Version 1.7.3

Version 1.7.3

What's New?

• Fixed an issue with longer queries that wasn't properly addressed in the previous release. Now, all sizes of kmers are supported on 64-bit architectures
• updated the number of threads argument for openMP from 8-bit to 32-bit. This allows for more than 255 threads in a parallel search.

Version 1.7.2

What's New?

Bug Fixes

• Fixed a bug where long queries could create undefined behavior.

Version 1.7.1

Version 1.7.1

What's new?

• add awFmFindDatabaseHitPositionSingle() function for finding the global sequence position from a single bwt position.
• add awFmNucleotideBacktraceReturnPreviousLetterIndex() function to find the previous letter while backtracing.
• improved the performance of the awFmGetLocalSequencePositionFromIndexPosition() function.

BugFixes

• Fix issue where ambiguity characters in the suffix of a query that gets looked up with the kmer seed table could cause undefined behavior.

Version 1.7.0

Version 1.7.0

What's new?

• An issue was fixed where the index could report hits that span the boundaries between multiple sequences when built from a fasta. Now, ambiguity characters are inserted between each sequence. This ensures that any reported hits belong to an individual sequence. While existing .awfmi files will still function, we recommend rebuilding your indexes to address this issue.

Version 1.6.3

Version 1.6.3

Changes:

• improve documentation for backward search functions to make the arguments clearer
• The Cmake build system, as well as the legacy makefile's static build now copy all necessary headers and libs to the build/ directory.
• Renamed awFmDatabaseSingleKmerExactMatch() function to awFmFindSearchRangeForString()
• Added awFmCreateInitialQueryRangeFromChar() function to init a search range for a single character

Bug Fixes:

• Fixed issues with the build system that would leave the static library missing dependencies

Version 1.6.2

What's new?

• Publicly expose awFmFindSearchRangeForString() function, allowing users to find the search range for a single kmer string.

Version 1.6.1

Version 1.6.1

What's new?

• Added function for starting a new search range with a single character

Bug Fixes

• Fixed an issue where locate positions read from a sampled suffix array weren't being incorrectly indexed.
• When attempting to read a sequence header, the index now correctly checks to see if it was made from a fasta. If not, there is no header to read and the function returns an error code.

Version 1.6

What's new?

Added awFmGetNumSequences() function to return count of sequences used to build the index when built from a fasta.
Updated 'restrict' keyword usage in the header file to allow c++ compilers to work with the library.

Bugfixes:

Fixed an issue where the fasta headers and metadata weren't loading when the index was read from file.

Version 1.5

What's new?

• The AwFmAlphabetNucleotide enum type has been replaced with AwFmAlphabetDna and AwFmAlphabetRna. This allows AwFmIndex to handle RNA sequences.
• Better C++ support. When compiled under a cxx-context, the 'restrict' keyword is replaced with the c++-supported 'restrict' keyword.

Version 1.4

What's new?

• Added functions awFmReturnCodeIsFailure() and awFmReturnCodeIsSuccess() to public header for easily checking return codes.
• Removed "allowOverwrite" function argument from AwFmCreateIndex() and AwFmCreateIndexFromFasta(). The boolean switch to allow or deny overwriting an index file was never actually useful, but could create very specific conditions where if the user ignored failure return codes, they could perform locate() operations and receive the incorrect results. Now, the create() functions will always allow overwriting of the index file.

Bug fixes:

• Fixed bug where locate() operations could fail if the index was built from fasta and the suffix array was stored in memory.