qsrlib_qstag.qstag module

Qualitative Spatio-Temporal Activity Graph module

class qsrlib_qstag.qstag.Activity_Graph(world, world_qsr, object_types={}, params={})

Activity Graph class: Lower level is a set of only nodes of type ‘object’. Middle level nodes are only of type ‘spatial_relation’. Top level nodes are only of type ‘temporal_relation’.

Accepts a QSRLib.World_Trace and QSRLib.QSR_World_Trace as input.

abstract_graph

Getter.

Returns:self.abstract_graph Return type:igraph.Graph

abstract_object_nodes

Getter.

Returns:object_nodes from abstract_graph Return type:list

episodes

Getter.

Returns:self.__episodes Return type:list

static get_objects_types(objects_types, world)

Generates a dictionary of object name and object type pairs Using both the dynamic_args dictionary where key = objects_types, and the **kwargs value [object_type] in the World Trace object

Parameters:

• objects_types (dictionary) – Uses the dynamic_args dictionary where key = objects_types if provided
• world (World_Trace) – Otherwise, looks at the **kwargs value [object_type] in the World Trace object

Returns:

A dictionary with the object name as keys and the generic object type as value.

Return type:

dict

graph = None

igraph.Graph: An igraph graph object containing all the object, spatial and temporal nodes. list: A list of edges connecting the spatial nodes to the object nodes. list: A list of edges connecting the spatial nodes to the temporal nodes.

graphlets = None

Creates a Graphlets object containing lists of, unique graphlets, hashes and histogram of graphlets.

histogram

Getter.

Returns:self.__histogram Return type:dict

object_nodes

Getter.

Returns:object_nodes Return type:list

spatial_nodes

Getter.

Returns:spatial_nodes Return type:list

spatial_obj_edges

Getter.

Returns:self.__spatial_obj_edges Return type:list

temp_spatial_edges

Getter.

Returns:self.__temp_spatial_edges Return type:list

temporal_nodes

Getter.

Returns:temporal_nodes Return type:list

class qsrlib_qstag.qstag.Graphlets(episodes, params, object_types)

Graphlet class: Minimal subgraphs of the same structure as the Activity Graph.

code_book = None

list: The list of graphlet hashes (zip with histogram for count of each).

graphlets = None

dict: dictionary of the graphlet hash as key, and the iGraph object as value.

histogram = None

list: The list of graphlet counts (zip with codebook for a hash of each, and check graphlets for the iGraph).

qsrlib_qstag.qstag.get_graph(episodes, object_types={})

Generates a graph from a set of input episode QSRs.

Parameters:

• episodes (list) – list of episodes, where one episode = [[obj_list], {QSR dict}, (start, end_tuple)]
• object_types (dict) – a dictionary of object ID and object type.

Returns:

igraph.Graph: An igraph graph object containing all the object, spatial and temporal nodes.

Return type:

igraph.Graph

qsrlib_qstag.qstag.get_graphlet_selections(episodes, params, object_types, vis=False)

This function implements Sridar’s validity criteria to select all valid graphlets from an activity graph: see Sridar_AAAI_2010 for more details.

Parameters:

• episodes (list) – list of episodes, where one episode = [[obj_list], {QSR dict}, (start, end_tuple)]
• params (dict) – a dictionary containing parameters for the generation of the QSTAG. i.e. “min_rows”, “max_rows”, “max_eps”

Return list_of_graphlets:  

a list of iGraph graphlet objects

Return type:

list

Return list_of_graphlet_hashes:  

a list of hashes, relating to the graphlets

Return type:

list