org.openimaj.ml.timeseries

Class TimeSeries<DATA,SINGLE_TYPE,RETURNTYPE extends TimeSeries<DATA,SINGLE_TYPE,RETURNTYPE>>

java.lang.Object

org.openimaj.ml.timeseries.TimeSeries<DATA,SINGLE_TYPE,RETURNTYPE>

Type Parameters:

DATA - the type of the data at each point in time

SINGLE_TYPE - the type of the an element at a single point in time

RETURNTYPE - the time series returned by the get

All Implemented Interfaces:

Iterable<IndependentPair<Long,SINGLE_TYPE>>

Direct Known Subclasses:

ConcreteTimeSeries, DoubleTimeSeries, TimeSeriesCollection

public abstract class TimeSeries<DATA,SINGLE_TYPE,RETURNTYPE extends TimeSeries<DATA,SINGLE_TYPE,RETURNTYPE>>
extends Object
implements Iterable<IndependentPair<Long,SINGLE_TYPE>>

A time series defines data at discrete points in time. The time series has the ability to return data at a specific point in time, return neighbours within some window, closest neighbours or n neighbours before and after a time.

These values can be used by a TimeSeriesInterpolation to get specific moments in time

Author:

Sina Samangooei (ss@ecs.soton.ac.uk)

Constructor Summary

Constructors

Constructor and Description
TimeSeries()

Method Summary

Modifier and Type	Method and Description
<OUTDATA,OUTSING,OUTRET extends TimeSeries<OUTDATA,OUTSING,OUTRET>> OUTRET	convert(TimeSeriesConverter<DATA,SINGLE_TYPE,RETURNTYPE,OUTDATA,OUTSING,OUTRET> converter) Convert a TimeSeries
<OUTDATA,OUTSING,OUTRET extends TimeSeries<OUTDATA,OUTSING,OUTRET>> OUTRET	convert(TimeSeriesConverter<DATA,SINGLE_TYPE,RETURNTYPE,OUTDATA,OUTSING,OUTRET> converter, TimeSeriesProcessor<OUTDATA,OUTSING,OUTRET> tsp) Convert a TimeSeries
RETURNTYPE	copy()
RETURNTYPE	get(long time) Same as calling get(long, int, int) with spans as 0
abstract RETURNTYPE	get(long time, int nbefore, int nafter) returns the DATA at a specific point in time and those before and after to the number requested.
abstract RETURNTYPE	get(long time, int nbefore, int nafter, RETURNTYPE output) Same as get(long, int, int) but instead of createing the output DATA instance, an existing data instance is handed which is filled.
abstract RETURNTYPE	get(long start, long end) returns the RETURNTYPE between the specified time periods.
abstract RETURNTYPE	get(long time, long threshbefore, long threshafter) returns the RETURNTYPE at a specific point in time and those before and after within the specified thresholds.
abstract DATA	getData()
abstract long[]	getTimes()
void	internalAssign(long[] times, DATA data)
abstract void	internalAssign(RETURNTYPE interpolate)
abstract RETURNTYPE	newInstance()
RETURNTYPE	process(TimeSeriesProcessor<DATA,SINGLE_TYPE,RETURNTYPE> tsp) process using the provided processor, return
RETURNTYPE	processInplace(TimeSeriesProcessor<DATA,SINGLE_TYPE,RETURNTYPE> tsp) Process using the provided processor
abstract void	set(long[] time, DATA data) Set the data associated with each time.
abstract int	size()
abstract String	toString()

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, wait, wait, wait

Methods inherited from interface java.lang.Iterable

forEach, iterator, spliterator

Constructor Detail

TimeSeries

public TimeSeries()

Method Detail

get

public RETURNTYPE get(long time)

Same as calling get(long, int, int) with spans as 0

Parameters:

time -

Returns:

the requested data or null.

get

public abstract RETURNTYPE get(long time,
                               int nbefore,
                               int nafter)

returns the DATA at a specific point in time and those before and after to the number requested. This method may not return data for the specific requested time if it does not exists. If the time series is completely empty this function may return at an empty array however if at least 1 data point exists and either nbefore and nafter are bigger than 1 then at least 1 datapoint will be returned.

Data should be returned in order

Parameters:

time -

nbefore -

nafter -

Returns:

all data found with these parameters

get

public abstract RETURNTYPE get(long time,
                               int nbefore,
                               int nafter,
                               RETURNTYPE output)

Same as get(long, int, int) but instead of createing the output DATA instance, an existing data instance is handed which is filled. For convenience this output is also returned

Data should be returned in order

Parameters:

time -

nbefore -

nafter -

output -

Returns:

all data found with these parameters

get

public abstract RETURNTYPE get(long time,
                               long threshbefore,
                               long threshafter)

returns the RETURNTYPE at a specific point in time and those before and after within the specified thresholds. This method may not return data for the specific requested time if it does not exists. Similarly this method may return an empty array if no time data is available within the window specified.

Data should be returned in order

Parameters:

time -

threshbefore -

threshafter -

Returns:

all data found with these parameters

get

public abstract RETURNTYPE get(long start,
                               long end)

returns the RETURNTYPE between the specified time periods. This method may not return data for the specific requested time if it does not exists. Similarly this method may return an empty array if no time data is available within the window specified.

Data should be returned in order

Parameters:

start -

end -

Returns:

all data found with these parameters

set

public abstract void set(long[] time,
                         DATA data)
                  throws TimeSeriesSetException

Set the data associated with each time. This function explicitly assumes that time.length == data.length and there exists a single data instance per time instance

Parameters:

time - instances of time

data - instances of data

Throws:

TimeSeriesSetException

getTimes

public abstract long[] getTimes()

Returns:

all times

getData

public abstract DATA getData()

Returns:

all data

newInstance

public abstract RETURNTYPE newInstance()

Returns:

an empty new instance of this timeseries type

size

public abstract int size()

Returns:

the number of valid time steps in this timeseries

internalAssign

public abstract void internalAssign(RETURNTYPE interpolate)

Parameters:

interpolate - assign this timeseries to the internal one, efforts should be made to copy the data, not simply assign it

internalAssign

public void internalAssign(long[] times,
                           DATA data)

Parameters:

times -

data -

copy

public RETURNTYPE copy()

Returns:

clone this time series

process

public RETURNTYPE process(TimeSeriesProcessor<DATA,SINGLE_TYPE,RETURNTYPE> tsp)

process using the provided processor, return

Parameters:

tsp -

Returns:

a new instance processed

processInplace

public RETURNTYPE processInplace(TimeSeriesProcessor<DATA,SINGLE_TYPE,RETURNTYPE> tsp)

Process using the provided processor

Parameters:

tsp -

Returns:

this object processed inplace

convert

public <OUTDATA,OUTSING,OUTRET extends TimeSeries<OUTDATA,OUTSING,OUTRET>> OUTRET convert(TimeSeriesConverter<DATA,SINGLE_TYPE,RETURNTYPE,OUTDATA,OUTSING,OUTRET> converter)

Convert a TimeSeries

Type Parameters:

OUTDATA -

OUTSING -

OUTRET -

Parameters:

converter - the converter

Returns:

the converted timeseries

convert

public <OUTDATA,OUTSING,OUTRET extends TimeSeries<OUTDATA,OUTSING,OUTRET>> OUTRET convert(TimeSeriesConverter<DATA,SINGLE_TYPE,RETURNTYPE,OUTDATA,OUTSING,OUTRET> converter,
                                                                                          TimeSeriesProcessor<OUTDATA,OUTSING,OUTRET> tsp)

Convert a TimeSeries

Type Parameters:

OUTDATA -

OUTSING -

OUTRET -

Parameters:

converter - the converter

tsp - the processor

Returns:

the converted timeseries

toString

public abstract String toString()

Overrides:

toString in class Object