Changelog¶

v1.2.0 (2023-12-11)¶

#124: Added support for chunking FASTA reads to read_paired_chunks. Previously, only FASTQ was supported.

#116: Added experimental support for reading unaligned BAM files (single-end only at the moment). This uses a custom, highly efficient BAM parser, making dnaio faster than htslib in this particular case.

#99: SequenceRecord initialization is now faster, which also provides a speed boost to FASTQ iteration. SequenceRecord.__new__ cannot be used anymore to initialize SequenceRecord objects.
#96: open_threads and compression_level are now added to open as arguments. By default dnaio now uses compression level 1 and does not utilize external programs to speed up gzip (de)compression.
#87: open can now open more than two files. The file1 and file2 arguments are now deprecated.

#79: Added a records_are_mates function to be used for checking whether three or more records are mates of each other (by checking the ID).
#74, #68: Made FASTQ parsing faster by implementing the check for ASCII using SSE vector instructions.
#72: Added a tutorial.

Preliminary documentation is available at <https://dnaio.readthedocs.io/>.
#53: Renamed Sequence to SequenceRecord. The previous name is still available as an alias so that existing code will continue to work.
When reading a FASTQ file, there is now a check that ensures that all characters are ASCII.
Function record_names_match is deprecated, use is_mate instead.
Added reverse_complement.
Dropped Python 3.6 support as it is end-of-life.

@rhpvorderman contributed many performance improvements in #15, #17, #18, #20, #21, #22, #23. Reading and writing FASTQ files and reading of paired-end FASTQ files was sped up significantly. For example, reading uncompressed FASTQ is 50% faster (!) than before.
#28: Windows support added