Hi all,
I have a use case where I want to have FTS (Full-Text-Search) over my corpus on ES which includes few documents over a single index.
This part is easy, but I have another index (BTW, there may be that my design of indexes is wrong, I would love to hear better design) where I set another metadata per each term in the document (e.g. term is a verb/noun/name/etc. and more..). This I do by indexing the term's index/position within the document (a.k.a "reverse index").
Now, assuming ES query's results returns some match documents, I would like also to return the metadata of processed terms.
For example:
Assuming querying the phrase "trip to London" and ES found 2 match documents (assuming ids are Doc_1 & Doc_2).
I would like also to return to the user these document ids and the metadata of processed query (e.g. "trip" & "london") within these two docs. For example:
Response would be by relevancy:
Doc_1, metadatas: [{16, verb}, {22, noun}]
Doc_2, metadatas: [{55, name}, {88, noun}]
note: numbers above 16,22,55,88 are the terms' index within the documents.
How can I make such kind of search efficiency?
I though using the "highlighting" feature of ES but not sure it would help.
Thanks.