All notable changes to this project will be documented in this file.
The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.
- Fixed mistake in KrakenUniq 'supported profilers' page, where the wrong parameter was listed in regards to which output file is supported by taxpasta (#152)
- Changed issues detected with compositionality constraints of profiles, i.e., relative abundances should always sum up to unity, to warnings rather than errors (#147).
- Modified the BIOM output to insert the global ranks into observation group metadata rather than creating it for each taxon (#149).
- Ensured that columns of profiles are read in correctly (#147).
- Used recent taxopy features to extract correct rank lineages (#149).
- Repaired the logic for adding taxonomic information to the output table (#139).
- Added an option
--ignore-errorsto thetaxpasta mergecommand. This allows ignoring individual profiles that cause errors, like empty profiles (#136).
- Created a special error message for empty profiles, making the cause of the error much clearer (#136).
- Internal restructuring of input validation and transformation services (#136).
- Added classifier support for KMCP profiles (#129).
- Added a command-line option
--add-rank-lineageto thestandardiseandmergecommands, which inserts a new columnrank_lineageto results that contains semi-colon-separated strings with the ranks (#130). - Added a taxonomy table to the BIOM output format when the
--taxonomyoption is used (#134).
- Reversed the order of lineages printed to output files (#131).
- Forced reading the
duplicatesandcoveragecolumns of KrakenUniq profiles as float (#123).
- Provided curl commands in the intro to quickly fetch example files to try out (#97).
- Added new tutorial on merging across classifiers, and warnings why taxpasta does not currently do this natively (#98).
- Added classifier support for ganon report files (#109).
- Made profile validation stricter, such that providing an input from another than the chosen profiler will usually result in an error (#101). The only exception is that Kraken2 standard profiles can be provided to the Centrifuge reader, and vice versa, since they are identical in format.
- Fixed a few broken internal links (#89).
- Simplified getting started tutorial and moved original to complex tutorial (#95).
- Improved help text for
--output-formatoption to clearly state that it disables any automatic detection of the output file format (#96). - Extended the MetaPhlAn profile reader to support version 3 & 4 profiles (#107).
- Inserted columns with taxonomic information like the name, rank, or lineage of a taxon at the beginning of the table rather than the end (#86).
- Enabled handling of MetaPhlAn profiles with taxonomy identifier -1 and combined all unclassified entries into one with ID 0 (#85).
- Generally increased the tolerance in the compositionality checks. Up to 1% deviation are now allowed; 2% for Bracken profiles (#84).
- Fixed the version string according to hatch-vcs output (#74).
- Increased the tolerance when comparing floats in the Bracken profile validation (#76).
- Fixed how the
taxopy.TaxDbis loaded. It was previously deleting local files and downloading a new copy from NCBI instead (#67).
- Handled metaphlan profiles with unclassified taxa in the lineage (#61).
- Created command line options to expand the standard profile with name, rank, and/or lineage taxonomic information (#60).
- Made the Krakenuniq reader more accepting (#57).
- Documented which columns are used in final standardised output profiles.
- Documentation on where the taxpasta-supported Kaiju output comes from (i.e.,
kaiju2tablerather than Kaiju itself) (#54). - Corrected the assumption that the first two entries in a kraken2 profile are
always unclassified and root node percentage of abundances. This is not the
case for profiles where 100% of reads are assigned to taxa. Kraken2 profile
validation now looks for the
'U'and'R'taxonomy levels explicitly and checks their sum.
- First release