Class PagedChannelRandomAccessSource

java.lang.Object
com.itextpdf.text.io.GroupedRandomAccessSource
com.itextpdf.text.io.PagedChannelRandomAccessSource
All Implemented Interfaces:
RandomAccessSource
class PagedChannelRandomAccessSource extends GroupedRandomAccessSource implements RandomAccessSource
A RandomAccessSource that is based on an underlying FileChannel. The channel is mapped into memory using a paging scheme to allow for efficient reads of very large files. As an implementation detail, we use GroupedRandomAccessSource functionality, but override to make determination of the underlying mapped page more efficient - and to close each page as another is opened
Since:
5.3.5

  • Field Details

    • DEFAULT_TOTAL_BUFSIZE

      public static final int DEFAULT_TOTAL_BUFSIZE
      See Also:

    • DEFAULT_MAX_OPEN_BUFFERS

      public static final int DEFAULT_MAX_OPEN_BUFFERS
      See Also:

    • bufferSize

      private final int bufferSize
      The size of each of the buffers to use when mapping files into memory. This must be greater than 0 and less than Integer.MAX_VALUE

    • channel

      private final FileChannel channel
      The channel this source is based on

    • mru

      private final PagedChannelRandomAccessSource.MRU<RandomAccessSource> mru
      Most recently used list used to hold a number of mapped pages open at a time

  • Constructor Details

    • PagedChannelRandomAccessSource

      public PagedChannelRandomAccessSource(FileChannel channel) throws IOException
      Constructs a new PagedChannelRandomAccessSource based on the specified FileChannel, with a default buffer configuration. The default buffer configuration is currently 2^26 total paged bytes, spread across a maximum of 16 active buffers. This arrangement resulted in a 24% speed improvement over the single buffer case in parametric tests extracting text from a 2.3 GB file.
      Parameters:
      channel - the channel to use as the backing store
      Throws:
      IOException - if the channel cannot be opened or mapped

    • PagedChannelRandomAccessSource

      public PagedChannelRandomAccessSource(FileChannel channel, int totalBufferSize, int maxOpenBuffers) throws IOException
      Constructs a new PagedChannelRandomAccessSource based on the specified FileChannel, with a specific buffer size
      Parameters:
      channel - the channel to use as the backing store
      bufferSize - the size of the buffers to use
      Throws:
      IOException - if the channel cannot be opened or mapped

  • Method Details

    • buildSources

      private static RandomAccessSource[] buildSources(FileChannel channel, int bufferSize) throws IOException
      Constructs a set of MappedChannelRandomAccessSources for each page (of size bufferSize) of the underlying channel
      Parameters:
      channel - the underlying channel
      bufferSize - the size of each page (the last page may be shorter)
      Returns:
      a list of sources that represent the pages of the channel
      Throws:
      IOException - if IO fails for any reason

    • getStartingSourceIndex

      protected int getStartingSourceIndex(long offset)
      Description copied from class: GroupedRandomAccessSource
      For a given offset, return the index of the source that contains the specified offset. This is an optimization feature to help optimize the access of the correct source without having to iterate through every single source each time. It is safe to always return 0, in which case the full set of sources will be searched. Subclasses should override this method if they are able to compute the source index more efficiently (for example FileChannelRandomAccessSource takes advantage of fixed size page buffers to compute the index)
      Overrides:
      getStartingSourceIndex in class GroupedRandomAccessSource
      Parameters:
      offset - the offset
      Returns:
      the index of the input source that contains the specified offset, or 0 if unknown

    • sourceReleased

      protected void sourceReleased(RandomAccessSource source) throws IOException
      Description copied from class: GroupedRandomAccessSource
      Called when a given source is no longer the active source. This gives subclasses the abilty to release resources, if appropriate.
      Overrides:
      sourceReleased in class GroupedRandomAccessSource
      Parameters:
      source - the source that is no longer the active source
      Throws:
      IOException - if there are any problems

    • sourceInUse

      protected void sourceInUse(RandomAccessSource source) throws IOException
      Description copied from class: GroupedRandomAccessSource
      Called when a given source is about to become the active source. This gives subclasses the abilty to retrieve resources, if appropriate.
      Overrides:
      sourceInUse in class GroupedRandomAccessSource
      Parameters:
      source - the source that is about to become the active source
      Throws:
      IOException - if there are any problems

    • close

      public void close() throws IOException
      Description copied from class: GroupedRandomAccessSource
      Closes this source. The underlying data structure or source (if any) will also be closed Closes all of the underlying sources
      Specified by:
      close in interface RandomAccessSource
      Overrides:
      close in class GroupedRandomAccessSource
      Throws:
      IOException