Alpha-0.3.2

v0.3.2

• alpha
• changes
  • rename cluster classes and methods IDBCAN and SIDBCAN to DBICAN and SDBICAN respectively
  • minor changes for compatibility with python 3.4
  • added Makefile
• bug fixes
  • simplify and fix travis CI
• documentation
  • add MIT licence
  • improve method docs
  • add graphics to method docs #74
  • document requirements

Alpha-0.3.1

v0.3.1

• alpha
• new features
  • maximum count proportion # 115
    • added 'max_count_proportion' field to output (was already used for determining colour)
    • renamed '--no-colour' switch to '--no-max-count-proportion'
• changes
• package version now stored in single location 'version.py'

Alpha-0.3.0

• new features
  • informative reads
    • option to include soft-clipped tips in place of their mate informative read
    • option to exclude full-length informative reads completely
  • output formats
    • outputs with the .gz extension are compressed with block-gzip rather than regular gzip #104
    • option to output tab-delimited plain text files suitable for tabix indexing when block-gzipped #103
  • extract-informative temporary files #106
    • temporary files are now written to the same location as the output by default
    • option to not remove these files
    • option to write them to a temp location provided by the operating system
• changes
  • added Biopython as a dependency to support block-gzip compression
  • refactor submodules to hide application specific code
  • refactor clustering classes/methods to use new names
    • non-hierarchical method named IDBCAN (Interval Density Based Clustering of Applications with Noise)
    • conservative-hierarchical method named SIDBCAN (Splitting-IDBCAN)
    • aggressive-hierarchical method named SIDBCAN-aggressive and deprecated
• bug fixes
  • python 3 shebang lines #98
  • add MANIFEST.in to allow for github install with pip #89
• documentation
  • named clustering algorithms loosely based on DBSCAN
    • non-hierarchical method named IDBCAN (Interval Density Based Clustering of Applications with Noise)
    • hierarchical method named SIDBCAN (Splitting-IDBCAN)
  • many corrections to documentation of clustering algorithms #99
  • specify tabix command to index block-gzipped output

Alpha-0.2.0

• new features
  • new 'conservative' clustering method #91
• changes
  • new 'conservative' clustering method used by default #91
  • simplified clustering/splitting method selection with flag '--splitting-method'
  • changed library imports to include loci, fingerprint, fingerprintio and cluster
  • tidied cluster.py submodule code
• bug fixes
  • fixed miscalculation of initial cluster support #90
  • fixed miscalculation of parent vs child support #95
  • fixed name collision between sub module cluster and function loci.cluster
• documentation
  • updated method.rst with new clustering method
  • updated usage.rst with new clustering method
  • updated CLI help text
  • updated cluster.py docstrings with new method
  • added metadata for PyPI

Alpha-0.1.4

Substantial changes to code base in order to accommodate new features.
Command-line-tool names and arguments have changed - see the documentation.
Core clustering algorithms are unchanged and will identify the same clusters/features as before.

• Renamed tools
  • remove tef wrapper as it is likely to have a name collision with something
    • tef fingerprint/compare are combined into single tool tefingerprint #63
    • tef preprocess is now tef-extract-informative
    • tef filter-gff is now tef-filter-gff (note: this tools behavior has changes substantially to handle the new gff output)
• Refactored modules
  • core sub-module loci.py re-written to be more flexible #82
    • single data structure for for representing collections of loci
    • arbitrary string length limit removed (use of python string in place numpy strings)
    • split tool logic into separate sub-module fingerprint.py
• New Features
  • tefingerprint
    • trim buffered clusters to extent of read tips
    • count n most common elements per sample in each bin #63 #81
    • use gff annotation for tagging known elements #80
    • join paired clusters using gff annotation #78
    • output files (gff, csv) are optionally pipe-able
    • output files (gff, csv) contain more detailed data
    • can read(anotation gff)/write files compressed with gzip or bz2 #86
    • escape special characters in gff files with percent encodings
  • tef-filter-gff
    • changed to handle new gff output #84
    • use of --any and --all contexts for combining filters
    • unix style wild cards for matching multiple fields
    • read and write gz abd bz2 compressed files #86
    • escape special characters with percent encodings
    • read gff from standard in

Alpha-0.1.3

• Performance improvements:
  • Significantly reduced memory usage and improved speed for preprocess #66 #70
  • Reduced memory usage for fingerprint/compare io operations #68
  • Reduced memory usage for filter-gff #67
• Fixes
  • Corrected feature-csv output on large arrays #58 #72
• Documentation
  • Split readme into multiple documents
  • Switched to .rst for documentation
  • Added description of methods

Alpha-0.1.2

• Faster output of CSV files (#58)
• Tidied up CLI arguments (#55 #62)
  • Replace underscores with dash
  • Use of flags for Boolean arguments
  • Updated Readme with changes
• Tests
  • Integration tests for fingerprint, compare and filter-gff
  • Further tests for preprocess

Alpha-0.1.1

• Minor fixes for for cluster support calculation #53
• Default selection of starting epsilon value now matches Campello et al 2015
• Update terminology to reflect that used in Campello et al 2015
• Tests

Alpha-0.1.0

• Pre-Processing
  • Corrections for reverse-complemented reads when extracting dangler reads
  • Inclusion of soft-clipped sections from the outer end of proper-mapped pairs
  • Separated mapping reads to repeats from pre-processing script
• Fingerprinting/Compare
  • Quality score filtering options
  • New output format options
• Other
  • Bump version to 0.1.0
  • Change status to alpha
  • Updated documentation
  • Additional tests

Tabular Data Structures

Improvements:

• Nicer data structures for use as a library #42
• Easy inter-op with Numpy, Pandas and Plotting libraries in python #42
• Improved buffering of comparative bins
• Multi-process pipelines return results to parent process rather than printing directly to file
• GFF output is sorted by chromosome, start position, stop position #32
• Added better examples to readme #40
• improved testing of fingerprint and compare data methods #6
• Compare output is no longer nested and contains normalised read counts #4 and is coloured by read count proportions #43
• filter_gff program is simplified (no longer deals with nested gff)
• renamed module to tefingerprint and CLI to tef #3