IRMA-core v0.10.1

Release Notes for v0.10.1 - 2026-07-24

Added

• Added preliminary implementations for FastX views (within irma-records public API)
• A ValidatePaths trait is now available in the library portion for ensuring no path is passed as both an input and an output, and that all output paths are distinct
• WriteRecord is now compatible with FastX

Changed

• Integrated phase process now labels variant phase clusters using thresholded
  connected components. Phase labels are now ordered deterministically by
  cluster size (descending), lowest variant position, then lowest minority
  allele byte value.
• The format of the tallies in aligner has been altered when
  --tally-diagnostics is used.
• Changes trimmer behavior when masking primers to mask multiple discontinuous
  matches, if present
• Passing the same input path twice is now allowed
• When searching for equal paths in input and output files, paths are now
  canonicalized and soft links are handled
• Added image signing to CI and updated default image to Trixie
• Renamed GzipReaderPiped to GzipReaderInThread and altered API for it
• Renamed ReadFileZipPipe to ReadFileZipInThread
• use_file_or_zip_threaded has been removed in favor of a separate
  decode_in_thread method
• Removed helper methods for adapter trimming from ReadTransforms, and added
  process_adapter (within irma-records public API)
• Improves preprocess error output via new PrintWarning trait
• preprocess, merge-sam, xflate, and aligner now include checks for the
  input and output paths
• Introduces a Finish trait for closing writers, and modifies the
  WriteRecords trait to use this

Fixes

• Fixed bug with masking primers in trimmer where overlapping ranges of
  matched primer k-mers could cause no masking to occur
• Fixes sequence count estimation in sampler to account for multiline FASTA
  files
• merge_sam no longer panics on SRA qnames that are parseable but lack an
  explicit read side
• Fixes trimmer bug with 2-in-1-out reads with widow filtering disabled where
  extra read(s) in either file would not get handled
• phase no longer truncates variants table file on malformed or corrupted
  variants matrix files
• Errors while writing the footer for zipped outputs are now properly propagated
• DPKG metadata is no longer incorrectly deleted for container images (bug
  exposed in v0.9.1)
• Fixed incorrect size_hint and try_fold impls for ZipReads iterator
• An error for a non-ascii byte is no longer encountered in
  ReadTransforms::fix_header (within irma-records public API)
• An erroneous unreachable! has been replaced with a descriptive panic in
  process_barcode (within irma-records public API)

IRMA-core v0.10.0

Release Notes for v0.10.0 - 2026-06-05

Added

• IRMA-core has been refactored into a workspace and now allows git-based
  consumption of some functionality via an initial irma-records crate

Changed

• Bare metal releases for macOS + Intel compile with x86-64-v3 (avx2) explicitly

Fixes

• Fixes error message when an ID for paired read headers fail to be parsed
• Fixes a bug in xflate and preprocess where tab characters corrupt
  deduplication formatting

IRMA-core v0.9.1

Release Notes for v0.9.1 - 2026-04-15

Fixes

• Fixes error in aligner tally diagnostics when reverse complement is not enabled
• Fixes error in phase when trying to read matrix file when no variants are present

Changed

• Adds more context to paired read errors in trimmer, xleave, preprocess,
  and sampler processes
• Hardens Dockerfile using dhi.io (requires login)

IRMA-core v0.9.0

Release Notes for v0.9.0 - 2026-03-06

Added

• Adds verbose flag to sampler for printing total number of reads in
  original input and number of downsampled records to stderr
• Adds verbose flag to trimmer for printing diagnostic output on number of
  reads trimmed to stderr
• Adds phase process for phase clustering assignment and labeling of genes
• Adds more detailed error messages with more context, including file paths for
  many IO errors

Fixes

• Adds requirement in sampler arguments to provide either a
  --subsample-target or a --percent-target, fixing unreachable error when
  neither is provided
• Fixes a bug where leading whitespace in a record file can cause the program to
  stall (in sampler and aligner)
• Allows repeated input/output files from processes if they are device files
  (e.g., /dev/null)
• merge-sam now respects the seed set in the environmental variable
  IRMA_SEED
• Fixes a bug where not all paired headers were being checked in preprocess,
  sampler, trimmer, and xleave

IRMA-core v0.8.1

Release Notes for v0.8.1 - 2025-12-02

Fixes

• aligner with --rev-comp and --best-match now produces the correct positions and CIGAR strings

IRMA-core v0.8.0

Release Notes for v0.8.0 - 2025-11-21

Added

• Adds standalone aligner process for perform pairwise local sequence
  alignment with the Striped Smith Waterman algorithm

Changed

• The trimmer process now can perform deinterleaving, as well as widow filtering with interleaved inputs and outputs

IRMA-core v0.7.0

Release Notes for v0.7.0 - 2025-10-09

Added

• Adds standalone xleave process for interleaving and de-interleaving
  paired-end FastQ and FASTA files

IRMA-core v0.6.2

Release Notes for v0.6.2 - 2025-09-25

Fixes

• xflate recently produced invalid FASTQ during inflation. This has been
  fixed. This bug was not surfaced in IRMA.

IRMA-core v0.6.1

Release Notes for v0.6.1 - 2025-09-09

Changed

• The fastq-converter process has been removed.
• More context is added for IO errors in sampler

Fixes

• Fixes bug where SRA paired-end headers could be could be incorrectly flagged
  as unmated when --filter-widows was selected in trimmer, or with paired
  reads in sampler
• Fixes a reproducibility issue in sampler where the seed wasn't being used in
  one case

IRMA-core v0.6.0

Release Notes for v0.6.0 - 2025-08-29

Added

• Adds a new standalone process called sampler that provides support for
  randomly downsampling FastQ or FASTA files. Deinterleaving is also supported.

Changed

• Use threads and anonymous pipes (in Rust) when decompressing zipped inputs
• Improve error messages to include file names in trimmer and preprocess
• Other dependencies updated, such as a faster foldhash.

Fixes

• mergeSAMpairs: updated Zoe dependency (in v0.0.19, using v0.0.20) fixes hard clipping in merged reads. Reader
  error messages are also improved generally.