Nucleaze v1.5.0-alpha

• Added optional approximate mode with --fpr using a Bloom filter
• Accelerated reference indexing by decreasing chunk sizes
• Collapses repetitive reference data prior to indexing
• Bounded how many read chunks run at once to avoid maximum memory spikes

Nucleaze v1.4.2

• Improved error handling and messaging for:
  • Missing essential arguments
  • Invalid output paths
  • K-mer length
• Switch from u32 to u64 to prevent overflow w/ many reads

Nucleaze v1.4.1

• Switch to FxHashSet intermediate for reference indexing
  • Large decrease in peak memory usage
  • Moderate decrease in reference indexing performance
  • Processing speed unchanged

Nucleaze v1.4.0

• Remove Bloom filter and switch to sharded HashSet for k-mer storage
  • Fully O(1) lookups
  • Improved performance
  • Lower memory usage
• Nested vectors for reference k-mer indexing
  • Improved performance
• Better error handling for empty/unusable data files

Nucleaze v1.3.0

• Add Bloom filter for quick O(1) rejections
  • Configurable memory (--bloomsize)
  • Linear reference index scaling
• Nested vector reference index
  • Improved performance
• Configurable memory cap (--maxmem)

Nuclease v1.2.0-beta

• Switch to region memory management system, reducing allocation function call and memory management overhead
  - Speed and peak memory usage imporved
• Add streaming batch output, which outputs sequence chunks as they are processed
• --ref marker is now optional, one of --ref or --binref is necessary, however
• Optimized for Linux (x86)

Nucleaze v1.1.1-beta

• Increased performance via improved encoding function
  • Both sides of program are encoded simultaneously instead of using the canonical function from the Needletail crate
• Memory limiting has been (temporarily) removed

Nucleaze v1.1.0-alpha

• --saveref args to save serialized reference index
• Portable memory limiting with cap crate
• Better help documentation (no more duplicates)
• --order tag to receive ordered output
• Failure assertions upon omitting the --interleaved flag for two output files and 1 input file
• Output paths are optional
• Comprehensive test suite
• Fix processing mode errors

Nucleaze v1.0.1-alpha

• Removed memory limiting because of compilation errors on Linux

Nucleaze v1.0.0-alpha

• Memory limiting only works on Linux
• Piping from stdin and to stdout
• Memory limiting
• FASTQ format for outputs
• Paired read input/output support
  • With multiple files or singular interleaved file

Request new features!