Series vs items in the Cruncher input parameter file

[a-s-russo]

Regarding the output dictionary for the input parameter file for Cruncher (https://github.com/jdemetra/jwsacruncher/wiki/Output-dictionaries), could a column please be added to clarify which keys refer to series as opposed to items?

That is, which tags correspond to the <matrix> items block, and which tags correspond to the <tsmatrix> series block, as per the example given at: https://github.com/jdemetra/jwsacruncher/wiki#parameters-file ? Are series tags simply those with a Java class of ec.tstoolkit.timeseries.simplets.TsData, and item tags are everything else?

[Immurb]

The short answer is yes.

You can see how it is decided at

jdplus-main/jdplus-main-cli/jdplus-main-cli-bin/src/main/java/jdplus/cruncher/WsaConfig.java

Lines 166 to 183 in 69be4a2

	Map<String, Class> dic = new LinkedHashMap<>();
	// for TramoSeats
	Dictionary tsdic = TramoSeatsDictionaries.TRAMOSEATSDICTIONARY;
	tsdic.entries().forEachOrdered(entry->dic.put(entry.fullName(), entry.getOutputClass()));
	// // for X13
	Dictionary x13dic = X13Dictionaries.X13DICTIONARY;
	x13dic.entries().forEachOrdered(entry->dic.put(entry.fullName(), entry.getOutputClass()));
	// series
	Set<Type> types = CsvInformationFormatter.formattedTypes();
	dic.entrySet().forEach(entry -> {
	if (entry.getValue() == TsData.class){
	tlist.add(entry.getKey());
	}else if (types.contains(entry.getValue())){
	mlist.add(entry.getKey());
	}
	});
	result.TSMatrix=tlist.toArray(result.TSMatrix);
	result.Matrix=mlist.toArray(result.Matrix);

Internally series have another output processor than the other information.
We are thinking about creating a new parameter file for the cruncher (without breaking the old one). Then the users don't have to make the separation because the program will be able to make it.