functional programming - How to make a python organic chemistry retro-synthesis generator? -

I am trying to learn a dragon by creating a simple program that creates a specific type of practice problem, biological chemistry Students usually face exams: retro-synthesis questions.

For people unfamiliar with this type of question: The elementary and final species of a series of chemical reactions are given to the student, then it is said to determine which reagent / reaction was the last Performed for the initial reactor to obtain the product.

Sometimes you are given only the final product and are asked to list the necessary reactions to synthesize some parameters (start with only 5 carbon or less complex, only Use alcohol, etc.)

By now, I have done some research, and I think that RDkit is a good place to start w / dragon. My plan is to use the smiley format to read the molecules (because I can manipulate it as a string), then define the work for each response, in the end I will need a database of chemical species, The program which randomly selects the species (inital and end species for the problem) program then a random database The selection of species, applies a bunch of responses to it (3-5, user specified) then displays the final product. The user then resolves the question itself, and the program then shows this path (print using the intermediate images and the reagents used to obtain them). easy. In theory

But once I had actually participated in some problems, then I actually started, firstly, it is very difficult to write a function for every single reaction, secondly, while smiling All molecular complications can be handled, it has been thrown on (stereo-chemistry, geometry, etc.), there are several variations for some molecules and I have trouble keeping responses specific. I am using the "replacement" method to manipulate the smiley string, and when I get regio-political responses I want to make it universal

For example: SN2 responses are primary alkal helms, but Not all tertiary people (prevention of disturbance), how do I create a function for this reaction?

There is another problem, I want to tag responses by my respective reagents, in this way I have taken the reagents used to name the function. However, it becomes problematic when there are reagents that can take many different forms (for example, Ginderhard Reagent).

I think there is a better, less repetitive and difficult way to deal with this. In search of a null in the right direction

This is a very ambitious task and you are not the first who The main examples are /

1. , originally E.J. Developed into a group of

2. Developed in J. Gasteiger's group at Erlangen University,

3. , developed at the group of Harvard University at the Corey Montreal University s. Hansian

These projects have seen the development of a few decades, but I do not have any reliable information on their current situation.