feat: graph-level annotations, proto in wheel, and schema folder consolidation

[harsh-sikhwal]

1. Graph-level annotation support — New graph field on Annotation allows annotating the graph itself (not just nodes/edges/links). Attributes are stored in networkx's graph.graph dict and surfaced in both infragraph and networkx outputs from get_graph. Immutable attribute guard applies. 7 tests added in test_graph_annotations.py.

2. Proto bundled in wheel — infragraph.proto is now generated into src/infragraph/ and included in the sdist/wheel via pyproject.toml. CI commits the generated proto alongside other artifacts.

3. Schema folder consolidation — api/ (endpoint definitions) and models/ (component schemas) merged into a single schema/ folder. All cross-references in schema/api.yaml updated. generate.py and CI workflow updated accordingly.

Graph-level annotation support

Adds a new graph target to the Annotation API so key-value metadata can be attached to the whole graph, not just nodes/edges/links. Also refactors models/ + api/ into a single schema/ folder.

Annotate the graph

annotation = Annotation()
annotation.graph.add(attribute="cluster", value="rack-a")
annotation.graph.add(attribute="experiment", value="run-42")

service.annotate_graph(annotation)

Query graph-level attributes

request = QueryRequest()
filter = request.graph_filters.add(name="cluster filter")
filter.attribute_filter.name = "cluster"
filter.attribute_filter.operator = QueryAttribute.EQ
filter.attribute_filter.value = "rack-a"

response = service.query_graph(request)
# response.graph_matches -> [NameValue(name="cluster", value="rack-a")]

Schema changes

• schema/annotation.yaml: Annotation gains a graph field (x-field-uid: 4, array of Annotation.Attribute)
• schema/api.yaml: relocated from api/api.yaml
• models/*.yaml → schema/*.yaml (folder rename, no schema content change)

Rendered spec: https://keysight.github.io/infragraph/openapi.html

[ajbalogh]

Please add code snippets to the PR summary showing how with the API a user would annotate the graph, query those graph attributes, and what it would look like when retrieving the full/partial graph. Also if there is a change to the schema please include a link to the redocly documentation.