here by first of all i would like to know do you want requirements written in english language to automatically get converted into the code for the same? or something else?
Hello Amine, it would be nice if you could provide us with more details. What exactly is your goal? Do you want to transform a text in English to a program in java? Will this text be a set of steps that can be somehow translated into commands or an arbitrary text?
If there is no structure in the text, then it seems a very difficult and vague task to solve. For instance for historical text, it doesn't even make sense to me to convert it into a program, because the text doesn't describe a solution of a problem. If there was a text explaining a mathematical task, let's say a summation, then it might make sense to use NLP and try to convert the text into a java program that implements the summation. But in the case of arbitrary text, I can't think of a use case for this kind of translation. The only relevant thing I found online is Betty (attached), where the system is trained to transform your keywords (that describe a question) into a linux command. Best, Konstantina
In fact, concerning historical texts, the direct application could be simulation of historical events (for example expansion of civilisations (over space and time, etc.), in a visual way, this could be used in historical documentaries or shows.
I agree with Sudhakar. If you can parse a sentence into a kind of internal sentence diagram, you can then translate that to a Java sentence diagram and then synthesize that diagram into Java code. That's pretty much what I do with my translators- see my language page,
mokennon.albion.edu
For instance, type "i love you" into the English to Irish translator. The first step makes a formal structure which is the syntactical representation of the sentence. Next a translator turns that sentence diagram into the sentence diagram for Irish. The last step formats that diagram using the grammatical rules of Irish into the desired Irish sentence. I'm supposing that the Stanford parser works a lot like that. You can take the parsed sentence diagram and turn in into the Java code, since Java code is just another kind of language.
So, if I understand correctly, there is no need to actually generate Java code or any kind of code. You only want to make some sense of your text data, for instance to create timelines or to perform topic detection/summarization. You can do this kind of analysis by using/implementing text mining algorithms in any programming language you wish. You can have a look at "latent dirichlet allocation", which is a technique to discover the topics of your text data. Once you know the topics for a number of years, then you can visualize them as a timeline and see the topic evolution.
Dear Amine, If i understood correctly, you can work out with Standford parser to identify the internal structure of the sentences and extract the stem of the sentence. With that stem, you can name the events or group it automatically. Since Standford parser uses JAVA, it will be simple to implement.