I assume that first part of your project is to identify English sentence. For that you need to handle morphology, syntax and semantics of the sentence. Better you use Stanford parser (or some else) to analyse an input text. then extract actual meaning of the text

Second part of the project is the generate reverent SQL code to handle data

parsar? what is that?

I have built a system that converts natural language to linked data queries. You can see it in action here http://fsasaki.github.io/stuff/linked-data-access/ and the source code here https://github.com/fsasaki/stuff/tree/gh-pages/linked-data-access . The system deploys a natural language understanding component (API.AI) that identifies speaker intents and the variables needed for a specific intent, e.g.: intent to query the birthdate of a person, and the variable "name of a person". The usage of API.AI is kind of arbitrary - other natural language understanding frameworks can do the same. I am generating SPARQL queries, but adapting the approach for SQL shouldn't be too difficult.   

I didn't compare it to the suggested softwares, but this could fit your needs: https://github.com/machinalis/quepy