Problem
The dependency graph's data currently lives inside an executable
graph-data.js — model and code in one file. That makes the data hard to
diff, hard to hand-edit safely, and impossible to consume from anything
other than the renderer.
Proposal
Split it the way the CMDB models nodes: declarative YAML, one file per
project / package boundary, plus a manifest —
Documentation/DependencyGraph/
graph.yml manifest: projects, root order, events, column map
projects/*.yml one per boundary: components, methods, outbound flows
index.html renderer only — fetches and assembles the data files
Same change as NHSISL/LondonFhirService#178, verified equivalent to the
current data (identical components, edges and rendered counts in both
views). Includes the LoggingBroker regeneration from the unit-tests branch
so those graph commits can be dropped on rebase.
Problem
The dependency graph's data currently lives inside an executable
graph-data.js— model and code in one file. That makes the data hard todiff, hard to hand-edit safely, and impossible to consume from anything
other than the renderer.
Proposal
Split it the way the CMDB models nodes: declarative YAML, one file per
project / package boundary, plus a manifest —
Same change as NHSISL/LondonFhirService#178, verified equivalent to the
current data (identical components, edges and rendered counts in both
views). Includes the LoggingBroker regeneration from the unit-tests branch
so those graph commits can be dropped on rebase.