graph

graph#

Classes#

SparkNodeViewer

Custom reimplementation of NodeViwer use to solve certain visual bugs.

SparkNodeGraph

NodeGraphQt object for building/managing Spark models.

Module Contents#

class graph.SparkNodeViewer(parent=None)#

Bases: NodeGraphQt.widgets.viewer.NodeViewer

Custom reimplementation of NodeViwer use to solve certain visual bugs.

mousePressEvent(event)#
class graph.SparkNodeGraph(parent=None, **kwargs)#

Bases: NodeGraphQt.NodeGraph

NodeGraphQt object for building/managing Spark models.

context_menu_prompt#
broadcast_message#
stateChanged#
name_exist(name)#
Parameters:

name (str)

Return type:

bool

update_node_name(id, name)#
Parameters:
Return type:

None

create_node(node_type, name=None, selected=True, color=None, text_color=None, pos=None, push_undo=True)#
Return type:

spark.graph_editor.models.nodes.AbstractNode

delete_node(node, push_undo=True)#
Return type:

None

build_raw_graph()#
Return type:

networkx.DiGraph

get_nodes_by_map()#
Return type:

Tuple[List[spark.graph_editor.models.nodes.AbstractNode], List[spark.graph_editor.models.nodes.AbstractNode], List[spark.graph_editor.models.nodes.AbstractNode]]

load_from_model(config)#
Parameters:

config (spark.nn.brain.BrainConfig)

validate_graph()#

Simple graph validation.

Ensures that graph has a single connected component and at least one source and one sink is present in the model.

Return type:

bool

build_brain_config()#

Build the model from the graph state.

Return type:

spark.nn.brain.BrainConfig

Contents