Core Crate Modules

hotspots-core is the main library crate. The CLI and GitHub Action exist primarily to drive this crate and present its output.

Public entry points

hotspots-core/src/lib.rs exports the public analysis entry points:

• analyze(path, options)
• analyze_with_config(path, options, resolved_config)
• analyze_with_progress(path, options, resolved_config, progress)

It also re-exports commonly used types such as FunctionRiskReport, ResolvedConfig, GitContext, CallGraph, and TouchMode.

Module groups

Source analysis

Module	Responsibility
analysis.rs	Per-file orchestration: parse, discover functions, compute reports.
language/	Language abstraction layer and parser/CFG implementations.
parser.rs	ECMAScript parser compatibility layer and parser tests.
discover.rs	Function discovery helpers and legacy ECMAScript discovery.
cfg.rs, cfg/builder.rs	Control-flow graph representation and ECMAScript CFG construction.
metrics.rs	Raw metric extraction: cyclomatic complexity, nesting, fan-out, exits.
suppression.rs	hotspots-ignore suppression comment handling.

Risk and prioritization

Module	Responsibility
risk.rs	LRS weights, thresholds, bands, and base risk calculations.
scoring.rs	Activity-weighted scoring and composite risk logic.
patterns.rs	Pattern classification such as god function, churn magnet, hub function.
aggregates.rs	File and directory aggregates.
compact.rs	Compact/default output data shaping.

Git, snapshots, and deltas

Module	Responsibility
git.rs	Git context, refs, churn extraction, commit metadata.
snapshot.rs	Snapshot construction, enrichment, persistence model.
delta.rs	Snapshot/function comparison.
policy.rs	Policy evaluation over deltas.
touch_cache.rs	Cache for expensive function touch calculations.
db/	SQLite-backed storage helpers.
trends.rs	Historical trend analysis.
prune.rs	Snapshot/cache pruning.

Graph and model enrichment

Module	Responsibility
callgraph.rs	Call graph, fan-in/fan-out, PageRank, betweenness approximations.
imports.rs	Import extraction and module relationship helpers.
models.rs	Model/entity declaration extraction and association.

Rendering and output

Module	Responsibility
report.rs	Text/JSON rendering for function reports.
html.rs	HTML report generation.
sarif.rs	SARIF output for code scanning integrations.

Language abstraction

The language layer normalizes TypeScript, JavaScript, Go, Java, Python, Rust, C#, and Vue-flavored inputs into common concepts:

• Language
• LanguageParser
• ParsedModule
• FunctionNode
• FunctionBody
• CfgBuilder
• SourceSpan

Most new language work should begin in hotspots-core/src/language/ and then add golden tests under hotspots-core/tests/fixtures/ and hotspots-core/tests/golden_tests.rs.

Where not to put logic

Avoid putting core analysis behavior in hotspots-cli. CLI code should parse arguments, load config, call hotspots-core, and render command-specific output. If behavior needs tests independent of terminal invocation, it probably belongs in hotspots-core.