in self.parameters if p.is_type('header')]
self.has_header_parameters = self.header_parameters and True
self.has_parameters = self.has_query_parameters or \
self.has_path_parameters or self.has_header_parameters
# Body param is different, since there's at most one
self.body_parameter = [
p for p in self.parameters if p.is_type('body')]
if len(self.body_parameter) > 1:
raise SwaggerError("Cannot have more than one body param", context)
self.body_parameter = self.body_parameter and self.body_parameter[0]
self.summary = op_json.get('summary')
self.notes = op_json.get('notes')
err_json = op_json.get('errorResponses') or []
self.error_responses = [
ErrorResponse().load(j, processor, context) for j in err_json]
Update events to use Swagger 1.3 subtyping, and related aftermath This patch started with the simple idea of changing the /events data model to be more sane. The original model would send out events like: { "stasis_start": { "args": [], "channel": { ... } } } The event discriminator was the field name instead of being a value in the object, due to limitations in how Swagger 1.1 could model objects. While technically sufficient in communicating event information, it was really difficult to deal with in terms of client side JSON handling. This patch takes advantage of a proposed extension[1] to Swagger which allows type variance through the use of a discriminator field. This had a domino effect that made this a surprisingly large patch. [1]: https://groups.google.com/d/msg/wordnik-api/EC3rGajE0os/ey_5dBI_jWcJ In changing the models, I also had to change the swagger_model.py processor so it can handle the type discriminator and subtyping. I took that a big step forward, and using that information to generate an ari_model module, which can validate a JSON object against the Swagger model. The REST and WebSocket generators were changed to take advantage of the validators. If compiled with AST_DEVMODE enabled, JSON objects that don't match their corresponding models will not be sent out. For REST API calls, a 500 Internal Server response is sent. For WebSockets, the invalid JSON message is replaced with an error message. Since this took over about half of the job of the existing JSON generators, and the .to_json virtual function on messages took over the other half, I reluctantly removed the generators. The validators turned up all sorts of errors and inconsistencies in our data models, and the code. These were cleaned up, with checks in the code generator avoid some of the consistency problems in the future. * The model for a channel snapshot was trimmed down to match the information sent via AMI. Many of the field being sent were not useful in the general case. * The model for a bridge snapshot was updated to be more consistent with the other ARI models. Another impact of introducing subtyping was that the swagger-codegen documentation generator was insufficient (at least until it catches up with Swagger 1.2). I wanted it to be easier to generate docs for the API anyways, so I ported the wiki pages to use the Asterisk Swagger generator. In the process, I was able to clean up many of the model links, which would occasionally give inconsistent results on the wiki. I also added error responses to the wiki docs, making the wiki documentation more complete. Finally, since Stasis-HTTP will now be named Asterisk REST Interface (ARI), any new functions and files I created carry the ari_ prefix. I changed a few stasis_http references to ari where it was non-intrusive and made sense. (closes issue ASTERISK-21885) Review: https://reviewboard.asterisk.org/r/2639/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@393529 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-07-03 16:32:41 +00:00
self.has_error_responses = self.error_responses != []
processor.process_operation(self, context)
return self
class Api(Stringify):
"""Model of a single API in an API declaration.
See https://github.com/wordnik/swagger-core/wiki/API-Declaration
"""
required_fields = ['path', 'operations']
def __init__(self,):
self.path = None
self.description = None
self.operations = []
def load(self, api_json, processor, context):
Update events to use Swagger 1.3 subtyping, and related aftermath This patch started with the simple idea of changing the /events data model to be more sane. The original model would send out events like: { "stasis_start": { "args": [], "channel": { ... } } } The event discriminator was the field name instead of being a value in the object, due to limitations in how Swagger 1.1 could model objects. While technically sufficient in communicating event information, it was really difficult to deal with in terms of client side JSON handling. This patch takes advantage of a proposed extension[1] to Swagger which allows type variance through the use of a discriminator field. This had a domino effect that made this a surprisingly large patch. [1]: https://groups.google.com/d/msg/wordnik-api/EC3rGajE0os/ey_5dBI_jWcJ In changing the models, I also had to change the swagger_model.py processor so it can handle the type discriminator and subtyping. I took that a big step forward, and using that information to generate an ari_model module, which can validate a JSON object against the Swagger model. The REST and WebSocket generators were changed to take advantage of the validators. If compiled with AST_DEVMODE enabled, JSON objects that don't match their corresponding models will not be sent out. For REST API calls, a 500 Internal Server response is sent. For WebSockets, the invalid JSON message is replaced with an error message. Since this took over about half of the job of the existing JSON generators, and the .to_json virtual function on messages took over the other half, I reluctantly removed the generators. The validators turned up all sorts of errors and inconsistencies in our data models, and the code. These were cleaned up, with checks in the code generator avoid some of the consistency problems in the future. * The model for a channel snapshot was trimmed down to match the information sent via AMI. Many of the field being sent were not useful in the general case. * The model for a bridge snapshot was updated to be more consistent with the other ARI models. Another impact of introducing subtyping was that the swagger-codegen documentation generator was insufficient (at least until it catches up with Swagger 1.2). I wanted it to be easier to generate docs for the API anyways, so I ported the wiki pages to use the Asterisk Swagger generator. In the process, I was able to clean up many of the model links, which would occasionally give inconsistent results on the wiki. I also added error responses to the wiki docs, making the wiki documentation more complete. Finally, since Stasis-HTTP will now be named Asterisk REST Interface (ARI), any new functions and files I created carry the ari_ prefix. I changed a few stasis_http references to ari where it was non-intrusive and made sense. (closes issue ASTERISK-21885) Review: https://reviewboard.asterisk.org/r/2639/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@393529 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-07-03 16:32:41 +00:00
context = context.next_stack(api_json, 'path')
validate_required_fields(api_json, self.required_fields, context)
self.path = api_json.get('path')
self.description = api_json.get('description')
op_json = api_json.get('operations')
self.operations = [
Operation().load(j, processor, context) for j in op_json]
self.has_websocket = \
filter(lambda op: op.is_websocket, self.operations) != []
Update events to use Swagger 1.3 subtyping, and related aftermath This patch started with the simple idea of changing the /events data model to be more sane. The original model would send out events like: { "stasis_start": { "args": [], "channel": { ... } } } The event discriminator was the field name instead of being a value in the object, due to limitations in how Swagger 1.1 could model objects. While technically sufficient in communicating event information, it was really difficult to deal with in terms of client side JSON handling. This patch takes advantage of a proposed extension[1] to Swagger which allows type variance through the use of a discriminator field. This had a domino effect that made this a surprisingly large patch. [1]: https://groups.google.com/d/msg/wordnik-api/EC3rGajE0os/ey_5dBI_jWcJ In changing the models, I also had to change the swagger_model.py processor so it can handle the type discriminator and subtyping. I took that a big step forward, and using that information to generate an ari_model module, which can validate a JSON object against the Swagger model. The REST and WebSocket generators were changed to take advantage of the validators. If compiled with AST_DEVMODE enabled, JSON objects that don't match their corresponding models will not be sent out. For REST API calls, a 500 Internal Server response is sent. For WebSockets, the invalid JSON message is replaced with an error message. Since this took over about half of the job of the existing JSON generators, and the .to_json virtual function on messages took over the other half, I reluctantly removed the generators. The validators turned up all sorts of errors and inconsistencies in our data models, and the code. These were cleaned up, with checks in the code generator avoid some of the consistency problems in the future. * The model for a channel snapshot was trimmed down to match the information sent via AMI. Many of the field being sent were not useful in the general case. * The model for a bridge snapshot was updated to be more consistent with the other ARI models. Another impact of introducing subtyping was that the swagger-codegen documentation generator was insufficient (at least until it catches up with Swagger 1.2). I wanted it to be easier to generate docs for the API anyways, so I ported the wiki pages to use the Asterisk Swagger generator. In the process, I was able to clean up many of the model links, which would occasionally give inconsistent results on the wiki. I also added error responses to the wiki docs, making the wiki documentation more complete. Finally, since Stasis-HTTP will now be named Asterisk REST Interface (ARI), any new functions and files I created carry the ari_ prefix. I changed a few stasis_http references to ari where it was non-intrusive and made sense. (closes issue ASTERISK-21885) Review: https://reviewboard.asterisk.org/r/2639/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@393529 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-07-03 16:32:41 +00:00
processor.process_api(self, context)
return self
Update events to use Swagger 1.3 subtyping, and related aftermath This patch started with the simple idea of changing the /events data model to be more sane. The original model would send out events like: { "stasis_start": { "args": [], "channel": { ... } } } The event discriminator was the field name instead of being a value in the object, due to limitations in how Swagger 1.1 could model objects. While technically sufficient in communicating event information, it was really difficult to deal with in terms of client side JSON handling. This patch takes advantage of a proposed extension[1] to Swagger which allows type variance through the use of a discriminator field. This had a domino effect that made this a surprisingly large patch. [1]: https://groups.google.com/d/msg/wordnik-api/EC3rGajE0os/ey_5dBI_jWcJ In changing the models, I also had to change the swagger_model.py processor so it can handle the type discriminator and subtyping. I took that a big step forward, and using that information to generate an ari_model module, which can validate a JSON object against the Swagger model. The REST and WebSocket generators were changed to take advantage of the validators. If compiled with AST_DEVMODE enabled, JSON objects that don't match their corresponding models will not be sent out. For REST API calls, a 500 Internal Server response is sent. For WebSockets, the invalid JSON message is replaced with an error message. Since this took over about half of the job of the existing JSON generators, and the .to_json virtual function on messages took over the other half, I reluctantly removed the generators. The validators turned up all sorts of errors and inconsistencies in our data models, and the code. These were cleaned up, with checks in the code generator avoid some of the consistency problems in the future. * The model for a channel snapshot was trimmed down to match the information sent via AMI. Many of the field being sent were not useful in the general case. * The model for a bridge snapshot was updated to be more consistent with the other ARI models. Another impact of introducing subtyping was that the swagger-codegen documentation generator was insufficient (at least until it catches up with Swagger 1.2). I wanted it to be easier to generate docs for the API anyways, so I ported the wiki pages to use the Asterisk Swagger generator. In the process, I was able to clean up many of the model links, which would occasionally give inconsistent results on the wiki. I also added error responses to the wiki docs, making the wiki documentation more complete. Finally, since Stasis-HTTP will now be named Asterisk REST Interface (ARI), any new functions and files I created carry the ari_ prefix. I changed a few stasis_http references to ari where it was non-intrusive and made sense. (closes issue ASTERISK-21885) Review: https://reviewboard.asterisk.org/r/2639/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@393529 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-07-03 16:32:41 +00:00
def get_list_parameter_type(type_string):
"""Returns the type parameter if the given type_string is List[].
@param type_string: Type string to parse
@returns Type parameter of the list, or None if not a List.
"""
list_match = re.match('^List\[(.*)\]$', type_string)
return list_match and list_match.group(1)
class Property(Stringify):
"""Model of a Swagger property.
See https://github.com/wordnik/swagger-core/wiki/datatypes
"""
required_fields = ['type']
def __init__(self, name):
self.name = name
self.type = None
self.description = None
self.required = None
def load(self, property_json, processor, context):
validate_required_fields(property_json, self.required_fields, context)
Update events to use Swagger 1.3 subtyping, and related aftermath This patch started with the simple idea of changing the /events data model to be more sane. The original model would send out events like: { "stasis_start": { "args": [], "channel": { ... } } } The event discriminator was the field name instead of being a value in the object, due to limitations in how Swagger 1.1 could model objects. While technically sufficient in communicating event information, it was really difficult to deal with in terms of client side JSON handling. This patch takes advantage of a proposed extension[1] to Swagger which allows type variance through the use of a discriminator field. This had a domino effect that made this a surprisingly large patch. [1]: https://groups.google.com/d/msg/wordnik-api/EC3rGajE0os/ey_5dBI_jWcJ In changing the models, I also had to change the swagger_model.py processor so it can handle the type discriminator and subtyping. I took that a big step forward, and using that information to generate an ari_model module, which can validate a JSON object against the Swagger model. The REST and WebSocket generators were changed to take advantage of the validators. If compiled with AST_DEVMODE enabled, JSON objects that don't match their corresponding models will not be sent out. For REST API calls, a 500 Internal Server response is sent. For WebSockets, the invalid JSON message is replaced with an error message. Since this took over about half of the job of the existing JSON generators, and the .to_json virtual function on messages took over the other half, I reluctantly removed the generators. The validators turned up all sorts of errors and inconsistencies in our data models, and the code. These were cleaned up, with checks in the code generator avoid some of the consistency problems in the future. * The model for a channel snapshot was trimmed down to match the information sent via AMI. Many of the field being sent were not useful in the general case. * The model for a bridge snapshot was updated to be more consistent with the other ARI models. Another impact of introducing subtyping was that the swagger-codegen documentation generator was insufficient (at least until it catches up with Swagger 1.2). I wanted it to be easier to generate docs for the API anyways, so I ported the wiki pages to use the Asterisk Swagger generator. In the process, I was able to clean up many of the model links, which would occasionally give inconsistent results on the wiki. I also added error responses to the wiki docs, making the wiki documentation more complete. Finally, since Stasis-HTTP will now be named Asterisk REST Interface (ARI), any new functions and files I created carry the ari_ prefix. I changed a few stasis_http references to ari where it was non-intrusive and made sense. (closes issue ASTERISK-21885) Review: https://reviewboard.asterisk.org/r/2639/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@393529 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-07-03 16:32:41 +00:00
# Bit of a hack, but properties do not self-identify
context = context.next_stack({'name': self.name}, 'name')
self.description = property_json.get('description') or ''
self.required = property_json.get('required') or False
Update events to use Swagger 1.3 subtyping, and related aftermath This patch started with the simple idea of changing the /events data model to be more sane. The original model would send out events like: { "stasis_start": { "args": [], "channel": { ... } } } The event discriminator was the field name instead of being a value in the object, due to limitations in how Swagger 1.1 could model objects. While technically sufficient in communicating event information, it was really difficult to deal with in terms of client side JSON handling. This patch takes advantage of a proposed extension[1] to Swagger which allows type variance through the use of a discriminator field. This had a domino effect that made this a surprisingly large patch. [1]: https://groups.google.com/d/msg/wordnik-api/EC3rGajE0os/ey_5dBI_jWcJ In changing the models, I also had to change the swagger_model.py processor so it can handle the type discriminator and subtyping. I took that a big step forward, and using that information to generate an ari_model module, which can validate a JSON object against the Swagger model. The REST and WebSocket generators were changed to take advantage of the validators. If compiled with AST_DEVMODE enabled, JSON objects that don't match their corresponding models will not be sent out. For REST API calls, a 500 Internal Server response is sent. For WebSockets, the invalid JSON message is replaced with an error message. Since this took over about half of the job of the existing JSON generators, and the .to_json virtual function on messages took over the other half, I reluctantly removed the generators. The validators turned up all sorts of errors and inconsistencies in our data models, and the code. These were cleaned up, with checks in the code generator avoid some of the consistency problems in the future. * The model for a channel snapshot was trimmed down to match the information sent via AMI. Many of the field being sent were not useful in the general case. * The model for a bridge snapshot was updated to be more consistent with the other ARI models. Another impact of introducing subtyping was that the swagger-codegen documentation generator was insufficient (at least until it catches up with Swagger 1.2). I wanted it to be easier to generate docs for the API anyways, so I ported the wiki pages to use the Asterisk Swagger generator. In the process, I was able to clean up many of the model links, which would occasionally give inconsistent results on the wiki. I also added error responses to the wiki docs, making the wiki documentation more complete. Finally, since Stasis-HTTP will now be named Asterisk REST Interface (ARI), any new functions and files I created carry the ari_ prefix. I changed a few stasis_http references to ari where it was non-intrusive and made sense. (closes issue ASTERISK-21885) Review: https://reviewboard.asterisk.org/r/2639/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@393529 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-07-03 16:32:41 +00:00
type = property_json.get('type')
self.type = type and SwaggerType().load(type, processor, context)
processor.process_property(self, context)
return self
class Model(Stringify):
"""Model of a Swagger model.
See https://github.com/wordnik/swagger-core/wiki/datatypes
"""
Update events to use Swagger 1.3 subtyping, and related aftermath This patch started with the simple idea of changing the /events data model to be more sane. The original model would send out events like: { "stasis_start": { "args": [], "channel": { ... } } } The event discriminator was the field name instead of being a value in the object, due to limitations in how Swagger 1.1 could model objects. While technically sufficient in communicating event information, it was really difficult to deal with in terms of client side JSON handling. This patch takes advantage of a proposed extension[1] to Swagger which allows type variance through the use of a discriminator field. This had a domino effect that made this a surprisingly large patch. [1]: https://groups.google.com/d/msg/wordnik-api/EC3rGajE0os/ey_5dBI_jWcJ In changing the models, I also had to change the swagger_model.py processor so it can handle the type discriminator and subtyping. I took that a big step forward, and using that information to generate an ari_model module, which can validate a JSON object against the Swagger model. The REST and WebSocket generators were changed to take advantage of the validators. If compiled with AST_DEVMODE enabled, JSON objects that don't match their corresponding models will not be sent out. For REST API calls, a 500 Internal Server response is sent. For WebSockets, the invalid JSON message is replaced with an error message. Since this took over about half of the job of the existing JSON generators, and the .to_json virtual function on messages took over the other half, I reluctantly removed the generators. The validators turned up all sorts of errors and inconsistencies in our data models, and the code. These were cleaned up, with checks in the code generator avoid some of the consistency problems in the future. * The model for a channel snapshot was trimmed down to match the information sent via AMI. Many of the field being sent were not useful in the general case. * The model for a bridge snapshot was updated to be more consistent with the other ARI models. Another impact of introducing subtyping was that the swagger-codegen documentation generator was insufficient (at least until it catches up with Swagger 1.2). I wanted it to be easier to generate docs for the API anyways, so I ported the wiki pages to use the Asterisk Swagger generator. In the process, I was able to clean up many of the model links, which would occasionally give inconsistent results on the wiki. I also added error responses to the wiki docs, making the wiki documentation more complete. Finally, since Stasis-HTTP will now be named Asterisk REST Interface (ARI), any new functions and files I created carry the ari_ prefix. I changed a few stasis_http references to ari where it was non-intrusive and made sense. (closes issue ASTERISK-21885) Review: https://reviewboard.asterisk.org/r/2639/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@393529 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-07-03 16:32:41 +00:00
required_fields = ['description', 'properties']
def __init__(self):
self.id = None
self.subtypes = []
self.__subtype_types = []
self.notes = None
self.description = None
Update events to use Swagger 1.3 subtyping, and related aftermath This patch started with the simple idea of changing the /events data model to be more sane. The original model would send out events like: { "stasis_start": { "args": [], "channel": { ... } } } The event discriminator was the field name instead of being a value in the object, due to limitations in how Swagger 1.1 could model objects. While technically sufficient in communicating event information, it was really difficult to deal with in terms of client side JSON handling. This patch takes advantage of a proposed extension[1] to Swagger which allows type variance through the use of a discriminator field. This had a domino effect that made this a surprisingly large patch. [1]: https://groups.google.com/d/msg/wordnik-api/EC3rGajE0os/ey_5dBI_jWcJ In changing the models, I also had to change the swagger_model.py processor so it can handle the type discriminator and subtyping. I took that a big step forward, and using that information to generate an ari_model module, which can validate a JSON object against the Swagger model. The REST and WebSocket generators were changed to take advantage of the validators. If compiled with AST_DEVMODE enabled, JSON objects that don't match their corresponding models will not be sent out. For REST API calls, a 500 Internal Server response is sent. For WebSockets, the invalid JSON message is replaced with an error message. Since this took over about half of the job of the existing JSON generators, and the .to_json virtual function on messages took over the other half, I reluctantly removed the generators. The validators turned up all sorts of errors and inconsistencies in our data models, and the code. These were cleaned up, with checks in the code generator avoid some of the consistency problems in the future. * The model for a channel snapshot was trimmed down to match the information sent via AMI. Many of the field being sent were not useful in the general case. * The model for a bridge snapshot was updated to be more consistent with the other ARI models. Another impact of introducing subtyping was that the swagger-codegen documentation generator was insufficient (at least until it catches up with Swagger 1.2). I wanted it to be easier to generate docs for the API anyways, so I ported the wiki pages to use the Asterisk Swagger generator. In the process, I was able to clean up many of the model links, which would occasionally give inconsistent results on the wiki. I also added error responses to the wiki docs, making the wiki documentation more complete. Finally, since Stasis-HTTP will now be named Asterisk REST Interface (ARI), any new functions and files I created carry the ari_ prefix. I changed a few stasis_http references to ari where it was non-intrusive and made sense. (closes issue ASTERISK-21885) Review: https://reviewboard.asterisk.org/r/2639/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@393529 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-07-03 16:32:41 +00:00
self.__properties = None
self.__discriminator = None
self.__extends_type = None
def load(self, id, model_json, processor, context):
Update events to use Swagger 1.3 subtyping, and related aftermath This patch started with the simple idea of changing the /events data model to be more sane. The original model would send out events like: { "stasis_start": { "args": [], "channel": { ... } } } The event discriminator was the field name instead of being a value in the object, due to limitations in how Swagger 1.1 could model objects. While technically sufficient in communicating event information, it was really difficult to deal with in terms of client side JSON handling. This patch takes advantage of a proposed extension[1] to Swagger which allows type variance through the use of a discriminator field. This had a domino effect that made this a surprisingly large patch. [1]: https://groups.google.com/d/msg/wordnik-api/EC3rGajE0os/ey_5dBI_jWcJ In changing the models, I also had to change the swagger_model.py processor so it can handle the type discriminator and subtyping. I took that a big step forward, and using that information to generate an ari_model module, which can validate a JSON object against the Swagger model. The REST and WebSocket generators were changed to take advantage of the validators. If compiled with AST_DEVMODE enabled, JSON objects that don't match their corresponding models will not be sent out. For REST API calls, a 500 Internal Server response is sent. For WebSockets, the invalid JSON message is replaced with an error message. Since this took over about half of the job of the existing JSON generators, and the .to_json virtual function on messages took over the other half, I reluctantly removed the generators. The validators turned up all sorts of errors and inconsistencies in our data models, and the code. These were cleaned up, with checks in the code generator avoid some of the consistency problems in the future. * The model for a channel snapshot was trimmed down to match the information sent via AMI. Many of the field being sent were not useful in the general case. * The model for a bridge snapshot was updated to be more consistent with the other ARI models. Another impact of introducing subtyping was that the swagger-codegen documentation generator was insufficient (at least until it catches up with Swagger 1.2). I wanted it to be easier to generate docs for the API anyways, so I ported the wiki pages to use the Asterisk Swagger generator. In the process, I was able to clean up many of the model links, which would occasionally give inconsistent results on the wiki. I also added error responses to the wiki docs, making the wiki documentation more complete. Finally, since Stasis-HTTP will now be named Asterisk REST Interface (ARI), any new functions and files I created carry the ari_ prefix. I changed a few stasis_http references to ari where it was non-intrusive and made sense. (closes issue ASTERISK-21885) Review: https://reviewboard.asterisk.org/r/2639/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@393529 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-07-03 16:32:41 +00:00
context = context.next_stack(model_json, 'id')
validate_required_fields(model_json, self.required_fields, context)
# The duplication of the model's id is required by the Swagger spec.
self.id = model_json.get('id')
if id != self.id:
Update events to use Swagger 1.3 subtyping, and related aftermath This patch started with the simple idea of changing the /events data model to be more sane. The original model would send out events like: { "stasis_start": { "args": [], "channel": { ... } } } The event discriminator was the field name instead of being a value in the object, due to limitations in how Swagger 1.1 could model objects. While technically sufficient in communicating event information, it was really difficult to deal with in terms of client side JSON handling. This patch takes advantage of a proposed extension[1] to Swagger which allows type variance through the use of a discriminator field. This had a domino effect that made this a surprisingly large patch. [1]: https://groups.google.com/d/msg/wordnik-api/EC3rGajE0os/ey_5dBI_jWcJ In changing the models, I also had to change the swagger_model.py processor so it can handle the type discriminator and subtyping. I took that a big step forward, and using that information to generate an ari_model module, which can validate a JSON object against the Swagger model. The REST and WebSocket generators were changed to take advantage of the validators. If compiled with AST_DEVMODE enabled, JSON objects that don't match their corresponding models will not be sent out. For REST API calls, a 500 Internal Server response is sent. For WebSockets, the invalid JSON message is replaced with an error message. Since this took over about half of the job of the existing JSON generators, and the .to_json virtual function on messages took over the other half, I reluctantly removed the generators. The validators turned up all sorts of errors and inconsistencies in our data models, and the code. These were cleaned up, with checks in the code generator avoid some of the consistency problems in the future. * The model for a channel snapshot was trimmed down to match the information sent via AMI. Many of the field being sent were not useful in the general case. * The model for a bridge snapshot was updated to be more consistent with the other ARI models. Another impact of introducing subtyping was that the swagger-codegen documentation generator was insufficient (at least until it catches up with Swagger 1.2). I wanted it to be easier to generate docs for the API anyways, so I ported the wiki pages to use the Asterisk Swagger generator. In the process, I was able to clean up many of the model links, which would occasionally give inconsistent results on the wiki. I also added error responses to the wiki docs, making the wiki documentation more complete. Finally, since Stasis-HTTP will now be named Asterisk REST Interface (ARI), any new functions and files I created carry the ari_ prefix. I changed a few stasis_http references to ari where it was non-intrusive and made sense. (closes issue ASTERISK-21885) Review: https://reviewboard.asterisk.org/r/2639/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@393529 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-07-03 16:32:41 +00:00
raise SwaggerError("Model id doesn't match name", context)
self.subtypes = model_json.get('subTypes') or []
if self.subtypes and context.version_less_than("1.2"):
raise SwaggerError("Type extension support added in Swagger 1.2",
Update events to use Swagger 1.3 subtyping, and related aftermath This patch started with the simple idea of changing the /events data model to be more sane. The original model would send out events like: { "stasis_start": { "args": [], "channel": { ... } } } The event discriminator was the field name instead of being a value in the object, due to limitations in how Swagger 1.1 could model objects. While technically sufficient in communicating event information, it was really difficult to deal with in terms of client side JSON handling. This patch takes advantage of a proposed extension[1] to Swagger which allows type variance through the use of a discriminator field. This had a domino effect that made this a surprisingly large patch. [1]: https://groups.google.com/d/msg/wordnik-api/EC3rGajE0os/ey_5dBI_jWcJ In changing the models, I also had to change the swagger_model.py processor so it can handle the type discriminator and subtyping. I took that a big step forward, and using that information to generate an ari_model module, which can validate a JSON object against the Swagger model. The REST and WebSocket generators were changed to take advantage of the validators. If compiled with AST_DEVMODE enabled, JSON objects that don't match their corresponding models will not be sent out. For REST API calls, a 500 Internal Server response is sent. For WebSockets, the invalid JSON message is replaced with an error message. Since this took over about half of the job of the existing JSON generators, and the .to_json virtual function on messages took over the other half, I reluctantly removed the generators. The validators turned up all sorts of errors and inconsistencies in our data models, and the code. These were cleaned up, with checks in the code generator avoid some of the consistency problems in the future. * The model for a channel snapshot was trimmed down to match the information sent via AMI. Many of the field being sent were not useful in the general case. * The model for a bridge snapshot was updated to be more consistent with the other ARI models. Another impact of introducing subtyping was that the swagger-codegen documentation generator was insufficient (at least until it catches up with Swagger 1.2). I wanted it to be easier to generate docs for the API anyways, so I ported the wiki pages to use the Asterisk Swagger generator. In the process, I was able to clean up many of the model links, which would occasionally give inconsistent results on the wiki. I also added error responses to the wiki docs, making the wiki documentation more complete. Finally, since Stasis-HTTP will now be named Asterisk REST Interface (ARI), any new functions and files I created carry the ari_ prefix. I changed a few stasis_http references to ari where it was non-intrusive and made sense. (closes issue ASTERISK-21885) Review: https://reviewboard.asterisk.org/r/2639/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@393529 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-07-03 16:32:41 +00:00
context)
self.description = model_json.get('description')
props = model_json.get('properties').items() or []
Update events to use Swagger 1.3 subtyping, and related aftermath This patch started with the simple idea of changing the /events data model to be more sane. The original model would send out events like: { "stasis_start": { "args": [], "channel": { ... } } } The event discriminator was the field name instead of being a value in the object, due to limitations in how Swagger 1.1 could model objects. While technically sufficient in communicating event information, it was really difficult to deal with in terms of client side JSON handling. This patch takes advantage of a proposed extension[1] to Swagger which allows type variance through the use of a discriminator field. This had a domino effect that made this a surprisingly large patch. [1]: https://groups.google.com/d/msg/wordnik-api/EC3rGajE0os/ey_5dBI_jWcJ In changing the models, I also had to change the swagger_model.py processor so it can handle the type discriminator and subtyping. I took that a big step forward, and using that information to generate an ari_model module, which can validate a JSON object against the Swagger model. The REST and WebSocket generators were changed to take advantage of the validators. If compiled with AST_DEVMODE enabled, JSON objects that don't match their corresponding models will not be sent out. For REST API calls, a 500 Internal Server response is sent. For WebSockets, the invalid JSON message is replaced with an error message. Since this took over about half of the job of the existing JSON generators, and the .to_json virtual function on messages took over the other half, I reluctantly removed the generators. The validators turned up all sorts of errors and inconsistencies in our data models, and the code. These were cleaned up, with checks in the code generator avoid some of the consistency problems in the future. * The model for a channel snapshot was trimmed down to match the information sent via AMI. Many of the field being sent were not useful in the general case. * The model for a bridge snapshot was updated to be more consistent with the other ARI models. Another impact of introducing subtyping was that the swagger-codegen documentation generator was insufficient (at least until it catches up with Swagger 1.2). I wanted it to be easier to generate docs for the API anyways, so I ported the wiki pages to use the Asterisk Swagger generator. In the process, I was able to clean up many of the model links, which would occasionally give inconsistent results on the wiki. I also added error responses to the wiki docs, making the wiki documentation more complete. Finally, since Stasis-HTTP will now be named Asterisk REST Interface (ARI), any new functions and files I created carry the ari_ prefix. I changed a few stasis_http references to ari where it was non-intrusive and made sense. (closes issue ASTERISK-21885) Review: https://reviewboard.asterisk.org/r/2639/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@393529 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-07-03 16:32:41 +00:00
self.__properties = [
Property(k).load(j, processor, context) for (k, j) in props]
Update events to use Swagger 1.3 subtyping, and related aftermath This patch started with the simple idea of changing the /events data model to be more sane. The original model would send out events like: { "stasis_start": { "args": [], "channel": { ... } } } The event discriminator was the field name instead of being a value in the object, due to limitations in how Swagger 1.1 could model objects. While technically sufficient in communicating event information, it was really difficult to deal with in terms of client side JSON handling. This patch takes advantage of a proposed extension[1] to Swagger which allows type variance through the use of a discriminator field. This had a domino effect that made this a surprisingly large patch. [1]: https://groups.google.com/d/msg/wordnik-api/EC3rGajE0os/ey_5dBI_jWcJ In changing the models, I also had to change the swagger_model.py processor so it can handle the type discriminator and subtyping. I took that a big step forward, and using that information to generate an ari_model module, which can validate a JSON object against the Swagger model. The REST and WebSocket generators were changed to take advantage of the validators. If compiled with AST_DEVMODE enabled, JSON objects that don't match their corresponding models will not be sent out. For REST API calls, a 500 Internal Server response is sent. For WebSockets, the invalid JSON message is replaced with an error message. Since this took over about half of the job of the existing JSON generators, and the .to_json virtual function on messages took over the other half, I reluctantly removed the generators. The validators turned up all sorts of errors and inconsistencies in our data models, and the code. These were cleaned up, with checks in the code generator avoid some of the consistency problems in the future. * The model for a channel snapshot was trimmed down to match the information sent via AMI. Many of the field being sent were not useful in the general case. * The model for a bridge snapshot was updated to be more consistent with the other ARI models. Another impact of introducing subtyping was that the swagger-codegen documentation generator was insufficient (at least until it catches up with Swagger 1.2). I wanted it to be easier to generate docs for the API anyways, so I ported the wiki pages to use the Asterisk Swagger generator. In the process, I was able to clean up many of the model links, which would occasionally give inconsistent results on the wiki. I also added error responses to the wiki docs, making the wiki documentation more complete. Finally, since Stasis-HTTP will now be named Asterisk REST Interface (ARI), any new functions and files I created carry the ari_ prefix. I changed a few stasis_http references to ari where it was non-intrusive and made sense. (closes issue ASTERISK-21885) Review: https://reviewboard.asterisk.org/r/2639/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@393529 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-07-03 16:32:41 +00:00
self.__properties = sorted(self.__properties, key=lambda p: p.name)
discriminator = model_json.get('discriminator')
if discriminator:
if context.version_less_than("1.2"):
raise SwaggerError("Discriminator support added in Swagger 1.2",
Update events to use Swagger 1.3 subtyping, and related aftermath This patch started with the simple idea of changing the /events data model to be more sane. The original model would send out events like: { "stasis_start": { "args": [], "channel": { ... } } } The event discriminator was the field name instead of being a value in the object, due to limitations in how Swagger 1.1 could model objects. While technically sufficient in communicating event information, it was really difficult to deal with in terms of client side JSON handling. This patch takes advantage of a proposed extension[1] to Swagger which allows type variance through the use of a discriminator field. This had a domino effect that made this a surprisingly large patch. [1]: https://groups.google.com/d/msg/wordnik-api/EC3rGajE0os/ey_5dBI_jWcJ In changing the models, I also had to change the swagger_model.py processor so it can handle the type discriminator and subtyping. I took that a big step forward, and using that information to generate an ari_model module, which can validate a JSON object against the Swagger model. The REST and WebSocket generators were changed to take advantage of the validators. If compiled with AST_DEVMODE enabled, JSON objects that don't match their corresponding models will not be sent out. For REST API calls, a 500 Internal Server response is sent. For WebSockets, the invalid JSON message is replaced with an error message. Since this took over about half of the job of the existing JSON generators, and the .to_json virtual function on messages took over the other half, I reluctantly removed the generators. The validators turned up all sorts of errors and inconsistencies in our data models, and the code. These were cleaned up, with checks in the code generator avoid some of the consistency problems in the future. * The model for a channel snapshot was trimmed down to match the information sent via AMI. Many of the field being sent were not useful in the general case. * The model for a bridge snapshot was updated to be more consistent with the other ARI models. Another impact of introducing subtyping was that the swagger-codegen documentation generator was insufficient (at least until it catches up with Swagger 1.2). I wanted it to be easier to generate docs for the API anyways, so I ported the wiki pages to use the Asterisk Swagger generator. In the process, I was able to clean up many of the model links, which would occasionally give inconsistent results on the wiki. I also added error responses to the wiki docs, making the wiki documentation more complete. Finally, since Stasis-HTTP will now be named Asterisk REST Interface (ARI), any new functions and files I created carry the ari_ prefix. I changed a few stasis_http references to ari where it was non-intrusive and made sense. (closes issue ASTERISK-21885) Review: https://reviewboard.asterisk.org/r/2639/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@393529 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-07-03 16:32:41 +00:00
context)
discr_props = [p for p in self.__properties if p.name == discriminator]
if not discr_props:
raise SwaggerError(
"Discriminator '%s' does not name a property of '%s'" % (
discriminator, self.id),
context)
self.__discriminator = discr_props[0]
self.model_json = json.dumps(model_json,
indent=2, separators=(',', ': '))
processor.process_model(self, context)
return self
def extends(self):
return self.__extends_type and self.__extends_type.id
Update events to use Swagger 1.3 subtyping, and related aftermath This patch started with the simple idea of changing the /events data model to be more sane. The original model would send out events like: { "stasis_start": { "args": [], "channel": { ... } } } The event discriminator was the field name instead of being a value in the object, due to limitations in how Swagger 1.1 could model objects. While technically sufficient in communicating event information, it was really difficult to deal with in terms of client side JSON handling. This patch takes advantage of a proposed extension[1] to Swagger which allows type variance through the use of a discriminator field. This had a domino effect that made this a surprisingly large patch. [1]: https://groups.google.com/d/msg/wordnik-api/EC3rGajE0os/ey_5dBI_jWcJ In changing the models, I also had to change the swagger_model.py processor so it can handle the type discriminator and subtyping. I took that a big step forward, and using that information to generate an ari_model module, which can validate a JSON object against the Swagger model. The REST and WebSocket generators were changed to take advantage of the validators. If compiled with AST_DEVMODE enabled, JSON objects that don't match their corresponding models will not be sent out. For REST API calls, a 500 Internal Server response is sent. For WebSockets, the invalid JSON message is replaced with an error message. Since this took over about half of the job of the existing JSON generators, and the .to_json virtual function on messages took over the other half, I reluctantly removed the generators. The validators turned up all sorts of errors and inconsistencies in our data models, and the code. These were cleaned up, with checks in the code generator avoid some of the consistency problems in the future. * The model for a channel snapshot was trimmed down to match the information sent via AMI. Many of the field being sent were not useful in the general case. * The model for a bridge snapshot was updated to be more consistent with the other ARI models. Another impact of introducing subtyping was that the swagger-codegen documentation generator was insufficient (at least until it catches up with Swagger 1.2). I wanted it to be easier to generate docs for the API anyways, so I ported the wiki pages to use the Asterisk Swagger generator. In the process, I was able to clean up many of the model links, which would occasionally give inconsistent results on the wiki. I also added error responses to the wiki docs, making the wiki documentation more complete. Finally, since Stasis-HTTP will now be named Asterisk REST Interface (ARI), any new functions and files I created carry the ari_ prefix. I changed a few stasis_http references to ari where it was non-intrusive and made sense. (closes issue ASTERISK-21885) Review: https://reviewboard.asterisk.org/r/2639/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@393529 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-07-03 16:32:41 +00:00
def set_extends_type(self, extends_type):
self.__extends_type = extends_type
def set_subtype_types(self, subtype_types):
self.__subtype_types = subtype_types
Update events to use Swagger 1.3 subtyping, and related aftermath This patch started with the simple idea of changing the /events data model to be more sane. The original model would send out events like: { "stasis_start": { "args": [], "channel": { ... } } } The event discriminator was the field name instead of being a value in the object, due to limitations in how Swagger 1.1 could model objects. While technically sufficient in communicating event information, it was really difficult to deal with in terms of client side JSON handling. This patch takes advantage of a proposed extension[1] to Swagger which allows type variance through the use of a discriminator field. This had a domino effect that made this a surprisingly large patch. [1]: https://groups.google.com/d/msg/wordnik-api/EC3rGajE0os/ey_5dBI_jWcJ In changing the models, I also had to change the swagger_model.py processor so it can handle the type discriminator and subtyping. I took that a big step forward, and using that information to generate an ari_model module, which can validate a JSON object against the Swagger model. The REST and WebSocket generators were changed to take advantage of the validators. If compiled with AST_DEVMODE enabled, JSON objects that don't match their corresponding models will not be sent out. For REST API calls, a 500 Internal Server response is sent. For WebSockets, the invalid JSON message is replaced with an error message. Since this took over about half of the job of the existing JSON generators, and the .to_json virtual function on messages took over the other half, I reluctantly removed the generators. The validators turned up all sorts of errors and inconsistencies in our data models, and the code. These were cleaned up, with checks in the code generator avoid some of the consistency problems in the future. * The model for a channel snapshot was trimmed down to match the information sent via AMI. Many of the field being sent were not useful in the general case. * The model for a bridge snapshot was updated to be more consistent with the other ARI models. Another impact of introducing subtyping was that the swagger-codegen documentation generator was insufficient (at least until it catches up with Swagger 1.2). I wanted it to be easier to generate docs for the API anyways, so I ported the wiki pages to use the Asterisk Swagger generator. In the process, I was able to clean up many of the model links, which would occasionally give inconsistent results on the wiki. I also added error responses to the wiki docs, making the wiki documentation more complete. Finally, since Stasis-HTTP will now be named Asterisk REST Interface (ARI), any new functions and files I created carry the ari_ prefix. I changed a few stasis_http references to ari where it was non-intrusive and made sense. (closes issue ASTERISK-21885) Review: https://reviewboard.asterisk.org/r/2639/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@393529 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-07-03 16:32:41 +00:00
def discriminator(self):
"""Returns the discriminator, digging through base types if needed.
"""
return self.__discriminator or \
self.__extends_type and self.__extends_type.discriminator()
Update events to use Swagger 1.3 subtyping, and related aftermath This patch started with the simple idea of changing the /events data model to be more sane. The original model would send out events like: { "stasis_start": { "args": [], "channel": { ... } } } The event discriminator was the field name instead of being a value in the object, due to limitations in how Swagger 1.1 could model objects. While technically sufficient in communicating event information, it was really difficult to deal with in terms of client side JSON handling. This patch takes advantage of a proposed extension[1] to Swagger which allows type variance through the use of a discriminator field. This had a domino effect that made this a surprisingly large patch. [1]: https://groups.google.com/d/msg/wordnik-api/EC3rGajE0os/ey_5dBI_jWcJ In changing the models, I also had to change the swagger_model.py processor so it can handle the type discriminator and subtyping. I took that a big step forward, and using that information to generate an ari_model module, which can validate a JSON object against the Swagger model. The REST and WebSocket generators were changed to take advantage of the validators. If compiled with AST_DEVMODE enabled, JSON objects that don't match their corresponding models will not be sent out. For REST API calls, a 500 Internal Server response is sent. For WebSockets, the invalid JSON message is replaced with an error message. Since this took over about half of the job of the existing JSON generators, and the .to_json virtual function on messages took over the other half, I reluctantly removed the generators. The validators turned up all sorts of errors and inconsistencies in our data models, and the code. These were cleaned up, with checks in the code generator avoid some of the consistency problems in the future. * The model for a channel snapshot was trimmed down to match the information sent via AMI. Many of the field being sent were not useful in the general case. * The model for a bridge snapshot was updated to be more consistent with the other ARI models. Another impact of introducing subtyping was that the swagger-codegen documentation generator was insufficient (at least until it catches up with Swagger 1.2). I wanted it to be easier to generate docs for the API anyways, so I ported the wiki pages to use the Asterisk Swagger generator. In the process, I was able to clean up many of the model links, which would occasionally give inconsistent results on the wiki. I also added error responses to the wiki docs, making the wiki documentation more complete. Finally, since Stasis-HTTP will now be named Asterisk REST Interface (ARI), any new functions and files I created carry the ari_ prefix. I changed a few stasis_http references to ari where it was non-intrusive and made sense. (closes issue ASTERISK-21885) Review: https://reviewboard.asterisk.org/r/2639/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@393529 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-07-03 16:32:41 +00:00
def properties(self):
base_props = []
if self.__extends_type:
base_props = self.__extends_type.properties()
Update events to use Swagger 1.3 subtyping, and related aftermath This patch started with the simple idea of changing the /events data model to be more sane. The original model would send out events like: { "stasis_start": { "args": [], "channel": { ... } } } The event discriminator was the field name instead of being a value in the object, due to limitations in how Swagger 1.1 could model objects. While technically sufficient in communicating event information, it was really difficult to deal with in terms of client side JSON handling. This patch takes advantage of a proposed extension[1] to Swagger which allows type variance through the use of a discriminator field. This had a domino effect that made this a surprisingly large patch. [1]: https://groups.google.com/d/msg/wordnik-api/EC3rGajE0os/ey_5dBI_jWcJ In changing the models, I also had to change the swagger_model.py processor so it can handle the type discriminator and subtyping. I took that a big step forward, and using that information to generate an ari_model module, which can validate a JSON object against the Swagger model. The REST and WebSocket generators were changed to take advantage of the validators. If compiled with AST_DEVMODE enabled, JSON objects that don't match their corresponding models will not be sent out. For REST API calls, a 500 Internal Server response is sent. For WebSockets, the invalid JSON message is replaced with an error message. Since this took over about half of the job of the existing JSON generators, and the .to_json virtual function on messages took over the other half, I reluctantly removed the generators. The validators turned up all sorts of errors and inconsistencies in our data models, and the code. These were cleaned up, with checks in the code generator avoid some of the consistency problems in the future. * The model for a channel snapshot was trimmed down to match the information sent via AMI. Many of the field being sent were not useful in the general case. * The model for a bridge snapshot was updated to be more consistent with the other ARI models. Another impact of introducing subtyping was that the swagger-codegen documentation generator was insufficient (at least until it catches up with Swagger 1.2). I wanted it to be easier to generate docs for the API anyways, so I ported the wiki pages to use the Asterisk Swagger generator. In the process, I was able to clean up many of the model links, which would occasionally give inconsistent results on the wiki. I also added error responses to the wiki docs, making the wiki documentation more complete. Finally, since Stasis-HTTP will now be named Asterisk REST Interface (ARI), any new functions and files I created carry the ari_ prefix. I changed a few stasis_http references to ari where it was non-intrusive and made sense. (closes issue ASTERISK-21885) Review: https://reviewboard.asterisk.org/r/2639/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@393529 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-07-03 16:32:41 +00:00
return base_props + self.__properties
def has_properties(self):
return len(self.properties()) > 0
def all_subtypes(self):
"""Returns the full list of all subtypes, including sub-subtypes.
Update events to use Swagger 1.3 subtyping, and related aftermath This patch started with the simple idea of changing the /events data model to be more sane. The original model would send out events like: { "stasis_start": { "args": [], "channel": { ... } } } The event discriminator was the field name instead of being a value in the object, due to limitations in how Swagger 1.1 could model objects. While technically sufficient in communicating event information, it was really difficult to deal with in terms of client side JSON handling. This patch takes advantage of a proposed extension[1] to Swagger which allows type variance through the use of a discriminator field. This had a domino effect that made this a surprisingly large patch. [1]: https://groups.google.com/d/msg/wordnik-api/EC3rGajE0os/ey_5dBI_jWcJ In changing the models, I also had to change the swagger_model.py processor so it can handle the type discriminator and subtyping. I took that a big step forward, and using that information to generate an ari_model module, which can validate a JSON object against the Swagger model. The REST and WebSocket generators were changed to take advantage of the validators. If compiled with AST_DEVMODE enabled, JSON objects that don't match their corresponding models will not be sent out. For REST API calls, a 500 Internal Server response is sent. For WebSockets, the invalid JSON message is replaced with an error message. Since this took over about half of the job of the existing JSON generators, and the .to_json virtual function on messages took over the other half, I reluctantly removed the generators. The validators turned up all sorts of errors and inconsistencies in our data models, and the code. These were cleaned up, with checks in the code generator avoid some of the consistency problems in the future. * The model for a channel snapshot was trimmed down to match the information sent via AMI. Many of the field being sent were not useful in the general case. * The model for a bridge snapshot was updated to be more consistent with the other ARI models. Another impact of introducing subtyping was that the swagger-codegen documentation generator was insufficient (at least until it catches up with Swagger 1.2). I wanted it to be easier to generate docs for the API anyways, so I ported the wiki pages to use the Asterisk Swagger generator. In the process, I was able to clean up many of the model links, which would occasionally give inconsistent results on the wiki. I also added error responses to the wiki docs, making the wiki documentation more complete. Finally, since Stasis-HTTP will now be named Asterisk REST Interface (ARI), any new functions and files I created carry the ari_ prefix. I changed a few stasis_http references to ari where it was non-intrusive and made sense. (closes issue ASTERISK-21885) Review: https://reviewboard.asterisk.org/r/2639/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@393529 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-07-03 16:32:41 +00:00
"""
res = self.__subtype_types + \
[subsubtypes for subtype in self.__subtype_types
for subsubtypes in subtype.all_subtypes()]
Update events to use Swagger 1.3 subtyping, and related aftermath This patch started with the simple idea of changing the /events data model to be more sane. The original model would send out events like: { "stasis_start": { "args": [], "channel": { ... } } } The event discriminator was the field name instead of being a value in the object, due to limitations in how Swagger 1.1 could model objects. While technically sufficient in communicating event information, it was really difficult to deal with in terms of client side JSON handling. This patch takes advantage of a proposed extension[1] to Swagger which allows type variance through the use of a discriminator field. This had a domino effect that made this a surprisingly large patch. [1]: https://groups.google.com/d/msg/wordnik-api/EC3rGajE0os/ey_5dBI_jWcJ In changing the models, I also had to change the swagger_model.py processor so it can handle the type discriminator and subtyping. I took that a big step forward, and using that information to generate an ari_model module, which can validate a JSON object against the Swagger model. The REST and WebSocket generators were changed to take advantage of the validators. If compiled with AST_DEVMODE enabled, JSON objects that don't match their corresponding models will not be sent out. For REST API calls, a 500 Internal Server response is sent. For WebSockets, the invalid JSON message is replaced with an error message. Since this took over about half of the job of the existing JSON generators, and the .to_json virtual function on messages took over the other half, I reluctantly removed the generators. The validators turned up all sorts of errors and inconsistencies in our data models, and the code. These were cleaned up, with checks in the code generator avoid some of the consistency problems in the future. * The model for a channel snapshot was trimmed down to match the information sent via AMI. Many of the field being sent were not useful in the general case. * The model for a bridge snapshot was updated to be more consistent with the other ARI models. Another impact of introducing subtyping was that the swagger-codegen documentation generator was insufficient (at least until it catches up with Swagger 1.2). I wanted it to be easier to generate docs for the API anyways, so I ported the wiki pages to use the Asterisk Swagger generator. In the process, I was able to clean up many of the model links, which would occasionally give inconsistent results on the wiki. I also added error responses to the wiki docs, making the wiki documentation more complete. Finally, since Stasis-HTTP will now be named Asterisk REST Interface (ARI), any new functions and files I created carry the ari_ prefix. I changed a few stasis_http references to ari where it was non-intrusive and made sense. (closes issue ASTERISK-21885) Review: https://reviewboard.asterisk.org/r/2639/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@393529 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-07-03 16:32:41 +00:00
return sorted(res, key=lambda m: m.id)
def has_subtypes(self):
"""Returns True if type has any subtypes.
"""
return len(self.subtypes) > 0
Update events to use Swagger 1.3 subtyping, and related aftermath This patch started with the simple idea of changing the /events data model to be more sane. The original model would send out events like: { "stasis_start": { "args": [], "channel": { ... } } } The event discriminator was the field name instead of being a value in the object, due to limitations in how Swagger 1.1 could model objects. While technically sufficient in communicating event information, it was really difficult to deal with in terms of client side JSON handling. This patch takes advantage of a proposed extension[1] to Swagger which allows type variance through the use of a discriminator field. This had a domino effect that made this a surprisingly large patch. [1]: https://groups.google.com/d/msg/wordnik-api/EC3rGajE0os/ey_5dBI_jWcJ In changing the models, I also had to change the swagger_model.py processor so it can handle the type discriminator and subtyping. I took that a big step forward, and using that information to generate an ari_model module, which can validate a JSON object against the Swagger model. The REST and WebSocket generators were changed to take advantage of the validators. If compiled with AST_DEVMODE enabled, JSON objects that don't match their corresponding models will not be sent out. For REST API calls, a 500 Internal Server response is sent. For WebSockets, the invalid JSON message is replaced with an error message. Since this took over about half of the job of the existing JSON generators, and the .to_json virtual function on messages took over the other half, I reluctantly removed the generators. The validators turned up all sorts of errors and inconsistencies in our data models, and the code. These were cleaned up, with checks in the code generator avoid some of the consistency problems in the future. * The model for a channel snapshot was trimmed down to match the information sent via AMI. Many of the field being sent were not useful in the general case. * The model for a bridge snapshot was updated to be more consistent with the other ARI models. Another impact of introducing subtyping was that the swagger-codegen documentation generator was insufficient (at least until it catches up with Swagger 1.2). I wanted it to be easier to generate docs for the API anyways, so I ported the wiki pages to use the Asterisk Swagger generator. In the process, I was able to clean up many of the model links, which would occasionally give inconsistent results on the wiki. I also added error responses to the wiki docs, making the wiki documentation more complete. Finally, since Stasis-HTTP will now be named Asterisk REST Interface (ARI), any new functions and files I created carry the ari_ prefix. I changed a few stasis_http references to ari where it was non-intrusive and made sense. (closes issue ASTERISK-21885) Review: https://reviewboard.asterisk.org/r/2639/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@393529 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-07-03 16:32:41 +00:00
class ApiDeclaration(Stringify):
"""Model class for an API Declaration.
See https://github.com/wordnik/swagger-core/wiki/API-Declaration
"""
required_fields = [
'swaggerVersion', '_author', '_copyright', 'apiVersion', 'basePath',
'resourcePath', 'apis', 'models'
]
def __init__(self):
self.swagger_version = None
self.author = None
self.copyright = None
self.api_version = None
self.base_path = None
self.resource_path = None
self.apis = []
self.models = []
Update events to use Swagger 1.3 subtyping, and related aftermath This patch started with the simple idea of changing the /events data model to be more sane. The original model would send out events like: { "stasis_start": { "args": [], "channel": { ... } } } The event discriminator was the field name instead of being a value in the object, due to limitations in how Swagger 1.1 could model objects. While technically sufficient in communicating event information, it was really difficult to deal with in terms of client side JSON handling. This patch takes advantage of a proposed extension[1] to Swagger which allows type variance through the use of a discriminator field. This had a domino effect that made this a surprisingly large patch. [1]: https://groups.google.com/d/msg/wordnik-api/EC3rGajE0os/ey_5dBI_jWcJ In changing the models, I also had to change the swagger_model.py processor so it can handle the type discriminator and subtyping. I took that a big step forward, and using that information to generate an ari_model module, which can validate a JSON object against the Swagger model. The REST and WebSocket generators were changed to take advantage of the validators. If compiled with AST_DEVMODE enabled, JSON objects that don't match their corresponding models will not be sent out. For REST API calls, a 500 Internal Server response is sent. For WebSockets, the invalid JSON message is replaced with an error message. Since this took over about half of the job of the existing JSON generators, and the .to_json virtual function on messages took over the other half, I reluctantly removed the generators. The validators turned up all sorts of errors and inconsistencies in our data models, and the code. These were cleaned up, with checks in the code generator avoid some of the consistency problems in the future. * The model for a channel snapshot was trimmed down to match the information sent via AMI. Many of the field being sent were not useful in the general case. * The model for a bridge snapshot was updated to be more consistent with the other ARI models. Another impact of introducing subtyping was that the swagger-codegen documentation generator was insufficient (at least until it catches up with Swagger 1.2). I wanted it to be easier to generate docs for the API anyways, so I ported the wiki pages to use the Asterisk Swagger generator. In the process, I was able to clean up many of the model links, which would occasionally give inconsistent results on the wiki. I also added error responses to the wiki docs, making the wiki documentation more complete. Finally, since Stasis-HTTP will now be named Asterisk REST Interface (ARI), any new functions and files I created carry the ari_ prefix. I changed a few stasis_http references to ari where it was non-intrusive and made sense. (closes issue ASTERISK-21885) Review: https://reviewboard.asterisk.org/r/2639/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@393529 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-07-03 16:32:41 +00:00
def load_file(self, api_declaration_file, processor):
context = ParsingContext(None, [api_declaration_file])
try:
return self.__load_file(api_declaration_file, processor, context)
except SwaggerError:
raise
except Exception as e:
print >> sys.stderr, "Error: ", traceback.format_exc()
raise SwaggerError(
"Error loading %s" % api_declaration_file, context, e)
def __load_file(self, api_declaration_file, processor, context):
with open(api_declaration_file) as fp:
self.load(json.load(fp), processor, context)
expected_resource_path = '/api-docs/' + \
os.path.basename(api_declaration_file) \
.replace(".json", ".{format}")
if self.resource_path != expected_resource_path:
print >> sys.stderr, \
"%s != %s" % (self.resource_path, expected_resource_path)
raise SwaggerError("resourcePath has incorrect value", context)
return self
def load(self, api_decl_json, processor, context):
"""Loads a resource from a single Swagger resource.json file.
"""
# If the version doesn't match, all bets are off.
self.swagger_version = api_decl_json.get('swaggerVersion')
Update events to use Swagger 1.3 subtyping, and related aftermath This patch started with the simple idea of changing the /events data model to be more sane. The original model would send out events like: { "stasis_start": { "args": [], "channel": { ... } } } The event discriminator was the field name instead of being a value in the object, due to limitations in how Swagger 1.1 could model objects. While technically sufficient in communicating event information, it was really difficult to deal with in terms of client side JSON handling. This patch takes advantage of a proposed extension[1] to Swagger which allows type variance through the use of a discriminator field. This had a domino effect that made this a surprisingly large patch. [1]: https://groups.google.com/d/msg/wordnik-api/EC3rGajE0os/ey_5dBI_jWcJ In changing the models, I also had to change the swagger_model.py processor so it can handle the type discriminator and subtyping. I took that a big step forward, and using that information to generate an ari_model module, which can validate a JSON object against the Swagger model. The REST and WebSocket generators were changed to take advantage of the validators. If compiled with AST_DEVMODE enabled, JSON objects that don't match their corresponding models will not be sent out. For REST API calls, a 500 Internal Server response is sent. For WebSockets, the invalid JSON message is replaced with an error message. Since this took over about half of the job of the existing JSON generators, and the .to_json virtual function on messages took over the other half, I reluctantly removed the generators. The validators turned up all sorts of errors and inconsistencies in our data models, and the code. These were cleaned up, with checks in the code generator avoid some of the consistency problems in the future. * The model for a channel snapshot was trimmed down to match the information sent via AMI. Many of the field being sent were not useful in the general case. * The model for a bridge snapshot was updated to be more consistent with the other ARI models. Another impact of introducing subtyping was that the swagger-codegen documentation generator was insufficient (at least until it catches up with Swagger 1.2). I wanted it to be easier to generate docs for the API anyways, so I ported the wiki pages to use the Asterisk Swagger generator. In the process, I was able to clean up many of the model links, which would occasionally give inconsistent results on the wiki. I also added error responses to the wiki docs, making the wiki documentation more complete. Finally, since Stasis-HTTP will now be named Asterisk REST Interface (ARI), any new functions and files I created carry the ari_ prefix. I changed a few stasis_http references to ari where it was non-intrusive and made sense. (closes issue ASTERISK-21885) Review: https://reviewboard.asterisk.org/r/2639/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@393529 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-07-03 16:32:41 +00:00
context = context.next(version=self.swagger_version)
if not self.swagger_version in SWAGGER_VERSIONS:
raise SwaggerError(
Update events to use Swagger 1.3 subtyping, and related aftermath This patch started with the simple idea of changing the /events data model to be more sane. The original model would send out events like: { "stasis_start": { "args": [], "channel": { ... } } } The event discriminator was the field name instead of being a value in the object, due to limitations in how Swagger 1.1 could model objects. While technically sufficient in communicating event information, it was really difficult to deal with in terms of client side JSON handling. This patch takes advantage of a proposed extension[1] to Swagger which allows type variance through the use of a discriminator field. This had a domino effect that made this a surprisingly large patch. [1]: https://groups.google.com/d/msg/wordnik-api/EC3rGajE0os/ey_5dBI_jWcJ In changing the models, I also had to change the swagger_model.py processor so it can handle the type discriminator and subtyping. I took that a big step forward, and using that information to generate an ari_model module, which can validate a JSON object against the Swagger model. The REST and WebSocket generators were changed to take advantage of the validators. If compiled with AST_DEVMODE enabled, JSON objects that don't match their corresponding models will not be sent out. For REST API calls, a 500 Internal Server response is sent. For WebSockets, the invalid JSON message is replaced with an error message. Since this took over about half of the job of the existing JSON generators, and the .to_json virtual function on messages took over the other half, I reluctantly removed the generators. The validators turned up all sorts of errors and inconsistencies in our data models, and the code. These were cleaned up, with checks in the code generator avoid some of the consistency problems in the future. * The model for a channel snapshot was trimmed down to match the information sent via AMI. Many of the field being sent were not useful in the general case. * The model for a bridge snapshot was updated to be more consistent with the other ARI models. Another impact of introducing subtyping was that the swagger-codegen documentation generator was insufficient (at least until it catches up with Swagger 1.2). I wanted it to be easier to generate docs for the API anyways, so I ported the wiki pages to use the Asterisk Swagger generator. In the process, I was able to clean up many of the model links, which would occasionally give inconsistent results on the wiki. I also added error responses to the wiki docs, making the wiki documentation more complete. Finally, since Stasis-HTTP will now be named Asterisk REST Interface (ARI), any new functions and files I created carry the ari_ prefix. I changed a few stasis_http references to ari where it was non-intrusive and made sense. (closes issue ASTERISK-21885) Review: https://reviewboard.asterisk.org/r/2639/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@393529 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-07-03 16:32:41 +00:00
"Unsupported Swagger version %s" % self.swagger_version, context)
validate_required_fields(api_decl_json, self.required_fields, context)
self.author = api_decl_json.get('_author')
self.copyright = api_decl_json.get('_copyright')
self.api_version = api_decl_json.get('apiVersion')
self.base_path = api_decl_json.get('basePath')
self.resource_path = api_decl_json.get('resourcePath')
api_json = api_decl_json.get('apis') or []
self.apis = [
Api().load(j, processor, context) for j in api_json]
paths = set()
for api in self.apis:
if api.path in paths:
raise SwaggerError("API with duplicated path: %s" % api.path, context)
paths.add(api.path)
optional_api: Fix linking problems between modules that export global symbols With the new work in Asterisk 12, there are some uses of the optional_api that are prone to failure. The details are rather involved, and captured on [the wiki][1]. This patch addresses the issue by removing almost all of the magic from the optional API implementation. Instead of relying on weak symbol resolution, a new optional_api.c module was added to Asterisk core. For modules providing an optional API, the pointer to the implementation function is registered with the core. For modules that use an optional API, a pointer to a stub function, along with a optional_ref function pointer are registered with the core. The optional_ref function pointers is set to the implementation function when it's provided, or the stub function when it's now. Since the implementation no longer relies on magic, it is now supported on all platforms. In the spirit of choice, an OPTIONAL_API flag was added, so we can disable the optional_api if needed (maybe it's buggy on some bizarre platform I haven't tested on) The AST_OPTIONAL_API*() macros themselves remained unchanged, so existing code could remain unchanged. But to help with debugging the optional_api, the patch limits the #include of optional API's to just the modules using the API. This also reduces resource waste maintaining optional_ref pointers that aren't used. Other changes made as a part of this patch: * The stubs for http_websocket that wrap system calls set errno to ENOSYS. * res_http_websocket now properly increments module use count. * In loader.c, the while() wrappers around dlclose() were removed. The while(!dlclose()) is actually an anti-pattern, which can lead to infinite loops if the module you're attempting to unload exports a symbol that was directly linked to. * The special handling of nonoptreq on systems without weak symbol support was removed, since we no longer rely on weak symbols for optional_api. [1]: https://wiki.asterisk.org/wiki/x/wACUAQ (closes issue ASTERISK-22296) Reported by: Matt Jordan Review: https://reviewboard.asterisk.org/r/2797/ ........ Merged revisions 397989 from http://svn.asterisk.org/svn/asterisk/branches/12 git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@397990 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-08-30 13:40:27 +00:00
self.has_websocket = filter(lambda api: api.has_websocket,
self.apis) == []
models = api_decl_json.get('models').items() or []
Update events to use Swagger 1.3 subtyping, and related aftermath This patch started with the simple idea of changing the /events data model to be more sane. The original model would send out events like: { "stasis_start": { "args": [], "channel": { ... } } } The event discriminator was the field name instead of being a value in the object, due to limitations in how Swagger 1.1 could model objects. While technically sufficient in communicating event information, it was really difficult to deal with in terms of client side JSON handling. This patch takes advantage of a proposed extension[1] to Swagger which allows type variance through the use of a discriminator field. This had a domino effect that made this a surprisingly large patch. [1]: https://groups.google.com/d/msg/wordnik-api/EC3rGajE0os/ey_5dBI_jWcJ In changing the models, I also had to change the swagger_model.py processor so it can handle the type discriminator and subtyping. I took that a big step forward, and using that information to generate an ari_model module, which can validate a JSON object against the Swagger model. The REST and WebSocket generators were changed to take advantage of the validators. If compiled with AST_DEVMODE enabled, JSON objects that don't match their corresponding models will not be sent out. For REST API calls, a 500 Internal Server response is sent. For WebSockets, the invalid JSON message is replaced with an error message. Since this took over about half of the job of the existing JSON generators, and the .to_json virtual function on messages took over the other half, I reluctantly removed the generators. The validators turned up all sorts of errors and inconsistencies in our data models, and the code. These were cleaned up, with checks in the code generator avoid some of the consistency problems in the future. * The model for a channel snapshot was trimmed down to match the information sent via AMI. Many of the field being sent were not useful in the general case. * The model for a bridge snapshot was updated to be more consistent with the other ARI models. Another impact of introducing subtyping was that the swagger-codegen documentation generator was insufficient (at least until it catches up with Swagger 1.2). I wanted it to be easier to generate docs for the API anyways, so I ported the wiki pages to use the Asterisk Swagger generator. In the process, I was able to clean up many of the model links, which would occasionally give inconsistent results on the wiki. I also added error responses to the wiki docs, making the wiki documentation more complete. Finally, since Stasis-HTTP will now be named Asterisk REST Interface (ARI), any new functions and files I created carry the ari_ prefix. I changed a few stasis_http references to ari where it was non-intrusive and made sense. (closes issue ASTERISK-21885) Review: https://reviewboard.asterisk.org/r/2639/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@393529 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-07-03 16:32:41 +00:00
self.models = [Model().load(id, json, processor, context)
for (id, json) in models]
self.models = sorted(self.models, key=lambda m: m.id)
# Now link all base/extended types
model_dict = dict((m.id, m) for m in self.models)
for m in self.models:
def link_subtype(name):
res = model_dict.get(subtype)
if not res:
raise SwaggerError("%s has non-existing subtype %s",
m.id, name)
res.set_extends_type(m)
return res;
if m.subtypes:
m.set_subtype_types([
link_subtype(subtype) for subtype in m.subtypes])
return self
class ResourceApi(Stringify):
"""Model of an API listing in the resources.json file.
"""
required_fields = ['path', 'description']
def __init__(self):
self.path = None
self.description = None
self.api_declaration = None
def load(self, api_json, processor, context):
Update events to use Swagger 1.3 subtyping, and related aftermath This patch started with the simple idea of changing the /events data model to be more sane. The original model would send out events like: { "stasis_start": { "args": [], "channel": { ... } } } The event discriminator was the field name instead of being a value in the object, due to limitations in how Swagger 1.1 could model objects. While technically sufficient in communicating event information, it was really difficult to deal with in terms of client side JSON handling. This patch takes advantage of a proposed extension[1] to Swagger which allows type variance through the use of a discriminator field. This had a domino effect that made this a surprisingly large patch. [1]: https://groups.google.com/d/msg/wordnik-api/EC3rGajE0os/ey_5dBI_jWcJ In changing the models, I also had to change the swagger_model.py processor so it can handle the type discriminator and subtyping. I took that a big step forward, and using that information to generate an ari_model module, which can validate a JSON object against the Swagger model. The REST and WebSocket generators were changed to take advantage of the validators. If compiled with AST_DEVMODE enabled, JSON objects that don't match their corresponding models will not be sent out. For REST API calls, a 500 Internal Server response is sent. For WebSockets, the invalid JSON message is replaced with an error message. Since this took over about half of the job of the existing JSON generators, and the .to_json virtual function on messages took over the other half, I reluctantly removed the generators. The validators turned up all sorts of errors and inconsistencies in our data models, and the code. These were cleaned up, with checks in the code generator avoid some of the consistency problems in the future. * The model for a channel snapshot was trimmed down to match the information sent via AMI. Many of the field being sent were not useful in the general case. * The model for a bridge snapshot was updated to be more consistent with the other ARI models. Another impact of introducing subtyping was that the swagger-codegen documentation generator was insufficient (at least until it catches up with Swagger 1.2). I wanted it to be easier to generate docs for the API anyways, so I ported the wiki pages to use the Asterisk Swagger generator. In the process, I was able to clean up many of the model links, which would occasionally give inconsistent results on the wiki. I also added error responses to the wiki docs, making the wiki documentation more complete. Finally, since Stasis-HTTP will now be named Asterisk REST Interface (ARI), any new functions and files I created carry the ari_ prefix. I changed a few stasis_http references to ari where it was non-intrusive and made sense. (closes issue ASTERISK-21885) Review: https://reviewboard.asterisk.org/r/2639/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@393529 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-07-03 16:32:41 +00:00
context = context.next_stack(api_json, 'path')
validate_required_fields(api_json, self.required_fields, context)
self.path = api_json['path']
self.description = api_json['description']
if not self.path or self.path[0] != '/':
raise SwaggerError("Path must start with /", context)
Update events to use Swagger 1.3 subtyping, and related aftermath This patch started with the simple idea of changing the /events data model to be more sane. The original model would send out events like: { "stasis_start": { "args": [], "channel": { ... } } } The event discriminator was the field name instead of being a value in the object, due to limitations in how Swagger 1.1 could model objects. While technically sufficient in communicating event information, it was really difficult to deal with in terms of client side JSON handling. This patch takes advantage of a proposed extension[1] to Swagger which allows type variance through the use of a discriminator field. This had a domino effect that made this a surprisingly large patch. [1]: https://groups.google.com/d/msg/wordnik-api/EC3rGajE0os/ey_5dBI_jWcJ In changing the models, I also had to change the swagger_model.py processor so it can handle the type discriminator and subtyping. I took that a big step forward, and using that information to generate an ari_model module, which can validate a JSON object against the Swagger model. The REST and WebSocket generators were changed to take advantage of the validators. If compiled with AST_DEVMODE enabled, JSON objects that don't match their corresponding models will not be sent out. For REST API calls, a 500 Internal Server response is sent. For WebSockets, the invalid JSON message is replaced with an error message. Since this took over about half of the job of the existing JSON generators, and the .to_json virtual function on messages took over the other half, I reluctantly removed the generators. The validators turned up all sorts of errors and inconsistencies in our data models, and the code. These were cleaned up, with checks in the code generator avoid some of the consistency problems in the future. * The model for a channel snapshot was trimmed down to match the information sent via AMI. Many of the field being sent were not useful in the general case. * The model for a bridge snapshot was updated to be more consistent with the other ARI models. Another impact of introducing subtyping was that the swagger-codegen documentation generator was insufficient (at least until it catches up with Swagger 1.2). I wanted it to be easier to generate docs for the API anyways, so I ported the wiki pages to use the Asterisk Swagger generator. In the process, I was able to clean up many of the model links, which would occasionally give inconsistent results on the wiki. I also added error responses to the wiki docs, making the wiki documentation more complete. Finally, since Stasis-HTTP will now be named Asterisk REST Interface (ARI), any new functions and files I created carry the ari_ prefix. I changed a few stasis_http references to ari where it was non-intrusive and made sense. (closes issue ASTERISK-21885) Review: https://reviewboard.asterisk.org/r/2639/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@393529 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-07-03 16:32:41 +00:00
processor.process_resource_api(self, context)
return self
def load_api_declaration(self, base_dir, processor):
self.file = (base_dir + self.path).replace('{format}', 'json')
self.api_declaration = ApiDeclaration().load_file(self.file, processor)
Update events to use Swagger 1.3 subtyping, and related aftermath This patch started with the simple idea of changing the /events data model to be more sane. The original model would send out events like: { "stasis_start": { "args": [], "channel": { ... } } } The event discriminator was the field name instead of being a value in the object, due to limitations in how Swagger 1.1 could model objects. While technically sufficient in communicating event information, it was really difficult to deal with in terms of client side JSON handling. This patch takes advantage of a proposed extension[1] to Swagger which allows type variance through the use of a discriminator field. This had a domino effect that made this a surprisingly large patch. [1]: https://groups.google.com/d/msg/wordnik-api/EC3rGajE0os/ey_5dBI_jWcJ In changing the models, I also had to change the swagger_model.py processor so it can handle the type discriminator and subtyping. I took that a big step forward, and using that information to generate an ari_model module, which can validate a JSON object against the Swagger model. The REST and WebSocket generators were changed to take advantage of the validators. If compiled with AST_DEVMODE enabled, JSON objects that don't match their corresponding models will not be sent out. For REST API calls, a 500 Internal Server response is sent. For WebSockets, the invalid JSON message is replaced with an error message. Since this took over about half of the job of the existing JSON generators, and the .to_json virtual function on messages took over the other half, I reluctantly removed the generators. The validators turned up all sorts of errors and inconsistencies in our data models, and the code. These were cleaned up, with checks in the code generator avoid some of the consistency problems in the future. * The model for a channel snapshot was trimmed down to match the information sent via AMI. Many of the field being sent were not useful in the general case. * The model for a bridge snapshot was updated to be more consistent with the other ARI models. Another impact of introducing subtyping was that the swagger-codegen documentation generator was insufficient (at least until it catches up with Swagger 1.2). I wanted it to be easier to generate docs for the API anyways, so I ported the wiki pages to use the Asterisk Swagger generator. In the process, I was able to clean up many of the model links, which would occasionally give inconsistent results on the wiki. I also added error responses to the wiki docs, making the wiki documentation more complete. Finally, since Stasis-HTTP will now be named Asterisk REST Interface (ARI), any new functions and files I created carry the ari_ prefix. I changed a few stasis_http references to ari where it was non-intrusive and made sense. (closes issue ASTERISK-21885) Review: https://reviewboard.asterisk.org/r/2639/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@393529 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-07-03 16:32:41 +00:00
processor.process_resource_api(self, [self.file])
class ResourceListing(Stringify):
"""Model of Swagger's resources.json file.
"""
required_fields = ['apiVersion', 'basePath', 'apis']
def __init__(self):
self.swagger_version = None
self.api_version = None
self.base_path = None
self.apis = None
def load_file(self, resource_file, processor):
Update events to use Swagger 1.3 subtyping, and related aftermath This patch started with the simple idea of changing the /events data model to be more sane. The original model would send out events like: { "stasis_start": { "args": [], "channel": { ... } } } The event discriminator was the field name instead of being a value in the object, due to limitations in how Swagger 1.1 could model objects. While technically sufficient in communicating event information, it was really difficult to deal with in terms of client side JSON handling. This patch takes advantage of a proposed extension[1] to Swagger which allows type variance through the use of a discriminator field. This had a domino effect that made this a surprisingly large patch. [1]: https://groups.google.com/d/msg/wordnik-api/EC3rGajE0os/ey_5dBI_jWcJ In changing the models, I also had to change the swagger_model.py processor so it can handle the type discriminator and subtyping. I took that a big step forward, and using that information to generate an ari_model module, which can validate a JSON object against the Swagger model. The REST and WebSocket generators were changed to take advantage of the validators. If compiled with AST_DEVMODE enabled, JSON objects that don't match their corresponding models will not be sent out. For REST API calls, a 500 Internal Server response is sent. For WebSockets, the invalid JSON message is replaced with an error message. Since this took over about half of the job of the existing JSON generators, and the .to_json virtual function on messages took over the other half, I reluctantly removed the generators. The validators turned up all sorts of errors and inconsistencies in our data models, and the code. These were cleaned up, with checks in the code generator avoid some of the consistency problems in the future. * The model for a channel snapshot was trimmed down to match the information sent via AMI. Many of the field being sent were not useful in the general case. * The model for a bridge snapshot was updated to be more consistent with the other ARI models. Another impact of introducing subtyping was that the swagger-codegen documentation generator was insufficient (at least until it catches up with Swagger 1.2). I wanted it to be easier to generate docs for the API anyways, so I ported the wiki pages to use the Asterisk Swagger generator. In the process, I was able to clean up many of the model links, which would occasionally give inconsistent results on the wiki. I also added error responses to the wiki docs, making the wiki documentation more complete. Finally, since Stasis-HTTP will now be named Asterisk REST Interface (ARI), any new functions and files I created carry the ari_ prefix. I changed a few stasis_http references to ari where it was non-intrusive and made sense. (closes issue ASTERISK-21885) Review: https://reviewboard.asterisk.org/r/2639/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@393529 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-07-03 16:32:41 +00:00
context = ParsingContext(None, [resource_file])
try:
return self.__load_file(resource_file, processor, context)
except SwaggerError:
raise
except Exception as e:
print >> sys.stderr, "Error: ", traceback.format_exc()
raise SwaggerError(
"Error loading %s" % resource_file, context, e)
def __load_file(self, resource_file, processor, context):
with open(resource_file) as fp:
return self.load(json.load(fp), processor, context)
def load(self, resources_json, processor, context):
# If the version doesn't match, all bets are off.
self.swagger_version = resources_json.get('swaggerVersion')
Update events to use Swagger 1.3 subtyping, and related aftermath This patch started with the simple idea of changing the /events data model to be more sane. The original model would send out events like: { "stasis_start": { "args": [], "channel": { ... } } } The event discriminator was the field name instead of being a value in the object, due to limitations in how Swagger 1.1 could model objects. While technically sufficient in communicating event information, it was really difficult to deal with in terms of client side JSON handling. This patch takes advantage of a proposed extension[1] to Swagger which allows type variance through the use of a discriminator field. This had a domino effect that made this a surprisingly large patch. [1]: https://groups.google.com/d/msg/wordnik-api/EC3rGajE0os/ey_5dBI_jWcJ In changing the models, I also had to change the swagger_model.py processor so it can handle the type discriminator and subtyping. I took that a big step forward, and using that information to generate an ari_model module, which can validate a JSON object against the Swagger model. The REST and WebSocket generators were changed to take advantage of the validators. If compiled with AST_DEVMODE enabled, JSON objects that don't match their corresponding models will not be sent out. For REST API calls, a 500 Internal Server response is sent. For WebSockets, the invalid JSON message is replaced with an error message. Since this took over about half of the job of the existing JSON generators, and the .to_json virtual function on messages took over the other half, I reluctantly removed the generators. The validators turned up all sorts of errors and inconsistencies in our data models, and the code. These were cleaned up, with checks in the code generator avoid some of the consistency problems in the future. * The model for a channel snapshot was trimmed down to match the information sent via AMI. Many of the field being sent were not useful in the general case. * The model for a bridge snapshot was updated to be more consistent with the other ARI models. Another impact of introducing subtyping was that the swagger-codegen documentation generator was insufficient (at least until it catches up with Swagger 1.2). I wanted it to be easier to generate docs for the API anyways, so I ported the wiki pages to use the Asterisk Swagger generator. In the process, I was able to clean up many of the model links, which would occasionally give inconsistent results on the wiki. I also added error responses to the wiki docs, making the wiki documentation more complete. Finally, since Stasis-HTTP will now be named Asterisk REST Interface (ARI), any new functions and files I created carry the ari_ prefix. I changed a few stasis_http references to ari where it was non-intrusive and made sense. (closes issue ASTERISK-21885) Review: https://reviewboard.asterisk.org/r/2639/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@393529 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-07-03 16:32:41 +00:00
if not self.swagger_version in SWAGGER_VERSIONS:
raise SwaggerError(
"Unsupported Swagger version %s" % swagger_version, context)
validate_required_fields(resources_json, self.required_fields, context)
self.api_version = resources_json['apiVersion']
self.base_path = resources_json['basePath']
apis_json = resources_json['apis']
self.apis = [
ResourceApi().load(j, processor, context) for j in apis_json]
Update events to use Swagger 1.3 subtyping, and related aftermath This patch started with the simple idea of changing the /events data model to be more sane. The original model would send out events like: { "stasis_start": { "args": [], "channel": { ... } } } The event discriminator was the field name instead of being a value in the object, due to limitations in how Swagger 1.1 could model objects. While technically sufficient in communicating event information, it was really difficult to deal with in terms of client side JSON handling. This patch takes advantage of a proposed extension[1] to Swagger which allows type variance through the use of a discriminator field. This had a domino effect that made this a surprisingly large patch. [1]: https://groups.google.com/d/msg/wordnik-api/EC3rGajE0os/ey_5dBI_jWcJ In changing the models, I also had to change the swagger_model.py processor so it can handle the type discriminator and subtyping. I took that a big step forward, and using that information to generate an ari_model module, which can validate a JSON object against the Swagger model. The REST and WebSocket generators were changed to take advantage of the validators. If compiled with AST_DEVMODE enabled, JSON objects that don't match their corresponding models will not be sent out. For REST API calls, a 500 Internal Server response is sent. For WebSockets, the invalid JSON message is replaced with an error message. Since this took over about half of the job of the existing JSON generators, and the .to_json virtual function on messages took over the other half, I reluctantly removed the generators. The validators turned up all sorts of errors and inconsistencies in our data models, and the code. These were cleaned up, with checks in the code generator avoid some of the consistency problems in the future. * The model for a channel snapshot was trimmed down to match the information sent via AMI. Many of the field being sent were not useful in the general case. * The model for a bridge snapshot was updated to be more consistent with the other ARI models. Another impact of introducing subtyping was that the swagger-codegen documentation generator was insufficient (at least until it catches up with Swagger 1.2). I wanted it to be easier to generate docs for the API anyways, so I ported the wiki pages to use the Asterisk Swagger generator. In the process, I was able to clean up many of the model links, which would occasionally give inconsistent results on the wiki. I also added error responses to the wiki docs, making the wiki documentation more complete. Finally, since Stasis-HTTP will now be named Asterisk REST Interface (ARI), any new functions and files I created carry the ari_ prefix. I changed a few stasis_http references to ari where it was non-intrusive and made sense. (closes issue ASTERISK-21885) Review: https://reviewboard.asterisk.org/r/2639/ git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@393529 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2013-07-03 16:32:41 +00:00
processor.process_resource_listing(self, context)
return self
def validate_required_fields(json, required_fields, context):
"""Checks a JSON object for a set of required fields.
If any required field is missing, a SwaggerError is raised.
@param json: JSON object to check.
@param required_fields: List of required fields.
@param context: Current context in the API.
"""
missing_fields = [f for f in required_fields if not f in json]
if missing_fields:
raise SwaggerError(
"Missing fields: %s" % ', '.join(missing_fields), context)