Struct ParseConfig 

pub struct ParseConfig {
    pub max_vector_length: usize,
    pub max_allocation_bytes: usize,
    pub mode: ParseMode,
    pub lazy_threshold: usize,
    pub bytecode_lazy_threshold: usize,
}

Configuration for parsing RDS files.

Allows customization of memory allocation limits to handle large files or enforce stricter safety constraints.

Safety Guardrails

Note: mode does NOT override safety guardrails (max_vector_length, max_allocation_bytes). These limits are enforced even in LazyMetadata mode to protect against corrupt headers.

Fields

max_vector_length: usize

Maximum number of elements allowed in a vector (default: 50,000,000)

max_allocation_bytes: usize

Maximum bytes that can be allocated for a single vector (default: 128 MB)

mode: ParseMode

Parsing mode (default: Full)

lazy_threshold: usize

In lazy mode, vectors smaller than this are always loaded (default: 10 elements)

Small vectors are typically metadata (dimensions, names, etc.) and should be loaded even in lazy mode for efficient structure inspection.

bytecode_lazy_threshold: usize

In lazy mode, bytecode constants smaller than this are always loaded (default: 1000 elements)

Bytecode parsing requires extracting tag/symbol names from character vectors. This threshold is higher than lazy_threshold because:

• Tags/symbols are almost always small (class names, function names, etc.)
• Loading a 1000-element character vector is cheap (typically <50 KB)
• If a tag exceeds this threshold, it will be skipped gracefully (placeholder used)
• This prevents failures when parsing S4 objects with embedded command history

Implementations

impl ParseConfig

pub fn new() -> Self

Create a new ParseConfig with default values.

pub fn with_max_vector_length(self, max: usize) -> Self

Set the maximum vector length.

pub fn with_max_allocation_bytes(self, max: usize) -> Self

Set the maximum allocation bytes.

pub fn with_mode(self, mode: ParseMode) -> Self

Set the parsing mode.

pub fn with_lazy_threshold(self, threshold: usize) -> Self

Set the lazy threshold (vectors smaller than this are always loaded in lazy mode).

pub fn with_bytecode_lazy_threshold(self, threshold: usize) -> Self

Set the bytecode lazy threshold (bytecode constants smaller than this are always loaded).

pub fn lazy_metadata() -> Self

Create a config for lazy metadata parsing.

This mode parses only structure and metadata without allocating vector/matrix data, enabling:

• Fast file inspection
• Handling files larger than available RAM
• Metadata extraction with minimal memory overhead

Note

Safety guardrails (max_vector_length, max_allocation_bytes) are still enforced to protect against corrupt headers.

pub fn large_data() -> Self

Create a config suitable for large scientific datasets (e.g., genomics).

Sets higher limits:

• max_vector_length: 500,000,000 (500M elements)
• max_allocation_bytes: 2 GB

pub fn unlimited() -> Self

Create a config with unlimited size (use with caution).

Only use this when you trust the input files and have sufficient memory.

Trait Implementations

impl Clone for ParseConfig

fn clone(&self) -> ParseConfig

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for ParseConfig

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

Formats the value using the given formatter.

impl Default for ParseConfig

fn default() -> Self

Returns the “default value” for a type.