Struct EmbedSettings 

pub struct EmbedSettings {

    pub max_tokens: u32,
    pub min_tokens: u32,
    pub overlap_tokens: u32,
    pub context_lines: u32,
    pub include_imports: bool,
    pub include_top_level: bool,
    pub token_model: String,
    pub algorithm_version: u32,
    pub scan_secrets: bool,
    pub fail_on_secrets: bool,
    pub redact_secrets: bool,
    pub include_patterns: Vec<String>,
    pub exclude_patterns: Vec<String>,
    pub include_tests: bool,
    pub enable_hierarchy: bool,
    pub hierarchy_min_children: usize,
}

Settings that control chunk generation

These settings affect the output of chunk generation. Changing settings will result in different chunk IDs, so the manifest tracks settings to detect when a full rebuild is needed.

Fields

max_tokens: u32

Maximum tokens per chunk (default: 1000 for code models)

min_tokens: u32

Minimum tokens per chunk (smaller merged, default: 50)

overlap_tokens: u32

Overlap tokens between sequential chunks (default: 100)

context_lines: u32

Lines of context around symbols (default: 5)

include_imports: bool

Include import statements as separate chunks

include_top_level: bool

Include top-level code outside symbols

token_model: String

Token counting model

algorithm_version: u32

Version of chunking algorithm (for compatibility)

scan_secrets: bool

Enable secret scanning

fail_on_secrets: bool

Fail if secrets detected (CI mode)

redact_secrets: bool

Redact detected secrets

include_patterns: Vec<String>

Include glob patterns (e.g., [“*.rs”, “src/**”]) Note: skip_serializing_if removed for bincode compatibility (requires all fields)

exclude_patterns: Vec<String>

Exclude glob patterns (e.g., [“tests/”, “.test.*”]) Note: skip_serializing_if removed for bincode compatibility (requires all fields)

include_tests: bool

Include test files (default: false)

enable_hierarchy: bool

Enable hierarchical chunking for improved RAG recall

When enabled, generates summary chunks for container types (classes, structs) that list their children with signatures and brief descriptions. This enables RAG systems to retrieve both high-level overviews and specific implementations.

Recommended for object-oriented codebases (Java, Python, TypeScript).

hierarchy_min_children: usize

Minimum number of children required to generate a summary chunk (default: 2, only relevant when enable_hierarchy is true)

Implementations

impl EmbedSettings

pub const CURRENT_ALGORITHM_VERSION: u32 = 1u32

Current algorithm version

pub const MAX_TOKENS_LIMIT: u32 = 100_000u32

Maximum tokens limit (DoS protection)

pub fn for_embedding_model(model: &str) -> Self

Get recommended settings for specific embedding model

Different embedding models have different optimal chunk sizes:

• voyage-code-2/3: 1500 tokens (large context window)
• cohere-embed-v3: 400 tokens (smaller model)
• openai-text-embedding-3: 800 tokens (balanced)
• sentence-transformers: 384 tokens (BERT-based)

pub fn validate(&self) -> Result<(), EmbedError>

Validate settings, return error if invalid

pub fn for_ci() -> Self

Create settings optimized for CI/CD pipelines

These settings fail on secrets and use stricter validation.

Trait Implementations

impl Clone for EmbedSettings

fn clone(&self) -> EmbedSettings

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for EmbedSettings

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

Formats the value using the given formatter.

impl Default for EmbedSettings

fn default() -> Self

Returns the “default value” for a type.

impl<'de> Deserialize<'de> for EmbedSettings

fn deserialize<__D>(__deserializer: __D) -> Result<Self, __D::Error>

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl PartialEq for EmbedSettings

fn eq(&self, other: &EmbedSettings) -> bool

Tests for self and other values to be equal, and is used by ==.

fn ne(&self, other: &Rhs) -> bool

Tests for !=. The default implementation is almost always sufficient, and should not be overridden without very good reason.

impl Serialize for EmbedSettings

fn serialize<__S>(&self, __serializer: __S) -> Result<__S::Ok, __S::Error>

where __S: Serializer,

Serialize this value into the given Serde serializer.

impl Eq for EmbedSettings

impl StructuralPartialEq for EmbedSettings