Search based on FullText Meta Information

Elastic Stack Elasticsearch
1

Hi There,

we've written a plugin that adds some linguistic information to the fulltext index like Stem and POS.
Now we wanna search based on this information.

Example:

Indexed Text: 
  a) "The cars are driving around the roundabout"
  b) "A car crashed into the wall"

When I now search for the term "car" example b) should be higher ranked than example a).

Also for special cases I wan't to search for constructs like "A noun followed by a verb".

Are such searches possible? Maybe by using scripts?

Bests
Pascal

2

No idea?

3

if you wrote your own plugins, that stores these information somewhere, you also need to write your own query, that makes sense of that stored information.

4

We are storing these information directly in the text index. But there seems to be no way to access them using a query.

5

indeed. It would be super inefficient to extract this on query time from the JSON. You would need to enrich the lucene index and have custom queries for this.

Dont be fooled by the seemingly simple approach. There are companies who have evolved their business model around exactly that, for more than a decade :slight_smile:

6

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.