Update change log [skip ci]

CHANGELOG.md

@@ -2,6 +2,29 @@
 All notable changes to this project will be documented in this file.
 This project adheres to [Semantic Versioning](http://semver.org/).
 
+## [Unreleased]
+
+### Added
+- The `kevlar count` operation now supports masks and 8-, 4-, or 1-bit counters (see #277 and #291).
+- A Jupyter notebook and supporting code and data for evaluating kevlar's performance on a simulated data set (see #271).
+- New flags for filtering gDNA cutouts or calls from specified sequences (see #285).
+- New filter that discards any contig/gDNA alignment with more than 4 mismatches (see #288).
+- A new feature that generates a Nodetable containing only variant-spanning k-mers to support re-counting k-mers and computing likelihood scores in low memory (see #289, #292).
+- A new `kevlar cutout` command that mimics the behavior of `kevlar localize` but does so much more efficiently (see #294).
+
+### Changed
+- Ported augfastx handling from `kevlar.seqio` module to a new Cython module (see #279).
+- Dynamic error model for likelihood calculations is now an configurable option (see #286).
+- Cleaned up overlap-related code with a new `ReadPair` class (see #283).
+- Updated `kevlar assemble`, `kevlar localize`, and `kevlar call` to accept streams of partitioned reads—previously, only reads for a single partition were permitted (see #294).
+
+### Fixed
+- Minor bug with .gml output due to a change in the networkx package (see #278).
+
+### Removed
+- Buggy home-grown greedy assembler dropped (see #279). Some parts of the overlap code retained and refactored (see #283).
+
+
 ## [0.5.0] 2018-06-14
 
 ### Fixed