Understanding natural language queries?

Question

data mining

I'm new to this area and I apologize in advance if my question may seem stupid or incorrect, I'm hoping for any help to point me in the right direction.
Now I am doing my thesis, an important part of which will be a natural language query parsing system. Previously, I was simply engaged in search technologies, without deepening into the “understanding” of the request, it was possible to get by with well-known simple algorithms on metrics like tf-idf and probability theory with mathematical statistics. But now the system should "understand" some types of queries and be able to convert it, roughly speaking, into a query to the database. To simplify it completely, for the query “what are my friends living in Moscow” or “my acquaintances living in Moscow”, the system should refer to the “people” table and select everyone who has “city = Moscow”. I hope I explained clearly.
Recently, more and more such systems began to appear (although in the literature “communication with the system in natural language” has been mentioned for the last 50 years). All sorts of Siri, Google Now, Facebook Graph Search, are able to "understand" the request and give not just ranked results like classic search engines, but the correct design for the requested type of information.
I would be interested in any information, any links, books, and even the right Google queries that would help me in studying such systems and writing my own (of course, not so powerful). I will accept any sources (you can in a personal, if you are embarrassed to “blurt out”), I will select it myself, and if necessary and with enough information found, I can write a summary post on Habr.
PS: I already googled on the topics “natural language query parsing”, “natural language database interface”, “syntax / semantic analysis”, started reading the old, old book “communication with computers in natural language” (I quit because it seemed that “ we’re not going there”), but I still didn’t find exhaustive information. There is not enough such a “push” to understand how to combine all this in the system.

Answer 1

Introduction to information retrieval. Christopher Manning
If English is fine, then there is a free version, as far as I remember.