Struct FileTuples 

pub struct FileTuples { /* private fields */ }

Handle to the merged external sort file for sequential buffered access.

Unlike the previous MmapTuples, this does NOT mmap the file. Instead, each access pass opens a fresh BufReader, reading tuples sequentially with a ~4 MB buffer. This keeps RSS proportional to the buffer size instead of the file size (~8-10 GB for large datasets).

Implementations

impl FileTuples

pub fn num_tuples(&self) -> usize

Total number of tuples in the merged file.

pub fn path(&self) -> &Path

Path to the underlying merged file (for deriving temp file names).

pub fn bucket_iter(&self) -> Result<FileBucketIter>

Create an iterator that yields one BucketScan per unique minimizer, scanning the file sequentially via buffered I/O.

Each call opens a fresh BufReader — no state is retained between passes, and no mmap pages linger in memory.

pub fn sequential_reader(&self) -> Result<SequentialTupleReader>

Create a sequential tuple reader for forward-only access.

Returns tuples one at a time via read_next(). Used by the sparse/skew index builder for its sequential pass over all tuples.

Trait Implementations

impl Drop for FileTuples

fn drop(&mut self)

Executes the destructor for this type.