KMP-IMPACT

ImpactGraph

Phase 2 aggregate output. Written to phase2/impact_graph.json and consumed by Phase 4.

Fields

Field Type Default Notes
dependency_group str — Maven group of the bump being analysed.
version_before str — BEFORE version.
version_after str — AFTER version.
seed_files list[str] [] Files that directly import the bumped dependency. Every FileImpact with relation = "direct" corresponds to an entry here.
impacted_files list[FileImpact] [] One entry per impacted Kotlin file.
expect_actual_pairs list[ExpectActualPair] [] Detected expect/actual pairs whose expect file is impacted.
total_project_files int 0 Total .kt files scanned (denominator for coverage statistics).
total_impacted int 0 len(impacted_files). Persisted to avoid recomputing in downstream consumers.
ksp_edges_added int 0 Extra edges contributed by the KSP-driven symbol index, if KSP output is available. 0 when the analyzer falls back to the Tree-sitter-only mode.

Example

{
  "dependency_group": "io.ktor",
  "version_before": "2.3.8",
  "version_after": "2.3.11",
  "seed_files": [
    "shared/src/commonMain/kotlin/com/example/api/PokedexClient.kt"
  ],
  "impacted_files": [ /* FileImpact[] */ ],
  "expect_actual_pairs": [ /* ExpectActualPair[] */ ],
  "total_project_files": 85,
  "total_impacted": 7,
  "ksp_edges_added": 12
}

Aggregating per source set

The model does not pre-compute counters per source set; the renderer iterates impacted_files and groups by source_set to build the sunburst preview. Consumers that need per-source-set counts can do the same:

from collections import Counter

graph = ImpactGraph.model_validate_json(open("output/phase2/impact_graph.json").read())
by_source_set = Counter(f.source_set for f in graph.impacted_files)
This site is open source. Improve this page.