This is a record of changes made between each kMer release.
Release date to be decided.
Released on August 14th, 2015.
- Option to create a k-mer profile per FASTA record instead of per FASTA
kpal count --by-recordon the command line or kpal.klib.Profile.from_fasta_by_record in the Python API).
- GitHub project moved to LUMC/kPAL.
- Change default precision to 10 decimals.
Released on November 21st, 2014.
- Save profiles from several files to one file (
Released on November 21st, 2014.
- Fixed a major bug that made the command line interface unable to start.
Released on November 18th, 2014.
- Rename from kMer to kPAL (k-mer profile analysis library). The Python
package is now kpal (was k_mer). The command line interface is now
Released on October 3rd, 2014.
- Fix typo in setuptools trove classifier which made it impossible to push to PyPI.
Released on October 2nd, 2014.
- Also count k-mers if k equals the length of the string.
- Python 2.6 compatibility.
- Added unit tests and a tox configuration.
- Use a NumPy ndarray for storing k-mer counts.
- New multi-profile HDF5 file format (see k-mer profile file format).
- Fix splitting a profile for calculating balance. Palindromes were previously not taken into account when splitting a profile. We now double all counts, so palindrome counts can be evenly distributed over both sides (see GitLab #1).
- Our own implementation of a vector’s median contained two bugs. Better to use a library for this.
- Fix Euclidean distance between two vectors. Don’t add one to the sum of squares. The distance between two empty vectors should be 0, not 1.
- Rename k_mer.klib.kMer to k_mer.klib.Profile.
- Support Python 3.3 and 3.4 (see below).
- Generalize custom function arguments (see below).
- Travis CI configuration.
- Sphinx documentation including a user guide and API reference.
- Renamed the index command to count.
- Renamed the diff command to distance and the builtin pairwise distance functions from diff-prod and diff-sum to just prod and sum.
Support Python 3.3 and 3.4¶
TL;DR: kMer supports Python 2 and 3 and every module has the following line at the top:
>>> from __future__ import (absolute_import, division, print_function, unicode_literals)
We now support Python versions 2.6, 2.7, 3.3, and 3.4 in a single codebase without using 2to3. We don’t support Python 3.2 because BioPython does not.
We use the Python future package as a compatibility layer between Python 2 and Python 3. The goal is to use a single, clean Python 3.x-compatible codebase to support both Python 2 and Python 3 with minimal overhead.
Most changes are quite straightforward (e.g., absolute imports, print statement, division operator). The main painpoint is of course the bytestring versus unicode story. We now import unicode_literals in each module and maintain that all text in kMer is unicode (unicode in Python 2, str in Python 3).
Generalize custom function arguments¶
Custom function arguments in the command line interface can now be either a Python expression or importable name. For example, all commands accepting a summary function argument, also except a custom summary function argument which should be one of:
- A Python expression over the NumPy ndarray values (e.g.,
- An importable name (e.g.,
package.module.summary) that can be called with an ndarray as argument.
Likewise for custom merger and pairwise functions (here the expression is over the two NumPy ndarrays left and right).
Released on July 3rd, 2014.
- Usage of the Euclidean distance is now handled differently, breaking backwards compatibility.
- Added Cosine similarity measure and generalised distance parameters.
- Fixed broken setup script.
- Added custom merging functionality.
Released on March 23rd, 2014.
- New command line interface, using positional arguments for required parameters.
- Added checking for existing files to prevent overwriting them.
- Fixed a bug in the scale subcommand that prevented scaling.
- Added a version parameter.
- Updated the homepage.
- Made code PEP 8 compliant.
- Switched to Sphynx docstrings.
- Added keyword selection for distance and smoothing functions.
- Added support for custom distance and smoothing functions.
- Added CHANGELOG and README.
Released on September 24th, 2013.
- Start of log.