Answer the question
In order to leave comments, you need to log in
How to make parser and search by plain text?
Hello.
I became interested in the question of what means can be used to implement a search in simple words, i.e. "What time is the next train to Moscow" or "When is the next train to Moscow" and similar requests. There is, of course, an option to simply try to extract keywords from the text using dictionaries, but maybe there are specialized tools for this?
In general, share who knows what on this topic.
Answer the question
In order to leave comments, you need to log in
Still sphinxsearch.
It has enough tools to solve this problem. Dictionaries are needed in any case to construct a selection of the resulting document/page or a set of them. Also, small dictionaries are usually needed to parse the direction of the sampling restriction, all sorts of "to / from, more / less, older / younger, farther / closer, to / from, etc."
Repeatedly solved the problem of search queries in natural language to various structured data. And I note that if the source of the request is a text field, then users are stupid, they don’t use it. they want to see a complex form with a bunch of select boxes, checkboxes, etc. No one wants to write: "black Lexus 2 liters not older than 2000". But if the input source is a recognized speech on a mobile phone, then it works.
Didn't find what you were looking for?
Ask your questionAsk a Question
731 491 924 answers to any question