Algorithms

Traversal

annnet.algorithms.traversal.Traversal

Functions

neighbors(entity_id)

Neighbors of an entity (vertex or edge-entity).

Parameters:

Name	Type	Description	Default
entity_id	str		required

Returns:

Type	Description
list[str]	Adjacent entities. For hyperedges, uses head/tail orientation.

out_neighbors(vertex_id)

Out-neighbors of a vertex under directed semantics.

Parameters:

Name	Type	Description	Default
vertex_id	str		required

Returns:

Type	Description
list[str]

successors(vertex_id)

Successors of a vertex under directed semantics.

Parameters:

Name	Type	Description	Default
vertex_id	str		required

Returns:

Type	Description
list[str]

in_neighbors(vertex_id)

In-neighbors of a vertex under directed semantics.

Parameters:

Name	Type	Description	Default
vertex_id	str		required

Returns:

Type	Description
list[str]

predecessors(vertex_id)

In-neighbors of a vertex under directed semantics.

Parameters:

Name	Type	Description	Default
vertex_id	str		required

Returns:

Type	Description
list[str]

Internal helpers

Functions

neighbors(entity_id)

Neighbors of an entity (vertex or edge-entity).

Parameters:

Name	Type	Description	Default
entity_id	str		required

Returns:

Type	Description
list[str]	Adjacent entities. For hyperedges, uses head/tail orientation.

out_neighbors(vertex_id)

Out-neighbors of a vertex under directed semantics.

Parameters:

Name	Type	Description	Default
vertex_id	str		required

Returns:

Type	Description
list[str]

successors(vertex_id)

Successors of a vertex under directed semantics.

Parameters:

Name	Type	Description	Default
vertex_id	str		required

Returns:

Type	Description
list[str]

in_neighbors(vertex_id)

In-neighbors of a vertex under directed semantics.

Parameters:

Name	Type	Description	Default
vertex_id	str		required

Returns:

Type	Description
list[str]

predecessors(vertex_id)

In-neighbors of a vertex under directed semantics.

Parameters:

Name	Type	Description	Default
vertex_id	str		required

Returns:

Type	Description
list[str]

Lazy Proxies (NetworkX, igraph, graph-tool)

These proxies provide a thin, lazy bridge to optional backends. Internal proxy classes and helper functions are intentionally hidden from this reference; see the source if you need implementation details.

annnet.core.lazy_proxies.nx_lazyproxy._LazyNXProxy

Lazy NetworkX proxy attached to an AnnNet instance.

This proxy lets you call NetworkX algorithms as G.nx.<algo>(...). On first use (or after a graph mutation), AnnNet is converted to a NetworkX graph and cached; subsequent calls reuse the cached backend until the AnnNet version changes.

Conversion produces a manifest dictionary that preserves information NetworkX cannot represent (e.g., hyperedges, per-edge directedness, slices, multilayer metadata, stable edge IDs). The manifest is JSON-serializable and can be persisted with annnet.adapters.networkx_adapter.save_manifest.

Notes

• Requires the optional networkx dependency.
• The typical usage pattern is G.nx.algorithm(...), which lazily converts, runs the NetworkX algorithm, and returns its output.

annnet.core.lazy_proxies.ig_lazyproxy._LazyIGProxy

Lazy igraph proxy attached to an AnnNet instance.

This proxy lets you call igraph algorithms as G.ig.<algo>(...). On first use (or after a graph mutation), AnnNet is converted to an igraph graph and cached; subsequent calls reuse the cached backend until the AnnNet version changes.

Conversion is optimized for speed and does not build a manifest. This means features that require a manifest (slices, reified hyperedges, multilayer metadata, stable edge IDs) are not preserved on the proxy path. Use adapters.igraph_adapter.to_igraph(...) for full-fidelity export.

Notes

• Requires the optional python-igraph dependency.
• The typical usage pattern is G.ig.algorithm(...), which lazily converts, runs the igraph algorithm, and returns its output.
• _ig_simple=True collapses parallel edges to a simple graph; _ig_edge_aggs controls aggregation (e.g., {"weight":"min","capacity":"sum"}).

annnet.core.lazy_proxies.gt_lazyproxy._LazyGTProxy

Lazy graph-tool proxy attached to an AnnNet instance.

This proxy lets you call graph-tool algorithms via namespaces such as: G.gt.topology.shortest_distance(...), G.gt.centrality.betweenness(...), and G.gt.flow.push_relabel_max_flow(...).

On first use (or after a graph mutation), AnnNet is converted to a graph-tool graph and cached; subsequent calls reuse the cached backend until the AnnNet version changes. Conversion produces a manifest dictionary that preserves information graph-tool cannot represent (e.g., hyperedges, per-edge directedness, slices, multilayer metadata, stable edge IDs). The manifest is JSON-serializable and can be persisted by the adapter.

Notes

• Requires the optional graph-tool dependency (not on PyPI).
• The typical usage pattern is G.gt.<namespace>.<algo>(...), which lazily converts, runs the graph-tool algorithm, and returns its output.