Representing nested terms

Currently dates are tagged in separate parts ie. 12th November 1981 is tagged this way:

[
	[
		{
			"text": "12th",
			"normal": "12th",
			"tags": {
				"Ordinal": true,
				"Value": true,
				"Date": true,
				"NumericValue": true
			}
		},
		{
			"text": "November",
			"normal": "november",
			"tags": {
				"TitleCase": true,
				"Month": true,
				"Date": true,
				"Singular": true,
				"Noun": true
			}
		},
		{
			"text": "1981",
			"normal": "1981",
			"tags": {
				"NumericValue": true,
				"Value": true,
				"Cardinal": true,
				"Date": true,
				"Year": true
			}
		}
	]
]

So each part of the date is tagged as “Date”. I think it would be better if the tag was changed to “DatePart” and an additional lumped/combined term would be returned for the whole date and tagged as “Date”.

Another related idea would be to have the concept of “LumpTerm”, a “Term” which contains an ordered collection of Terms and no text of its own… This would be useful in instances other than Dates - cantractions, money (ie. $35) and the other lump-terms for example.

What do you think?

[spencermountain]

Right. Maybe it should support that (#Person|#Date)+ should work. But if you dont want combinations of people and dates, you’ll have to make two separate queries and do some kinda diff

[leoseccia]

Sorry this should probably be on a separate issue but it is related 😃 ‘(#Person|#Date)’ works as I would expect while ‘(#Person+|#Date+)’ does not return anything at all.