core/homeassistant/components/rest/sensor.py

306 lines
10 KiB
Python
Raw Normal View History

"""Support for RESTful API sensors."""
Unpacking RESTful sensor JSON results into attributes. (#10753) * Added support for extracting JSON attributes from RESTful values Setting the json_attributes configuration option to true on the RESTful sensor will cause the result of the REST request to be parsed as a JSON string and if successful the resulting dictionary will be used for the attributes of the sensor. * Added support for extracting JSON attributes from RESTful values Setting the json_attributes configuration option to true on the RESTful sensor will cause the result of the REST request to be parsed as a JSON string and if successful the resulting dictionary will be used for the attributes of the sensor. * Added requirement that RESTful JSON results used as attributes must be objects, not lists. * Expanded test coverage to test REFTful JSON attributes with and without a value template. * Added support for extracting JSON attributes from RESTful values Setting the json_attributes configuration option to true on the RESTful sensor will cause the result of the REST request to be parsed as a JSON string and if successful the resulting dictionary will be used for the attributes of the sensor. * Added requirement that RESTful JSON results used as attributes must be objects, not lists. * Expanded test coverage to test REFTful JSON attributes with and without a value template. * sensor.envirophat: add missing requirement (#7451) Adding requirements that is not explicitly pulled in by the library that manages the Enviro pHAT. * PyPI Openzwave (#7415) * Remove default zwave config path PYOZW now has much more comprehensive default handling for the config path (in src-lib/libopenzwave/libopenzwave.pyx:getConfig()). It looks in the same place we were looking, plus _many_ more. It will certainly do a much better job of finding the config files than we will (and will be updated as the library is changed, so we don't end up chasing it). The getConfig() method has been there for a while, but was subsntially improved recently. This change simply leaves the config_path as None if it is not specified, which will trigger the default handling in PYOZW. * Install python-openzwave from PyPI As of version 0.4, python-openzwave supports installation from PyPI, which means we can use our 'normal' dependency management tooling to install it. Yay. This uses the default 'embed' build (which goes and downloads statically sources to avoid having to compile anything locally). Check out the python-openzwave readme for more details. * Add python-openzwave deps to .travis.yml Python OpenZwave require the libudev headers to build. This adds the libudev-dev package to Travis runs via the 'apt' addon for Travis. Thanks to @MartinHjelmare for this fix. * Update docker build for PyPI openzwave Now that PYOZW can be install from PyPI, the docker image build process can be simplified to remove the explicit compilation of PYOZW. * Add datadog component (#7158) * Add datadog component * Improve test_invalid_config datadog test * Use assert_setup_component for test setup * Fix object type for default KNX port #7429 describes a TypeError that is raised if the port is omitted in the config for the KNX component (integer is required (got type str)). This commit changes the default port from a string to an integer. I expect this will resolve that issue... * Added support for extracting JSON attributes from RESTful values Setting the json_attributes configuration option to true on the RESTful sensor will cause the result of the REST request to be parsed as a JSON string and if successful the resulting dictionary will be used for the attributes of the sensor. * Added requirement that RESTful JSON results used as attributes must be objects, not lists. * Expanded test coverage to test REFTful JSON attributes with and without a value template. * Added support for extracting JSON attributes from RESTful values Setting the json_attributes configuration option to true on the RESTful sensor will cause the result of the REST request to be parsed as a JSON string and if successful the resulting dictionary will be used for the attributes of the sensor. * Added requirement that RESTful JSON results used as attributes must be objects, not lists. * Expanded test coverage to test REFTful JSON attributes with and without a value template. * Added support for extracting JSON attributes from RESTful values Setting the json_attributes configuration option to true on the RESTful sensor will cause the result of the REST request to be parsed as a JSON string and if successful the resulting dictionary will be used for the attributes of the sensor. * Added requirement that RESTful JSON results used as attributes must be objects, not lists. * Expanded test coverage to test REFTful JSON attributes with and without a value template. * Fixed breaks cause by manual upstream merge. * Added one extra blank line to make PyLint happy. * Switched json_attributes to be a list of keys rather than a boolean. The value of json_attributes can now be either a comma sepaated list of key names or a YAML list of key names. Only matching keys in a retuned JSON dictionary will be mapped to sensor attributes. Updated test cases to handle json_attributes being a list. Also fixed two minor issues arrising from manual merge with 0.58 master. * Added an explicit default value to the json_attributes config entry. * Removed self.update() from __init__() body. * Expended unit tests for error cases of json_attributes processing. * Align quotes
2017-12-03 15:30:25 +00:00
import json
import logging
from xml.parsers.expat import ExpatError
2016-02-19 05:27:50 +00:00
from jsonpath import jsonpath
import requests
from requests import Session
from requests.auth import HTTPBasicAuth, HTTPDigestAuth
import voluptuous as vol
import xmltodict
2015-09-13 20:41:16 +00:00
from homeassistant.components.sensor import DEVICE_CLASSES_SCHEMA, PLATFORM_SCHEMA
from homeassistant.const import (
2019-07-31 19:25:30 +00:00
CONF_AUTHENTICATION,
CONF_DEVICE_CLASS,
2019-07-31 19:25:30 +00:00
CONF_FORCE_UPDATE,
CONF_HEADERS,
CONF_METHOD,
CONF_NAME,
2019-07-31 19:25:30 +00:00
CONF_PASSWORD,
CONF_PAYLOAD,
CONF_RESOURCE,
CONF_RESOURCE_TEMPLATE,
CONF_TIMEOUT,
2019-07-31 19:25:30 +00:00
CONF_UNIT_OF_MEASUREMENT,
CONF_USERNAME,
CONF_VALUE_TEMPLATE,
CONF_VERIFY_SSL,
HTTP_BASIC_AUTHENTICATION,
HTTP_DIGEST_AUTHENTICATION,
)
from homeassistant.exceptions import PlatformNotReady
import homeassistant.helpers.config_validation as cv
from homeassistant.helpers.entity import Entity
2015-09-13 20:41:16 +00:00
2016-09-03 23:45:49 +00:00
_LOGGER = logging.getLogger(__name__)
2019-07-31 19:25:30 +00:00
DEFAULT_METHOD = "GET"
DEFAULT_NAME = "REST Sensor"
2016-09-03 23:45:49 +00:00
DEFAULT_VERIFY_SSL = True
DEFAULT_FORCE_UPDATE = False
DEFAULT_TIMEOUT = 10
2019-07-31 19:25:30 +00:00
CONF_JSON_ATTRS = "json_attributes"
CONF_JSON_ATTRS_PATH = "json_attributes_path"
2019-07-31 19:25:30 +00:00
METHODS = ["POST", "GET"]
PLATFORM_SCHEMA = PLATFORM_SCHEMA.extend(
{
vol.Exclusive(CONF_RESOURCE, CONF_RESOURCE): cv.url,
vol.Exclusive(CONF_RESOURCE_TEMPLATE, CONF_RESOURCE): cv.template,
2019-07-31 19:25:30 +00:00
vol.Optional(CONF_AUTHENTICATION): vol.In(
[HTTP_BASIC_AUTHENTICATION, HTTP_DIGEST_AUTHENTICATION]
),
vol.Optional(CONF_HEADERS): vol.Schema({cv.string: cv.string}),
vol.Optional(CONF_JSON_ATTRS, default=[]): cv.ensure_list_csv,
vol.Optional(CONF_METHOD, default=DEFAULT_METHOD): vol.In(METHODS),
vol.Optional(CONF_NAME, default=DEFAULT_NAME): cv.string,
vol.Optional(CONF_PASSWORD): cv.string,
vol.Optional(CONF_PAYLOAD): cv.string,
vol.Optional(CONF_UNIT_OF_MEASUREMENT): cv.string,
vol.Optional(CONF_DEVICE_CLASS): DEVICE_CLASSES_SCHEMA,
vol.Optional(CONF_USERNAME): cv.string,
vol.Optional(CONF_JSON_ATTRS_PATH): cv.string,
2019-07-31 19:25:30 +00:00
vol.Optional(CONF_VALUE_TEMPLATE): cv.template,
vol.Optional(CONF_VERIFY_SSL, default=DEFAULT_VERIFY_SSL): cv.boolean,
vol.Optional(CONF_FORCE_UPDATE, default=DEFAULT_FORCE_UPDATE): cv.boolean,
vol.Optional(CONF_TIMEOUT, default=DEFAULT_TIMEOUT): cv.positive_int,
}
)
PLATFORM_SCHEMA = vol.All(
cv.has_at_least_one_key(CONF_RESOURCE, CONF_RESOURCE_TEMPLATE), PLATFORM_SCHEMA
)
2015-09-13 20:41:16 +00:00
def setup_platform(hass, config, add_entities, discovery_info=None):
"""Set up the RESTful sensor."""
name = config.get(CONF_NAME)
resource = config.get(CONF_RESOURCE)
resource_template = config.get(CONF_RESOURCE_TEMPLATE)
method = config.get(CONF_METHOD)
payload = config.get(CONF_PAYLOAD)
2016-09-03 23:45:49 +00:00
verify_ssl = config.get(CONF_VERIFY_SSL)
username = config.get(CONF_USERNAME)
password = config.get(CONF_PASSWORD)
headers = config.get(CONF_HEADERS)
unit = config.get(CONF_UNIT_OF_MEASUREMENT)
device_class = config.get(CONF_DEVICE_CLASS)
value_template = config.get(CONF_VALUE_TEMPLATE)
Unpacking RESTful sensor JSON results into attributes. (#10753) * Added support for extracting JSON attributes from RESTful values Setting the json_attributes configuration option to true on the RESTful sensor will cause the result of the REST request to be parsed as a JSON string and if successful the resulting dictionary will be used for the attributes of the sensor. * Added support for extracting JSON attributes from RESTful values Setting the json_attributes configuration option to true on the RESTful sensor will cause the result of the REST request to be parsed as a JSON string and if successful the resulting dictionary will be used for the attributes of the sensor. * Added requirement that RESTful JSON results used as attributes must be objects, not lists. * Expanded test coverage to test REFTful JSON attributes with and without a value template. * Added support for extracting JSON attributes from RESTful values Setting the json_attributes configuration option to true on the RESTful sensor will cause the result of the REST request to be parsed as a JSON string and if successful the resulting dictionary will be used for the attributes of the sensor. * Added requirement that RESTful JSON results used as attributes must be objects, not lists. * Expanded test coverage to test REFTful JSON attributes with and without a value template. * sensor.envirophat: add missing requirement (#7451) Adding requirements that is not explicitly pulled in by the library that manages the Enviro pHAT. * PyPI Openzwave (#7415) * Remove default zwave config path PYOZW now has much more comprehensive default handling for the config path (in src-lib/libopenzwave/libopenzwave.pyx:getConfig()). It looks in the same place we were looking, plus _many_ more. It will certainly do a much better job of finding the config files than we will (and will be updated as the library is changed, so we don't end up chasing it). The getConfig() method has been there for a while, but was subsntially improved recently. This change simply leaves the config_path as None if it is not specified, which will trigger the default handling in PYOZW. * Install python-openzwave from PyPI As of version 0.4, python-openzwave supports installation from PyPI, which means we can use our 'normal' dependency management tooling to install it. Yay. This uses the default 'embed' build (which goes and downloads statically sources to avoid having to compile anything locally). Check out the python-openzwave readme for more details. * Add python-openzwave deps to .travis.yml Python OpenZwave require the libudev headers to build. This adds the libudev-dev package to Travis runs via the 'apt' addon for Travis. Thanks to @MartinHjelmare for this fix. * Update docker build for PyPI openzwave Now that PYOZW can be install from PyPI, the docker image build process can be simplified to remove the explicit compilation of PYOZW. * Add datadog component (#7158) * Add datadog component * Improve test_invalid_config datadog test * Use assert_setup_component for test setup * Fix object type for default KNX port #7429 describes a TypeError that is raised if the port is omitted in the config for the KNX component (integer is required (got type str)). This commit changes the default port from a string to an integer. I expect this will resolve that issue... * Added support for extracting JSON attributes from RESTful values Setting the json_attributes configuration option to true on the RESTful sensor will cause the result of the REST request to be parsed as a JSON string and if successful the resulting dictionary will be used for the attributes of the sensor. * Added requirement that RESTful JSON results used as attributes must be objects, not lists. * Expanded test coverage to test REFTful JSON attributes with and without a value template. * Added support for extracting JSON attributes from RESTful values Setting the json_attributes configuration option to true on the RESTful sensor will cause the result of the REST request to be parsed as a JSON string and if successful the resulting dictionary will be used for the attributes of the sensor. * Added requirement that RESTful JSON results used as attributes must be objects, not lists. * Expanded test coverage to test REFTful JSON attributes with and without a value template. * Added support for extracting JSON attributes from RESTful values Setting the json_attributes configuration option to true on the RESTful sensor will cause the result of the REST request to be parsed as a JSON string and if successful the resulting dictionary will be used for the attributes of the sensor. * Added requirement that RESTful JSON results used as attributes must be objects, not lists. * Expanded test coverage to test REFTful JSON attributes with and without a value template. * Fixed breaks cause by manual upstream merge. * Added one extra blank line to make PyLint happy. * Switched json_attributes to be a list of keys rather than a boolean. The value of json_attributes can now be either a comma sepaated list of key names or a YAML list of key names. Only matching keys in a retuned JSON dictionary will be mapped to sensor attributes. Updated test cases to handle json_attributes being a list. Also fixed two minor issues arrising from manual merge with 0.58 master. * Added an explicit default value to the json_attributes config entry. * Removed self.update() from __init__() body. * Expended unit tests for error cases of json_attributes processing. * Align quotes
2017-12-03 15:30:25 +00:00
json_attrs = config.get(CONF_JSON_ATTRS)
json_attrs_path = config.get(CONF_JSON_ATTRS_PATH)
force_update = config.get(CONF_FORCE_UPDATE)
timeout = config.get(CONF_TIMEOUT)
Unpacking RESTful sensor JSON results into attributes. (#10753) * Added support for extracting JSON attributes from RESTful values Setting the json_attributes configuration option to true on the RESTful sensor will cause the result of the REST request to be parsed as a JSON string and if successful the resulting dictionary will be used for the attributes of the sensor. * Added support for extracting JSON attributes from RESTful values Setting the json_attributes configuration option to true on the RESTful sensor will cause the result of the REST request to be parsed as a JSON string and if successful the resulting dictionary will be used for the attributes of the sensor. * Added requirement that RESTful JSON results used as attributes must be objects, not lists. * Expanded test coverage to test REFTful JSON attributes with and without a value template. * Added support for extracting JSON attributes from RESTful values Setting the json_attributes configuration option to true on the RESTful sensor will cause the result of the REST request to be parsed as a JSON string and if successful the resulting dictionary will be used for the attributes of the sensor. * Added requirement that RESTful JSON results used as attributes must be objects, not lists. * Expanded test coverage to test REFTful JSON attributes with and without a value template. * sensor.envirophat: add missing requirement (#7451) Adding requirements that is not explicitly pulled in by the library that manages the Enviro pHAT. * PyPI Openzwave (#7415) * Remove default zwave config path PYOZW now has much more comprehensive default handling for the config path (in src-lib/libopenzwave/libopenzwave.pyx:getConfig()). It looks in the same place we were looking, plus _many_ more. It will certainly do a much better job of finding the config files than we will (and will be updated as the library is changed, so we don't end up chasing it). The getConfig() method has been there for a while, but was subsntially improved recently. This change simply leaves the config_path as None if it is not specified, which will trigger the default handling in PYOZW. * Install python-openzwave from PyPI As of version 0.4, python-openzwave supports installation from PyPI, which means we can use our 'normal' dependency management tooling to install it. Yay. This uses the default 'embed' build (which goes and downloads statically sources to avoid having to compile anything locally). Check out the python-openzwave readme for more details. * Add python-openzwave deps to .travis.yml Python OpenZwave require the libudev headers to build. This adds the libudev-dev package to Travis runs via the 'apt' addon for Travis. Thanks to @MartinHjelmare for this fix. * Update docker build for PyPI openzwave Now that PYOZW can be install from PyPI, the docker image build process can be simplified to remove the explicit compilation of PYOZW. * Add datadog component (#7158) * Add datadog component * Improve test_invalid_config datadog test * Use assert_setup_component for test setup * Fix object type for default KNX port #7429 describes a TypeError that is raised if the port is omitted in the config for the KNX component (integer is required (got type str)). This commit changes the default port from a string to an integer. I expect this will resolve that issue... * Added support for extracting JSON attributes from RESTful values Setting the json_attributes configuration option to true on the RESTful sensor will cause the result of the REST request to be parsed as a JSON string and if successful the resulting dictionary will be used for the attributes of the sensor. * Added requirement that RESTful JSON results used as attributes must be objects, not lists. * Expanded test coverage to test REFTful JSON attributes with and without a value template. * Added support for extracting JSON attributes from RESTful values Setting the json_attributes configuration option to true on the RESTful sensor will cause the result of the REST request to be parsed as a JSON string and if successful the resulting dictionary will be used for the attributes of the sensor. * Added requirement that RESTful JSON results used as attributes must be objects, not lists. * Expanded test coverage to test REFTful JSON attributes with and without a value template. * Added support for extracting JSON attributes from RESTful values Setting the json_attributes configuration option to true on the RESTful sensor will cause the result of the REST request to be parsed as a JSON string and if successful the resulting dictionary will be used for the attributes of the sensor. * Added requirement that RESTful JSON results used as attributes must be objects, not lists. * Expanded test coverage to test REFTful JSON attributes with and without a value template. * Fixed breaks cause by manual upstream merge. * Added one extra blank line to make PyLint happy. * Switched json_attributes to be a list of keys rather than a boolean. The value of json_attributes can now be either a comma sepaated list of key names or a YAML list of key names. Only matching keys in a retuned JSON dictionary will be mapped to sensor attributes. Updated test cases to handle json_attributes being a list. Also fixed two minor issues arrising from manual merge with 0.58 master. * Added an explicit default value to the json_attributes config entry. * Removed self.update() from __init__() body. * Expended unit tests for error cases of json_attributes processing. * Align quotes
2017-12-03 15:30:25 +00:00
if value_template is not None:
value_template.hass = hass
if resource_template is not None:
resource_template.hass = hass
resource = resource_template.render()
if username and password:
if config.get(CONF_AUTHENTICATION) == HTTP_DIGEST_AUTHENTICATION:
auth = HTTPDigestAuth(username, password)
else:
auth = HTTPBasicAuth(username, password)
else:
auth = None
2019-07-31 19:25:30 +00:00
rest = RestData(method, resource, auth, headers, payload, verify_ssl, timeout)
2016-01-02 21:29:33 +00:00
rest.update()
if rest.data is None:
raise PlatformNotReady
# Must update the sensor now (including fetching the rest resource) to
# ensure it's updating its state.
2019-07-31 19:25:30 +00:00
add_entities(
[
RestSensor(
hass,
rest,
name,
unit,
device_class,
value_template,
json_attrs,
force_update,
resource_template,
json_attrs_path,
2019-07-31 19:25:30 +00:00
)
],
True,
)
2015-09-13 20:41:16 +00:00
class RestSensor(Entity):
2016-03-08 15:46:34 +00:00
"""Implementation of a REST sensor."""
2015-09-13 20:41:16 +00:00
2019-07-31 19:25:30 +00:00
def __init__(
self,
hass,
rest,
name,
unit_of_measurement,
device_class,
value_template,
json_attrs,
force_update,
resource_template,
json_attrs_path,
2019-07-31 19:25:30 +00:00
):
2016-09-03 23:45:49 +00:00
"""Initialize the REST sensor."""
2015-12-11 22:19:49 +00:00
self._hass = hass
2015-09-13 20:41:16 +00:00
self.rest = rest
self._name = name
self._state = None
2015-09-13 20:41:16 +00:00
self._unit_of_measurement = unit_of_measurement
self._device_class = device_class
2015-12-11 22:19:49 +00:00
self._value_template = value_template
Unpacking RESTful sensor JSON results into attributes. (#10753) * Added support for extracting JSON attributes from RESTful values Setting the json_attributes configuration option to true on the RESTful sensor will cause the result of the REST request to be parsed as a JSON string and if successful the resulting dictionary will be used for the attributes of the sensor. * Added support for extracting JSON attributes from RESTful values Setting the json_attributes configuration option to true on the RESTful sensor will cause the result of the REST request to be parsed as a JSON string and if successful the resulting dictionary will be used for the attributes of the sensor. * Added requirement that RESTful JSON results used as attributes must be objects, not lists. * Expanded test coverage to test REFTful JSON attributes with and without a value template. * Added support for extracting JSON attributes from RESTful values Setting the json_attributes configuration option to true on the RESTful sensor will cause the result of the REST request to be parsed as a JSON string and if successful the resulting dictionary will be used for the attributes of the sensor. * Added requirement that RESTful JSON results used as attributes must be objects, not lists. * Expanded test coverage to test REFTful JSON attributes with and without a value template. * sensor.envirophat: add missing requirement (#7451) Adding requirements that is not explicitly pulled in by the library that manages the Enviro pHAT. * PyPI Openzwave (#7415) * Remove default zwave config path PYOZW now has much more comprehensive default handling for the config path (in src-lib/libopenzwave/libopenzwave.pyx:getConfig()). It looks in the same place we were looking, plus _many_ more. It will certainly do a much better job of finding the config files than we will (and will be updated as the library is changed, so we don't end up chasing it). The getConfig() method has been there for a while, but was subsntially improved recently. This change simply leaves the config_path as None if it is not specified, which will trigger the default handling in PYOZW. * Install python-openzwave from PyPI As of version 0.4, python-openzwave supports installation from PyPI, which means we can use our 'normal' dependency management tooling to install it. Yay. This uses the default 'embed' build (which goes and downloads statically sources to avoid having to compile anything locally). Check out the python-openzwave readme for more details. * Add python-openzwave deps to .travis.yml Python OpenZwave require the libudev headers to build. This adds the libudev-dev package to Travis runs via the 'apt' addon for Travis. Thanks to @MartinHjelmare for this fix. * Update docker build for PyPI openzwave Now that PYOZW can be install from PyPI, the docker image build process can be simplified to remove the explicit compilation of PYOZW. * Add datadog component (#7158) * Add datadog component * Improve test_invalid_config datadog test * Use assert_setup_component for test setup * Fix object type for default KNX port #7429 describes a TypeError that is raised if the port is omitted in the config for the KNX component (integer is required (got type str)). This commit changes the default port from a string to an integer. I expect this will resolve that issue... * Added support for extracting JSON attributes from RESTful values Setting the json_attributes configuration option to true on the RESTful sensor will cause the result of the REST request to be parsed as a JSON string and if successful the resulting dictionary will be used for the attributes of the sensor. * Added requirement that RESTful JSON results used as attributes must be objects, not lists. * Expanded test coverage to test REFTful JSON attributes with and without a value template. * Added support for extracting JSON attributes from RESTful values Setting the json_attributes configuration option to true on the RESTful sensor will cause the result of the REST request to be parsed as a JSON string and if successful the resulting dictionary will be used for the attributes of the sensor. * Added requirement that RESTful JSON results used as attributes must be objects, not lists. * Expanded test coverage to test REFTful JSON attributes with and without a value template. * Added support for extracting JSON attributes from RESTful values Setting the json_attributes configuration option to true on the RESTful sensor will cause the result of the REST request to be parsed as a JSON string and if successful the resulting dictionary will be used for the attributes of the sensor. * Added requirement that RESTful JSON results used as attributes must be objects, not lists. * Expanded test coverage to test REFTful JSON attributes with and without a value template. * Fixed breaks cause by manual upstream merge. * Added one extra blank line to make PyLint happy. * Switched json_attributes to be a list of keys rather than a boolean. The value of json_attributes can now be either a comma sepaated list of key names or a YAML list of key names. Only matching keys in a retuned JSON dictionary will be mapped to sensor attributes. Updated test cases to handle json_attributes being a list. Also fixed two minor issues arrising from manual merge with 0.58 master. * Added an explicit default value to the json_attributes config entry. * Removed self.update() from __init__() body. * Expended unit tests for error cases of json_attributes processing. * Align quotes
2017-12-03 15:30:25 +00:00
self._json_attrs = json_attrs
self._attributes = None
self._force_update = force_update
self._resource_template = resource_template
self._json_attrs_path = json_attrs_path
2015-09-13 20:41:16 +00:00
@property
def name(self):
2016-03-08 15:46:34 +00:00
"""Return the name of the sensor."""
2015-09-13 20:41:16 +00:00
return self._name
@property
def unit_of_measurement(self):
2016-03-08 15:46:34 +00:00
"""Return the unit the value is expressed in."""
2015-09-13 20:41:16 +00:00
return self._unit_of_measurement
@property
def device_class(self):
"""Return the class of this sensor."""
return self._device_class
@property
def available(self):
"""Return if the sensor data are available."""
return self.rest.data is not None
2015-09-13 20:41:16 +00:00
@property
def state(self):
2016-03-08 15:46:34 +00:00
"""Return the state of the device."""
2015-09-13 20:41:16 +00:00
return self._state
@property
def force_update(self):
"""Force update."""
return self._force_update
2015-09-13 20:41:16 +00:00
def update(self):
2016-03-08 15:46:34 +00:00
"""Get the latest data from REST API and update the state."""
if self._resource_template is not None:
self.rest.set_url(self._resource_template.render())
2015-09-13 20:41:16 +00:00
self.rest.update()
value = self.rest.data
_LOGGER.debug("Data fetched from resource: %s", value)
if self.rest.headers is not None:
# If the http request failed, headers will be None
content_type = self.rest.headers.get("content-type")
if content_type and (
content_type.startswith("text/xml")
or content_type.startswith("application/xml")
):
try:
value = json.dumps(xmltodict.parse(value))
_LOGGER.debug("JSON converted from XML: %s", value)
except ExpatError:
_LOGGER.warning(
"REST xml result could not be parsed and converted to JSON"
)
_LOGGER.debug("Erroneous XML: %s", value)
2015-09-13 20:41:16 +00:00
Unpacking RESTful sensor JSON results into attributes. (#10753) * Added support for extracting JSON attributes from RESTful values Setting the json_attributes configuration option to true on the RESTful sensor will cause the result of the REST request to be parsed as a JSON string and if successful the resulting dictionary will be used for the attributes of the sensor. * Added support for extracting JSON attributes from RESTful values Setting the json_attributes configuration option to true on the RESTful sensor will cause the result of the REST request to be parsed as a JSON string and if successful the resulting dictionary will be used for the attributes of the sensor. * Added requirement that RESTful JSON results used as attributes must be objects, not lists. * Expanded test coverage to test REFTful JSON attributes with and without a value template. * Added support for extracting JSON attributes from RESTful values Setting the json_attributes configuration option to true on the RESTful sensor will cause the result of the REST request to be parsed as a JSON string and if successful the resulting dictionary will be used for the attributes of the sensor. * Added requirement that RESTful JSON results used as attributes must be objects, not lists. * Expanded test coverage to test REFTful JSON attributes with and without a value template. * sensor.envirophat: add missing requirement (#7451) Adding requirements that is not explicitly pulled in by the library that manages the Enviro pHAT. * PyPI Openzwave (#7415) * Remove default zwave config path PYOZW now has much more comprehensive default handling for the config path (in src-lib/libopenzwave/libopenzwave.pyx:getConfig()). It looks in the same place we were looking, plus _many_ more. It will certainly do a much better job of finding the config files than we will (and will be updated as the library is changed, so we don't end up chasing it). The getConfig() method has been there for a while, but was subsntially improved recently. This change simply leaves the config_path as None if it is not specified, which will trigger the default handling in PYOZW. * Install python-openzwave from PyPI As of version 0.4, python-openzwave supports installation from PyPI, which means we can use our 'normal' dependency management tooling to install it. Yay. This uses the default 'embed' build (which goes and downloads statically sources to avoid having to compile anything locally). Check out the python-openzwave readme for more details. * Add python-openzwave deps to .travis.yml Python OpenZwave require the libudev headers to build. This adds the libudev-dev package to Travis runs via the 'apt' addon for Travis. Thanks to @MartinHjelmare for this fix. * Update docker build for PyPI openzwave Now that PYOZW can be install from PyPI, the docker image build process can be simplified to remove the explicit compilation of PYOZW. * Add datadog component (#7158) * Add datadog component * Improve test_invalid_config datadog test * Use assert_setup_component for test setup * Fix object type for default KNX port #7429 describes a TypeError that is raised if the port is omitted in the config for the KNX component (integer is required (got type str)). This commit changes the default port from a string to an integer. I expect this will resolve that issue... * Added support for extracting JSON attributes from RESTful values Setting the json_attributes configuration option to true on the RESTful sensor will cause the result of the REST request to be parsed as a JSON string and if successful the resulting dictionary will be used for the attributes of the sensor. * Added requirement that RESTful JSON results used as attributes must be objects, not lists. * Expanded test coverage to test REFTful JSON attributes with and without a value template. * Added support for extracting JSON attributes from RESTful values Setting the json_attributes configuration option to true on the RESTful sensor will cause the result of the REST request to be parsed as a JSON string and if successful the resulting dictionary will be used for the attributes of the sensor. * Added requirement that RESTful JSON results used as attributes must be objects, not lists. * Expanded test coverage to test REFTful JSON attributes with and without a value template. * Added support for extracting JSON attributes from RESTful values Setting the json_attributes configuration option to true on the RESTful sensor will cause the result of the REST request to be parsed as a JSON string and if successful the resulting dictionary will be used for the attributes of the sensor. * Added requirement that RESTful JSON results used as attributes must be objects, not lists. * Expanded test coverage to test REFTful JSON attributes with and without a value template. * Fixed breaks cause by manual upstream merge. * Added one extra blank line to make PyLint happy. * Switched json_attributes to be a list of keys rather than a boolean. The value of json_attributes can now be either a comma sepaated list of key names or a YAML list of key names. Only matching keys in a retuned JSON dictionary will be mapped to sensor attributes. Updated test cases to handle json_attributes being a list. Also fixed two minor issues arrising from manual merge with 0.58 master. * Added an explicit default value to the json_attributes config entry. * Removed self.update() from __init__() body. * Expended unit tests for error cases of json_attributes processing. * Align quotes
2017-12-03 15:30:25 +00:00
if self._json_attrs:
self._attributes = {}
if value:
try:
json_dict = json.loads(value)
if self._json_attrs_path is not None:
json_dict = jsonpath(json_dict, self._json_attrs_path)
# jsonpath will always store the result in json_dict[0]
# so the next line happens to work exactly as needed to
# find the result
if isinstance(json_dict, list):
json_dict = json_dict[0]
if isinstance(json_dict, dict):
2019-07-31 19:25:30 +00:00
attrs = {
k: json_dict[k] for k in self._json_attrs if k in json_dict
}
self._attributes = attrs
else:
_LOGGER.warning(
"JSON result was not a dictionary"
" or list with 0th element a dictionary"
)
except ValueError:
_LOGGER.warning("REST result could not be parsed as JSON")
_LOGGER.debug("Erroneous JSON: %s", value)
else:
_LOGGER.warning("Empty reply found when expecting JSON data")
if value is not None and self._value_template is not None:
2019-07-31 19:25:30 +00:00
value = self._value_template.render_with_possible_json_value(value, None)
2015-12-25 21:34:30 +00:00
self._state = value
2015-09-13 20:41:16 +00:00
Unpacking RESTful sensor JSON results into attributes. (#10753) * Added support for extracting JSON attributes from RESTful values Setting the json_attributes configuration option to true on the RESTful sensor will cause the result of the REST request to be parsed as a JSON string and if successful the resulting dictionary will be used for the attributes of the sensor. * Added support for extracting JSON attributes from RESTful values Setting the json_attributes configuration option to true on the RESTful sensor will cause the result of the REST request to be parsed as a JSON string and if successful the resulting dictionary will be used for the attributes of the sensor. * Added requirement that RESTful JSON results used as attributes must be objects, not lists. * Expanded test coverage to test REFTful JSON attributes with and without a value template. * Added support for extracting JSON attributes from RESTful values Setting the json_attributes configuration option to true on the RESTful sensor will cause the result of the REST request to be parsed as a JSON string and if successful the resulting dictionary will be used for the attributes of the sensor. * Added requirement that RESTful JSON results used as attributes must be objects, not lists. * Expanded test coverage to test REFTful JSON attributes with and without a value template. * sensor.envirophat: add missing requirement (#7451) Adding requirements that is not explicitly pulled in by the library that manages the Enviro pHAT. * PyPI Openzwave (#7415) * Remove default zwave config path PYOZW now has much more comprehensive default handling for the config path (in src-lib/libopenzwave/libopenzwave.pyx:getConfig()). It looks in the same place we were looking, plus _many_ more. It will certainly do a much better job of finding the config files than we will (and will be updated as the library is changed, so we don't end up chasing it). The getConfig() method has been there for a while, but was subsntially improved recently. This change simply leaves the config_path as None if it is not specified, which will trigger the default handling in PYOZW. * Install python-openzwave from PyPI As of version 0.4, python-openzwave supports installation from PyPI, which means we can use our 'normal' dependency management tooling to install it. Yay. This uses the default 'embed' build (which goes and downloads statically sources to avoid having to compile anything locally). Check out the python-openzwave readme for more details. * Add python-openzwave deps to .travis.yml Python OpenZwave require the libudev headers to build. This adds the libudev-dev package to Travis runs via the 'apt' addon for Travis. Thanks to @MartinHjelmare for this fix. * Update docker build for PyPI openzwave Now that PYOZW can be install from PyPI, the docker image build process can be simplified to remove the explicit compilation of PYOZW. * Add datadog component (#7158) * Add datadog component * Improve test_invalid_config datadog test * Use assert_setup_component for test setup * Fix object type for default KNX port #7429 describes a TypeError that is raised if the port is omitted in the config for the KNX component (integer is required (got type str)). This commit changes the default port from a string to an integer. I expect this will resolve that issue... * Added support for extracting JSON attributes from RESTful values Setting the json_attributes configuration option to true on the RESTful sensor will cause the result of the REST request to be parsed as a JSON string and if successful the resulting dictionary will be used for the attributes of the sensor. * Added requirement that RESTful JSON results used as attributes must be objects, not lists. * Expanded test coverage to test REFTful JSON attributes with and without a value template. * Added support for extracting JSON attributes from RESTful values Setting the json_attributes configuration option to true on the RESTful sensor will cause the result of the REST request to be parsed as a JSON string and if successful the resulting dictionary will be used for the attributes of the sensor. * Added requirement that RESTful JSON results used as attributes must be objects, not lists. * Expanded test coverage to test REFTful JSON attributes with and without a value template. * Added support for extracting JSON attributes from RESTful values Setting the json_attributes configuration option to true on the RESTful sensor will cause the result of the REST request to be parsed as a JSON string and if successful the resulting dictionary will be used for the attributes of the sensor. * Added requirement that RESTful JSON results used as attributes must be objects, not lists. * Expanded test coverage to test REFTful JSON attributes with and without a value template. * Fixed breaks cause by manual upstream merge. * Added one extra blank line to make PyLint happy. * Switched json_attributes to be a list of keys rather than a boolean. The value of json_attributes can now be either a comma sepaated list of key names or a YAML list of key names. Only matching keys in a retuned JSON dictionary will be mapped to sensor attributes. Updated test cases to handle json_attributes being a list. Also fixed two minor issues arrising from manual merge with 0.58 master. * Added an explicit default value to the json_attributes config entry. * Removed self.update() from __init__() body. * Expended unit tests for error cases of json_attributes processing. * Align quotes
2017-12-03 15:30:25 +00:00
@property
def device_state_attributes(self):
"""Return the state attributes."""
return self._attributes
2015-09-13 20:41:16 +00:00
class RestData:
2016-01-02 21:29:33 +00:00
"""Class for handling the data retrieval."""
2015-09-13 20:41:16 +00:00
2019-07-31 19:25:30 +00:00
def __init__(
self, method, resource, auth, headers, data, verify_ssl, timeout=DEFAULT_TIMEOUT
):
2016-03-08 15:46:34 +00:00
"""Initialize the data object."""
self._method = method
self._resource = resource
self._auth = auth
self._headers = headers
self._request_data = data
self._verify_ssl = verify_ssl
self._timeout = timeout
self._http_session = Session()
2015-12-25 21:34:30 +00:00
self.data = None
self.headers = None
def __del__(self):
"""Destroy the http session on destroy."""
self._http_session.close()
def set_url(self, url):
"""Set url."""
self._resource = url
def update(self):
2016-10-31 04:51:03 +00:00
"""Get the latest data from REST service with provided method."""
_LOGGER.debug("Updating from %s", self._resource)
try:
response = self._http_session.request(
self._method,
self._resource,
headers=self._headers,
auth=self._auth,
data=self._request_data,
timeout=self._timeout,
verify=self._verify_ssl,
)
2015-12-11 22:19:49 +00:00
self.data = response.text
self.headers = response.headers
2018-05-11 10:28:28 +00:00
except requests.exceptions.RequestException as ex:
_LOGGER.error("Error fetching data: %s failed with %s", self._resource, ex)
2015-12-25 21:34:30 +00:00
self.data = None
self.headers = None