Convert doubles and floats to decimals

[dr-shorthair]

Fixes #307

[dr-shorthair]

Used the following SPARQL

INSERT { ?u qudt:conversionMultiplier ?md }
WHERE {
	?u a qudt:Unit ;
		qudt:conversionMultiplier ?m .
	BIND( DATATYPE(?m) as ?mt )
	FILTER( ?mt != xsd:decimal )
	BIND( xsd:decimal(?m) as ?md )
}

DELETE { ?u qudt:conversionMultiplier ?m }
WHERE {
	?u a qudt:Unit ;
		qudt:conversionMultiplier ?m .
	BIND( DATATYPE(?m) as ?mt )
	FILTER( ?mt != xsd:decimal )
}

INSERT { ?u qudt:conversionOffset ?od }
WHERE {
	?u a qudt:Unit ;
		qudt:conversionOffset ?o .
	BIND( DATATYPE(?o) as ?ot )
	FILTER( ?ot != xsd:decimal )
	BIND( xsd:decimal(?o) as ?od )
}

DELETE { ?u qudt:conversionOffset ?o }
WHERE {
	?u a qudt:Unit ;
		qudt:conversionOffset ?o .
	BIND( DATATYPE(?o) as ?ot )
	FILTER( ?ot != xsd:decimal )
}

Also - remove all zero offets

DELETE { ?u qudt:conversionOffset ?o }
WHERE {
	?u a qudt:Unit ;
		qudt:conversionOffset ?o .
	FILTER( ?o = 0.0 )
}

[nicholascar]

remove all zero offets

Is this because it is inferred to be zero if not present by calculation queries?

[dr-shorthair]

Offsets apply in very few cases - and only on interval scales (not ratio scales).
As they are the exceptional case, they need not be present in general.
(AFAICT there are only 3 valid offsets in the current catalogue.)

[steveraysteveray]

OK.
I hope the multiplier for PlanckDensity doesn't break anything!