dtlpy 1.115.44__py3-none-any.whl → 1.116.6__py3-none-any.whl

dtlpy/entities/node.py

@@ -1,1005 +1,1005 @@
-import warnings
-
-import inspect
-import json
-import logging
-import uuid
-from typing import Callable
-from enum import Enum
-from typing import List
-import datetime
-
-from .. import entities, assets, repositories, PlatformException
-
-NODE_SIZE = (200, 87)
-
-logger = logging.getLogger(name='dtlpy')
-
-
-class PipelineConnectionPort:
-    def __init__(self, node_id: str, port_id: str):
-        self.node_id = node_id
-        self.port_id = port_id
-
-    @staticmethod
-    def from_json(_json: dict):
-        return PipelineConnectionPort(
-            node_id=_json.get('nodeId', None),
-            port_id=_json.get('portId', None),
-        )
-
-    def to_json(self):
-        _json = {
-            'nodeId': self.node_id,
-            'portId': self.port_id,
-        }
-        return _json
-
-
-class PipelineConnection:
-    def __init__(self,
-                 source: PipelineConnectionPort,
-                 target: PipelineConnectionPort,
-                 filters: entities.Filters,
-                 action: str = None
-                 ):
-        """
-        :param PipelineConnectionPort source: the source pipeline connection
-        :param PipelineConnectionPort target: the target pipeline connection
-        :param entities.Filters filters: condition for the connection between the nodes
-        :param str action: the action that move the input when it happen
-        """
-        self.source = source
-        self.target = target
-        self.filters = filters
-        self.action = action
-
-    @staticmethod
-    def from_json(_json: dict):
-        condition = _json.get('condition', None)
-        if condition:
-            condition = json.loads(condition)
-        return PipelineConnection(
-            source=PipelineConnectionPort.from_json(_json=_json.get('src', None)),
-            target=PipelineConnectionPort.from_json(_json=_json.get('tgt', None)),
-            filters=condition,
-            action=_json.get('action', None)
-        )
-
-    def to_json(self):
-        _json = {
-            'src': self.source.to_json(),
-            'tgt': self.target.to_json(),
-        }
-        if self.action:
-            _json['action'] = self.action
-        if self.filters:
-            if isinstance(self.filters, entities.Filters):
-                filters = self.filters.prepare(query_only=True).get('filter', dict())
-            else:
-                filters = self.filters
-
-            _json['condition'] = json.dumps(filters)
-        return _json
-
-
-class PipelineNodeIO:
-    def __init__(self,
-                 input_type: entities.PackageInputType,
-                 name: str,
-                 display_name: str,
-                 port_id: str = None,
-                 color: tuple = None,
-                 port_percentage: int = None,
-                 default_value=None,
-                 variable_name: str = None,
-                 actions: list = None,
-                 description: str = None):
-        """
-        Pipeline Node
-
-        :param entities.PackageInputType input_type: entities.PackageInputType of the input type of the pipeline
-        :param str name: name of the input
-        :param str display_name: of the input
-        :param str port_id: port id
-        :param tuple color: tuple the display the color
-        :param int port_percentage: port percentage
-        :param str action: the action that move the input when it happen
-        :param default_value: default value of the input
-        :param list actions: the actions list that move the input when it happen
-        """
-        self.port_id = port_id if port_id else str(uuid.uuid4())
-        self.input_type = input_type
-        self.name = name
-        self.color = color
-        self.display_name = display_name
-        self.port_percentage = port_percentage
-        self.default_value = default_value
-        self.variable_name = variable_name
-        self.description = description
-        self.actions = actions
-
-    @staticmethod
-    def from_json(_json: dict):
-        return PipelineNodeIO(
-            port_id=_json.get('portId', None),
-            input_type=_json.get('type', None),
-            name=_json.get('name', None),
-            color=_json.get('color', None),
-            display_name=_json.get('displayName', None),
-            port_percentage=_json.get('portPercentage', None),
-            default_value=_json.get('defaultValue', None),
-            variable_name=_json.get('variableName', None),
-            actions=_json.get('actions', None),
-            description=_json.get('description', None),
-        )
-
-    def to_json(self):
-        _json = {
-            'portId': self.port_id,
-            'type': self.input_type,
-            'name': self.name,
-            'color': self.color,
-            'displayName': self.display_name,
-            'variableName': self.variable_name,
-            'portPercentage': self.port_percentage,
-        }
-
-        if self.actions:
-            _json['actions'] = self.actions
-        if self.default_value:
-            _json['defaultValue'] = self.default_value
-        if self.description:
-            _json['description'] = self.description
-        return _json
-
-
-class PipelineNodeType(str, Enum):
-    TASK = 'task'
-    CODE = 'code'
-    FUNCTION = 'function'
-    STORAGE = 'storage'
-    ML = 'ml'
-
-
-class PipelineNameSpace:
-    def __init__(self, function_name, project_name, module_name=None, service_name=None, package_name=None):
-        self.function_name = function_name
-        self.project_name = project_name
-        self.module_name = module_name
-        self.service_name = service_name
-        self.package_name = package_name
-
-    def to_json(self):
-        _json = {
-            "functionName": self.function_name,
-            "projectName": self.project_name
-        }
-        if self.module_name:
-            _json['moduleName'] = self.module_name
-
-        if self.service_name:
-            _json['serviceName'] = self.service_name
-
-        if self.package_name:
-            _json['packageName'] = self.package_name
-        return _json
-
-    @staticmethod
-    def from_json(_json: dict):
-        return PipelineNameSpace(
-            function_name=_json.get('functionName'),
-            project_name=_json.get('projectName'),
-            module_name=_json.get('moduleName', None),
-            service_name=_json.get('serviceName', None),
-            package_name=_json.get('packageName', None)
-        )
-
-
-class PipelineNode:
-    def __init__(self,
-                 name: str,
-                 node_id: str,
-                 outputs: list,
-                 inputs: list,
-                 node_type: PipelineNodeType,
-                 namespace: PipelineNameSpace,
-                 project_id: str,
-                 metadata: dict = None,
-                 config: dict = None,
-                 position: tuple = (1, 1),
-                 app_id: str = None,
-                 dpk_name: str = None,
-                 app_name: str = None,
-                 ):
-        """
-        :param str name: node name
-        :param str node_id: node id
-        :param list outputs: list of PipelineNodeIO outputs
-        :param list inputs: list of PipelineNodeIO inputs
-        :param dict metadata: dict of the metadata of the node
-        :param PipelineNodeType node_type: task, code, function
-        :param PipelineNameSpace namespace: PipelineNameSpace of the node space
-        :param str project_id: project id
-        :param dict config: for the code node dict in format { package: {code : the_code}}
-        :param tuple position: tuple of the node place
-        :param str app_id: app id
-        :param str dpk_name: dpk name
-        :param str app_name: app name
-        """
-        self.name = name
-        self.node_id = node_id
-        self.outputs = outputs
-        self.inputs = inputs
-        self.metadata = metadata if metadata is not None else {}
-        self.node_type = node_type
-        self.namespace = namespace
-        self.project_id = project_id
-        self.config = config
-        self.position = position
-        self.app_id = app_id
-        self.dpk_name = dpk_name
-        self.app_name = app_name
-        self._pipeline = None
-
-    @property
-    def position(self):
-        position_tuple = (self.metadata['position']['x'],
-                          self.metadata['position']['y'])
-        return position_tuple
-
-    @position.setter
-    def position(self, position):
-        self.metadata['position'] = \
-            {
-                "x": position[0] * 1.7 * NODE_SIZE[0] + NODE_SIZE[0] / 2,
-                "y": position[1] * 1.5 * NODE_SIZE[1] + NODE_SIZE[1],
-                "z": 0
-            }
-
-    def _default_io(self, actions: list = None) -> PipelineNodeIO:
-        """
-        Create a default item pipeline input
-
-        :param str actions:  the action that move the input when it happen
-        :return PipelineNodeIO: the default item PipelineNodeIO
-        """
-        default_io = PipelineNodeIO(port_id=str(uuid.uuid4()),
-                                    input_type=entities.PackageInputType.ITEM,
-                                    name='item',
-                                    color=None,
-                                    display_name=actions[0] if actions else 'item',
-                                    actions=actions)
-        return default_io
-
-    @staticmethod
-    def from_json(_json: dict):
-        inputs = [PipelineNodeIO.from_json(_json=i_input) for i_input in _json.get('inputs', list())]
-        outputs = [PipelineNodeIO.from_json(_json=i_output) for i_output in _json.get('outputs', list())]
-        namespace = PipelineNameSpace.from_json(_json.get('namespace', {}))
-        metadata = _json.get('metadata', {})
-        position = ((metadata['position']['x'] - NODE_SIZE[0] / 2) / (1.7 * NODE_SIZE[0]),
-                    (metadata['position']['y'] - NODE_SIZE[1]) / (1.5 * NODE_SIZE[1]))
-        return PipelineNode(
-            name=_json.get('name', None),
-            node_id=_json.get('id', None),
-            outputs=outputs,
-            inputs=inputs,
-            metadata=metadata,
-            node_type=_json.get('type', None),
-            namespace=namespace,
-            project_id=_json.get('projectId', None),
-            config=_json.get('config', None),
-            position=position,
-            app_id=_json.get('appId', None),
-            dpk_name=_json.get('dpkName', None),
-            app_name=_json.get('appName', None),
-        )
-
-    def to_json(self):
-        _json = {
-            'name': self.name,
-            'id': self.node_id,
-            'outputs': [_io.to_json() for _io in self.outputs],
-            'inputs': [_io.to_json() for _io in self.inputs],
-            'metadata': self.metadata,
-            'type': self.node_type,
-            'namespace': self.namespace.to_json(),
-            'projectId': self.project_id,
-            'dpkName': self.dpk_name,
-            'appName': self.app_name,
-        }
-        if self.config is not None:
-            _json['config'] = self.config
-        if self.app_id is not None:
-            _json['appId'] = self.app_id
-        return _json
-
-    def is_root(self):
-        if self._pipeline is not None:
-            for node in self._pipeline.start_nodes:
-                if self.node_id == node.get('nodeId', None) and node.get('type', None) == 'root':
-                    return True
-        return False
-
-    def _build_connection(self,
-                          node,
-                          source_port: PipelineNodeIO = None,
-                          target_port: PipelineNodeIO = None,
-                          filters: entities.Filters = None,
-                          action: str = None) -> PipelineConnection:
-        """
-        Build connection between the current node and the target node use the given ports
-
-        :param PipelineNode node: the node to connect to it
-        :param PipelineNodeIO source_port: the source PipelineNodeIO input port
-        :param PipelineNodeIO target_port: the target PipelineNodeIO output port
-        :param entities.Filters filters: condition for the connection between the nodes
-        :param str action:  the action that move the input when it happen
-        :return: the connection between the nodes
-        """
-        if source_port is None and self.outputs:
-            source_port = self.outputs[0]
-
-        if target_port is None and node.inputs:
-            target_port = node.inputs[0]
-
-        if node.is_root():
-            self._pipeline.set_start_node(self)
-
-        source_connection = PipelineConnectionPort(node_id=self.node_id, port_id=source_port.port_id)
-        target_connection = PipelineConnectionPort(node_id=node.node_id, port_id=target_port.port_id)
-        if action is None and source_port.actions is not None and source_port.actions != []:
-            action = source_port.actions[0]
-        connection = PipelineConnection(source=source_connection, target=target_connection, filters=filters,
-                                        action=action)
-        return connection
-
-    def connect(self,
-                node,
-                source_port: PipelineNodeIO = None,
-                target_port: PipelineNodeIO = None,
-                filters=None,
-                action: str = None):
-        """
-        Build connection between the current node and the target node use the given ports
-
-        :param PipelineNode node: the node to connect to it
-        :param PipelineNodeIO source_port: the source PipelineNodeIO input port
-        :param PipelineNodeIO target_port: the target PipelineNodeIO output port
-        :param entities.Filters filters: condition for the connection between the nodes
-        :param str action:  the action that move the input when it happen
-        :return: the connected node
-        """
-        if self._pipeline is None:
-            raise Exception("must add the node to the pipeline first, e.g pipeline.nodes.add(node)")
-        connection = self._build_connection(node=node,
-                                            source_port=source_port,
-                                            target_port=target_port,
-                                            filters=filters,
-                                            action=action)
-        self._pipeline.connections.append(connection)
-        self._pipeline.nodes.add(node)
-        return node
-
-    def disconnect(self,
-                   node,
-                   source_port: PipelineNodeIO = None,
-                   target_port: PipelineNodeIO = None) -> bool:
-        """
-        remove connection between the current node and the target node use the given ports
-
-        :param PipelineNode node: the node to connect to it
-        :param PipelineNodeIO source_port: the source PipelineNodeIO input port
-        :param PipelineNodeIO target_port: the target PipelineNodeIO output port
-        :return: true if success and false if not
-        """
-        if self._pipeline is None:
-            raise Exception("must add the node to the pipeline first, e.g pipeline.nodes.add(node)")
-        connection = self._build_connection(node=node,
-                                            source_port=source_port,
-                                            target_port=target_port,
-                                            filters=None)
-
-        current_connection = connection.to_json()
-        if 'condition' in current_connection:
-            current_connection = current_connection.pop('condition')
-
-        for connection_index in range(len(self._pipeline.connections)):
-            pipeline_connection = self._pipeline.connections[connection_index].to_json()
-            if 'condition' in pipeline_connection:
-                pipeline_connection = pipeline_connection.pop('condition')
-
-            if current_connection == pipeline_connection:
-                self._pipeline.connections.pop(connection_index)
-                return True
-        logger.warning('do not found a connection')
-        return False
-
-    def add_trigger(self,
-                    trigger_type: entities.TriggerType = entities.TriggerType.EVENT,
-                    filters=None,
-                    resource: entities.TriggerResource = entities.TriggerResource.ITEM,
-                    actions: entities.TriggerAction = entities.TriggerAction.CREATED,
-                    execution_mode: entities.TriggerExecutionMode = entities.TriggerExecutionMode.ONCE,
-                    cron: str = None,
-                    ):
-        """
-        Create a Trigger. Can create two types: a cron trigger or an event trigger.
-        Inputs are different for each type
-
-        Inputs for all types:
-
-        :param trigger_type: can be cron or event. use enum dl.TriggerType for the full list
-
-        Inputs for event trigger:
-        :param filters: optional - Item/Annotation metadata filters, default = none
-        :param resource: optional - Dataset/Item/Annotation/ItemStatus, default = Item
-        :param actions: optional - Created/Updated/Deleted, default = create
-        :param execution_mode: how many time trigger should be activate. default is "Once". enum dl.TriggerExecutionMode
-
-        Inputs for cron trigger:
-        :param str cron: cron spec specifying when it should run. more information: https://en.wikipedia.org/wiki/Cron
-
-        :return: Trigger entity
-        """
-        if self._pipeline is None:
-            raise Exception("must add the node to the pipeline first, e.g pipeline.nodes.add(node)")
-
-        if not isinstance(actions, list):
-            actions = [actions]
-
-        if filters is None:
-            filters = {}
-        else:
-            filters = json.dumps(filters.prepare(query_only=True).get('filter', dict()))
-
-        if trigger_type == entities.TriggerType.EVENT:
-            spec = {
-                'filter': filters,
-                'resource': resource,
-                'executionMode': execution_mode,
-                'actions': actions
-            }
-        elif trigger_type == entities.TriggerType.CRON:
-            spec = {
-                'cron': cron,
-            }
-        else:
-            raise ValueError('Unknown trigger type: "{}". Use dl.TriggerType for known types'.format(trigger_type))
-
-        trigger = {
-            "type": trigger_type,
-            "spec": spec
-        }
-
-        set_trigger = False
-        for pipe_node in self._pipeline.start_nodes:
-            if pipe_node['nodeId'] == self.node_id:
-                set_trigger = True
-                pipe_node['trigger'] = trigger
-
-        if not set_trigger:
-            self._pipeline.start_nodes.append(
-                {
-                    "nodeId": self.node_id,
-                    "type": "trigger",
-                    'trigger': trigger
-                }
-            )
-
-
-class CodeNode(PipelineNode):
-    def __init__(self,
-                 name: str,
-                 project_id: str,
-                 project_name: str,
-                 method: Callable,
-                 outputs: List[PipelineNodeIO] = None,
-                 inputs: List[PipelineNodeIO] = None,
-                 position: tuple = (1, 1),
-                 ):
-        """
-        :param str name: node name
-        :param str project_id: project id
-        :param str project_name: project name
-        :param Callable method: function to deploy
-        :param list outputs: list of PipelineNodeIO outputs
-        :param list inputs: list of PipelineNodeIO inputs
-        :param tuple position: tuple of the node place
-        """
-        if inputs is None:
-            inputs = [self._default_io()]
-        if outputs is None:
-            outputs = [self._default_io()]
-
-        if method is None or not isinstance(method, Callable):
-            raise Exception('must provide a function as input')
-        else:
-            function_code = self._build_code_from_func(method)
-            function_name = method.__name__
-
-        super().__init__(name=name,
-                         node_id=str(uuid.uuid4()),
-                         outputs=outputs,
-                         inputs=inputs,
-                         metadata={},
-                         node_type=PipelineNodeType.CODE,
-                         namespace=PipelineNameSpace(function_name=function_name, project_name=project_name),
-                         project_id=project_id,
-                         position=position)
-
-        self.config = {
-            "package":
-                {
-                    "code": function_code,
-                    "name": function_name,
-                    "type": "code"
-                }
-        }
-
-    def _build_code_from_func(self, func: Callable) -> str:
-        """
-        Build a code format from the given function
-
-        :param Callable func: function to deploy
-        :return: a string the display the code with the package format
-        """
-        with open(assets.paths.PARTIAL_MAIN_FILEPATH, 'r') as f:
-            main_string = f.read()
-        lines = inspect.getsourcelines(func)
-
-        tabs_diff = lines[0][0].count('    ') - 1
-        for line_index in range(len(lines[0])):
-            line_tabs = lines[0][line_index].count('    ') - tabs_diff
-            lines[0][line_index] = ('    ' * line_tabs) + lines[0][line_index].strip() + '\n'
-
-        method_func_string = "".join(lines[0])
-
-        code = '{}\n{}\n    @staticmethod\n{}'.format('', main_string,
-                                                      method_func_string)
-        return code
-
-    @staticmethod
-    def from_json(_json: dict):
-        parent = PipelineNode.from_json(_json)
-        parent.__class__ = CodeNode
-        return parent
-
-
-class TaskNode(PipelineNode):
-    def __init__(self,
-                 name: str,
-                 project_id: str,
-                 dataset_id: str,
-                 recipe_title: str,
-                 recipe_id: str,
-                 task_owner: str,
-                 workload: List[entities.WorkloadUnit],
-                 task_type: str = 'annotation',
-                 position: tuple = (1, 1),
-                 actions: list = None,
-                 repeatable: bool = True,
-                 batch_size=None,
-                 max_batch_workload=None,
-                 priority=entities.TaskPriority.MEDIUM,
-                 due_date=None,
-                 consensus_task_type=None,
-                 consensus_percentage=None,
-                 consensus_assignees=None,
-                 groups=None
-                 ):
-        """
-        :param str name: node name
-        :param str project_id: project id
-        :param str dataset_id: dataset id
-        :param str recipe_title: recipe title
-        :param str recipe_id: recipe id
-        :param str task_owner: email of task owner
-        :param List[WorkloadUnit] workload: list of WorkloadUnit
-        :param str task_type: 'annotation' or 'qa'
-        :param tuple position: tuple of the node place
-        :param list actions: list of task actions
-        :param bool repeatable: can repeat in the item
-        :param int groups: groups to assign the task to
-        :param int batch_size: Pulling batch size (items) . Restrictions - Min 3, max 100 - for create pulling task
-        :param int max_batch_workload: Max items in assignment . Restrictions - Min batchSize + 2 , max batchSize * 2 - for create pulling task
-        :param entities.TaskPriority priority: priority of the task options in entities.TaskPriority
-        :param float due_date: date by which the task should be finished; for example, due_date = datetime.datetime(day= 1, month= 1, year= 2029).timestamp()
-        :param entities.ConsensusTaskType consensus_task_type: consensus_task_type of the task options in entities.ConsensusTaskType
-        :param int consensus_percentage: percentage of items to be copied to multiple annotators (consensus items)
-        :param int consensus_assignees: the number of different annotators per item (number of copies per item)
-        """
-        if actions is None or actions == []:
-            actions = []
-            if task_type == 'qa':
-                if 'approve' not in actions:
-                    actions.append('approve')
-            else:
-                if 'complete' not in actions:
-                    actions.append('complete')
-            actions.append('discard')
-        else:
-            logger.warning(
-                "The 'actions' field was updated to override the system default actions for task (complete/approve, discard) if provided, due to a bug fix.")
-
-        inputs = [self._default_io()]
-
-        outputs = [self._default_io(actions=actions)]
-
-        if groups is not None:
-            if not isinstance(groups, list) or not all(isinstance(group, str) for group in groups):
-                raise ValueError('groups must be a list of strings')
-
-        super().__init__(name=name,
-                         node_id=str(uuid.uuid4()),
-                         outputs=outputs,
-                         inputs=inputs,
-                         metadata=dict(),
-                         node_type=PipelineNodeType.TASK,
-                         namespace=PipelineNameSpace(function_name="move_to_task",
-                                                     project_name="DataloopTasks",
-                                                     service_name="pipeline-utils"),
-                         project_id=project_id,
-                         position=position)
-
-        self.dataset_id = dataset_id
-        self.recipe_title = recipe_title
-        self.recipe_id = recipe_id
-        self.task_owner = task_owner
-        self.task_type = task_type
-        if not isinstance(workload, list):
-            workload = [workload]
-        self.workload = workload
-        self.repeatable = repeatable
-        if max_batch_workload:
-            self.max_batch_workload = max_batch_workload
-        if batch_size:
-            self.batch_size = batch_size
-        if consensus_task_type:
-            self.consensus_task_type = consensus_task_type
-        if consensus_percentage:
-            self.consensus_percentage = consensus_percentage
-        if consensus_assignees:
-            self.consensus_assignees = consensus_assignees
-        self.priority = priority
-        if due_date is None:
-            due_date = (datetime.datetime.now() + datetime.timedelta(days=7)).timestamp() * 1000
-        self.due_date = due_date
-        self.groups = groups
-
-    @property
-    def dataset_id(self):
-        return self.metadata['datasetId']
-
-    @dataset_id.setter
-    def dataset_id(self, dataset_id: str):
-        if not isinstance(dataset_id, str):
-            raise PlatformException('400', 'Param dataset_id must be of type string')
-        self.metadata['datasetId'] = dataset_id
-
-    @property
-    def groups(self):
-        return self.metadata.get('groups')
-
-    @groups.setter
-    def groups(self, groups: List[str]):
-        if groups is not None:
-            self.metadata['groups'] = groups
-
-    @property
-    def repeatable(self):
-        return self.metadata['repeatable']
-
-    @repeatable.setter
-    def repeatable(self, repeatable: bool):
-        if not isinstance(repeatable, bool):
-            raise PlatformException('400', 'Param repeatable must be of type bool')
-        self.metadata['repeatable'] = repeatable
-
-    @property
-    def recipe_title(self):
-        return self.metadata['recipeTitle']
-
-    @recipe_title.setter
-    def recipe_title(self, recipe_title: str):
-        if not isinstance(recipe_title, str):
-            raise PlatformException('400', 'Param recipe_title must be of type string')
-        self.metadata['recipeTitle'] = recipe_title
-
-    @property
-    def recipe_id(self):
-        return self.metadata['recipeId']
-
-    @recipe_id.setter
-    def recipe_id(self, recipe_id: str):
-        if not isinstance(recipe_id, str):
-            raise PlatformException('400', 'Param recipe_id must be of type string')
-        self.metadata['recipeId'] = recipe_id
-
-    @property
-    def task_owner(self):
-        return self.metadata['taskOwner']
-
-    @task_owner.setter
-    def task_owner(self, task_owner: str):
-        if not isinstance(task_owner, str):
-            raise PlatformException('400', 'Param task_owner must be of type string')
-        self.metadata['taskOwner'] = task_owner
-
-    @property
-    def task_type(self):
-        return self.metadata['taskType']
-
-    @task_type.setter
-    def task_type(self, task_type: str):
-        if not isinstance(task_type, str):
-            raise PlatformException('400', 'Param task_type must be of type string')
-        self.metadata['taskType'] = task_type
-
-    @property
-    def workload(self):
-        return self.metadata['workload']
-
-    @workload.setter
-    def workload(self, workload: list):
-        if not isinstance(workload, list):
-            workload = [workload]
-        self.metadata['workload'] = [val.to_json() for val in workload]
-
-    @property
-    def batch_size(self):
-        return self.metadata['batchSize']
-
-    @batch_size.setter
-    def batch_size(self, batch_size: int):
-        if not isinstance(batch_size, int):
-            raise PlatformException('400', 'Param batch_size must be of type int')
-        self.metadata['batchSize'] = batch_size
-
-    @property
-    def max_batch_workload(self):
-        return self.metadata['maxBatchWorkload']
-
-    @max_batch_workload.setter
-    def max_batch_workload(self, max_batch_workload: int):
-        if not isinstance(max_batch_workload, int):
-            raise PlatformException('400', 'Param max_batch_workload must be of type int')
-        self.metadata['maxBatchWorkload'] = max_batch_workload
-
-    @property
-    def consensus_task_type(self):
-        return self.metadata['consensusTaskType']
-
-    @consensus_task_type.setter
-    def consensus_task_type(self, consensus_task_type: entities.ConsensusTaskType):
-        if not isinstance(consensus_task_type, str) and not isinstance(consensus_task_type, entities.ConsensusTaskType):
-            raise PlatformException('400', 'Param consensus_task_type must be of type entities.ConsensusTaskType')
-        self.metadata['consensusTaskType'] = consensus_task_type
-
-    @property
-    def consensus_percentage(self):
-        return self.metadata['consensusPercentage']
-
-    @consensus_percentage.setter
-    def consensus_percentage(self, consensus_percentage: int):
-        if not isinstance(consensus_percentage, int):
-            raise PlatformException('400', 'Param consensus_percentage must be of type int')
-        self.metadata['consensusPercentage'] = consensus_percentage
-
-    @property
-    def consensus_assignees(self):
-        return self.metadata['consensusAssignees']
-
-    @consensus_assignees.setter
-    def consensus_assignees(self, consensus_assignees: int):
-        if not isinstance(consensus_assignees, int):
-            raise PlatformException('400', 'Param consensus_assignees must be of type int')
-        self.metadata['consensusAssignees'] = consensus_assignees
-
-    @property
-    def priority(self):
-        return self.metadata['priority']
-
-    @priority.setter
-    def priority(self, priority: entities.TaskPriority):
-        if not isinstance(priority, int) and not isinstance(priority, entities.TaskPriority):
-            raise PlatformException('400', 'Param priority must be of type entities.TaskPriority')
-        self.metadata['priority'] = priority
-
-    @property
-    def due_date(self):
-        return self.metadata['dueDate']
-
-    @due_date.setter
-    def due_date(self, due_date: float):
-        if not isinstance(due_date, float) and not isinstance(due_date, int):
-            raise PlatformException('400', 'Param due_date must be of type float or int')
-        self.metadata['dueDate'] = due_date
-
-    @staticmethod
-    def from_json(_json: dict):
-        parent = PipelineNode.from_json(_json)
-        parent.__class__ = TaskNode
-        return parent
-
-
-class FunctionNode(PipelineNode):
-    def __init__(self,
-                 name: str,
-                 service: entities.Service,
-                 function_name,
-                 position: tuple = (1, 1),
-                 project_id=None,
-                 project_name=None
-                 ):
-        """
-        :param str name: node name
-        :param entities.Service service: service to deploy
-        :param str function_name: function name
-        :param tuple position: tuple of the node place
-        """
-        self.service = service
-
-        if project_id is None:
-            project_id = service.project_id
-        if project_id != service.project_id:
-            logger.warning("the project id that provide different from the service project id")
-
-        if project_name is None:
-            try:
-                project = repositories.Projects(client_api=self.service._client_api).get(project_id=project_id,
-                                                                                         log_error=False)
-                project_name = project.name
-            except:
-                logger.warning(
-                    'Service project not found using DataloopTasks project.'
-                    ' If this is incorrect please provide project_name param.')
-                project_name = 'DataloopTasks'
-        inputs = []
-        outputs = []
-        package = self.service.package
-        modules = []
-        if isinstance(package, entities.Package):
-            modules = package.modules
-        elif isinstance(package, entities.Dpk):
-            modules = package.components.modules
-        for model in modules:
-            if model.name == self.service.module_name:
-                for func in model.functions:
-                    if func.name == function_name:
-                        inputs = self._convert_from_function_io_to_pipeline_io(func.inputs)
-                        outputs = self._convert_from_function_io_to_pipeline_io(func.outputs)
-
-        namespace = PipelineNameSpace(
-            function_name=function_name,
-            service_name=self.service.name,
-            module_name=self.service.module_name,
-            package_name=self.service.package.name,
-            project_name=project_name
-        )
-        super().__init__(name=name,
-                         node_id=str(uuid.uuid4()),
-                         outputs=outputs,
-                         inputs=inputs,
-                         metadata={},
-                         node_type=PipelineNodeType.FUNCTION,
-                         namespace=namespace,
-                         project_id=service.project_id,
-                         position=position)
-
-    def _convert_from_function_io_to_pipeline_io(self, function_io: List[entities.FunctionIO]) -> List[PipelineNodeIO]:
-        """
-        Get a list of FunctionIO and convert them to PipelineIO
-        :param List[entities.FunctionIO] function_io: list of functionIO
-        :return: list of PipelineIO
-        """
-        pipeline_io = []
-        for single_input in function_io:
-            pipeline_io.append(
-                PipelineNodeIO(port_id=str(uuid.uuid4()),
-                               input_type=single_input.type,
-                               name=single_input.name,
-                               color=None,
-                               display_name=single_input.name,
-                               default_value=single_input.value,
-                               actions=single_input.actions if single_input.actions is not None else []))
-        return pipeline_io
-
-    @staticmethod
-    def from_json(_json: dict):
-        parent = PipelineNode.from_json(_json)
-        parent.__class__ = FunctionNode
-        return parent
-
-
-class DatasetNode(PipelineNode):
-    def __init__(self,
-                 name: str,
-                 project_id: str,
-                 dataset_id: str,
-                 dataset_folder: str = None,
-                 load_existing_data: bool = False,
-                 data_filters: entities.Filters = None,
-                 position: tuple = (1, 1)):
-        """
-        :param str name: node name
-        :param str project_id: project id
-        :param str dataset_id: dataset id
-        :param str dataset_folder: folder in dataset to work in it
-        :param bool load_existing_data: optional - enable to automatically load existing data into the
-                                        pipeline (executions) upon activation, based on the defined dataset,
-                                        folder, and data_filters.
-        :param entities.Filters data_filters: optional - filters entity or a dictionary containing filters parameters.
-                                              Use to filter the data items to be loaded when load_existing_data
-                                              is enabled.
-        :param tuple position: tuple of the node place
-        """
-        inputs = [self._default_io()]
-        outputs = [self._default_io()]
-        super().__init__(name=name,
-                         node_id=str(uuid.uuid4()),
-                         outputs=outputs,
-                         inputs=inputs,
-                         metadata={},
-                         node_type=PipelineNodeType.STORAGE,
-                         namespace=PipelineNameSpace(function_name="dataset_handler",
-                                                     project_name="DataloopTasks",
-                                                     service_name="pipeline-utils"),
-                         project_id=project_id,
-                         position=position)
-        self.dataset_id = dataset_id
-        self.dataset_folder = dataset_folder
-        self.load_existing_data = load_existing_data
-        self.data_filters = data_filters
-
-    @property
-    def dataset_id(self):
-        return self.metadata['datasetId']
-
-    @dataset_id.setter
-    def dataset_id(self, dataset_id: str):
-        self.metadata['datasetId'] = dataset_id
-
-    @property
-    def dataset_folder(self):
-        return self.metadata.get('dir', None)
-
-    @dataset_folder.setter
-    def dataset_folder(self, dataset_folder: str):
-        if dataset_folder is not None:
-            if not dataset_folder.startswith("/"):
-                dataset_folder = '/' + dataset_folder
-            self.metadata['dir'] = dataset_folder
-
-    @property
-    def load_existing_data(self):
-        return self.metadata.get('triggerToPipeline', {}).get('active', False)
-
-    @load_existing_data.setter
-    def load_existing_data(self, load_existing_data: bool):
-        if load_existing_data:
-            self.metadata.setdefault('triggerToPipeline', {})['active'] = True
-        else:
-            self.metadata.pop('triggerToPipeline', None)
-
-    @property
-    def data_filters(self):
-        data_filters = self.metadata.get('triggerToPipeline', {}).get('filter', None)
-        if data_filters:
-            data_filters = entities.Filters(custom_filter=json.loads(data_filters))
-        return data_filters
-
-    @data_filters.setter
-    def data_filters(self, data_filters: entities.Filters):
-        if data_filters is None:
-            filters = None
-        else:
-            filters = json.dumps(data_filters.prepare(query_only=True).get('filter'))
-        self.metadata.setdefault('triggerToPipeline', {})['filter'] = filters
-
-    @staticmethod
-    def from_json(_json: dict):
-        parent = PipelineNode.from_json(_json)
-        parent.__class__ = DatasetNode
-        return parent
+import warnings
+
+import inspect
+import json
+import logging
+import uuid
+from typing import Callable
+from enum import Enum
+from typing import List
+import datetime
+
+from .. import entities, assets, repositories, PlatformException
+
+NODE_SIZE = (200, 87)
+
+logger = logging.getLogger(name='dtlpy')
+
+
+class PipelineConnectionPort:
+    def __init__(self, node_id: str, port_id: str):
+        self.node_id = node_id
+        self.port_id = port_id
+
+    @staticmethod
+    def from_json(_json: dict):
+        return PipelineConnectionPort(
+            node_id=_json.get('nodeId', None),
+            port_id=_json.get('portId', None),
+        )
+
+    def to_json(self):
+        _json = {
+            'nodeId': self.node_id,
+            'portId': self.port_id,
+        }
+        return _json
+
+
+class PipelineConnection:
+    def __init__(self,
+                 source: PipelineConnectionPort,
+                 target: PipelineConnectionPort,
+                 filters: entities.Filters,
+                 action: str = None
+                 ):
+        """
+        :param PipelineConnectionPort source: the source pipeline connection
+        :param PipelineConnectionPort target: the target pipeline connection
+        :param entities.Filters filters: condition for the connection between the nodes
+        :param str action: the action that move the input when it happen
+        """
+        self.source = source
+        self.target = target
+        self.filters = filters
+        self.action = action
+
+    @staticmethod
+    def from_json(_json: dict):
+        condition = _json.get('condition', None)
+        if condition:
+            condition = json.loads(condition)
+        return PipelineConnection(
+            source=PipelineConnectionPort.from_json(_json=_json.get('src', None)),
+            target=PipelineConnectionPort.from_json(_json=_json.get('tgt', None)),
+            filters=condition,
+            action=_json.get('action', None)
+        )
+
+    def to_json(self):
+        _json = {
+            'src': self.source.to_json(),
+            'tgt': self.target.to_json(),
+        }
+        if self.action:
+            _json['action'] = self.action
+        if self.filters:
+            if isinstance(self.filters, entities.Filters):
+                filters = self.filters.prepare(query_only=True).get('filter', dict())
+            else:
+                filters = self.filters
+
+            _json['condition'] = json.dumps(filters)
+        return _json
+
+
+class PipelineNodeIO:
+    def __init__(self,
+                 input_type: entities.PackageInputType,
+                 name: str,
+                 display_name: str,
+                 port_id: str = None,
+                 color: tuple = None,
+                 port_percentage: int = None,
+                 default_value=None,
+                 variable_name: str = None,
+                 actions: list = None,
+                 description: str = None):
+        """
+        Pipeline Node
+
+        :param entities.PackageInputType input_type: entities.PackageInputType of the input type of the pipeline
+        :param str name: name of the input
+        :param str display_name: of the input
+        :param str port_id: port id
+        :param tuple color: tuple the display the color
+        :param int port_percentage: port percentage
+        :param str action: the action that move the input when it happen
+        :param default_value: default value of the input
+        :param list actions: the actions list that move the input when it happen
+        """
+        self.port_id = port_id if port_id else str(uuid.uuid4())
+        self.input_type = input_type
+        self.name = name
+        self.color = color
+        self.display_name = display_name
+        self.port_percentage = port_percentage
+        self.default_value = default_value
+        self.variable_name = variable_name
+        self.description = description
+        self.actions = actions
+
+    @staticmethod
+    def from_json(_json: dict):
+        return PipelineNodeIO(
+            port_id=_json.get('portId', None),
+            input_type=_json.get('type', None),
+            name=_json.get('name', None),
+            color=_json.get('color', None),
+            display_name=_json.get('displayName', None),
+            port_percentage=_json.get('portPercentage', None),
+            default_value=_json.get('defaultValue', None),
+            variable_name=_json.get('variableName', None),
+            actions=_json.get('actions', None),
+            description=_json.get('description', None),
+        )
+
+    def to_json(self):
+        _json = {
+            'portId': self.port_id,
+            'type': self.input_type,
+            'name': self.name,
+            'color': self.color,
+            'displayName': self.display_name,
+            'variableName': self.variable_name,
+            'portPercentage': self.port_percentage,
+        }
+
+        if self.actions:
+            _json['actions'] = self.actions
+        if self.default_value:
+            _json['defaultValue'] = self.default_value
+        if self.description:
+            _json['description'] = self.description
+        return _json
+
+
+class PipelineNodeType(str, Enum):
+    TASK = 'task'
+    CODE = 'code'
+    FUNCTION = 'function'
+    STORAGE = 'storage'
+    ML = 'ml'
+
+
+class PipelineNameSpace:
+    def __init__(self, function_name, project_name, module_name=None, service_name=None, package_name=None):
+        self.function_name = function_name
+        self.project_name = project_name
+        self.module_name = module_name
+        self.service_name = service_name
+        self.package_name = package_name
+
+    def to_json(self):
+        _json = {
+            "functionName": self.function_name,
+            "projectName": self.project_name
+        }
+        if self.module_name:
+            _json['moduleName'] = self.module_name
+
+        if self.service_name:
+            _json['serviceName'] = self.service_name
+
+        if self.package_name:
+            _json['packageName'] = self.package_name
+        return _json
+
+    @staticmethod
+    def from_json(_json: dict):
+        return PipelineNameSpace(
+            function_name=_json.get('functionName'),
+            project_name=_json.get('projectName'),
+            module_name=_json.get('moduleName', None),
+            service_name=_json.get('serviceName', None),
+            package_name=_json.get('packageName', None)
+        )
+
+
+class PipelineNode:
+    def __init__(self,
+                 name: str,
+                 node_id: str,
+                 outputs: list,
+                 inputs: list,
+                 node_type: PipelineNodeType,
+                 namespace: PipelineNameSpace,
+                 project_id: str,
+                 metadata: dict = None,
+                 config: dict = None,
+                 position: tuple = (1, 1),
+                 app_id: str = None,
+                 dpk_name: str = None,
+                 app_name: str = None,
+                 ):
+        """
+        :param str name: node name
+        :param str node_id: node id
+        :param list outputs: list of PipelineNodeIO outputs
+        :param list inputs: list of PipelineNodeIO inputs
+        :param dict metadata: dict of the metadata of the node
+        :param PipelineNodeType node_type: task, code, function
+        :param PipelineNameSpace namespace: PipelineNameSpace of the node space
+        :param str project_id: project id
+        :param dict config: for the code node dict in format { package: {code : the_code}}
+        :param tuple position: tuple of the node place
+        :param str app_id: app id
+        :param str dpk_name: dpk name
+        :param str app_name: app name
+        """
+        self.name = name
+        self.node_id = node_id
+        self.outputs = outputs
+        self.inputs = inputs
+        self.metadata = metadata if metadata is not None else {}
+        self.node_type = node_type
+        self.namespace = namespace
+        self.project_id = project_id
+        self.config = config
+        self.position = position
+        self.app_id = app_id
+        self.dpk_name = dpk_name
+        self.app_name = app_name
+        self._pipeline = None
+
+    @property
+    def position(self):
+        position_tuple = (self.metadata['position']['x'],
+                          self.metadata['position']['y'])
+        return position_tuple
+
+    @position.setter
+    def position(self, position):
+        self.metadata['position'] = \
+            {
+                "x": position[0] * 1.7 * NODE_SIZE[0] + NODE_SIZE[0] / 2,
+                "y": position[1] * 1.5 * NODE_SIZE[1] + NODE_SIZE[1],
+                "z": 0
+            }
+
+    def _default_io(self, actions: list = None) -> PipelineNodeIO:
+        """
+        Create a default item pipeline input
+
+        :param str actions:  the action that move the input when it happen
+        :return PipelineNodeIO: the default item PipelineNodeIO
+        """
+        default_io = PipelineNodeIO(port_id=str(uuid.uuid4()),
+                                    input_type=entities.PackageInputType.ITEM,
+                                    name='item',
+                                    color=None,
+                                    display_name=actions[0] if actions else 'item',
+                                    actions=actions)
+        return default_io
+
+    @staticmethod
+    def from_json(_json: dict):
+        inputs = [PipelineNodeIO.from_json(_json=i_input) for i_input in _json.get('inputs', list())]
+        outputs = [PipelineNodeIO.from_json(_json=i_output) for i_output in _json.get('outputs', list())]
+        namespace = PipelineNameSpace.from_json(_json.get('namespace', {}))
+        metadata = _json.get('metadata', {})
+        position = ((metadata['position']['x'] - NODE_SIZE[0] / 2) / (1.7 * NODE_SIZE[0]),
+                    (metadata['position']['y'] - NODE_SIZE[1]) / (1.5 * NODE_SIZE[1]))
+        return PipelineNode(
+            name=_json.get('name', None),
+            node_id=_json.get('id', None),
+            outputs=outputs,
+            inputs=inputs,
+            metadata=metadata,
+            node_type=_json.get('type', None),
+            namespace=namespace,
+            project_id=_json.get('projectId', None),
+            config=_json.get('config', None),
+            position=position,
+            app_id=_json.get('appId', None),
+            dpk_name=_json.get('dpkName', None),
+            app_name=_json.get('appName', None),
+        )
+
+    def to_json(self):
+        _json = {
+            'name': self.name,
+            'id': self.node_id,
+            'outputs': [_io.to_json() for _io in self.outputs],
+            'inputs': [_io.to_json() for _io in self.inputs],
+            'metadata': self.metadata,
+            'type': self.node_type,
+            'namespace': self.namespace.to_json(),
+            'projectId': self.project_id,
+            'dpkName': self.dpk_name,
+            'appName': self.app_name,
+        }
+        if self.config is not None:
+            _json['config'] = self.config
+        if self.app_id is not None:
+            _json['appId'] = self.app_id
+        return _json
+
+    def is_root(self):
+        if self._pipeline is not None:
+            for node in self._pipeline.start_nodes:
+                if self.node_id == node.get('nodeId', None) and node.get('type', None) == 'root':
+                    return True
+        return False
+
+    def _build_connection(self,
+                          node,
+                          source_port: PipelineNodeIO = None,
+                          target_port: PipelineNodeIO = None,
+                          filters: entities.Filters = None,
+                          action: str = None) -> PipelineConnection:
+        """
+        Build connection between the current node and the target node use the given ports
+
+        :param PipelineNode node: the node to connect to it
+        :param PipelineNodeIO source_port: the source PipelineNodeIO input port
+        :param PipelineNodeIO target_port: the target PipelineNodeIO output port
+        :param entities.Filters filters: condition for the connection between the nodes
+        :param str action:  the action that move the input when it happen
+        :return: the connection between the nodes
+        """
+        if source_port is None and self.outputs:
+            source_port = self.outputs[0]
+
+        if target_port is None and node.inputs:
+            target_port = node.inputs[0]
+
+        if node.is_root():
+            self._pipeline.set_start_node(self)
+
+        source_connection = PipelineConnectionPort(node_id=self.node_id, port_id=source_port.port_id)
+        target_connection = PipelineConnectionPort(node_id=node.node_id, port_id=target_port.port_id)
+        if action is None and source_port.actions is not None and source_port.actions != []:
+            action = source_port.actions[0]
+        connection = PipelineConnection(source=source_connection, target=target_connection, filters=filters,
+                                        action=action)
+        return connection
+
+    def connect(self,
+                node,
+                source_port: PipelineNodeIO = None,
+                target_port: PipelineNodeIO = None,
+                filters=None,
+                action: str = None):
+        """
+        Build connection between the current node and the target node use the given ports
+
+        :param PipelineNode node: the node to connect to it
+        :param PipelineNodeIO source_port: the source PipelineNodeIO input port
+        :param PipelineNodeIO target_port: the target PipelineNodeIO output port
+        :param entities.Filters filters: condition for the connection between the nodes
+        :param str action:  the action that move the input when it happen
+        :return: the connected node
+        """
+        if self._pipeline is None:
+            raise Exception("must add the node to the pipeline first, e.g pipeline.nodes.add(node)")
+        connection = self._build_connection(node=node,
+                                            source_port=source_port,
+                                            target_port=target_port,
+                                            filters=filters,
+                                            action=action)
+        self._pipeline.connections.append(connection)
+        self._pipeline.nodes.add(node)
+        return node
+
+    def disconnect(self,
+                   node,
+                   source_port: PipelineNodeIO = None,
+                   target_port: PipelineNodeIO = None) -> bool:
+        """
+        remove connection between the current node and the target node use the given ports
+
+        :param PipelineNode node: the node to connect to it
+        :param PipelineNodeIO source_port: the source PipelineNodeIO input port
+        :param PipelineNodeIO target_port: the target PipelineNodeIO output port
+        :return: true if success and false if not
+        """
+        if self._pipeline is None:
+            raise Exception("must add the node to the pipeline first, e.g pipeline.nodes.add(node)")
+        connection = self._build_connection(node=node,
+                                            source_port=source_port,
+                                            target_port=target_port,
+                                            filters=None)
+
+        current_connection = connection.to_json()
+        if 'condition' in current_connection:
+            current_connection = current_connection.pop('condition')
+
+        for connection_index in range(len(self._pipeline.connections)):
+            pipeline_connection = self._pipeline.connections[connection_index].to_json()
+            if 'condition' in pipeline_connection:
+                pipeline_connection = pipeline_connection.pop('condition')
+
+            if current_connection == pipeline_connection:
+                self._pipeline.connections.pop(connection_index)
+                return True
+        logger.warning('do not found a connection')
+        return False
+
+    def add_trigger(self,
+                    trigger_type: entities.TriggerType = entities.TriggerType.EVENT,
+                    filters=None,
+                    resource: entities.TriggerResource = entities.TriggerResource.ITEM,
+                    actions: entities.TriggerAction = entities.TriggerAction.CREATED,
+                    execution_mode: entities.TriggerExecutionMode = entities.TriggerExecutionMode.ONCE,
+                    cron: str = None,
+                    ):
+        """
+        Create a Trigger. Can create two types: a cron trigger or an event trigger.
+        Inputs are different for each type
+
+        Inputs for all types:
+
+        :param trigger_type: can be cron or event. use enum dl.TriggerType for the full list
+
+        Inputs for event trigger:
+        :param filters: optional - Item/Annotation metadata filters, default = none
+        :param resource: optional - Dataset/Item/Annotation/ItemStatus, default = Item
+        :param actions: optional - Created/Updated/Deleted, default = create
+        :param execution_mode: how many time trigger should be activate. default is "Once". enum dl.TriggerExecutionMode
+
+        Inputs for cron trigger:
+        :param str cron: cron spec specifying when it should run. more information: https://en.wikipedia.org/wiki/Cron
+
+        :return: Trigger entity
+        """
+        if self._pipeline is None:
+            raise Exception("must add the node to the pipeline first, e.g pipeline.nodes.add(node)")
+
+        if not isinstance(actions, list):
+            actions = [actions]
+
+        if filters is None:
+            filters = {}
+        else:
+            filters = json.dumps(filters.prepare(query_only=True).get('filter', dict()))
+
+        if trigger_type == entities.TriggerType.EVENT:
+            spec = {
+                'filter': filters,
+                'resource': resource,
+                'executionMode': execution_mode,
+                'actions': actions
+            }
+        elif trigger_type == entities.TriggerType.CRON:
+            spec = {
+                'cron': cron,
+            }
+        else:
+            raise ValueError('Unknown trigger type: "{}". Use dl.TriggerType for known types'.format(trigger_type))
+
+        trigger = {
+            "type": trigger_type,
+            "spec": spec
+        }
+
+        set_trigger = False
+        for pipe_node in self._pipeline.start_nodes:
+            if pipe_node['nodeId'] == self.node_id:
+                set_trigger = True
+                pipe_node['trigger'] = trigger
+
+        if not set_trigger:
+            self._pipeline.start_nodes.append(
+                {
+                    "nodeId": self.node_id,
+                    "type": "trigger",
+                    'trigger': trigger
+                }
+            )
+
+
+class CodeNode(PipelineNode):
+    def __init__(self,
+                 name: str,
+                 project_id: str,
+                 project_name: str,
+                 method: Callable,
+                 outputs: List[PipelineNodeIO] = None,
+                 inputs: List[PipelineNodeIO] = None,
+                 position: tuple = (1, 1),
+                 ):
+        """
+        :param str name: node name
+        :param str project_id: project id
+        :param str project_name: project name
+        :param Callable method: function to deploy
+        :param list outputs: list of PipelineNodeIO outputs
+        :param list inputs: list of PipelineNodeIO inputs
+        :param tuple position: tuple of the node place
+        """
+        if inputs is None:
+            inputs = [self._default_io()]
+        if outputs is None:
+            outputs = [self._default_io()]
+
+        if method is None or not isinstance(method, Callable):
+            raise Exception('must provide a function as input')
+        else:
+            function_code = self._build_code_from_func(method)
+            function_name = method.__name__
+
+        super().__init__(name=name,
+                         node_id=str(uuid.uuid4()),
+                         outputs=outputs,
+                         inputs=inputs,
+                         metadata={},
+                         node_type=PipelineNodeType.CODE,
+                         namespace=PipelineNameSpace(function_name=function_name, project_name=project_name),
+                         project_id=project_id,
+                         position=position)
+
+        self.config = {
+            "package":
+                {
+                    "code": function_code,
+                    "name": function_name,
+                    "type": "code"
+                }
+        }
+
+    def _build_code_from_func(self, func: Callable) -> str:
+        """
+        Build a code format from the given function
+
+        :param Callable func: function to deploy
+        :return: a string the display the code with the package format
+        """
+        with open(assets.paths.PARTIAL_MAIN_FILEPATH, 'r') as f:
+            main_string = f.read()
+        lines = inspect.getsourcelines(func)
+
+        tabs_diff = lines[0][0].count('    ') - 1
+        for line_index in range(len(lines[0])):
+            line_tabs = lines[0][line_index].count('    ') - tabs_diff
+            lines[0][line_index] = ('    ' * line_tabs) + lines[0][line_index].strip() + '\n'
+
+        method_func_string = "".join(lines[0])
+
+        code = '{}\n{}\n    @staticmethod\n{}'.format('', main_string,
+                                                      method_func_string)
+        return code
+
+    @staticmethod
+    def from_json(_json: dict):
+        parent = PipelineNode.from_json(_json)
+        parent.__class__ = CodeNode
+        return parent
+
+
+class TaskNode(PipelineNode):
+    def __init__(self,
+                 name: str,
+                 project_id: str,
+                 dataset_id: str,
+                 recipe_title: str,
+                 recipe_id: str,
+                 task_owner: str,
+                 workload: List[entities.WorkloadUnit],
+                 task_type: str = 'annotation',
+                 position: tuple = (1, 1),
+                 actions: list = None,
+                 repeatable: bool = True,
+                 batch_size=None,
+                 max_batch_workload=None,
+                 priority=entities.TaskPriority.MEDIUM,
+                 due_date=None,
+                 consensus_task_type=None,
+                 consensus_percentage=None,
+                 consensus_assignees=None,
+                 groups=None
+                 ):
+        """
+        :param str name: node name
+        :param str project_id: project id
+        :param str dataset_id: dataset id
+        :param str recipe_title: recipe title
+        :param str recipe_id: recipe id
+        :param str task_owner: email of task owner
+        :param List[WorkloadUnit] workload: list of WorkloadUnit
+        :param str task_type: 'annotation' or 'qa'
+        :param tuple position: tuple of the node place
+        :param list actions: list of task actions
+        :param bool repeatable: can repeat in the item
+        :param int groups: groups to assign the task to
+        :param int batch_size: Pulling batch size (items) . Restrictions - Min 3, max 100 - for create pulling task
+        :param int max_batch_workload: Max items in assignment . Restrictions - Min batchSize + 2 , max batchSize * 2 - for create pulling task
+        :param entities.TaskPriority priority: priority of the task options in entities.TaskPriority
+        :param float due_date: date by which the task should be finished; for example, due_date = datetime.datetime(day= 1, month= 1, year= 2029).timestamp()
+        :param entities.ConsensusTaskType consensus_task_type: consensus_task_type of the task options in entities.ConsensusTaskType
+        :param int consensus_percentage: percentage of items to be copied to multiple annotators (consensus items)
+        :param int consensus_assignees: the number of different annotators per item (number of copies per item)
+        """
+        if actions is None or actions == []:
+            actions = []
+            if task_type == 'qa':
+                if 'approve' not in actions:
+                    actions.append('approve')
+            else:
+                if 'complete' not in actions:
+                    actions.append('complete')
+            actions.append('discard')
+        else:
+            logger.warning(
+                "The 'actions' field was updated to override the system default actions for task (complete/approve, discard) if provided, due to a bug fix.")
+
+        inputs = [self._default_io()]
+
+        outputs = [self._default_io(actions=actions)]
+
+        if groups is not None:
+            if not isinstance(groups, list) or not all(isinstance(group, str) for group in groups):
+                raise ValueError('groups must be a list of strings')
+
+        super().__init__(name=name,
+                         node_id=str(uuid.uuid4()),
+                         outputs=outputs,
+                         inputs=inputs,
+                         metadata=dict(),
+                         node_type=PipelineNodeType.TASK,
+                         namespace=PipelineNameSpace(function_name="move_to_task",
+                                                     project_name="DataloopTasks",
+                                                     service_name="pipeline-utils"),
+                         project_id=project_id,
+                         position=position)
+
+        self.dataset_id = dataset_id
+        self.recipe_title = recipe_title
+        self.recipe_id = recipe_id
+        self.task_owner = task_owner
+        self.task_type = task_type
+        if not isinstance(workload, list):
+            workload = [workload]
+        self.workload = workload
+        self.repeatable = repeatable
+        if max_batch_workload:
+            self.max_batch_workload = max_batch_workload
+        if batch_size:
+            self.batch_size = batch_size
+        if consensus_task_type:
+            self.consensus_task_type = consensus_task_type
+        if consensus_percentage:
+            self.consensus_percentage = consensus_percentage
+        if consensus_assignees:
+            self.consensus_assignees = consensus_assignees
+        self.priority = priority
+        if due_date is None:
+            due_date = (datetime.datetime.now() + datetime.timedelta(days=7)).timestamp() * 1000
+        self.due_date = due_date
+        self.groups = groups
+
+    @property
+    def dataset_id(self):
+        return self.metadata['datasetId']
+
+    @dataset_id.setter
+    def dataset_id(self, dataset_id: str):
+        if not isinstance(dataset_id, str):
+            raise PlatformException('400', 'Param dataset_id must be of type string')
+        self.metadata['datasetId'] = dataset_id
+
+    @property
+    def groups(self):
+        return self.metadata.get('groups')
+
+    @groups.setter
+    def groups(self, groups: List[str]):
+        if groups is not None:
+            self.metadata['groups'] = groups
+
+    @property
+    def repeatable(self):
+        return self.metadata['repeatable']
+
+    @repeatable.setter
+    def repeatable(self, repeatable: bool):
+        if not isinstance(repeatable, bool):
+            raise PlatformException('400', 'Param repeatable must be of type bool')
+        self.metadata['repeatable'] = repeatable
+
+    @property
+    def recipe_title(self):
+        return self.metadata['recipeTitle']
+
+    @recipe_title.setter
+    def recipe_title(self, recipe_title: str):
+        if not isinstance(recipe_title, str):
+            raise PlatformException('400', 'Param recipe_title must be of type string')
+        self.metadata['recipeTitle'] = recipe_title
+
+    @property
+    def recipe_id(self):
+        return self.metadata['recipeId']
+
+    @recipe_id.setter
+    def recipe_id(self, recipe_id: str):
+        if not isinstance(recipe_id, str):
+            raise PlatformException('400', 'Param recipe_id must be of type string')
+        self.metadata['recipeId'] = recipe_id
+
+    @property
+    def task_owner(self):
+        return self.metadata['taskOwner']
+
+    @task_owner.setter
+    def task_owner(self, task_owner: str):
+        if not isinstance(task_owner, str):
+            raise PlatformException('400', 'Param task_owner must be of type string')
+        self.metadata['taskOwner'] = task_owner
+
+    @property
+    def task_type(self):
+        return self.metadata['taskType']
+
+    @task_type.setter
+    def task_type(self, task_type: str):
+        if not isinstance(task_type, str):
+            raise PlatformException('400', 'Param task_type must be of type string')
+        self.metadata['taskType'] = task_type
+
+    @property
+    def workload(self):
+        return self.metadata['workload']
+
+    @workload.setter
+    def workload(self, workload: list):
+        if not isinstance(workload, list):
+            workload = [workload]
+        self.metadata['workload'] = [val.to_json() for val in workload]
+
+    @property
+    def batch_size(self):
+        return self.metadata['batchSize']
+
+    @batch_size.setter
+    def batch_size(self, batch_size: int):
+        if not isinstance(batch_size, int):
+            raise PlatformException('400', 'Param batch_size must be of type int')
+        self.metadata['batchSize'] = batch_size
+
+    @property
+    def max_batch_workload(self):
+        return self.metadata['maxBatchWorkload']
+
+    @max_batch_workload.setter
+    def max_batch_workload(self, max_batch_workload: int):
+        if not isinstance(max_batch_workload, int):
+            raise PlatformException('400', 'Param max_batch_workload must be of type int')
+        self.metadata['maxBatchWorkload'] = max_batch_workload
+
+    @property
+    def consensus_task_type(self):
+        return self.metadata['consensusTaskType']
+
+    @consensus_task_type.setter
+    def consensus_task_type(self, consensus_task_type: entities.ConsensusTaskType):
+        if not isinstance(consensus_task_type, str) and not isinstance(consensus_task_type, entities.ConsensusTaskType):
+            raise PlatformException('400', 'Param consensus_task_type must be of type entities.ConsensusTaskType')
+        self.metadata['consensusTaskType'] = consensus_task_type
+
+    @property
+    def consensus_percentage(self):
+        return self.metadata['consensusPercentage']
+
+    @consensus_percentage.setter
+    def consensus_percentage(self, consensus_percentage: int):
+        if not isinstance(consensus_percentage, int):
+            raise PlatformException('400', 'Param consensus_percentage must be of type int')
+        self.metadata['consensusPercentage'] = consensus_percentage
+
+    @property
+    def consensus_assignees(self):
+        return self.metadata['consensusAssignees']
+
+    @consensus_assignees.setter
+    def consensus_assignees(self, consensus_assignees: int):
+        if not isinstance(consensus_assignees, int):
+            raise PlatformException('400', 'Param consensus_assignees must be of type int')
+        self.metadata['consensusAssignees'] = consensus_assignees
+
+    @property
+    def priority(self):
+        return self.metadata['priority']
+
+    @priority.setter
+    def priority(self, priority: entities.TaskPriority):
+        if not isinstance(priority, int) and not isinstance(priority, entities.TaskPriority):
+            raise PlatformException('400', 'Param priority must be of type entities.TaskPriority')
+        self.metadata['priority'] = priority
+
+    @property
+    def due_date(self):
+        return self.metadata['dueDate']
+
+    @due_date.setter
+    def due_date(self, due_date: float):
+        if not isinstance(due_date, float) and not isinstance(due_date, int):
+            raise PlatformException('400', 'Param due_date must be of type float or int')
+        self.metadata['dueDate'] = due_date
+
+    @staticmethod
+    def from_json(_json: dict):
+        parent = PipelineNode.from_json(_json)
+        parent.__class__ = TaskNode
+        return parent
+
+
+class FunctionNode(PipelineNode):
+    def __init__(self,
+                 name: str,
+                 service: entities.Service,
+                 function_name,
+                 position: tuple = (1, 1),
+                 project_id=None,
+                 project_name=None
+                 ):
+        """
+        :param str name: node name
+        :param entities.Service service: service to deploy
+        :param str function_name: function name
+        :param tuple position: tuple of the node place
+        """
+        self.service = service
+
+        if project_id is None:
+            project_id = service.project_id
+        if project_id != service.project_id:
+            logger.warning("the project id that provide different from the service project id")
+
+        if project_name is None:
+            try:
+                project = repositories.Projects(client_api=self.service._client_api).get(project_id=project_id,
+                                                                                         log_error=False)
+                project_name = project.name
+            except:
+                logger.warning(
+                    'Service project not found using DataloopTasks project.'
+                    ' If this is incorrect please provide project_name param.')
+                project_name = 'DataloopTasks'
+        inputs = []
+        outputs = []
+        package = self.service.package
+        modules = []
+        if isinstance(package, entities.Package):
+            modules = package.modules
+        elif isinstance(package, entities.Dpk):
+            modules = package.components.modules
+        for model in modules:
+            if model.name == self.service.module_name:
+                for func in model.functions:
+                    if func.name == function_name:
+                        inputs = self._convert_from_function_io_to_pipeline_io(func.inputs)
+                        outputs = self._convert_from_function_io_to_pipeline_io(func.outputs)
+
+        namespace = PipelineNameSpace(
+            function_name=function_name,
+            service_name=self.service.name,
+            module_name=self.service.module_name,
+            package_name=self.service.package.name,
+            project_name=project_name
+        )
+        super().__init__(name=name,
+                         node_id=str(uuid.uuid4()),
+                         outputs=outputs,
+                         inputs=inputs,
+                         metadata={},
+                         node_type=PipelineNodeType.FUNCTION,
+                         namespace=namespace,
+                         project_id=service.project_id,
+                         position=position)
+
+    def _convert_from_function_io_to_pipeline_io(self, function_io: List[entities.FunctionIO]) -> List[PipelineNodeIO]:
+        """
+        Get a list of FunctionIO and convert them to PipelineIO
+        :param List[entities.FunctionIO] function_io: list of functionIO
+        :return: list of PipelineIO
+        """
+        pipeline_io = []
+        for single_input in function_io:
+            pipeline_io.append(
+                PipelineNodeIO(port_id=str(uuid.uuid4()),
+                               input_type=single_input.type,
+                               name=single_input.name,
+                               color=None,
+                               display_name=single_input.name,
+                               default_value=single_input.value,
+                               actions=single_input.actions if single_input.actions is not None else []))
+        return pipeline_io
+
+    @staticmethod
+    def from_json(_json: dict):
+        parent = PipelineNode.from_json(_json)
+        parent.__class__ = FunctionNode
+        return parent
+
+
+class DatasetNode(PipelineNode):
+    def __init__(self,
+                 name: str,
+                 project_id: str,
+                 dataset_id: str,
+                 dataset_folder: str = None,
+                 load_existing_data: bool = False,
+                 data_filters: entities.Filters = None,
+                 position: tuple = (1, 1)):
+        """
+        :param str name: node name
+        :param str project_id: project id
+        :param str dataset_id: dataset id
+        :param str dataset_folder: folder in dataset to work in it
+        :param bool load_existing_data: optional - enable to automatically load existing data into the
+                                        pipeline (executions) upon activation, based on the defined dataset,
+                                        folder, and data_filters.
+        :param entities.Filters data_filters: optional - filters entity or a dictionary containing filters parameters.
+                                              Use to filter the data items to be loaded when load_existing_data
+                                              is enabled.
+        :param tuple position: tuple of the node place
+        """
+        inputs = [self._default_io()]
+        outputs = [self._default_io()]
+        super().__init__(name=name,
+                         node_id=str(uuid.uuid4()),
+                         outputs=outputs,
+                         inputs=inputs,
+                         metadata={},
+                         node_type=PipelineNodeType.STORAGE,
+                         namespace=PipelineNameSpace(function_name="dataset_handler",
+                                                     project_name="DataloopTasks",
+                                                     service_name="pipeline-utils"),
+                         project_id=project_id,
+                         position=position)
+        self.dataset_id = dataset_id
+        self.dataset_folder = dataset_folder
+        self.load_existing_data = load_existing_data
+        self.data_filters = data_filters
+
+    @property
+    def dataset_id(self):
+        return self.metadata['datasetId']
+
+    @dataset_id.setter
+    def dataset_id(self, dataset_id: str):
+        self.metadata['datasetId'] = dataset_id
+
+    @property
+    def dataset_folder(self):
+        return self.metadata.get('dir', None)
+
+    @dataset_folder.setter
+    def dataset_folder(self, dataset_folder: str):
+        if dataset_folder is not None:
+            if not dataset_folder.startswith("/"):
+                dataset_folder = '/' + dataset_folder
+            self.metadata['dir'] = dataset_folder
+
+    @property
+    def load_existing_data(self):
+        return self.metadata.get('triggerToPipeline', {}).get('active', False)
+
+    @load_existing_data.setter
+    def load_existing_data(self, load_existing_data: bool):
+        if load_existing_data:
+            self.metadata.setdefault('triggerToPipeline', {})['active'] = True
+        else:
+            self.metadata.pop('triggerToPipeline', None)
+
+    @property
+    def data_filters(self):
+        data_filters = self.metadata.get('triggerToPipeline', {}).get('filter', None)
+        if data_filters:
+            data_filters = entities.Filters(custom_filter=json.loads(data_filters))
+        return data_filters
+
+    @data_filters.setter
+    def data_filters(self, data_filters: entities.Filters):
+        if data_filters is None:
+            filters = None
+        else:
+            filters = json.dumps(data_filters.prepare(query_only=True).get('filter'))
+        self.metadata.setdefault('triggerToPipeline', {})['filter'] = filters
+
+    @staticmethod
+    def from_json(_json: dict):
+        parent = PipelineNode.from_json(_json)
+        parent.__class__ = DatasetNode
+        return parent