Syntax Error Detection and Correction in Python Code using ML

A lot of methodologies have been developed to automatically localize and correct a code written in certain programming languages. Requirements of such automatic program repair for senior programmers allow them to utilize their time effectively on the logic of the code and not in manually localizing the syntactic bugs within the code. Machine Learning models hence are one of the major contender technologies being used for achieving the goal. Hence, a brief walk through various developments in syntactic correction would help in noticing the efforts made by the industry. Many researchers from Microsoft have also contributed to this development. This list comprises models developed using OpenAI Codex, the encoder decoder transformer, recurrent neural networks or extending the spelling correction algorithms for the target purpose. An inspiration from all these allowed us to build a model to correct the syntax for the Python codes. Our model combines the two different techniques, a model for generating real world buggy Python codes to enhance the accuracy and prefix and suffix consideration for prediction of the next expected token. A practical usage for the programmers, the extension of the model for various other highly used programming languages in future would have varied applications