0.9.0

Version 0.9.0

Released on Jul 11th 2019

Changed

• BedEntry.unpack logic has changed significantly. It now throws on missing fields. The exception can be
  used to determine the offending field and adjust the BED format correspondingly.
• BedEntry.unpack now controls the extra fields parsing with a flag instead of a number. The old signature
  remains available for compatibility sake, but is deprecated.
  New
• BedEntryUnpackException is thrown when BedEntry.unpack fails to parse an entry. The exception properties
  provide the reason and the number of the field which caused parsing to fail.
  Fixed
• ExtendedBedEntry.score is now Int instead of Short. The reason for this is that many
  BED file providers (e.g. MACS2, SICER) don't respect the UCSC standard which limits the score
  to 0..1000 range, and we want to be able to parse those files.

0.8.4

Version 0.8.4

Released on Feb 12th 2019

New

• ExtendedBedEntry.getField is a public method that returns the ith field of the entry
  as an instance of a correct type, whatever that might be, with no need for conversion
  to and from String.

0.8.3

Version 0.8.3

Released on Oct 24th 2018

New

• ExtendedBedEntry.rest is a public method to return the list of fields except for
  the first three obligatory BED columns. Is used during ExtendedBedEntry.pack.
  It could be convenient when we want to iterate through optional fields. For simplicity
  and to avoid code duplication the method returns a string representation of extended
  bed entry fields (the same values as joined by pack method.) instead of original values.
  Such behavior is consistent with values which user will see in a BED/BigBed file.

0.8.2

Version 0.8.2

Released on June 26th 2018

Fixed

• BigWig.read & BigBed.read should close factory in case of exception
  while creating big file
• Default buffer size for EndianSynchronizedBufferFactory and EndianBufferFactory fixed
  to BetterSeekableBufferedStream.DEFAULT_BUFFER_SIZE

0.8.1

Version 0.8.1

Released on June 20th 2018
New

• BigFile.source public property added

Changed

• Report stream source in exception messages from BetterSeekableBufferedStream

0.8.0

Version 0.8.0

Released on April 29th 2018

New

• HTTP/HTTPS urls support for BigWig, BigBed, TDF files. Server is
  supposed to support accept-ranges http header
• New buffer factories for files/urls reading based on hts-jdk
  SeekableStream:
  • Without concurrent BigFile access support: EndianBufferFactory
  • With concurrent BigFile access support: EndianSynchronizedBufferFactory,
    EndianThreadSafeBufferFactory
• Big files read and write methods accepts cancelledChecker closure to
  abort current operation, e.g. if you need to render another range.
  Just throw exception in this closure.
• 'BigFile#BigFile.determineFileType(src)' returns src file type, could be
  BigFile.Type.BIGBED, BigFile.Type.BIGWIG or null.

Changed

• Show 'Header extensions are unsupported' only in debug log
• Use RomBufferFactory to detect file endianness
• Changed prefetch option to int value in BigFile.read(),
  BigWigFile.read(), BigBedFile.read(). Supported values are:
  • BigFile.PREFETCH_LEVEL_OFF : do not prefetch
  • BigFile.PREFETCH_LEVEL_FAST : prefetch zoom level indexes tree,
    chromosomes list
  • BigFile.PREFETCH_LEVEL_DETAILED : prefetch w/o zoom level data tree
    indexes up to chromosomes level
    Larger prefetch values increase big file opening time and memory
    consumption but provides faster data access with fewer i/o operations

Removed:

• RandomAccessFile based factory and rom buffer
  removed. Please use EndianBufferFactory, EndianThreadSafeBufferFactory
  or EndianSynchronizedBufferFactory factories instead.

0.7.1

Version 0.7.1

Released on April 31th 2017

Changed

• Use maxHeapSize = "1024m" option for tests. Seems by default on
  windows heap size in about 256m and it isn't enough for tests
• RAFBufferFactory, RAFBuffer, RomBufferFactory moved to
  org.jetbrains.bio package
• RAFBufferFactory supports buffer size option for underlying
  random access file
• By default RAFBufferFactory uses default Random Access File
  buffers size (8092 bytes) instead of 125 kb.

0.7.0

Version 0.7.0

Released on April 30th 2017

Fixed

• RandomAccessFile (https://www.unidata.ucar.edu) copied to project and
  rewritten in Kotlin so as not to depend on heavy netcdf-java library.
  For simplicity writing operations were removed from file.

• Windows support #36, using RandomAccessFile
  (https://www.unidata.ucar.edu/software/thredds/v4.3/netcdf-java project)
  instead of memory mapped buffer. The previous impl (works only on
  Mac/Linux) is available using MMBRomBufferFactory in Big*.read().

Improved

• Show less warnings like "R+ tree leaves are overlapping", see issue #35

Changed

• BedFile was made closeable
• Default name field value in ExtendedBedEntry was changed to ".".
  also while parsing empty name value is converted to ".".
• Bed entry pack/unpack methods now supports custom delimiter and
• ExtendedBedEntry.unpack() method supports omitEmptyStrings option,
  which treats several consecutive separators as one, e.g. if TAB
  delimiter was converted by error to sequence of whitespace characters.
• Score range [0..1000] check disable because MACS2 output may contain
  scores > 1000
• ExtendedBedEntry.unpack() parses '.' values as fields default values

0.5.3

Version 0.5.3

Released on March 13th 2018

Improved

• Show less warnings like "R+ tree leaves are overlapping", see issue #35

0.6.0

Version 0.6.0

Released on December 13th 2017

Fixed

• Issue #33: Encode *.bigBed files: NumberFormatException: For input
  string: "5.30862042004096

Changed

• org.jetbrains.bio.big.BedEntry now is minimal bed entry impl and
  represents same info as records in bed file. Obligatory fields are
  chromosome, start, end, all other fields stored as '\t' separated string.
  BedEntry.unpack() allow to parse as BED+ format to ExtendedBedEntry
  which contain all BED12 fields plus optional extended fields.
  'ExtendedBedEntry.pack()' allows to pack back to minimal
  representation.
• Switched to Kotlin 1.2
• Source compatibility updated to Java 8

Removed:

• In class org.jetbrains.bio.big.BedEntry removed all fields except
  'chrom', 'start', 'end', 'rest'