# Copyright 2019 StreamSets Inc.
"""Models to be used by multiple StreamSets components."""
# fmt: off
import json
import warnings
from uuid import uuid4
import inflection
# fmt: on
json_to_python_style = lambda x: inflection.underscore(x)
python_to_json_style = lambda x: inflection.camelize(x, uppercase_first_letter=False)
[docs]class Configuration:
"""Abstraction for configurations.
This class enables easy access to and modification of data stored as a list of dictionaries. A Configuration is
stored in the form:
.. code-block:: none
[{"name" : "<name_1>","value" : "<value_1>"}, {"name" : "<name_2>", value" : "<value_2>"},...]
However, the passed in configuration parameter can be a list of Configurations such as:
.. code-block:: none
[[{"name" : "<name_1>","value" : "<value_1>"}, {"name" : "<name_2>", value" : "<value_2>"},...],
[{"name" : "<name_3>","value" : "<value_3>"}, {"name" : "<name_4>", value" : "<value_4>"},...],...]
Args:
compatibility_map (:obj:`dict`, optional): A dictionary mapping values used for backwards compatibility.
configuration (:obj:`list`): List of configurations (see above for format).
property_key (:obj:`str`, optional): The dictionary entry denoting the property key.
Default: ``name``
property_value (:obj:`str`, optional): The dictionary entry denoting the property value.
Default: ``value``
update_callable (optional): A callable to which ``self._data`` will be passed as part of ``__setitem__``.
update_callable_kwargs (:obj:`dict`, optional): A dictionary of kwargs to pass (along with a body)
to the callable.
id_to_remap (:obj:`dict`, optional): A dictionary mapping configuration IDs to human-readable container keys.
Example: {'custom_region':'googleCloudConfig.customRegion', ... }
"""
# Use an uber secret class attribute to specify whether other attributes can be assigned by __setattr__.
__frozen = False
def __init__(
self,
configuration=None,
compatibility_map=None,
property_key='name',
property_value='value',
update_callable=None,
update_callable_kwargs=None,
id_to_remap=None,
):
# Apply overrides to initial data
if compatibility_map:
for configuration_entry in configuration:
configuration_name = configuration_entry.get(property_key)
configuration_value = configuration_entry.get(property_value)
if configuration_name in compatibility_map:
overrides = compatibility_map[configuration_name]
override_values = overrides.get('values', {})
configuration_entry[property_key] = overrides['name']
if configuration_value in override_values:
configuration_entry[property_value] = override_values[configuration_value]
warnings.warn(
'Configuration {}={} has been deprecated. Please use {}={} instead.'.format(
configuration_name,
configuration_value,
overrides['name'],
override_values[configuration_value],
),
DeprecationWarning,
)
else:
warnings.warn(
'Configuration {} has been deprecated. Please use {} instead.'.format(
configuration_name, overrides['name']
),
DeprecationWarning,
)
self._compatibility_map = compatibility_map or {}
self.property_key = property_key
self.property_value = property_value
self._id_to_remap = id_to_remap or {}
self._update_callable = update_callable
self._update_callable_kwargs = update_callable_kwargs or {}
# Ensure the input 'configuration' is properly formatted, handling both single configurations and lists.
self._data = [configuration] if isinstance(configuration[0], dict) else configuration
self._configuration_index_map = self._create_configuration_index_map()
self.__frozen = True
def _create_configuration_index_map(self):
"""
Creates a mapping {config_item_name: (config_item_index, config_list_index)} for efficient lookups.
This method facilitates quick identification of the index of a configuration item within the
Configurations/self._data list.
Here,
1. config_list_index is the index within the Configurations list, i.e. the outer index.
2. config_item_index is the index within an individual Configuration, i.e. the inner index.
If our ._data is as follows:
[[{"name" : "<name_1>","value" : "<value_1>"}, {"name" : "<name_2>", value" : "<value_2>"}],
[{"name" : "<name_3>","value" : "<value_3>"}, {"name" : "<name_4>", value" : "<value_4>"}]]
Then configuration_index_map would be returned as:
{<name_1>:(0,0), <name_2>:(1,0),, <name_3>:(0,1),, <name_4>:(1,1),}
Returns:
A configuration index map (:obj:`dict`).
Raises:
TypeError: If the input data structure is not a list of configurations,
if a Configuration is not a list of dictionaries, or
if a Configuration does not contain the specified property_key and property_value.
"""
configuration_index_map = {}
for config_list_index, config_list in enumerate(self._data):
if not isinstance(config_list, list):
raise TypeError('Please pass in a list of configurations')
for config_item_index, config_item in enumerate(config_list):
if not isinstance(config_item, dict):
raise TypeError('A Configuration must be a list of dictionaries')
if self.property_key not in config_item or self.property_value not in config_item:
raise TypeError(
'Configuration {} does not contain property_key:{} or property_value:{}'.format(
config_item, self.property_key, self.property_value
)
)
configuration_index_map[config_item[self.property_key]] = (config_item_index, config_list_index)
return configuration_index_map
def __getattr__(self, key):
if not self.__frozen:
super().__getattr__(key)
return
return self.__getitem__(key)
def __getitem__(self, key):
if key in self._id_to_remap:
key = self._id_to_remap[key]
if key not in self._configuration_index_map:
raise AttributeError(key)
index, configuration_index = self._configuration_index_map[key]
config = self._data[configuration_index][index]
return self._convert_value(config)
def __setattr__(self, key, value):
self.__setitem__(key, value)
def __setitem__(self, key, value):
if not self.__frozen:
super().__setattr__(key, value)
return
if key in self._id_to_remap:
key = self._id_to_remap[key]
if key in self._compatibility_map:
overrides = self._compatibility_map[key]
if value in self._compatibility_map[key]['values']:
warnings.warn(
'Deprecation warning: Configuration {}={} is deprecated on this engine version. '
'Updating value to {}={}.'.format(key, value, overrides['name'], overrides['values'][value]),
DeprecationWarning,
)
value = overrides['values'][value]
else:
warnings.warn(
'Configuration {} has been deprecated. Please use {} instead.'.format(key, overrides['name']),
DeprecationWarning,
)
key = overrides['name']
if key not in self._configuration_index_map:
raise AttributeError(key)
index, configuration_index = self._configuration_index_map[key]
config = self._data[configuration_index][index]
config[self.property_value] = value
if self._update_callable:
kwargs = dict(body=[config])
kwargs.update(self._update_callable_kwargs)
self._update_callable(**kwargs)
def __contains__(self, item):
return item in self._id_to_remap or item in self._configuration_index_map
def __repr__(self):
configs = {}
for configuration in self._data:
for config in configuration:
key = config[self.property_key]
configs[key] = self._convert_value(config)
# If a key has a remapped key, delete the original key and add the remapped key into configs
for remapped_key, original_key in self._id_to_remap.items():
if original_key != remapped_key and original_key in configs:
configs[remapped_key] = configs[original_key]
del configs[original_key]
return '{{{}}}'.format(', '.join("'{}': {}".format(k, v) for k, v in configs.items()))
def __dir__(self):
# Stripping out any values that have multiple words in it. Example: 'Organization account type'
# in sch_models.Organization.configuration
id_to_remap_cleaned = [key for key in self._id_to_remap.keys() if ' ' not in key]
return sorted(list(dir(object)) + list(self.__dict__.keys()) + id_to_remap_cleaned)
[docs] def items(self):
"""Gets the configuration's items.
Returns:
A new view of the configuration’s items ((key, value) pairs).
"""
# To keep the behavior in line with a Python dict's, we'll generate one and then use its items method.
configuration_dict = {}
for configuration in self._data:
for config in configuration:
configuration_dict[config[self.property_key]] = self._convert_value(config)
for config_property in self._id_to_remap:
key = self._id_to_remap[config_property]
if key in configuration_dict:
configuration_dict[config_property] = configuration_dict[key]
del configuration_dict[key]
return configuration_dict.items()
[docs] def get(self, key, default=None):
"""Return the value of key or, if not in the configuration, the default value."""
try:
return self[key]
except KeyError:
return default
[docs] def update(self, configs):
"""Update instance with a collection of configurations.
Args:
configs (:obj:`dict`): Dictionary of configurations to use.
"""
for key, value in configs.items():
self[key] = value
def _convert_value(self, config):
if config.get('type') == 'boolean':
return json.loads(config[self.property_value])
elif config.get('type') == 'integer':
return int(config[self.property_value])
else:
return config[self.property_value]
class BaseModel:
"""Base class for StreamSets Accounts models that essentially just wrap a dictionary.
Args:
data (:obj:`dict`): The underlying JSON representation of the model.
attributes_to_ignore (:obj:`list`, optional): A list of string attributes to mask from being handled
by this class' __setattr__ method. Default: ``None``.
attributes_to_remap (:obj:`dict`, optional): A dictionary of attributes to remap with the desired attributes
as keys and the corresponding property name in the JSON representation as values. Default: ``None``.
repr_metadata (:obj:`list`, optional): A list of attributes to use in the model's __repr__ string.
Default: ``None``.
"""
def __new__(cls, *args, **kwargs):
instance = super(BaseModel, cls).__new__(cls)
super(BaseModel, instance).__setattr__('_data_internal', {})
super(BaseModel, instance).__setattr__('_attributes_to_ignore', [])
super(BaseModel, instance).__setattr__('_attributes_to_remap', {})
super(BaseModel, instance).__setattr__('_repr_metadata', [])
return instance
def __init__(self, data, attributes_to_ignore=None, attributes_to_remap=None, repr_metadata=None):
# _data_internal is introduced to help inherited classes that need to load _data when _data is accessed
# eg. Pipeline
super().__setattr__('_data_internal', data)
super().__setattr__('_attributes_to_ignore', attributes_to_ignore or [])
super().__setattr__('_attributes_to_remap', attributes_to_remap or {})
super().__setattr__('_repr_metadata', repr_metadata or [])
# By default these properties don't do anything by can be overrided by inherited classes to load something
@property
def _data_internal(self):
return self.__dict__['_data'] if '_data' in self.__dict__ else None
@_data_internal.setter
def _data_internal(self, data):
self.__dict__['_data'] = data
@property
def _data(self):
return self._data_internal
@_data.setter
def _data(self, data):
self._data_internal = data
def __getattr__(self, name):
name_ = python_to_json_style(name)
if name in self._attributes_to_remap:
remapped_name = self._attributes_to_remap[name]
return self._data_internal[remapped_name]
elif (
name_ in self._data_internal
and name not in self._attributes_to_ignore
and name not in self._attributes_to_remap.values()
):
return self._data_internal[name_]
raise AttributeError('Could not find attribute {}.'.format(name_))
def __setattr__(self, name, value):
name_ = python_to_json_style(name)
if name in self._attributes_to_remap:
remapped_name = self._attributes_to_remap[name]
self._data_internal[remapped_name] = value
elif (
name_ in self._data_internal
and name not in self._attributes_to_ignore
and name not in self._attributes_to_remap.values()
):
self._data_internal[name_] = value
else:
super().__setattr__(name, value)
def __dir__(self):
return sorted(
list(dir(object))
+ list(self.__dict__.keys())
+ list(
json_to_python_style(key)
for key in self._data_internal.keys()
if key not in (list(self._attributes_to_remap.values()) + self._attributes_to_ignore)
)
+ list(self._attributes_to_remap.keys())
)
def __eq__(self, other):
return self._data_internal == other._data_internal
def __repr__(self):
return '<{} ({})>'.format(
self.__class__.__name__, ', '.join('{}={}'.format(key, getattr(self, key)) for key in self._repr_metadata)
)
class ModelCollection:
"""Base class wrapper with Abstractions.
Args:
streamsets_entity: An instance of underlysing StreamSets entity
e.g. :py:class:`streamsets.sdk.next.Next` or :py:class:`streamsets.sdk.accounts.Accounts`.
"""
def __init__(self, streamsets_entity):
self._streamsets_entity = streamsets_entity
self._id_attr = 'id'
def _get_all_results_from_api(self, **kwargs):
"""Used to get multiple (all) results from api.
Args:
Optional arguments to be passed to filter the results.
Returns:
A :obj:`collections.namedtuple`: of
results (:py:class:`streamsets.sdk.utils.SeekableList`): a SeekableList of inherited instances of
:py:class:`streamsets.sdk.models.BaseModel` and
kwargs (:obj:`dict`): a dict of local variables not used in this function.
"""
pass
def __iter__(self):
"""Enables the list enumeration or iteration."""
for item in self._get_all_results_from_api().results:
yield item
def __getitem__(self, i):
"""Enables the user to fetch items by index.
Args:
i (:obj:`int`): Index of the item.
Returns:
An inherited instance of :py:class:`streamsets.sdk.models.BaseModel`.
"""
return self._get_all_results_from_api().results[i]
def __len__(self):
"""Provides length (count) of items.
Returns:
A :py:obj:`int` object.
"""
return len(self._get_all_results_from_api().results)
def __contains__(self, item_given):
"""Checks if given item is in the list of items by comparing the ids.
Returns:
A :py:obj:`boolean` object.
"""
return self.contains(**{self._id_attr: getattr(item_given, self._id_attr)})
def get(self, **kwargs):
"""
Args:
**kwargs: Optional arguments to be passed to filter the results offline.
Returns:
An inherited instance of :py:class:`streamsets.sdk.models.BaseModel`.
"""
result, new_kwargs = self._get_all_results_from_api(**kwargs)
return result.get(**new_kwargs)
def get_all(self, **kwargs):
"""
Args:
**kwargs: Optional other arguments to be passed to filter the results offline.
Returns:
A :py:obj:`streamsets.sdk.utils.SeekableList` of inherited instances of
:py:class:`streamsets.sdk.models.BaseModel`.
"""
result, new_kwargs = self._get_all_results_from_api(**kwargs)
return result.get_all(**new_kwargs)
def __repr__(self):
return str(self._get_all_results_from_api().results)
def contains(self, **kwargs):
"""
Args:
**kwargs: Optional arguments to be passed to filter the results offline.
Returns:
A :py:obj:`boolean` object.
"""
try:
self.get(**kwargs)
except ValueError:
return False
return True
class _StageWithPredicates:
"""Pipeline Stage extension to include a predicate configuration and handle the output lanes.
This class is expected to inherit from either SDC or ST Stage class.
Args:
variable_output_drive (:obj:`str`, optional): Configuration name that drives the output lanes available.
Default: ``None``.
"""
def __init__(self, variable_output_drive=None, *args, **kwargs):
self.variable_output_drive = variable_output_drive
super().__init__(*args, **kwargs)
if not self.predicates:
self.predicates = ['default']
def _prepare_predicates(self, predicates):
"""Validate a configuration dictionary and add an output lane if necessary.
Args:
predicates (:obj:`list`): A list of predicates in predicate/outputLane form or as predicates.
Returns:
A corresponding :obj:`list` of dictionaries containing the predicate and outputLane for each entry.
"""
formatted_predicates = []
for predicate in predicates:
if not isinstance(predicate, dict):
predicate = {'predicate': str(predicate)}
output_lane = predicate.get(
'outputLane', '{}OutputLane{}'.format(self._data['instanceName'], str(uuid4())).replace('-', '_')
)
if not predicate.get('predicate', None):
raise ValueError('Output Lane drives should have a predicate key.')
if not predicate.get('outputLane', None):
output_lane_config = {'outputLane': output_lane}
# We make sure that the predicate is in the form {'outputLane': 'value', 'predicate': 'value'} to be
# consistent with the UI
predicate = {**output_lane_config, **predicate}
formatted_predicates.append(predicate)
return formatted_predicates
def _set_config(self, config_property, value):
# Check whether the configuration asked for is driving the output lanes.
if self.variable_output_drive == config_property.config_name:
if isinstance(value, list):
self.predicates = value
else:
raise ValueError('Configuration must be a list of predicates or a list of dictionaries.')
else:
super()._set_config(config_property, value)
def _disconnect_and_remove_predicate(self, i):
deleted_predicate = self.predicates.pop(i)
output_lane = deleted_predicate['outputLane']
output_lane_index = self.output_lanes.index(output_lane)
self.disconnect_output_lanes(output_lane_index=output_lane_index)
self._data['outputLanes'].remove(output_lane)
self.output_streams -= 1
@property
def predicates(self):
"""
Get the predicate list for this stage.
Predicates define the output lanes number and behaviour.
Set it using either a list of predicates or a list of the full dictionary specifying the output lane.
Example:
stage.predicates = ['>0']
stage.predicates = ['>0', 'default']
stage.predicates = [{'predicate':'>0', 'outputLane': 'lane1'}]
Returns:
A :obj:`list` of the predicates.
"""
return self.configuration[self.variable_output_drive]
@predicates.setter
def predicates(self, predicates):
predicates = self._prepare_predicates(predicates)
# Create default condition if not present
if not [predicate for predicate in predicates if predicate['predicate'] == 'default']:
default_predicate = self._prepare_predicates(['default'])
predicates.extend(default_predicate)
if not predicates[-1]['predicate'] == 'default':
raise ValueError('The default predicate must be placed at the end of the list.')
self.disconnect_output_lanes(all_stages=True)
self.output_streams = 0
self._data['outputLanes'] = []
self._output_lane_idx = 0
# Create output lanes in parent stage for each predicate
for predicate in predicates:
self._data['outputLanes'].append(predicate['outputLane'])
self.output_streams += 1
self.configuration[self.variable_output_drive] = predicates
def add_predicates(self, predicates):
"""Add a predicate.
Example:
stage.add_predicates(['>0'])
stage.add_predicates([{'predicate':'>0', 'outputLane':'lane1'}])
stage.add_predicates(['>0' ,'=0'])
Args:
predicates (:obj:`list`): The list of predicates to add.
"""
if not isinstance(predicates, list):
raise ValueError('Predicates should be a list.')
formatted_predicates = self._prepare_predicates(predicates)
for new_predicate in formatted_predicates:
self._data['outputLanes'].insert(0, new_predicate['outputLane'])
self.predicates.insert(0, new_predicate)
self.output_streams += 1
def remove_predicate(self, predicate):
"""Remove a predicate.
Example:
stage.remove_predicate(stage.predicates[0])
stage.remove_predicate({'predicate':'>0', 'outputLane':'lane1'})
Args:
predicate (:obj:`dict`): The predicate to delete as a dictionary including the outputLane.
"""
clean_predicate = next(iter(self._prepare_predicates([predicate])), {})
if not clean_predicate:
raise ValueError("Need to specify a predicate")
if clean_predicate.get('predicate') == 'default':
raise ValueError("Can't delete the default predicate.")
for i, found_predicate in enumerate(self.predicates):
if found_predicate == clean_predicate:
self._disconnect_and_remove_predicate(i)
return
raise ValueError("Can't find target predicate in the predicates list.")
