Published on feb 24, 2023
Among the pioneers in this field, Tabnine is a code completion assistant that uses generative AI to predict and suggest the next lines of code based on its surrounding context. In order to generate unit tests, Tabnine has now opened beta access to new capabilities.
In 2014, Rails developer David Heinemeier Hansson declared unit testing dead. This may have been the moment when the developer community began to split into two camps: those who defended unit testing and those who advocated integration testing as a better method to ensure that software behaved as intended.
As Tabnine explains, unit testing is often regarded as a tedious and time-consuming task, and is therefore neglected by many developers. Consequently, they have added unit testing generation to their AI-based software development lifecycle
It supports several programming languages, including Python, Java, and JavaScript, and is integrated with Visual Studio Code and JetBrains IDEs. This tool is capable of not only generating unit tests, but also learning how to match them to the developer’s coding style and patterns.
Tabnine offers three levels of code completion: completing a line, completing a function, or converting natural language comments into code. Depending on the privacy and compliance requirements, it can be deployed either in the cloud or on-premises. Additionally, Tabnine code completion supports Rust, Go, and Bash in addition to unit testing generation.
In the past year, several services have been launched to generate code suggestions, including GitHub Copilot, AWS CodeWhisperer, OpenAI Codex, etc. Tabnine is the first product to offer unit testing generation as well.
Additionally, Tabnine has announced that it has reached over one million monthly users.
