426 lines
15 KiB
C++
426 lines
15 KiB
C++
|
/*
|
||
|
|
----
|
||
|
|
This file is part of SECONDO.
|
||
|
|
|
||
|
|
Copyright (C) 2020,
|
||
|
|
Faculty of Mathematics and Computer Science,
|
||
|
|
Database Systems for New Applications.
|
||
|
|
|
||
|
|
SECONDO is free software; you can redistribute it and/or modify
|
||
|
|
it under the terms of the GNU General Public License as published by
|
||
|
|
the Free Software Foundation; either version 2 of the License, or
|
||
|
|
(at your option) any later version.
|
||
|
|
|
||
|
|
SECONDO is distributed in the hope that it will be useful,
|
||
|
|
but WITHOUT ANY WARRANTY; without even the implied warranty of
|
||
|
|
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
||
|
|
GNU General Public License for more details.
|
||
|
|
|
||
|
|
You should have received a copy of the GNU General Public License
|
||
|
|
along with SECONDO; if not, write to the Free Software
|
||
|
|
Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
|
||
|
|
----
|
||
|
|
|
||
|
|
|
||
|
|
//[$][\$]
|
||
|
|
//[_][\_]
|
||
|
|
|
||
|
|
*/
|
||
|
|
#include "timeseriesalgebra.h"
|
||
|
|
|
||
|
|
namespace TSer {
|
||
|
|
|
||
|
|
ListExpr TimeSeriesAlgebra::genericTSTypeMapping(ListExpr timeSeries){
|
||
|
|
const string error_message = "OrderedRelation with an instant as"
|
||
|
|
" order criteria and an arbitrary number"
|
||
|
|
" of reals exspected.";
|
||
|
|
|
||
|
|
//Check type of arguments
|
||
|
|
if(!listutils::isOrelDescription(nl->First(timeSeries))){
|
||
|
|
return listutils::typeError(error_message);
|
||
|
|
}
|
||
|
|
|
||
|
|
ListExpr firstTSExpression = nl->First(timeSeries);
|
||
|
|
|
||
|
|
ListExpr tsTupleDescription = nl->Second(firstTSExpression);
|
||
|
|
|
||
|
|
|
||
|
|
if(!listutils::isTupleDescription(tsTupleDescription)){
|
||
|
|
return listutils::typeError(error_message +
|
||
|
|
" ( TupleDescription exspected.)");
|
||
|
|
}
|
||
|
|
|
||
|
|
if(!datetime::DateTime::checkType(nl
|
||
|
|
->Second(nl
|
||
|
|
->First(tsTupleDescription)))){
|
||
|
|
return listutils::typeError((error_message +
|
||
|
|
" ( First attribute should "
|
||
|
|
"be of type instant.)"));
|
||
|
|
}
|
||
|
|
if(!CcString::checkType(nl->Second(nl->Second(tsTupleDescription)))){
|
||
|
|
return listutils::typeError((error_message + "( Second attribute"
|
||
|
|
" should be of type string.)"));
|
||
|
|
}
|
||
|
|
|
||
|
|
int noOfAttributes = nl->ListLength(tsTupleDescription);
|
||
|
|
|
||
|
|
for(int indexOfAttribute = 2;
|
||
|
|
indexOfAttribute < noOfAttributes; indexOfAttribute++){
|
||
|
|
if(!CcReal::checkType(nl
|
||
|
|
->Second(nl
|
||
|
|
->Nth(indexOfAttribute, tsTupleDescription)))){
|
||
|
|
return listutils::typeError(error_message +
|
||
|
|
"( only attributes of type real exspected after"
|
||
|
|
" second attribute.)");
|
||
|
|
}
|
||
|
|
}
|
||
|
|
|
||
|
|
return timeSeries;
|
||
|
|
};
|
||
|
|
|
||
|
|
|
||
|
|
OperatorSpec timeSeriesDistanceFunSpec(
|
||
|
|
"orel x orel x(fun : tuple x tuple) -> real ",
|
||
|
|
"_ _ tsdistancefun[ _ ] ",
|
||
|
|
"computes the distance between the two "
|
||
|
|
"given timeseries utilizing the defined function",
|
||
|
|
"query [ const (orel(tuple((Time instant)(Value real)))"
|
||
|
|
" (Time)) value((1 1.0)) ] feed {a} [ const"
|
||
|
|
" (orel(tuple((Time instant)(Value real))) (Time)) value((1 2.0)) ]"
|
||
|
|
"feed {b} tsdistancefun[ "
|
||
|
|
"fun (tuple1: TUPLE2, tuple2: TUPLE2) "
|
||
|
|
"sqrt(pow(attr(tuple1, Value_a) - attr(tuple2, Value_b),2))];"
|
||
|
|
);
|
||
|
|
|
||
|
|
Operator tsDistanceFunOp(
|
||
|
|
"tsdistancefun",
|
||
|
|
timeSeriesDistanceFunSpec.getStr(),
|
||
|
|
TimeSeriesDistance::timeSeriesDistanceFunOpValueMapping,
|
||
|
|
//SimpleSelect ist für nicht überladene Operatoren,
|
||
|
|
//bei weiteren Distanzmetriken -> anpassen!
|
||
|
|
Operator::SimpleSelect,
|
||
|
|
TimeSeriesDistance::distanceOpFunTypeMap
|
||
|
|
);
|
||
|
|
|
||
|
|
OperatorSpec timeSeriesDistanceSpec(
|
||
|
|
"orel x orel -> real",
|
||
|
|
"tsdistance(_, _)",
|
||
|
|
"Computes the distance between two"
|
||
|
|
" timeseries utilizing the euklidian metric.",
|
||
|
|
"query dist([ const (orel(tuple((Time instant)(Value real)))"
|
||
|
|
" (Time)) value((1 1.0)) ],[ const (orel(tuple((Time instant)"
|
||
|
|
"(Value real))) (Time)) value((1 2.0)) ])"
|
||
|
|
);
|
||
|
|
|
||
|
|
|
||
|
|
Operator tsDistanceOp(
|
||
|
|
"tsdistance",
|
||
|
|
timeSeriesDistanceSpec.getStr(),
|
||
|
|
TimeSeriesDistance::timeSeriesDistanceValueMapping,
|
||
|
|
//SimpleSelect ist für nicht überladene Operatoren,
|
||
|
|
//bei weiteren Distanzmetriken -> anpassen!
|
||
|
|
Operator::SimpleSelect,
|
||
|
|
TimeSeriesDistance::distanceOpTypeMap
|
||
|
|
);
|
||
|
|
|
||
|
|
OperatorSpec tsMotifBfSpec(
|
||
|
|
"orel -> orel",
|
||
|
|
" _ tsmotifbf(_,_) or _ tsmotifbf(_,_,[FUN])",
|
||
|
|
"Computes the motif of the timeseries with the most not"
|
||
|
|
" trivial matches of length n within a range r, utilizing"
|
||
|
|
" the euklidian distance. The alternative utilizes an"
|
||
|
|
" individual distance measurement.",
|
||
|
|
"query tsmotifbf([ const (orel(tuple((Time instant"
|
||
|
|
")(Value real))) (Time)) value((1 1.0)(2 1.0)) ],1, 0 ) ");
|
||
|
|
|
||
|
|
OperatorSpec tsDTWSpec(
|
||
|
|
"orel x orel -> orel",
|
||
|
|
"_ _ tsdtw",
|
||
|
|
"Computes the warping path to match the first timeseries"
|
||
|
|
" as good as possible to the second timeseries.",
|
||
|
|
"query [ const (orel(tuple((Time instant)(Value real)))"
|
||
|
|
" (Time)) value((1 1.0)(2 1.0)) ] "
|
||
|
|
"[ const (orel(tuple((Time instant)(Value real))) (Time))"
|
||
|
|
" value((1 1.0)(2 1.0)) ] tsdtw count; ");
|
||
|
|
|
||
|
|
OperatorSpec tsDTWSakoeChibaSpec(
|
||
|
|
"orel x orel x int-> orel",
|
||
|
|
"dtwsc(_,_,_)",
|
||
|
|
"Computes the warping path to match the first timeseries as good "
|
||
|
|
"as possible to the second timeseries."
|
||
|
|
"The integer paramter defines the Sakoe-Chiba band in wich the"
|
||
|
|
" warping may deviate from the diagonal.",
|
||
|
|
"query [ const (orel(tuple((Time instant)(Value real))) (Time))"
|
||
|
|
" value((1 1.0)(2 1.0)) ] [ const (orel(tuple((Time instant)"
|
||
|
|
"(Value real))) (Time)) value((1 1.0)(2 1.0)) ] tsdtw count;");
|
||
|
|
|
||
|
|
OperatorSpec tsWhiteNoiseSpec(
|
||
|
|
"orel x real x real -> orel",
|
||
|
|
"_ tswhitenoise(_, _)",
|
||
|
|
"Adds white noise to all values of the timeseries."
|
||
|
|
" The parameters are the variance and mean of"
|
||
|
|
" a gaussian distribution.",
|
||
|
|
"query [ const (orel(tuple((Time instant)(Value real)))"
|
||
|
|
" (Time)) value((1 1.0)(2 1.0)) ] tswhitenoise(0.0, 1.0)"
|
||
|
|
);
|
||
|
|
|
||
|
|
OperatorSpec tsDifSpec(
|
||
|
|
"orel x int -> orel",
|
||
|
|
"_ tsdif(_)",
|
||
|
|
"Computes the differenzing of the timeseries n times"
|
||
|
|
" where n is the operation parameter",
|
||
|
|
"query [ const (orel(tuple((Time instant)(Value real))) "
|
||
|
|
"(Time)) value((1 1.0)(2 1.0)) ] tsdif(1)"
|
||
|
|
);
|
||
|
|
|
||
|
|
OperatorSpec predictMaSpec(
|
||
|
|
"orel x int x int -> orel",
|
||
|
|
"_ predictma[_,_]",
|
||
|
|
"Utilizes an Moving Average Process MA(p)"
|
||
|
|
" to predict the next n points in the time series."
|
||
|
|
" Order of the MA is the first parameter,"
|
||
|
|
" number of predictions the second. ",
|
||
|
|
"query [ const (orel(tuple((Time instant)(Value real))) (Time))"
|
||
|
|
" value((1 1.0)(2 1.0)) ] predictma[1,1]"
|
||
|
|
);
|
||
|
|
|
||
|
|
OperatorSpec predictArSpec(
|
||
|
|
"orel x int x int -> orel",
|
||
|
|
"_ predictar[_,_]",
|
||
|
|
"Utilizes an Autoregressive Process AR(q) to predict the next"
|
||
|
|
" n points in the time series. Order of the AR is the"
|
||
|
|
" first parameter, number of predictions the second. ",
|
||
|
|
"query [ const (orel(tuple((Time instant)(Value real))) (Time))"
|
||
|
|
" value((1 1.0)(2 1.0)) ] predictar[1,1]"
|
||
|
|
);
|
||
|
|
|
||
|
|
OperatorSpec predictARIMASpec(
|
||
|
|
"orel x int x int x int x int -> orel",
|
||
|
|
"_ predictarima[_,_,_,_]",
|
||
|
|
"Utilizes an Integrated Autoregressive Moving Average Prozess"
|
||
|
|
" ARIMA(p,q,i) to predict the next n points in the time series."
|
||
|
|
" Order of the AR is the first parameter,"
|
||
|
|
" the second the degree of differenzing, the third order of the"
|
||
|
|
" MA Prozess and the fourth the number of predictions. ",
|
||
|
|
"query [ const (orel(tuple((Time instant)(Value real))) (Time))"
|
||
|
|
" value((1 1.0)(2 1.0)) ] predictarima[1,0,1,1]"
|
||
|
|
);
|
||
|
|
|
||
|
|
OperatorSpec predictARMASpec(
|
||
|
|
"orel x int x int x int -> orel",
|
||
|
|
"_ predictarima[_,_,_]",
|
||
|
|
"Utilizes an Autoregressive Moving Average Prozess ARMA(p,q,i)"
|
||
|
|
" to predict the next n points in the time series. Order of the"
|
||
|
|
" AR is the first parameter,"
|
||
|
|
" the second the order of the MA Prozess and the third the number"
|
||
|
|
" of predictions. ",
|
||
|
|
"query [ const (orel(tuple((Time instant)(Value real))) (Time))"
|
||
|
|
" value((1 1.0)(2 1.0)) ] predictarima[1,0,1,1]"
|
||
|
|
);
|
||
|
|
|
||
|
|
OperatorSpec tsDDTWSpec(
|
||
|
|
"orel x orel -> orel",
|
||
|
|
"ddtw(_,_)",
|
||
|
|
"Utilizes an approximation of point distances to compute an"
|
||
|
|
" optimal warping path to match the first timeseries argument"
|
||
|
|
" to the second",
|
||
|
|
"query ddtw([ const (orel(tuple((Time instant)(Value real)))"
|
||
|
|
" (Time))"
|
||
|
|
" value((1 1.0)(2 1.0)) ],"
|
||
|
|
" ddtw([ const (orel(tuple((Time instant)(Value real))) (Time))"
|
||
|
|
" value((1 1.0)(2 1.0)(3 1.5)) ])"
|
||
|
|
);
|
||
|
|
|
||
|
|
OperatorSpec tsPaaSpec(
|
||
|
|
"orel x int -> orel",
|
||
|
|
"_ paa[_]",
|
||
|
|
"Uses the piecewise approximate aggregation described"
|
||
|
|
" by Keogh to produce a dimensionally reduced timeseries from"
|
||
|
|
" the original consisting of int frames.",
|
||
|
|
"query paa([ const (orel(tuple((Time instant)(Value real))) (Time))"
|
||
|
|
" value((1 1.0)(2 1.0)) ],1)"
|
||
|
|
);
|
||
|
|
|
||
|
|
OperatorSpec tsMotifMKSpec(
|
||
|
|
"orel x int -> orel",
|
||
|
|
"_ motifmk[_]",
|
||
|
|
"",
|
||
|
|
"query motifmk ([ const (orel(tuple((Time instant)(Value real)))"
|
||
|
|
"(Time)) value((1 1.0)(2 1.0)) ],1)"
|
||
|
|
);
|
||
|
|
|
||
|
|
OperatorSpec tsPacfSpec(
|
||
|
|
"orel x int -> int",
|
||
|
|
"_ pacf[_]",
|
||
|
|
"",
|
||
|
|
"query ([ const (orel(tuple((Time instant)(Value real))) (Time))"
|
||
|
|
" value((1 1.0)(2 1.0)) ] pacf[1]"
|
||
|
|
);
|
||
|
|
OperatorSpec tsAcfSpec(
|
||
|
|
"orel x int -> int",
|
||
|
|
"_ acf[_]",
|
||
|
|
"",
|
||
|
|
"query ([ const (orel(tuple((Time instant)(Value real))) (Time))"
|
||
|
|
" value((1 1.0)(2 1.0)) ] acf[1]"
|
||
|
|
);
|
||
|
|
|
||
|
|
Operator tsMotifOpFun(
|
||
|
|
"tsmotifbffun",
|
||
|
|
tsMotifBfSpec.getStr(),
|
||
|
|
TSMotifOp::tsMotifBfValueMap,
|
||
|
|
Operator::SimpleSelect,
|
||
|
|
TSMotifOp::tsMotifBfTypeMap
|
||
|
|
);
|
||
|
|
|
||
|
|
Operator tsMotifBfOp(
|
||
|
|
"tsmotifbf",
|
||
|
|
tsMotifBfSpec.getStr(),
|
||
|
|
TSMotifOp::tsMotifBfValueMap,
|
||
|
|
Operator::SimpleSelect,
|
||
|
|
TSMotifOp::tsMotifBfTypeMap
|
||
|
|
);
|
||
|
|
|
||
|
|
Operator tsDTWOp(
|
||
|
|
"tsdtw",
|
||
|
|
tsDTWSpec.getStr(),
|
||
|
|
TsDTWOp::tsDTWValueMap,
|
||
|
|
Operator::SimpleSelect,
|
||
|
|
TsDTWOp::tsDTWTypeMap
|
||
|
|
);
|
||
|
|
|
||
|
|
Operator tsDifOp(
|
||
|
|
"tsdif",
|
||
|
|
tsDifSpec.getStr(),
|
||
|
|
TsArima::tsDifValueMap,
|
||
|
|
Operator::SimpleSelect,
|
||
|
|
TsArima::tsDifTypeMap
|
||
|
|
);
|
||
|
|
|
||
|
|
Operator tsWhiteNoiseOp(
|
||
|
|
"tswhitenoise",
|
||
|
|
tsWhiteNoiseSpec.getStr(),
|
||
|
|
TsArima::tsWhiteNoiseValueMap,
|
||
|
|
Operator::SimpleSelect,
|
||
|
|
TsArima::tsWhiteNoiseTypeMap
|
||
|
|
);
|
||
|
|
|
||
|
|
Operator predictMA(
|
||
|
|
"predictma",
|
||
|
|
predictMaSpec.getStr(),
|
||
|
|
TsArima::tsMAValueMap,
|
||
|
|
Operator::SimpleSelect,
|
||
|
|
TsArima::tsMATypeMap
|
||
|
|
);
|
||
|
|
|
||
|
|
Operator predictAR(
|
||
|
|
"predictar",
|
||
|
|
predictMaSpec.getStr(),
|
||
|
|
TsArima::tsARValueMap,
|
||
|
|
Operator::SimpleSelect,
|
||
|
|
TsArima::tsARTypeMap
|
||
|
|
);
|
||
|
|
|
||
|
|
Operator predictARMA(
|
||
|
|
"predictarma",
|
||
|
|
predictMaSpec.getStr(),
|
||
|
|
TsArima::tsARMAValueMap,
|
||
|
|
Operator::SimpleSelect,
|
||
|
|
TsArima::tsARMATypeMap
|
||
|
|
);
|
||
|
|
|
||
|
|
Operator predictARIMA(
|
||
|
|
"predictarima",
|
||
|
|
predictMaSpec.getStr(),
|
||
|
|
TsArima::tsARIMAValueMap,
|
||
|
|
Operator::SimpleSelect,
|
||
|
|
TsArima::tsARIMATypeMap
|
||
|
|
);
|
||
|
|
|
||
|
|
Operator tsDDTWOp(
|
||
|
|
"ddtw",
|
||
|
|
tsDDTWSpec.getStr(),
|
||
|
|
TsDTWOp::tsDDTWValueMap,
|
||
|
|
Operator::SimpleSelect,
|
||
|
|
TsDTWOp::tsDTWTypeMap
|
||
|
|
);
|
||
|
|
|
||
|
|
Operator tsDTWSakoeChibaOp(
|
||
|
|
"dtwsc",
|
||
|
|
tsDTWSakoeChibaSpec.getStr(),
|
||
|
|
TsDTWOp::tsDTWSCValueMap,
|
||
|
|
Operator::SimpleSelect,
|
||
|
|
TsDTWOp::tsDTWSCTypeMap
|
||
|
|
);
|
||
|
|
|
||
|
|
|
||
|
|
Operator tsPaaOp(
|
||
|
|
"paa",
|
||
|
|
tsPaaSpec.getStr(),
|
||
|
|
TSMotifOp::tsPaaValueMap,
|
||
|
|
Operator::SimpleSelect,
|
||
|
|
TSMotifOp::tsPaaTypeMap
|
||
|
|
);
|
||
|
|
|
||
|
|
Operator motifMK(
|
||
|
|
"motifMK",
|
||
|
|
tsMotifMKSpec.getStr(),
|
||
|
|
TSMotifOp::tsMotifMueenKeoghValueMap,
|
||
|
|
Operator::SimpleSelect,
|
||
|
|
TSMotifOp::tsMotifMueenKeoghTypeMap
|
||
|
|
);
|
||
|
|
|
||
|
|
Operator pacf(
|
||
|
|
"pacf",
|
||
|
|
tsPacfSpec.getStr(),
|
||
|
|
TsArima::pacfValueMap,
|
||
|
|
Operator::SimpleSelect,
|
||
|
|
TsArima::pacfTypeMap
|
||
|
|
);
|
||
|
|
|
||
|
|
Operator acf(
|
||
|
|
"acf",
|
||
|
|
tsPacfSpec.getStr(),
|
||
|
|
TsArima::acfValueMap,
|
||
|
|
Operator::SimpleSelect,
|
||
|
|
TsArima::acfTypeMap
|
||
|
|
);
|
||
|
|
|
||
|
|
TimeSeriesAlgebra::TimeSeriesAlgebra() : Algebra(){
|
||
|
|
AddOperator(&tsDistanceOp);
|
||
|
|
AddOperator(&tsDistanceFunOp);
|
||
|
|
AddOperator(&tsMotifBfOp);
|
||
|
|
AddOperator(&tsDTWOp);
|
||
|
|
AddOperator(&tsDifOp);
|
||
|
|
AddOperator(&tsWhiteNoiseOp);
|
||
|
|
AddOperator(&predictAR);
|
||
|
|
AddOperator(&predictMA);
|
||
|
|
AddOperator(&predictARMA);
|
||
|
|
AddOperator(&predictARIMA);
|
||
|
|
AddOperator(&tsPaaOp);
|
||
|
|
AddOperator(&tsDTWSakoeChibaOp);
|
||
|
|
AddOperator(&tsDDTWOp);
|
||
|
|
AddOperator(&pacf);
|
||
|
|
AddOperator(&acf);
|
||
|
|
//Type Mapping noch nicht implementiert, Operator nicht getestet.
|
||
|
|
// AddOperator(&motifMK);
|
||
|
|
}
|
||
|
|
|
||
|
|
|
||
|
|
|
||
|
|
/*
|
||
|
|
5 Initialization
|
||
|
|
|
||
|
|
*/
|
||
|
|
|
||
|
|
|
||
|
|
|
||
|
|
extern "C"
|
||
|
|
Algebra*
|
||
|
|
InitializeTimeSeriesAlgebra( NestedList* nlRef,
|
||
|
|
QueryProcessor* qpRef )
|
||
|
|
{
|
||
|
|
// The C++ scope-operator :: must be used to qualify the full name
|
||
|
|
return new TSer::TimeSeriesAlgebra;
|
||
|
|
}
|
||
|
|
}
|