Add in-memory GFA-reduced output with parallel processing and optional input-order preservation#60

Merged

rob-p merged 38 commits into

cf1-for-piscemfrom

copilot/add-in-memory-unitigs-function

Jan 28, 2026

Commits on Jan 27, 2026

Initial plan
Copilot
committed
Add implementation of output_maximal_unitigs_gfa_reduced_in_mem function

Copilot
and
rob-p
committed
Fix code review issues - add path buffers and complete implementation

Copilot
and
rob-p
committed
Address code review feedback - use RAII lock guards and fix warnings

Copilot
and
rob-p
committed
Final implementation complete - ready for review

Copilot
and
rob-p
committed
Update gitignore to exclude CodeQL build directory
Copilot
committed
Redesign in-memory implementation to eliminate file-based buffering

Copilot
and
rob-p
committed
Fix segfault by calling reset_extreme_unitigs for each sequence

Copilot
and
rob-p
committed
Enable parallel sequence processing and add --collate-output-in-mem flag

Copilot
and
rob-p
committed
Fix segfault by properly initializing unitig tracking vectors

Copilot
and
rob-p
committed
Add retain_input_order parameter to enable completion-order output

Copilot
and
rob-p
committed
Fix segfault by preserving job data when threads are reused

Copilot
and
rob-p
committed
Use reset_extreme_unitigs() for proper initialization

Copilot
and
rob-p
committed
Fix race condition by updating job metadata before assigning task

Copilot
and
rob-p
committed

Commits on Jan 28, 2026