Version 2.1.0

• Fixes issues in docker packaging.
• Adds missing #include for to files that are using std::numeric_limits
• Fixes error message when piping export output.
• Allows modifying BGZF compression level and sets the default level to 6 for both BGZF and ZSTD.
• Fix for environments where char is unsigned.
• Fix for setting max block size of 65536.
• Updates shrinkwrap to fix broken zlib URL.
• Fixes bug in automatic updates to standard INFO fields.
• Validates data types when writing BCF.
• Crossplatform fix for sprintf with 64-bit integers.
• Improves speed of writing VCF individual data.
• Fixes for type thinning.
• Removes FORMAT column from header line with no individual data.
• Fixes index file check during import/export.
• Allows reading of VCF files that are missing INFO or FORMAT headers (assumes String type).
• Terminates immediately after write failure is encountered during import/export.
• Fixes bug where last record of VCF file was dropped.
• Fixes "Number" in SAV file GT header.
• Removes IDX header field when writing new files.
• Bug fix for header IDX being out of order.
• Fixes bug in copy_as_dense().
• Index query changes to support temporary files.
• Fixes serialization bugs related to missing and end_of_vector values.
• Fixes for is_missing() and is_end_of_vector().
• Fixes is_end_of_vector() for floating point types.
• Adds support for "end of vector" with stride_reduce().
• Fixes region compare bug.
• Adds support for INFO/END when performing index queries.
• Fixes BCF rlen and adds support for indexing INFO/END.
• Inherit site_info assignment operator to variant class.
• Fixes deserialize_vec when size and type are zero.
• Fixes parsing of VCF Flag INFO fields.
• Removes "no-tree-vectorize" attributes now that data vectors are aligned.
• Replaces flat buffer design for typed_value so that data vectors are memory aligned.
• Adds get-dependencies.cmake as alternative to cget.
• Fixes issues with Github workflow.
• Fixes big endian bugs.
• Updates endianess swapping behavior to reflect latest BCF spec.
• Fixes bug in PBWT unsort.
• Gracefully skips PBWT for vectors with bit widths greater than 16.
• Bug fixes for multi-format files.
• Updates shrinkwrap version to fix open_mode bug.
• Adds conda instructions to readme.
• Updates shrinkwrap to fix link error when liblzma is not available.
• Fix for sample subsetting of SAV v1 files.
• Add experimental sparse regression tool.
• Fixes examples in README.
• Fixes eval link error.
• Adds find_dependency to export config for proper import of shrinkwrap into downstream projects.
• Moves SAVVY_VERSION defintion from CLI to library target.
• Adds support for individual data of type String.
• Links shrinkwrap package instead of directly linking zlib and zstd libs.
• Reorders fields in auto-generated FORMAT headers to be consistent with VCFs in htsspec examples.
• Removes extra line from end of stat-merge output.
• Adds --filter option to stat sub-command and removes n_singletons column from per-sample output.
• Adds per-AC stats.
• Adds stat-merge sub-command and per-sample statistics to stat sub-command.
• Adds support for tabix (.tbi) index files.
• Fixes regression in CSI interval merging.
• Fixes bad flag being set when EOF is reached with sites-only VCFs.
• Updates CSI index parsing to support unbounded ends and mimic htslib behavior.
• Removes --data-format option from export sub-command.
• Fixes VCF parsing of sites-only files.
• Fixes exit status for head sub-command.
• Adds sparse_vector class (compressed_vector alternative).

v2.0.1

• Fixes critical bugs in stride_reduce routines.
• Adds support for reading files that have no samples.

v2.0.0

This is a major version update, which adds support for the SAV 2.0 file format. Some breaking API changes have been made to support the new format:

• There is now only one reader class and one writer class to support all file formats.
• site_info::alt() has been replaced with site_info::alts().
• site_info::prop has been replaced with site_info::get/set_info.
• id(), quality(), and filters() methods have been added to site_info. These fields were previously retrieved with site_info::prop.
• Individual level data is now retrieved and set via variant::get/set_format instead of read/write methods.

v1.3.0

• Adds eof() to generic reader.
• Adds set_policy to generic reader.
• Verifies that --sample-ids-file exists.
• Merges headers during sav concat.
• Adds --update-info option to CLI.
• Merges info fields during sav merge.
• Fixes sav merge merge for multi-allelic variants.
• Fixes indexing bug that excluded last block of variants.
• Fixes crash when closing vcf/bcf files.

v1.2.0

• Adds missing sub-commands to manuals.
• Adds manuals to packages - Fixes static linking of libc in cli packages.
• Adds experimental concat sub-command to cli.
• Adds 'sav head' and updates 'sav rehead'.
• Adds --headers option to export and handles phasing header for VCF output.
• Improves error handling in vcf reader.
• Automatically appends .s1r extension to .sav files when running stat-index.
• Adds manuals for CLI usage.
• Uses compressed vectors for some conversions.
• Fixes options not being passed to sav writer in export.
• Copies some CLI options from import to export.
• Fixes handling of special info fields in sav export.
• Adds default info headers when export to sav format.
• Adds experimental filter expressions.
• Adds missing license notice.
• Auto generates some property fields and changes read_if behavior.
• Fixes PASS being removed in filter field.
• Removes extra conditions for rounding in sav reader.

v1.1.0

• Fixes finding cget version of zlib.
• Improves support for packaging/distribution.
• Adds empty vector policy and improves handling of read failures in CLI.
• Corrects Flag info field output in VCF files.
• Improves float parsing.
• Ensures distinct info fields.

v1.0.0

Release Notes

• Adds more graceful failures with invalid input.
• Adds cpack packaging.
• Changes fmt enum names (breaking change).
• Adds method for retrieving position range from s1r index.
• Adds uuid and file level ploidy to SAV.
• Adds variant_group_iterator.
• Adds region::merge.
• Fixes chromosome name queries from vcf files.
• Adds bgzf compression to sav export.
• Adds --bounding-point to sav and updates enum names.
• Adds regions-file support to sav import/export.
• Fixes reset_region for vcf reader.
• Fixes bad vs eof logic in SAV readers.
• Adds stat-index sub-command to sav.
• Prevents fileformat header from being written to SAV format.
• Adds genotype matrix example.
• Removes dependency on size() method for vector classes.
• Prefers linking to static when BUILD_SHARED_LIBS is off.
• Adds support for exporting GP from SAV.
• Adds support for VCF writing of genotype probabilities.

Version 1.0.0-alpha2

• Removes variadic constructors from generic API.
• Updates s1r indexes.
• Adds sav CLI.

Version 1.0.0-alpha

First pre-release version.