nan-graph (Python)

Port en Python del core @carloscortezcloud/nan-graph (TypeScript) — el motor del grafo de SOFE Architecture Graph (fusión Ñan × BYaML v2). OSS, Apache-2.0, bajo breakingthecloud.

Misma API que el core TS; se mantiene en sync manual (patrón styrr → styrr-py).

Grafo dirigido en memoria con relaciones tipadas: traversal BFS/DFS, blast radius, cost chain, fan-in / SPOF, y filtrado por tipo de relación.

Install

pip install nan-graph

Uso

from nangraph import NanGraph, blast_radius, cost_chain, from_yaml

g = (
    NanGraph()
    .add_node({"id": "api", "type": "aws.apigateway", "attrs": {"monthly_cost": 20}})
    .add_node({"id": "lambda", "type": "aws.lambda", "attrs": {"monthly_cost": 5}})
    .add_node({"id": "ddb", "type": "aws.dynamodb", "attrs": {"monthly_cost": 30}})
    .add_edge("api", "lambda", "routes_to")
    .add_edge("lambda", "ddb", "reads_writes")
)

blast_radius(g, "api")   # ['lambda', 'ddb'] — qué se cae si api falla
cost_chain(g, "api")     # 55.0 — costo downstream total

Importar desde YAML / JSON

from nangraph import from_yaml, from_object

# estilo dependency-graph.yaml
g = from_yaml("""
nodes:
  sofe-engine: { type: python-library }
  sofe-server: { type: python-api }
edges:
  - { from: sofe-engine, to: sofe-server, relType: bundles }
""")

# estilo BYaML v0.3 (components + relationships)
g2 = from_object({
    "components": [{"id": "api", "type": "aws.apigateway", "monthly_cost": 20}],
    "relationships": [{"from": "api", "to": "lambda", "type": "routes_to"}],
})

API

Función	Descripción
NanGraph()	Grafo dirigido multi-borde en memoria
add_node(dict) / add_edge(from, to, rel_type)	Añadir nodo/arista (auto-crea nodos)
get_related(id, rel_type?, direction?)	Vecinos outgoing/incoming/both, con filtro
traverse_bfs(g, start, rel_types?, max_depth?, direction?)	BFS
traverse_dfs(g, start, ...)	DFS
blast_radius(g, start)	Nodos afectados downstream si start falla
cost_chain(g, start)	Suma attrs.monthly_cost de start + downstream
team_cost(g, owner)	Suma de costo por attrs.owner
fan_in(g, id) / single_points_of_failure(g, threshold)	SPOF por alta dependencia entrante
from_object(dict) / from_yaml(str) / from_json(str)	Construir grafo desde input

Paridad con el core TS

TS (@carloscortezcloud/nan-graph)	Python (nan-graph)
new NanGraph()	NanGraph()
g.addNode / g.addEdge	g.add_node / g.add_edge
g.getRelated	g.get_related
traverseBFS / traverseDFS	traverse_bfs / traverse_dfs
blastRadius / costChain	blast_radius / cost_chain
teamCost / fanIn / singlePointsOfFailure	team_cost / fan_in / single_points_of_failure
fromObject / fromYaml / fromJson	from_object / from_yaml / from_json

Licencia

Apache-2.0. Ver LICENSE.