Analysis of the search phrase for specific queries. How?

Question

Data processing

Imagine that there is a large form with fields for searching information, such as cars. Each time filling it out is long and dreary. As an alternative, one field was chosen, where the user can enter arbitrary text. The task was to recognize the search parameters in this text and map it to the query. Now I went head-on, regular regular expressions for stop words. But I have already regretted it a hundred times, there are always exceptions to the rules, too many options for writing a search query, etc.
Which way to look for a better selection of search parameters from an arbitrary phrase? Neural networks? Where to start if yes?

Answer 1

Look towards lexical analyzers , compiler theory - to obtain an abstract tree that can be turned into a query (s).
Neural networks, in a good way, should be left aside. It is unlikely that they will be needed at any stage.

Answer 2

Tomita parser and analysis of customer requests
To train the network, you will need many people to create a sample
In general, for one person, the task is not worth the time spent - it's easier to connect a search engine or use sphinxsearch.com