952 lines
39 KiB
Python
952 lines
39 KiB
Python
|
import os
|
||
|
|
import sys
|
||
|
|
import yaml
|
||
|
|
import inspect
|
||
|
|
import io
|
||
|
|
from datetime import datetime
|
||
|
|
from functools import wraps, partial
|
||
|
|
|
||
|
|
import aniso8601
|
||
|
|
from werkzeug.contrib.cache import SimpleCache
|
||
|
|
from werkzeug.local import LocalProxy, LocalStack
|
||
|
|
from jinja2 import BaseLoader, ChoiceLoader, TemplateNotFound
|
||
|
|
from flask import current_app, json, request as flask_request, _app_ctx_stack
|
||
|
|
|
||
|
|
from . import verifier, logger
|
||
|
|
from .convert import to_date, to_time, to_timedelta
|
||
|
|
from .cache import top_stream, set_stream
|
||
|
|
import collections
|
||
|
|
|
||
|
|
|
||
|
|
def find_ask():
|
||
|
|
"""
|
||
|
|
Find our instance of Ask, navigating Local's and possible blueprints.
|
||
|
|
|
||
|
|
Note: This only supports returning a reference to the first instance
|
||
|
|
of Ask found.
|
||
|
|
"""
|
||
|
|
if hasattr(current_app, 'ask'):
|
||
|
|
return getattr(current_app, 'ask')
|
||
|
|
else:
|
||
|
|
if hasattr(current_app, 'blueprints'):
|
||
|
|
blueprints = getattr(current_app, 'blueprints')
|
||
|
|
for blueprint_name in blueprints:
|
||
|
|
if hasattr(blueprints[blueprint_name], 'ask'):
|
||
|
|
return getattr(blueprints[blueprint_name], 'ask')
|
||
|
|
|
||
|
|
|
||
|
|
def dbgdump(obj, default=None, cls=None):
|
||
|
|
if current_app.config.get('ASK_PRETTY_DEBUG_LOGS', False):
|
||
|
|
indent = 2
|
||
|
|
else:
|
||
|
|
indent = None
|
||
|
|
msg = json.dumps(obj, indent=indent, default=default, cls=cls)
|
||
|
|
logger.debug(msg)
|
||
|
|
|
||
|
|
|
||
|
|
request = LocalProxy(lambda: find_ask().request)
|
||
|
|
session = LocalProxy(lambda: find_ask().session)
|
||
|
|
version = LocalProxy(lambda: find_ask().version)
|
||
|
|
context = LocalProxy(lambda: find_ask().context)
|
||
|
|
convert_errors = LocalProxy(lambda: find_ask().convert_errors)
|
||
|
|
current_stream = LocalProxy(lambda: find_ask().current_stream)
|
||
|
|
stream_cache = LocalProxy(lambda: find_ask().stream_cache)
|
||
|
|
|
||
|
|
from . import models
|
||
|
|
|
||
|
|
|
||
|
|
_converters = {'date': to_date, 'time': to_time, 'timedelta': to_timedelta}
|
||
|
|
|
||
|
|
|
||
|
|
class Ask(object):
|
||
|
|
"""The Ask object provides the central interface for interacting with the Alexa service.
|
||
|
|
|
||
|
|
Ask object maps Alexa Requests to flask view functions and handles Alexa sessions.
|
||
|
|
The constructor is passed a Flask App instance, and URL endpoint.
|
||
|
|
The Flask instance allows the convienient API of endpoints and their view functions,
|
||
|
|
so that Alexa requests may be mapped with syntax similar to a typical Flask server.
|
||
|
|
Route provides the entry point for the skill, and must be provided if an app is given.
|
||
|
|
|
||
|
|
Keyword Arguments:
|
||
|
|
app {Flask object} -- App instance - created with Flask(__name__) (default: {None})
|
||
|
|
route {str} -- entry point to which initial Alexa Requests are forwarded (default: {None})
|
||
|
|
blueprint {Flask blueprint} -- Flask Blueprint instance to use instead of Flask App (default: {None})
|
||
|
|
stream_cache {Werkzeug BasicCache} -- BasicCache-like object for storing Audio stream data (default: {SimpleCache})
|
||
|
|
path {str} -- path to templates yaml file for VUI dialog (default: {'templates.yaml'})
|
||
|
|
"""
|
||
|
|
|
||
|
|
def __init__(self, app=None, route=None, blueprint=None, stream_cache=None, path='templates.yaml'):
|
||
|
|
self.app = app
|
||
|
|
self._route = route
|
||
|
|
self._intent_view_funcs = {}
|
||
|
|
self._intent_converts = {}
|
||
|
|
self._intent_defaults = {}
|
||
|
|
self._intent_mappings = {}
|
||
|
|
self._launch_view_func = None
|
||
|
|
self._session_ended_view_func = None
|
||
|
|
self._on_session_started_callback = None
|
||
|
|
self._default_intent_view_func = None
|
||
|
|
self._player_request_view_funcs = {}
|
||
|
|
self._player_mappings = {}
|
||
|
|
self._player_converts = {}
|
||
|
|
if app is not None:
|
||
|
|
self.init_app(app, path)
|
||
|
|
elif blueprint is not None:
|
||
|
|
self.init_blueprint(blueprint, path)
|
||
|
|
if stream_cache is None:
|
||
|
|
self.stream_cache = SimpleCache()
|
||
|
|
else:
|
||
|
|
self.stream_cache = stream_cache
|
||
|
|
|
||
|
|
def init_app(self, app, path='templates.yaml'):
|
||
|
|
"""Initializes Ask app by setting configuration variables, loading templates, and maps Ask route to a flask view.
|
||
|
|
|
||
|
|
The Ask instance is given the following configuration variables by calling on Flask's configuration:
|
||
|
|
|
||
|
|
`ASK_APPLICATION_ID`:
|
||
|
|
|
||
|
|
Turn on application ID verification by setting this variable to an application ID or a
|
||
|
|
list of allowed application IDs. By default, application ID verification is disabled and a
|
||
|
|
warning is logged. This variable should be set in production to ensure
|
||
|
|
requests are being sent by the applications you specify.
|
||
|
|
Default: None
|
||
|
|
|
||
|
|
`ASK_VERIFY_REQUESTS`:
|
||
|
|
|
||
|
|
Enables or disables Alexa request verification, which ensures requests sent to your skill
|
||
|
|
are from Amazon's Alexa service. This setting should not be disabled in production.
|
||
|
|
It is useful for mocking JSON requests in automated tests.
|
||
|
|
Default: True
|
||
|
|
|
||
|
|
`ASK_VERIFY_TIMESTAMP_DEBUG`:
|
||
|
|
|
||
|
|
Turn on request timestamp verification while debugging by setting this to True.
|
||
|
|
Timestamp verification helps mitigate against replay attacks. It relies on the system clock
|
||
|
|
being synchronized with an NTP server. This setting should not be enabled in production.
|
||
|
|
Default: False
|
||
|
|
|
||
|
|
`ASK_PRETTY_DEBUG_LOGS`:
|
||
|
|
|
||
|
|
Add tabs and linebreaks to the Alexa request and response printed to the debug log.
|
||
|
|
This improves readability when printing to the console, but breaks formatting when logging to CloudWatch.
|
||
|
|
Default: False
|
||
|
|
"""
|
||
|
|
if self._route is None:
|
||
|
|
raise TypeError("route is a required argument when app is not None")
|
||
|
|
|
||
|
|
self.app = app
|
||
|
|
|
||
|
|
app.ask = self
|
||
|
|
|
||
|
|
app.add_url_rule(self._route, view_func=self._flask_view_func, methods=['POST'])
|
||
|
|
app.jinja_loader = ChoiceLoader([app.jinja_loader, YamlLoader(app, path)])
|
||
|
|
|
||
|
|
def init_blueprint(self, blueprint, path='templates.yaml'):
|
||
|
|
"""Initialize a Flask Blueprint, similar to init_app, but without the access
|
||
|
|
to the application config.
|
||
|
|
|
||
|
|
Keyword Arguments:
|
||
|
|
blueprint {Flask Blueprint} -- Flask Blueprint instance to initialize (Default: {None})
|
||
|
|
path {str} -- path to templates yaml file, relative to Blueprint (Default: {'templates.yaml'})
|
||
|
|
"""
|
||
|
|
if self._route is not None:
|
||
|
|
raise TypeError("route cannot be set when using blueprints!")
|
||
|
|
|
||
|
|
# we need to tuck our reference to this Ask instance into the blueprint object and find it later!
|
||
|
|
blueprint.ask = self
|
||
|
|
|
||
|
|
# BlueprintSetupState.add_url_rule gets called underneath the covers and
|
||
|
|
# concats the rule string, so we should set to an empty string to allow
|
||
|
|
# Blueprint('blueprint_api', __name__, url_prefix="/ask") to result in
|
||
|
|
# exposing the rule at "/ask" and not "/ask/".
|
||
|
|
blueprint.add_url_rule("", view_func=self._flask_view_func, methods=['POST'])
|
||
|
|
blueprint.jinja_loader = ChoiceLoader([YamlLoader(blueprint, path)])
|
||
|
|
|
||
|
|
@property
|
||
|
|
def ask_verify_requests(self):
|
||
|
|
return current_app.config.get('ASK_VERIFY_REQUESTS', True)
|
||
|
|
|
||
|
|
@property
|
||
|
|
def ask_verify_timestamp_debug(self):
|
||
|
|
return current_app.config.get('ASK_VERIFY_TIMESTAMP_DEBUG', False)
|
||
|
|
|
||
|
|
@property
|
||
|
|
def ask_application_id(self):
|
||
|
|
return current_app.config.get('ASK_APPLICATION_ID', None)
|
||
|
|
|
||
|
|
def on_session_started(self, f):
|
||
|
|
"""Decorator to call wrapped function upon starting a session.
|
||
|
|
|
||
|
|
@ask.on_session_started
|
||
|
|
def new_session():
|
||
|
|
log.info('new session started')
|
||
|
|
|
||
|
|
Because both launch and intent requests may begin a session, this decorator is used call
|
||
|
|
a function regardless of how the session began.
|
||
|
|
|
||
|
|
Arguments:
|
||
|
|
f {function} -- function to be called when session is started.
|
||
|
|
"""
|
||
|
|
self._on_session_started_callback = f
|
||
|
|
|
||
|
|
def launch(self, f):
|
||
|
|
"""Decorator maps a view function as the endpoint for an Alexa LaunchRequest and starts the skill.
|
||
|
|
|
||
|
|
@ask.launch
|
||
|
|
def launched():
|
||
|
|
return question('Welcome to Foo')
|
||
|
|
|
||
|
|
The wrapped function is registered as the launch view function and renders the response
|
||
|
|
for requests to the Launch URL.
|
||
|
|
A request to the launch URL is verified with the Alexa server before the payload is
|
||
|
|
passed to the view function.
|
||
|
|
|
||
|
|
Arguments:
|
||
|
|
f {function} -- Launch view function
|
||
|
|
"""
|
||
|
|
self._launch_view_func = f
|
||
|
|
|
||
|
|
@wraps(f)
|
||
|
|
def wrapper(*args, **kw):
|
||
|
|
self._flask_view_func(*args, **kw)
|
||
|
|
return f
|
||
|
|
|
||
|
|
def session_ended(self, f):
|
||
|
|
"""Decorator routes Alexa SessionEndedRequest to the wrapped view function to end the skill.
|
||
|
|
|
||
|
|
@ask.session_ended
|
||
|
|
def session_ended():
|
||
|
|
return "{}", 200
|
||
|
|
|
||
|
|
The wrapped function is registered as the session_ended view function
|
||
|
|
and renders the response for requests to the end of the session.
|
||
|
|
|
||
|
|
Arguments:
|
||
|
|
f {function} -- session_ended view function
|
||
|
|
"""
|
||
|
|
self._session_ended_view_func = f
|
||
|
|
|
||
|
|
@wraps(f)
|
||
|
|
def wrapper(*args, **kw):
|
||
|
|
self._flask_view_func(*args, **kw)
|
||
|
|
return f
|
||
|
|
|
||
|
|
def intent(self, intent_name, mapping={}, convert={}, default={}):
|
||
|
|
"""Decorator routes an Alexa IntentRequest and provides the slot parameters to the wrapped function.
|
||
|
|
|
||
|
|
Functions decorated as an intent are registered as the view function for the Intent's URL,
|
||
|
|
and provide the backend responses to give your Skill its functionality.
|
||
|
|
|
||
|
|
@ask.intent('WeatherIntent', mapping={'city': 'City'})
|
||
|
|
def weather(city):
|
||
|
|
return statement('I predict great weather for {}'.format(city))
|
||
|
|
|
||
|
|
Arguments:
|
||
|
|
intent_name {str} -- Name of the intent request to be mapped to the decorated function
|
||
|
|
|
||
|
|
Keyword Arguments:
|
||
|
|
mapping {dict} -- Maps parameters to intent slots of a different name
|
||
|
|
default: {}
|
||
|
|
|
||
|
|
convert {dict} -- Converts slot values to data types before assignment to parameters
|
||
|
|
default: {}
|
||
|
|
|
||
|
|
default {dict} -- Provides default values for Intent slots if Alexa reuqest
|
||
|
|
returns no corresponding slot, or a slot with an empty value
|
||
|
|
default: {}
|
||
|
|
"""
|
||
|
|
def decorator(f):
|
||
|
|
self._intent_view_funcs[intent_name] = f
|
||
|
|
self._intent_mappings[intent_name] = mapping
|
||
|
|
self._intent_converts[intent_name] = convert
|
||
|
|
self._intent_defaults[intent_name] = default
|
||
|
|
|
||
|
|
@wraps(f)
|
||
|
|
def wrapper(*args, **kw):
|
||
|
|
self._flask_view_func(*args, **kw)
|
||
|
|
return f
|
||
|
|
return decorator
|
||
|
|
|
||
|
|
def default_intent(self, f):
|
||
|
|
"""Decorator routes any Alexa IntentRequest that is not matched by any existing @ask.intent routing."""
|
||
|
|
self._default_intent_view_func = f
|
||
|
|
|
||
|
|
@wraps(f)
|
||
|
|
def wrapper(*args, **kw):
|
||
|
|
self._flask_view_func(*args, **kw)
|
||
|
|
return f
|
||
|
|
|
||
|
|
def display_element_selected(self, f):
|
||
|
|
"""Decorator routes Alexa Display.ElementSelected request to the wrapped view function.
|
||
|
|
|
||
|
|
@ask.display_element_selected
|
||
|
|
def eval_element():
|
||
|
|
return "", 200
|
||
|
|
|
||
|
|
The wrapped function is registered as the display_element_selected view function
|
||
|
|
and renders the response for requests.
|
||
|
|
|
||
|
|
Arguments:
|
||
|
|
f {function} -- display_element_selected view function
|
||
|
|
"""
|
||
|
|
self._display_element_selected_func = f
|
||
|
|
|
||
|
|
@wraps(f)
|
||
|
|
def wrapper(*args, **kw):
|
||
|
|
self._flask_view_func(*args, **kw)
|
||
|
|
return f
|
||
|
|
|
||
|
|
|
||
|
|
def on_purchase_completed(self, mapping={'payload': 'payload','name':'name','status':'status','token':'token'}, convert={}, default={}):
|
||
|
|
"""Decorator routes an Connections.Response to the wrapped function.
|
||
|
|
|
||
|
|
Request is sent when Alexa completes the purchase flow.
|
||
|
|
See https://developer.amazon.com/docs/in-skill-purchase/add-isps-to-a-skill.html#handle-results
|
||
|
|
|
||
|
|
|
||
|
|
The wrapped view function may accept parameters from the Request.
|
||
|
|
In addition to locale, requestId, timestamp, and type
|
||
|
|
|
||
|
|
|
||
|
|
@ask.on_purchase_completed( mapping={'payload': 'payload','name':'name','status':'status','token':'token'})
|
||
|
|
def completed(payload, name, status, token):
|
||
|
|
logger.info(payload)
|
||
|
|
logger.info(name)
|
||
|
|
logger.info(status)
|
||
|
|
logger.info(token)
|
||
|
|
|
||
|
|
"""
|
||
|
|
def decorator(f):
|
||
|
|
self._intent_view_funcs['Connections.Response'] = f
|
||
|
|
self._intent_mappings['Connections.Response'] = mapping
|
||
|
|
self._intent_converts['Connections.Response'] = convert
|
||
|
|
self._intent_defaults['Connections.Response'] = default
|
||
|
|
@wraps(f)
|
||
|
|
def wrapper(*args, **kwargs):
|
||
|
|
self._flask_view_func(*args, **kwargs)
|
||
|
|
return f
|
||
|
|
return decorator
|
||
|
|
|
||
|
|
|
||
|
|
def on_playback_started(self, mapping={'offset': 'offsetInMilliseconds'}, convert={}, default={}):
|
||
|
|
"""Decorator routes an AudioPlayer.PlaybackStarted Request to the wrapped function.
|
||
|
|
|
||
|
|
Request sent when Alexa begins playing the audio stream previously sent in a Play directive.
|
||
|
|
This lets your skill verify that playback began successfully.
|
||
|
|
This request is also sent when Alexa resumes playback after pausing it for a voice request.
|
||
|
|
|
||
|
|
The wrapped view function may accept parameters from the AudioPlayer Request.
|
||
|
|
In addition to locale, requestId, timestamp, and type
|
||
|
|
AudioPlayer Requests include:
|
||
|
|
offsetInMilliseconds - Position in stream when request was sent.
|
||
|
|
Not end of stream, often few ms after Play Directive offset.
|
||
|
|
This parameter is automatically mapped to 'offset' by default
|
||
|
|
|
||
|
|
token - token of the stream that is nearly finished.
|
||
|
|
|
||
|
|
@ask.on_playback_started()
|
||
|
|
def on_playback_start(token, offset):
|
||
|
|
logger.info('stream has token {}'.format(token))
|
||
|
|
logger.info('Current position within the stream is {} ms'.format(offset))
|
||
|
|
"""
|
||
|
|
def decorator(f):
|
||
|
|
self._intent_view_funcs['AudioPlayer.PlaybackStarted'] = f
|
||
|
|
self._intent_mappings['AudioPlayer.PlaybackStarted'] = mapping
|
||
|
|
self._intent_converts['AudioPlayer.PlaybackStarted'] = convert
|
||
|
|
self._intent_defaults['AudioPlayer.PlaybackStarted'] = default
|
||
|
|
|
||
|
|
@wraps(f)
|
||
|
|
def wrapper(*args, **kwargs):
|
||
|
|
self._flask_view_func(*args, **kwargs)
|
||
|
|
return f
|
||
|
|
return decorator
|
||
|
|
|
||
|
|
def on_playback_finished(self, mapping={'offset': 'offsetInMilliseconds'}, convert={}, default={}):
|
||
|
|
"""Decorator routes an AudioPlayer.PlaybackFinished Request to the wrapped function.
|
||
|
|
|
||
|
|
This type of request is sent when the stream Alexa is playing comes to an end on its own.
|
||
|
|
|
||
|
|
Note: If your skill explicitly stops the playback with the Stop directive,
|
||
|
|
Alexa sends PlaybackStopped instead of PlaybackFinished.
|
||
|
|
|
||
|
|
The wrapped view function may accept parameters from the AudioPlayer Request.
|
||
|
|
In addition to locale, requestId, timestamp, and type
|
||
|
|
AudioPlayer Requests include:
|
||
|
|
offsetInMilliseconds - Position in stream when request was sent.
|
||
|
|
Not end of stream, often few ms after Play Directive offset.
|
||
|
|
This parameter is automatically mapped to 'offset' by default.
|
||
|
|
|
||
|
|
token - token of the stream that is nearly finished.
|
||
|
|
|
||
|
|
Audioplayer Requests do not include the stream URL, it must be accessed from current_stream.url
|
||
|
|
"""
|
||
|
|
def decorator(f):
|
||
|
|
self._intent_view_funcs['AudioPlayer.PlaybackFinished'] = f
|
||
|
|
self._intent_mappings['AudioPlayer.PlaybackFinished'] = mapping
|
||
|
|
self._intent_converts['AudioPlayer.PlaybackFinished'] = convert
|
||
|
|
self._intent_defaults['AudioPlayer.PlaybackFinished'] = default
|
||
|
|
|
||
|
|
@wraps(f)
|
||
|
|
def wrapper(*args, **kwargs):
|
||
|
|
self._flask_view_func(*args, **kwargs)
|
||
|
|
return f
|
||
|
|
return decorator
|
||
|
|
|
||
|
|
def on_playback_stopped(self, mapping={'offset': 'offsetInMilliseconds'}, convert={}, default={}):
|
||
|
|
"""Decorator routes an AudioPlayer.PlaybackStopped Request to the wrapped function.
|
||
|
|
|
||
|
|
Sent when Alexa stops playing an audio stream in response to one of the following:
|
||
|
|
-AudioPlayer.Stop
|
||
|
|
-AudioPlayer.Play with a playBehavior of REPLACE_ALL.
|
||
|
|
-AudioPlayer.ClearQueue with a clearBehavior of CLEAR_ALL.
|
||
|
|
|
||
|
|
This request is also sent if the user makes a voice request to Alexa,
|
||
|
|
since this temporarily pauses the playback.
|
||
|
|
In this case, the playback begins automatically once the voice interaction is complete.
|
||
|
|
|
||
|
|
Note: If playback stops because the audio stream comes to an end on its own,
|
||
|
|
Alexa sends PlaybackFinished instead of PlaybackStopped.
|
||
|
|
|
||
|
|
The wrapped view function may accept parameters from the AudioPlayer Request.
|
||
|
|
In addition to locale, requestId, timestamp, and type
|
||
|
|
AudioPlayer Requests include:
|
||
|
|
offsetInMilliseconds - Position in stream when request was sent.
|
||
|
|
Not end of stream, often few ms after Play Directive offset.
|
||
|
|
This parameter is automatically mapped to 'offset' by default.
|
||
|
|
|
||
|
|
token - token of the stream that is nearly finished.
|
||
|
|
|
||
|
|
Audioplayer Requests do not include the stream URL, it must be accessed from current_stream.url
|
||
|
|
"""
|
||
|
|
def decorator(f):
|
||
|
|
self._intent_view_funcs['AudioPlayer.PlaybackStopped'] = f
|
||
|
|
self._intent_mappings['AudioPlayer.PlaybackStopped'] = mapping
|
||
|
|
self._intent_converts['AudioPlayer.PlaybackStopped'] = convert
|
||
|
|
self._intent_defaults['AudioPlayer.PlaybackStopped'] = default
|
||
|
|
|
||
|
|
@wraps(f)
|
||
|
|
def wrapper(*args, **kwargs):
|
||
|
|
self._flask_view_func(*args, **kwargs)
|
||
|
|
return f
|
||
|
|
return decorator
|
||
|
|
|
||
|
|
def on_playback_nearly_finished(self, mapping={'offset': 'offsetInMilliseconds'}, convert={}, default={}):
|
||
|
|
"""Decorator routes an AudioPlayer.PlaybackNearlyFinished Request to the wrapped function.
|
||
|
|
|
||
|
|
This AudioPlayer Request sent when the device is ready to receive a new stream.
|
||
|
|
To progress through a playlist, respond to this request with an enqueue or play_next audio response.
|
||
|
|
|
||
|
|
**Note** that this request is sent when Alexa is ready to receive a new stream to enqueue, and not
|
||
|
|
necessarily when the stream's offset is near the end.
|
||
|
|
The request may be sent by Alexa immediately after your skill sends a Play Directive.
|
||
|
|
|
||
|
|
The wrapped view function may accept parameters from the AudioPlayer Request.
|
||
|
|
In addition to locale, requestId, timestamp, and type
|
||
|
|
This AudioPlayer Request includes:
|
||
|
|
AudioPlayer Requests include:
|
||
|
|
offsetInMilliseconds - Position in stream when request was sent.
|
||
|
|
Not end of stream, often few ms after Play Directive offset.
|
||
|
|
This parameter is automatically mapped to 'offset' by default.
|
||
|
|
|
||
|
|
token - token of the stream that is nearly finished.
|
||
|
|
|
||
|
|
Audioplayer Requests do not include the stream URL, and must be accessed from current_stream
|
||
|
|
|
||
|
|
Example usage:
|
||
|
|
|
||
|
|
@ask.on_playback_nearly_finished()
|
||
|
|
def play_next_stream():
|
||
|
|
audio().enqueue(my_next_song)
|
||
|
|
|
||
|
|
# offsetInMilliseconds is mapped to offset by default for convenience
|
||
|
|
@ask.on_playback_nearly_finished()
|
||
|
|
def show_request_feedback(offset, token):
|
||
|
|
logging.info('Nearly Finished')
|
||
|
|
logging.info('Stream at {} ms when Playback Request sent'.format(offset))
|
||
|
|
logging.info('Stream holds the token {}'.format(token))
|
||
|
|
logging.info('Streaming from {}'.format(current_stream.url))
|
||
|
|
|
||
|
|
# example of changing the default parameter mapping
|
||
|
|
@ask.on_playback_nearly_finished(mapping={'pos': 'offsetInMilliseconds', 'stream_token': 'token'})
|
||
|
|
def show_request_feedback(pos, stream_token):
|
||
|
|
_infodump('Nearly Finished')
|
||
|
|
_infodump('Stream at {} ms when Playback Request sent'.format(pos))
|
||
|
|
_infodump('Stream holds the token {}'.format(stream_token))
|
||
|
|
"""
|
||
|
|
def decorator(f):
|
||
|
|
self._intent_view_funcs['AudioPlayer.PlaybackNearlyFinished'] = f
|
||
|
|
self._intent_mappings['AudioPlayer.PlaybackNearlyFinished'] = mapping
|
||
|
|
self._intent_converts['AudioPlayer.PlaybackNearlyFinished'] = convert
|
||
|
|
self._intent_defaults['AudioPlayer.PlaybackNearlyFinished'] = default
|
||
|
|
|
||
|
|
@wraps(f)
|
||
|
|
def wrapper(*args, **kwargs):
|
||
|
|
self._flask_view_func(*args, **kwargs)
|
||
|
|
return f
|
||
|
|
return decorator
|
||
|
|
|
||
|
|
def on_playback_failed(self, mapping={}, convert={}, default={}):
|
||
|
|
"""Decorator routes an AudioPlayer.PlaybackFailed Request to the wrapped function.
|
||
|
|
|
||
|
|
This AudioPlayer Request sent when Alexa encounters an error when attempting to play a stream.
|
||
|
|
|
||
|
|
The wrapped view function may accept parameters from the AudioPlayer Request.
|
||
|
|
In addition to locale, requestId, timestamp, and type
|
||
|
|
|
||
|
|
PlayBackFailed Requests include:
|
||
|
|
error - Contains error info under parameters type and message
|
||
|
|
|
||
|
|
token - represents the stream that failed to play.
|
||
|
|
|
||
|
|
currentPlaybackState - Details about the playback activity occurring at the time of the error
|
||
|
|
Contains the following parameters:
|
||
|
|
|
||
|
|
token - represents the audio stream currently playing when the error occurred.
|
||
|
|
Note that this may be different from the value of the request.token property.
|
||
|
|
|
||
|
|
offsetInMilliseconds - Position in stream when request was sent.
|
||
|
|
Not end of stream, often few ms after Play Directive offset.
|
||
|
|
This parameter is automatically mapped to 'offset' by default.
|
||
|
|
|
||
|
|
playerActivity - player state when the error occurred
|
||
|
|
"""
|
||
|
|
def decorator(f):
|
||
|
|
self._intent_view_funcs['AudioPlayer.PlaybackFailed'] = f
|
||
|
|
self._intent_mappings['AudioPlayer.PlaybackFailed'] = mapping
|
||
|
|
self._intent_converts['AudioPlayer.PlaybackFailed'] = convert
|
||
|
|
self._intent_defaults['AudioPlayer.PlaybackFailed'] = default
|
||
|
|
|
||
|
|
@wraps(f)
|
||
|
|
def wrapper(*args, **kwargs):
|
||
|
|
self._flask_view_func(*args, **kwargs)
|
||
|
|
return f
|
||
|
|
return decorator
|
||
|
|
|
||
|
|
@property
|
||
|
|
def request(self):
|
||
|
|
return getattr(_app_ctx_stack.top, '_ask_request', None)
|
||
|
|
|
||
|
|
@request.setter
|
||
|
|
def request(self, value):
|
||
|
|
_app_ctx_stack.top._ask_request = value
|
||
|
|
|
||
|
|
@property
|
||
|
|
def session(self):
|
||
|
|
return getattr(_app_ctx_stack.top, '_ask_session', models._Field())
|
||
|
|
|
||
|
|
@session.setter
|
||
|
|
def session(self, value):
|
||
|
|
_app_ctx_stack.top._ask_session = value
|
||
|
|
|
||
|
|
@property
|
||
|
|
def version(self):
|
||
|
|
return getattr(_app_ctx_stack.top, '_ask_version', None)
|
||
|
|
|
||
|
|
@version.setter
|
||
|
|
def version(self, value):
|
||
|
|
_app_ctx_stack.top._ask_version = value
|
||
|
|
|
||
|
|
@property
|
||
|
|
def context(self):
|
||
|
|
return getattr(_app_ctx_stack.top, '_ask_context', None)
|
||
|
|
|
||
|
|
@context.setter
|
||
|
|
def context(self, value):
|
||
|
|
_app_ctx_stack.top._ask_context = value
|
||
|
|
|
||
|
|
@property
|
||
|
|
def convert_errors(self):
|
||
|
|
return getattr(_app_ctx_stack.top, '_ask_convert_errors', None)
|
||
|
|
|
||
|
|
@convert_errors.setter
|
||
|
|
def convert_errors(self, value):
|
||
|
|
_app_ctx_stack.top._ask_convert_errors = value
|
||
|
|
|
||
|
|
@property
|
||
|
|
def current_stream(self):
|
||
|
|
#return getattr(_app_ctx_stack.top, '_ask_current_stream', models._Field())
|
||
|
|
user = self._get_user()
|
||
|
|
if user:
|
||
|
|
stream = top_stream(self.stream_cache, user)
|
||
|
|
if stream:
|
||
|
|
current = models._Field()
|
||
|
|
current.__dict__.update(stream)
|
||
|
|
return current
|
||
|
|
return models._Field()
|
||
|
|
|
||
|
|
@current_stream.setter
|
||
|
|
def current_stream(self, value):
|
||
|
|
# assumption 1 is we get a models._Field as value
|
||
|
|
# assumption 2 is if someone sets a value, it's resetting the stack
|
||
|
|
user = self._get_user()
|
||
|
|
if user:
|
||
|
|
set_stream(self.stream_cache, user, value.__dict__)
|
||
|
|
|
||
|
|
def run_aws_lambda(self, event):
|
||
|
|
"""Invoke the Flask Ask application from an AWS Lambda function handler.
|
||
|
|
|
||
|
|
Use this method to service AWS Lambda requests from a custom Alexa
|
||
|
|
skill. This method will invoke your Flask application providing a
|
||
|
|
WSGI-compatible environment that wraps the original Alexa event
|
||
|
|
provided to the AWS Lambda handler. Returns the output generated by
|
||
|
|
a Flask Ask application, which should be used as the return value
|
||
|
|
to the AWS Lambda handler function.
|
||
|
|
|
||
|
|
Example usage:
|
||
|
|
|
||
|
|
from flask import Flask
|
||
|
|
from flask_ask import Ask, statement
|
||
|
|
|
||
|
|
app = Flask(__name__)
|
||
|
|
ask = Ask(app, '/')
|
||
|
|
|
||
|
|
# This function name is what you defined when you create an
|
||
|
|
# AWS Lambda function. By default, AWS calls this function
|
||
|
|
# lambda_handler.
|
||
|
|
def lambda_handler(event, _context):
|
||
|
|
return ask.run_aws_lambda(event)
|
||
|
|
|
||
|
|
@ask.intent('HelloIntent')
|
||
|
|
def hello(firstname):
|
||
|
|
speech_text = "Hello %s" % firstname
|
||
|
|
return statement(speech_text).simple_card('Hello', speech_text)
|
||
|
|
"""
|
||
|
|
|
||
|
|
# We are guaranteed to be called by AWS as a Lambda function does not
|
||
|
|
# expose a public facing interface.
|
||
|
|
self.app.config['ASK_VERIFY_REQUESTS'] = False
|
||
|
|
|
||
|
|
# Convert an environment variable to a WSGI "bytes-as-unicode" string
|
||
|
|
enc, esc = sys.getfilesystemencoding(), 'surrogateescape'
|
||
|
|
def unicode_to_wsgi(u):
|
||
|
|
return u.encode(enc, esc).decode('iso-8859-1')
|
||
|
|
|
||
|
|
# Create a WSGI-compatible environ that can be passed to the
|
||
|
|
# application. It is loaded with the OS environment variables,
|
||
|
|
# mandatory CGI-like variables, as well as the mandatory WSGI
|
||
|
|
# variables.
|
||
|
|
environ = {k: unicode_to_wsgi(v) for k, v in os.environ.items()}
|
||
|
|
environ['REQUEST_METHOD'] = 'POST'
|
||
|
|
environ['PATH_INFO'] = '/'
|
||
|
|
environ['SERVER_NAME'] = 'AWS-Lambda'
|
||
|
|
environ['SERVER_PORT'] = '80'
|
||
|
|
environ['SERVER_PROTOCOL'] = 'HTTP/1.0'
|
||
|
|
environ['wsgi.version'] = (1, 0)
|
||
|
|
environ['wsgi.url_scheme'] = 'http'
|
||
|
|
environ['wsgi.errors'] = sys.stderr
|
||
|
|
environ['wsgi.multithread'] = False
|
||
|
|
environ['wsgi.multiprocess'] = False
|
||
|
|
environ['wsgi.run_once'] = True
|
||
|
|
|
||
|
|
# Convert the event provided by the AWS Lambda handler to a JSON
|
||
|
|
# string that can be read as the body of a HTTP POST request.
|
||
|
|
body = json.dumps(event)
|
||
|
|
environ['CONTENT_TYPE'] = 'application/json'
|
||
|
|
environ['CONTENT_LENGTH'] = len(body)
|
||
|
|
|
||
|
|
PY3 = sys.version_info[0] == 3
|
||
|
|
|
||
|
|
if PY3:
|
||
|
|
environ['wsgi.input'] = io.StringIO(body)
|
||
|
|
else:
|
||
|
|
environ['wsgi.input'] = io.BytesIO(body)
|
||
|
|
|
||
|
|
# Start response is a required callback that must be passed when
|
||
|
|
# the application is invoked. It is used to set HTTP status and
|
||
|
|
# headers. Read the WSGI spec for details (PEP3333).
|
||
|
|
headers = []
|
||
|
|
def start_response(status, response_headers, _exc_info=None):
|
||
|
|
headers[:] = [status, response_headers]
|
||
|
|
|
||
|
|
# Invoke the actual Flask application providing our environment,
|
||
|
|
# with our Alexa event as the body of the HTTP request, as well
|
||
|
|
# as the callback function above. The result will be an iterator
|
||
|
|
# that provides a serialized JSON string for our Alexa response.
|
||
|
|
result = self.app(environ, start_response)
|
||
|
|
try:
|
||
|
|
if not headers:
|
||
|
|
raise AssertionError("start_response() not called by WSGI app")
|
||
|
|
|
||
|
|
output = b"".join(result)
|
||
|
|
if not headers[0].startswith("2"):
|
||
|
|
raise AssertionError("Non-2xx from app: hdrs={}, body={}".format(headers, output))
|
||
|
|
|
||
|
|
# The Lambda handler expects a Python object that can be
|
||
|
|
# serialized as JSON, so we need to take the already serialized
|
||
|
|
# JSON and deserialize it.
|
||
|
|
return json.loads(output)
|
||
|
|
|
||
|
|
finally:
|
||
|
|
# Per the WSGI spec, we need to invoke the close method if it
|
||
|
|
# is implemented on the result object.
|
||
|
|
if hasattr(result, 'close'):
|
||
|
|
result.close()
|
||
|
|
|
||
|
|
|
||
|
|
def _get_user(self):
|
||
|
|
if self.context:
|
||
|
|
return self.context.get('System', {}).get('user', {}).get('userId')
|
||
|
|
return None
|
||
|
|
|
||
|
|
|
||
|
|
def _alexa_request(self, verify=True):
|
||
|
|
raw_body = flask_request.data
|
||
|
|
alexa_request_payload = json.loads(raw_body)
|
||
|
|
|
||
|
|
if verify:
|
||
|
|
cert_url = flask_request.headers['Signaturecertchainurl']
|
||
|
|
signature = flask_request.headers['Signature']
|
||
|
|
|
||
|
|
# load certificate - this verifies a the certificate url and format under the hood
|
||
|
|
cert = verifier.load_certificate(cert_url)
|
||
|
|
# verify signature
|
||
|
|
verifier.verify_signature(cert, signature, raw_body)
|
||
|
|
|
||
|
|
# verify timestamp
|
||
|
|
raw_timestamp = alexa_request_payload.get('request', {}).get('timestamp')
|
||
|
|
timestamp = self._parse_timestamp(raw_timestamp)
|
||
|
|
|
||
|
|
if not current_app.debug or self.ask_verify_timestamp_debug:
|
||
|
|
verifier.verify_timestamp(timestamp)
|
||
|
|
|
||
|
|
# verify application id
|
||
|
|
try:
|
||
|
|
application_id = alexa_request_payload['session']['application']['applicationId']
|
||
|
|
except KeyError:
|
||
|
|
application_id = alexa_request_payload['context'][
|
||
|
|
'System']['application']['applicationId']
|
||
|
|
if self.ask_application_id is not None:
|
||
|
|
verifier.verify_application_id(application_id, self.ask_application_id)
|
||
|
|
|
||
|
|
return alexa_request_payload
|
||
|
|
|
||
|
|
@staticmethod
|
||
|
|
def _parse_timestamp(timestamp):
|
||
|
|
"""
|
||
|
|
Parse a given timestamp value, raising ValueError if None or Flasey
|
||
|
|
"""
|
||
|
|
if timestamp:
|
||
|
|
try:
|
||
|
|
return aniso8601.parse_datetime(timestamp)
|
||
|
|
except AttributeError:
|
||
|
|
# raised by aniso8601 if raw_timestamp is not valid string
|
||
|
|
# in ISO8601 format
|
||
|
|
try:
|
||
|
|
return datetime.utcfromtimestamp(timestamp)
|
||
|
|
except:
|
||
|
|
# relax the timestamp a bit in case it was sent in millis
|
||
|
|
return datetime.utcfromtimestamp(timestamp/1000)
|
||
|
|
|
||
|
|
raise ValueError('Invalid timestamp value! Cannot parse from either ISO8601 string or UTC timestamp.')
|
||
|
|
|
||
|
|
|
||
|
|
def _update_stream(self):
|
||
|
|
fresh_stream = models._Field()
|
||
|
|
fresh_stream.__dict__.update(self.current_stream.__dict__) # keeps url attribute after stopping stream
|
||
|
|
fresh_stream.__dict__.update(self._from_directive())
|
||
|
|
|
||
|
|
context_info = self._from_context()
|
||
|
|
if context_info != None:
|
||
|
|
fresh_stream.__dict__.update(context_info)
|
||
|
|
|
||
|
|
self.current_stream = fresh_stream
|
||
|
|
dbgdump(current_stream.__dict__)
|
||
|
|
|
||
|
|
def _from_context(self):
|
||
|
|
return getattr(self.context, 'AudioPlayer', {})
|
||
|
|
|
||
|
|
def _from_directive(self):
|
||
|
|
from_buffer = top_stream(self.stream_cache, self._get_user())
|
||
|
|
if from_buffer:
|
||
|
|
if self.request.intent and 'PauseIntent' in self.request.intent.name:
|
||
|
|
return {}
|
||
|
|
return from_buffer
|
||
|
|
return {}
|
||
|
|
|
||
|
|
def _flask_view_func(self, *args, **kwargs):
|
||
|
|
ask_payload = self._alexa_request(verify=self.ask_verify_requests)
|
||
|
|
dbgdump(ask_payload)
|
||
|
|
request_body = models._Field(ask_payload)
|
||
|
|
|
||
|
|
self.request = request_body.request
|
||
|
|
self.version = request_body.version
|
||
|
|
self.context = getattr(request_body, 'context', models._Field())
|
||
|
|
self.session = getattr(request_body, 'session', self.session) # to keep old session.attributes through AudioRequests
|
||
|
|
|
||
|
|
if not self.session:
|
||
|
|
self.session = models._Field()
|
||
|
|
if not self.session.attributes:
|
||
|
|
self.session.attributes = models._Field()
|
||
|
|
|
||
|
|
self._update_stream()
|
||
|
|
|
||
|
|
# add current dialog state in session
|
||
|
|
try:
|
||
|
|
self.session["dialogState"] = request.dialogState
|
||
|
|
except KeyError:
|
||
|
|
self.session["dialogState"] = "unknown"
|
||
|
|
|
||
|
|
try:
|
||
|
|
if self.session.new and self._on_session_started_callback is not None:
|
||
|
|
self._on_session_started_callback()
|
||
|
|
except AttributeError:
|
||
|
|
pass
|
||
|
|
|
||
|
|
result = None
|
||
|
|
request_type = self.request.type
|
||
|
|
|
||
|
|
if request_type == 'LaunchRequest' and self._launch_view_func:
|
||
|
|
result = self._launch_view_func()
|
||
|
|
elif request_type == 'SessionEndedRequest':
|
||
|
|
if self._session_ended_view_func:
|
||
|
|
result = self._session_ended_view_func()
|
||
|
|
else:
|
||
|
|
result = "{}", 200
|
||
|
|
elif request_type == 'IntentRequest' and self._intent_view_funcs:
|
||
|
|
result = self._map_intent_to_view_func(self.request.intent)()
|
||
|
|
elif request_type == 'Display.ElementSelected' and self._display_element_selected_func:
|
||
|
|
result = self._display_element_selected_func()
|
||
|
|
elif 'AudioPlayer' in request_type:
|
||
|
|
result = self._map_player_request_to_func(self.request.type)()
|
||
|
|
# routes to on_playback funcs
|
||
|
|
# user can also access state of content.AudioPlayer with current_stream
|
||
|
|
elif 'Connections.Response' in request_type:
|
||
|
|
result = self._map_purchase_request_to_func(self.request.type)()
|
||
|
|
|
||
|
|
if result is not None:
|
||
|
|
if isinstance(result, models._Response):
|
||
|
|
return result.render_response()
|
||
|
|
return result
|
||
|
|
return "", 400
|
||
|
|
|
||
|
|
def _map_intent_to_view_func(self, intent):
|
||
|
|
"""Provides appropiate parameters to the intent functions."""
|
||
|
|
if intent.name in self._intent_view_funcs:
|
||
|
|
view_func = self._intent_view_funcs[intent.name]
|
||
|
|
elif self._default_intent_view_func is not None:
|
||
|
|
view_func = self._default_intent_view_func
|
||
|
|
else:
|
||
|
|
raise NotImplementedError('Intent "{}" not found and no default intent specified.'.format(intent.name))
|
||
|
|
|
||
|
|
PY3 = sys.version_info[0] == 3
|
||
|
|
if PY3:
|
||
|
|
argspec = inspect.getfullargspec(view_func)
|
||
|
|
else:
|
||
|
|
argspec = inspect.getargspec(view_func)
|
||
|
|
|
||
|
|
arg_names = argspec.args
|
||
|
|
arg_values = self._map_params_to_view_args(intent.name, arg_names)
|
||
|
|
|
||
|
|
return partial(view_func, *arg_values)
|
||
|
|
|
||
|
|
def _map_player_request_to_func(self, player_request_type):
|
||
|
|
"""Provides appropriate parameters to the on_playback functions."""
|
||
|
|
# calbacks for on_playback requests are optional
|
||
|
|
view_func = self._intent_view_funcs.get(player_request_type, lambda: None)
|
||
|
|
|
||
|
|
argspec = inspect.getargspec(view_func)
|
||
|
|
arg_names = argspec.args
|
||
|
|
arg_values = self._map_params_to_view_args(player_request_type, arg_names)
|
||
|
|
|
||
|
|
return partial(view_func, *arg_values)
|
||
|
|
|
||
|
|
def _map_purchase_request_to_func(self, purchase_request_type):
|
||
|
|
"""Provides appropriate parameters to the on_purchase functions."""
|
||
|
|
|
||
|
|
if purchase_request_type in self._intent_view_funcs:
|
||
|
|
view_func = self._intent_view_funcs[purchase_request_type]
|
||
|
|
else:
|
||
|
|
raise NotImplementedError('Request type "{}" not found and no default view specified.'.format(purchase_request_type))
|
||
|
|
|
||
|
|
argspec = inspect.getargspec(view_func)
|
||
|
|
arg_names = argspec.args
|
||
|
|
arg_values = self._map_params_to_view_args(purchase_request_type, arg_names)
|
||
|
|
|
||
|
|
print('_map_purchase_request_to_func', arg_names, arg_values, view_func, purchase_request_type)
|
||
|
|
return partial(view_func, *arg_values)
|
||
|
|
|
||
|
|
def _get_slot_value(self, slot_object):
|
||
|
|
slot_name = slot_object.name
|
||
|
|
slot_value = getattr(slot_object, 'value', None)
|
||
|
|
resolutions = getattr(slot_object, 'resolutions', None)
|
||
|
|
|
||
|
|
if resolutions is not None:
|
||
|
|
resolutions_per_authority = getattr(resolutions, 'resolutionsPerAuthority', None)
|
||
|
|
if resolutions_per_authority is not None and len(resolutions_per_authority) > 0:
|
||
|
|
values = resolutions_per_authority[0].get('values', None)
|
||
|
|
if values is not None and len(values) > 0:
|
||
|
|
value = values[0].get('value', None)
|
||
|
|
if value is not None:
|
||
|
|
slot_value = value.get('name', slot_value)
|
||
|
|
|
||
|
|
return slot_value
|
||
|
|
|
||
|
|
def _map_params_to_view_args(self, view_name, arg_names):
|
||
|
|
|
||
|
|
arg_values = []
|
||
|
|
convert = self._intent_converts.get(view_name)
|
||
|
|
default = self._intent_defaults.get(view_name)
|
||
|
|
mapping = self._intent_mappings.get(view_name)
|
||
|
|
|
||
|
|
convert_errors = {}
|
||
|
|
|
||
|
|
request_data = {}
|
||
|
|
intent = getattr(self.request, 'intent', None)
|
||
|
|
if intent is not None:
|
||
|
|
if intent.slots is not None:
|
||
|
|
for slot_key in intent.slots.keys():
|
||
|
|
slot_object = getattr(intent.slots, slot_key)
|
||
|
|
request_data[slot_object.name] = self._get_slot_value(slot_object=slot_object)
|
||
|
|
|
||
|
|
else:
|
||
|
|
for param_name in self.request:
|
||
|
|
request_data[param_name] = getattr(self.request, param_name, None)
|
||
|
|
|
||
|
|
for arg_name in arg_names:
|
||
|
|
param_or_slot = mapping.get(arg_name, arg_name)
|
||
|
|
arg_value = request_data.get(param_or_slot)
|
||
|
|
if arg_value is None or arg_value == "":
|
||
|
|
if arg_name in default:
|
||
|
|
default_value = default[arg_name]
|
||
|
|
if isinstance(default_value, collections.Callable):
|
||
|
|
default_value = default_value()
|
||
|
|
arg_value = default_value
|
||
|
|
elif arg_name in convert:
|
||
|
|
shorthand_or_function = convert[arg_name]
|
||
|
|
if shorthand_or_function in _converters:
|
||
|
|
shorthand = shorthand_or_function
|
||
|
|
convert_func = _converters[shorthand]
|
||
|
|
else:
|
||
|
|
convert_func = shorthand_or_function
|
||
|
|
try:
|
||
|
|
arg_value = convert_func(arg_value)
|
||
|
|
except Exception as e:
|
||
|
|
convert_errors[arg_name] = e
|
||
|
|
arg_values.append(arg_value)
|
||
|
|
self.convert_errors = convert_errors
|
||
|
|
return arg_values
|
||
|
|
|
||
|
|
|
||
|
|
class YamlLoader(BaseLoader):
|
||
|
|
|
||
|
|
def __init__(self, app, path):
|
||
|
|
self.path = app.root_path + os.path.sep + path
|
||
|
|
self.mapping = {}
|
||
|
|
self._reload_mapping()
|
||
|
|
|
||
|
|
def _reload_mapping(self):
|
||
|
|
if os.path.isfile(self.path):
|
||
|
|
self.last_mtime = os.path.getmtime(self.path)
|
||
|
|
with open(self.path) as f:
|
||
|
|
self.mapping = yaml.safe_load(f.read())
|
||
|
|
|
||
|
|
def get_source(self, environment, template):
|
||
|
|
if not os.path.isfile(self.path):
|
||
|
|
return None, None, None
|
||
|
|
if self.last_mtime != os.path.getmtime(self.path):
|
||
|
|
self._reload_mapping()
|
||
|
|
if template in self.mapping:
|
||
|
|
source = self.mapping[template]
|
||
|
|
return source, None, lambda: source == self.mapping.get(template)
|
||
|
|
raise TemplateNotFound(template)
|