Code Summarizer

Source Code Summarization implies generating summary in natural language from a given code snippet which can be helpful to developers for a platitude of reasons like Knowledge Training, to understand in brief about a newly imported project, to maintain precise summaries on the evolution of source code (using git history), etc. Instead of using state-of-art approaches like RNN and CNN, we propose an alternative approach that uses UAST (Universal Abstract Syntax Tree) of the source code to generate tokens and then use the Transformer model with self-attention mechanism that uses encoder-decoder, which unlike RNN method is helpful for capturing long-range dependencies. We have considered Java code snippets for generating the code summaries.