Package org.biojava.bio.seq.db.emblcd

Class EmblCDROMIndexReader

  • Direct Known Subclasses:
    AcnumHitReader, AcnumTrgReader, DivisionLkpReader, EntryNamIdxReader
    public abstract class EmblCDROMIndexReader
extends Object

    EmblCDROMIndexReader is an abstract class whose concrete subclasses read EMBL CD-ROM format indices from an underlying InputStream. This format is used by the EMBOSS package for database indexing (see programs dbiblast, dbifasta, dbiflat and dbigcg). Indexing produces four binary files with a simple format:

    • division.lkp : master index
    • entrynam.idx : sequence ID index
    • acnum.trg : accession number index
    • acnum.hit : accession number auxiliary index

    Internally EMBOSS checks for Big-endian architechtures and switches the byte order to Little-endian. This means trouble if you try to read the file using DataInputStream, but at least the binaries are consistent across architechtures. This class carries out the necessary conversion.

    The EMBL CD-ROM format stores the date in 4 bytes. One byte is unused (the first one), leaving one byte for the day, one for the month and one (!) for the year.

    For further information see the EMBOSS documentation, or for a full description, the source code of the dbi programs and the Ajax library.

    Since:
    1.2
    Author:
    Keith James

    • Constructor Detail

      • EmblCDROMIndexReader

        public EmblCDROMIndexReader​(InputStream input)
                     throws IOException
        Creates a new EmblCDROMIndexReader instance. A BufferedInputStream is probably the most suitable.
        Parameters:
        input - an InputStream.
        Throws:
        IOException - if an error occurs.

    • Method Detail

      • readFileLength

        public long readFileLength()
        readFileLength returns the file length in bytes (stored within the file's header by the indexing program). This may be called more than once as the value is cached.
        Returns:
        a long.

      • readRecordCount

        public long readRecordCount()
        readRecordCount returns the number of records in the file. This may be called more than once as the value is cached.
        Returns:
        a long.

      • readRecordLength

        public int readRecordLength()
        readRecordLength returns the record length (bytes). This may be called more than once as the value is cached.
        Returns:
        an int.

      • readDBName

        public String readDBName()
        readDBName returns the database name from the index header. This may be called more than once as the value is cached.
        Returns:
        a String.

      • readDBRelease

        public String readDBRelease()
        readDBRelease returns the database release from the index header. This may be called more than once as the value is cached.
        Returns:
        a String.

      • readDBDate

        public String readDBDate()
        readDBDate reads the date from the index header. The date is stored in 4 bytes: 0, unused; 1, year; 2, month; 3, day. With a 1 byte year it's not very much use and I'm not sure that the EMBOSS programs set the value correctly anyway.
        Returns:
        a String.

      • readRecord

        public abstract Object[] readRecord()
                             throws IOException
        readRecord returns an array of objects parsed from a single record. Its content will depend on the type of index file. Concrete subclasses must provide an implementation of this method.
        Returns:
        an Object [] array.
        Throws:
        IOException - if an error occurs.

      • readRawRecord

        public byte[] readRawRecord()
                     throws IOException
        readRawRecord returns the raw bytes of a single record from the index.
        Returns:
        a byte [] array.
        Throws:
        IOException - if an error occurs.