Struct ImpactConfig 

pub struct ImpactConfig {
    pub max_depth: usize,
    pub language: Option<String>,
    pub include_patterns: Vec<String>,
    pub exclude_patterns: Vec<String>,
    pub exclude_tests: bool,
    pub include_call_sites: bool,
    pub deduplicate_paths: bool,
}

Configuration for impact analysis.

Fields

max_depth: usize

Maximum depth to traverse (0 = NO traversal, direct callers only).

This controls how many levels of the call hierarchy are explored:

• max_depth = 0: NO exploration - returns only direct callers (distance 1)
• max_depth = 1: Explore direct callers to find distance 2 callers
• max_depth = 2: Explore up to distance 2, finding callers up to distance 3
• max_depth = N: Explore up to distance N, potentially finding callers at distance N+1

This matches Python semantics where depth <= 0 causes immediate truncation. Note: The depth controls exploration, not the maximum distance in results.

language: Option<String>

Filter by language (e.g., “python”, “rust”).

include_patterns: Vec<String>

Glob patterns to include (e.g., “src/**/*.rs”).

exclude_patterns: Vec<String>

Glob patterns to exclude (e.g., “/test/”).

exclude_tests: bool

Whether to exclude test files from results.

include_call_sites: bool

Whether to include call site information.

deduplicate_paths: bool

If true, each node appears only once (first path wins, BFS behavior). If false, all paths are explored with per-branch visited sets (DFS, Python semantics). When false, the minimum distance across all paths is recorded for each node.

For diamond-shaped call graphs like:

    A
   / \
  B   C
   \ /
    D (target)

• deduplicate_paths = true: A appears once at distance 2 (via first explored path)
• deduplicate_paths = false: All paths explored, A at minimum distance 2 with aggregated call sites

Implementations

impl ImpactConfig

pub fn new() -> Self

Create a new default configuration.

pub fn with_depth(self, depth: usize) -> Self

Set maximum traversal depth.

The depth controls how many levels of callers are explored. With depth N, the algorithm explores callers up to distance N from the target, which means callers at distance N+1 may be discovered but not explored further.

pub fn with_language(self, lang: &str) -> Self

Filter results by language.

pub fn with_includes(self, patterns: &[&str]) -> Self

Add file patterns to include.

pub fn with_excludes(self, patterns: &[&str]) -> Self

Add file patterns to exclude.

pub fn exclude_tests(self) -> Self

Exclude test files from results.

pub fn with_call_sites(self) -> Self

Include call site line numbers in results.

pub fn with_deduplicate_paths(self, deduplicate: bool) -> Self

Enable or disable path deduplication.

When enabled (default), each node appears only once using BFS traversal. When disabled, all paths are explored using DFS with per-branch visited sets, matching Python’s _build_caller_tree semantics where visited.copy() is used.

Use false for complete path exploration in diamond-shaped call graphs.

pub fn explore_all_paths(self) -> Self

Disable path deduplication to explore all paths (Python-compatible mode).

This is a convenience method equivalent to with_deduplicate_paths(false).

Trait Implementations

impl Clone for ImpactConfig

fn clone(&self) -> ImpactConfig

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for ImpactConfig

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl Default for ImpactConfig

fn default() -> Self

Returns the “default value” for a type.