Views

View helpers from annnet.core._Views.

Use G.view(...) for graph view workflows. Direct imports from underscore modules follow the internal API policy.

annnet.core._Views.GraphView

Lazy, filtered view into a graph; materialize() for a concrete subgraph.

Attributes

obs property

obs

Return the filtered node attribute table for this view.

Returns:

Type	Description
DataFrame - like

Notes

Materialized from the node table of the graph and filtered by the node ids of this view. It is a table for the caller, not the storage of the graph.

var property

var

Return the filtered edge attribute table for this view.

Returns:

Type	Description
DataFrame - like

Notes

Materialized from the edge table of the graph and filtered by the edge ids of this view. It is a table for the caller, not the storage of the graph.

B property

B

Return the filtered incidence matrix subview.

Named as the graph names it. X was the spelling of the incidence matrix before cycle 002 renamed it, and it stayed here after it went from the graph, so one object answered to a name the other refused.

Returns:

Type	Description
dok_matrix

node_ids property

node_ids

Get filtered node IDs (cached).

Returns:

Type	Description
set[str] | None	None means no node filter (full graph).

edge_ids property

edge_ids

Get filtered edge IDs (cached).

Returns:

Type	Description
set[str] | None	None means no edge filter (full graph).

node_count property

node_count

Return the number of nodes in this view.

Returns:

Type	Description
int

edge_count property

edge_count

Return the number of edges in this view.

Returns:

Type	Description
int

Functions

edges_df

edges_df(**kwargs)

Return an edge DataFrame view filtered to this view's edges.

Parameters:

Name	Type	Description	Default
**kwargs		Passed through to AnnNet.edges_view().	{}

Returns:

Type	Description
DataFrame - like

Notes

Uses AnnNet.edges_view() and then filters by the view's edge IDs.

nodes_df

nodes_df(**kwargs)

Return a node DataFrame view filtered to this view's nodes.

Parameters:

Name	Type	Description	Default
**kwargs		Passed through to AnnNet.nodes_view().	{}

Returns:

Type	Description
DataFrame - like

Notes

Uses AnnNet.nodes_view() and then filters by the view's node IDs.

materialize

materialize(copy_attributes=True)

Create a concrete subgraph from this view.

Parameters:

Name	Type	Description	Default
copy_attributes	bool	If True, copy node/edge attributes into the new graph.	True

Returns:

Type	Description
AnnNet	Materialized subgraph.

subview

subview(
    nodes=None, edges=None, slices=None, predicate=None
)

Create a new GraphView by further restricting this view.

Parameters:

Name	Type	Description	Default
nodes	Iterable[str] | callable | None	Node IDs or predicate; intersects with current view if provided.	None
edges	Iterable[str] | callable | None	Edge IDs or predicate; intersects with current view if provided.	None
slices	Iterable[str] | None	Slice IDs to include. Defaults to current view's slices if None.	None
predicate	callable | None	Additional node predicate applied in conjunction with existing filters.	None

Returns:

Type	Description
GraphView

Notes

Predicates are combined with logical AND.

summary

summary()

Return a human-readable summary of this view.

Returns:

Type	Description
str

annnet.core._Views.ViewsClass

Materialized table builders mixed into AnnNet.

Functions

edges_view

edges_view(
    slice=None,
    include_directed=True,
    include_weight=True,
    resolved_weight=True,
    copy=True,
    *,
    layer=None,
    in_slice=None,
    include_hyper=True,
    include_binary=True
)

Build a DataFrame view of edges with optional slice join.

Parameters:

Name	Type	Description	Default
slice	str	Slice id whose per-edge attributes are joined onto every row, as slice_* columns. This does not filter — see in_slice.	None
include_directed	bool	Include directedness column.	True
include_weight	bool	Include global weight column.	True
resolved_weight	bool	Include effective weight (slice override if present).	True
copy	bool	Return a cloned DataFrame if True.	True
layer	tuple[str, ...]	Keep only the edges of this layer, as :meth:LayerAccessor.layer_edge_set names them.	None
in_slice	str	Keep only the rows of this slice. Distinct from slice, which joins without filtering.	None
include_hyper	bool	Include hyperedges. False leaves a table whose rows are all binary, which is what an exporter that cannot hold a hyperedge wants.	True
include_binary	bool	Include binary edges. False with include_hyper leaves the hyperedges alone — which is what :meth:ViewsAccessor.hyperedges asks for.	True

Returns:

Type	Description
DataFrame - like	source and target are bare node ids; src_layer and dst_layer are the canonical layer ids of the two endpoints, and are null when the graph is flat or a side does not sit in one layer.

Notes

slice= joins, in_slice= filters. Both take a slice id and they do different things, so a call that means "only this slice's edges" wants the second.

nodes_view

nodes_view(copy=True)

Read-only node attribute table.

Parameters:

Name	Type	Description	Default
copy	bool	Return a cloned DataFrame.	True

Returns:

Type	Description
DataFrame - like	Columns include node_id plus pure attributes.

slices_view

slices_view(copy=True)

Read-only slice attribute table.

Parameters:

Name	Type	Description	Default
copy	bool	Return a cloned DataFrame.	True

Returns:

Type	Description
DataFrame - like	One row per slice (including the default slice), keyed by slice_id. User-set slice attributes appear as additional columns; slices without user attrs still appear, with null cells.

aspects_view

aspects_view(copy=True)

Return a view of Kivela aspects and their metadata.

Parameters:

Name	Type	Description	Default
copy	bool	Return a cloned DataFrame.	True

Returns:

Type	Description
DataFrame - like

Notes

Columns include aspect, elem_layers, and any aspect attribute keys.

layers_view

layers_view(copy=True)

Return a read-only table of multi-aspect layers.

Parameters:

Name	Type	Description	Default
copy	bool	Return a cloned DataFrame.	True

Returns:

Type	Description
DataFrame - like

Notes

Columns include layer_tuple, layer_id, aspect columns, layer attributes, and prefixed elementary layer attributes.

annnet.core._Views.ViewsAccessor

Namespace for materialized graph tables (G.views).

Functions

edges

edges(*args, **kwargs)

Materialize the edge table view.

hyperedges

hyperedges(
    slice=None, copy=True, *, layer=None, in_slice=None
)

Materialize the hyperedge table view.

The same table :meth:edges builds, holding only the rows whose kind is "hyper". head, tail and members are the columns that carry a hyperedge's shape, and they are null on every binary row — which is why reading hyperedges out of the full table means filtering it first.

Parameters:

Name	Type	Description	Default
slice	str	Slice id whose per-edge attributes are joined onto every row.	None
copy	bool	Return a cloned DataFrame if True.	True
layer	tuple[str, ...]	Keep only the hyperedges of this layer.	None
in_slice	str	Keep only the rows of this slice.	None

Returns:

Type	Description
DataFrame - like

entity_kinds

entity_kinds()

Return the kind of every entity, as a mapping from its id.

An entity is a node, or an edge that is a node in its own right. The answer is built on each call, so changing it changes nothing.

nodes

nodes(*args, **kwargs)

Materialize the node table view.

slices

slices(*args, **kwargs)

Materialize the slice table view.

aspects

aspects(*args, **kwargs)

Materialize the aspect table view.

layers

layers(*args, **kwargs)

Materialize the layer table view.

layers_view

layers_view(copy=True)

Materialize the layer table view.