TransCoder builds on advances in natural-language processing (NLP), in particular unsupervised NMT. However, transpilers are far from perfect, and creating one requires significant development effort (and often customization). Sometimes it is helpful to transpile entire codebases from source languages that are obsolete or deprecated for example the 2to3 transpiler tool used to port Python code from the deprecated version 2 to version 3. For example, some languages, such as CoffeeScript and TypeScript, are intentionally designed to use a transpiler to convert from a more developer-friendly language into a more broadly-supported one. Transpilers are useful in several scenarios. The AST is then converted back into source code in a different language, usually by applying re-write rules. Most of these tools work similar to a standard code compiler: they parse the source code into an abstract syntax tree (AST). Although the team restricted its work to only those three languages, they claim it can 'easily be extended to most programming languages.'Īutomated tools for translating source code from one language to another, also known as source-to-source compilers, transcompilers, or transpilers, have existed since the 1970s. Compared to existing systems, TransCoder performed better on this validation set than existing commercial solutions: by up to 33 percentage points compared to j2py, a Java-to-Python translator. To compare the performance of the model, the Facebook team collected a validation set of 852 functions and associated unit tests in each of the system's target languages: Java, Python, and C++. TransCoder is inspired by other neural machine translation (NMT) systems that use deep-learning to translate text from one natural language to another and is trained only on monolingual source data. The team described the system in a paper published on arXiv. TransCoder was trained on more than 2.8 million open source projects and outperforms existing code translation systems that use rule-based methods. Facebook AI Research has announced TransCoder, a system that uses unsupervised deep-learning to convert code from one programming language to another.

10010 code translator