org.gridfour.compress

Interface IPredictorModel

All Known Implementing Classes:

PredictorModelDifferencing, PredictorModelDifferencingWithNulls, PredictorModelLinear, PredictorModelTriangle

public interface IPredictorModel

Defines an interface for predictor-model implementations.

Method Summary

Modifier and Type	Method and Description
void	decode(int seed, int nRows, int nColumns, byte[] encoding, int offset, int length, int[] output) Decodes the specified data using the predictor model to compute output values from the specified input.
int	encode(int nRows, int nColumns, int[] values, byte[] encoding) Encodes the specified data using the predictor model to develop computed adjustments for the data grid.
PredictorModelType	getPredictorType() Gets the predictor type.
int	getSeed() Gets the seed value determined by the most recent call to the encoding method.
boolean	isNullDataSupported() Indicates that the predictor model is intended to handle null values.

Method Detail

decode

void decode(int seed,
            int nRows,
            int nColumns,
            byte[] encoding,
            int offset,
            int length,
            int[] output)

Decodes the specified data using the predictor model to compute output values from the specified input.

Parameters:

seed - a seed value to provide the initial value for the encoding

encoding - the input values for the encoding, usually dimensions to nRows*nColumns

nRows - the number of rows in the raster data grid

nColumns - the number of columns in the raster data grid

offset - the starting offset in the encoding array from which data is to be taken.

length - the length of the valid section of the encoding array from which data is to be taken.

output - array to accept the output data values, dimensions to at least nRows*nColumns

encode

int encode(int nRows,
           int nColumns,
           int[] values,
           byte[] encoding)

Encodes the specified data using the predictor model to develop computed adjustments for the data grid.

This routine returns the seed value detected in the data. This value is usually the first entry in the input value. In cases where the data contains nulls, the seed is sometimes the mean value of the non-null entries.

The length of the encoded output will be a maximum of 5 times the total number of entries in the input (it is usually much shorter than that). The encoding array should be dimensioned large enough to accept the maximum sized output.

Parameters:

nRows - the number of rows in the raster data grid

nColumns - the number of columns in the data data grid

values - the input values to be encoded

encoding - the output encoding

Returns:

if successful, the number of bytes in the output encoding; otherwise, a value -1 to indicate an encoding failure.

isNullDataSupported

boolean isNullDataSupported()

Indicates that the predictor model is intended to handle null values. In many real-world raster data products, the transition from regions of valid data to null data is frequently discontinuous. While most predictor models can handle the transition, the results often compress poorly. This method indicates if a predictor is intended to support cells with null data.

Returns:

true if null values are permitted; otherwise false.

getSeed

int getSeed()

Gets the seed value determined by the most recent call to the encoding method. This value is defined only if the encoding was successful.

Returns:

an arbitrary integer value

getPredictorType

PredictorModelType getPredictorType()

Gets the predictor type.

Returns:

a valid enumeration corresponding the the type of predictor