Establishes foundational infrastructure across multiple crates by introducing unified persistent bit matrix storage with columnar, packed, and implicit variants, alongside De Bruijn graph node encoding and unitig iteration logic. Adds a macro-driven parallel pipeline scheduler featuring NUMA-aware runners, bounded channels, and memory budgets to enforce concurrency limits. Implements streaming nucleotide parsers with pooled page buffers for FASTA, FASTQ, and Genbank formats, complemented by system resource monitoring, progress tracking, and stage profiling utilities. Collectively, these changes provide the core data models, execution frameworks, and I/O pipelines required for downstream k-mer indexing and analysis workloads.
- Added `use std::marker::PhantomData;` to prepare for generic scheduler implementations
- Ensures type safety and avoids unused lifetime/type parameters warnings
- Change `rope_tell()` return type from Option<usize> to usize, always returning cursor's absolute position (offset if unmoved).
- Update all call sites to remove `.unwrap_or(...)` around `rope_tell()`.
- Add new method `<Rope>::truncate(pos)`, replacing `split_off(...).map(|_| ())`.
- Refactor FASTA/FASTQ normalizers to use a single mutable write cursor (`wc`) and document the protocol.
- Simplify `end_segment()` logic: commit segment with 0x00 if length ≥ k, else reset.
- Improve documentation for write-cursor protocol and rope truncation semantics.
- Introduce SeekMode::Rope for absolute rope-index positioning
- Add CursorState.offset field to support local coordinate systems per cursor
- Implement ForwardCursor.cursor() and BackwardCursor(cursor()) to create sub-cursors with independent offsets
- Update tell(), get(i), set i, len() to use local coordinates (relative offset)
- Add rope_tell(), reset()—deprecate old absolute behavior in favor of offset-aware API
- Add comprehensive tests for sub-cursor semantics, including write/reset and bounds checking
Eric Coissac