This blog was written by Brandon Jung, VP Marketing at Tabnine.
We tested Tabnine as part of our efforts to review the latest code AI apps. Tabnine helps speed up your development process by delivering optimized code suggestions via a chat module and auto complete functionality.
Works with all popular IDEs. Sign up for a free trial.
AI is transforming how we create, maintain, and deploy software. AI coding assistants like Tabnine help developers work faster, develop higher quality code, and be more satisfied. Tabnine achieves this by automating repetitive tasks and generating code, documentation, and tests using a combination of auto-complete and chat-based agents embedded in the IDE.
Tabnine uses the context of your code stored in Bitbucket to deliver optimized recommendations for each of your developers. This blog will show you how to get started with Tabnine and how it works with your Bitbucket repos.
While tools like ChatGPT introduced many to the potential of generative AI for code, Tabnine is the originator of the category – now in use by over a million users across thousands of organizations. Tabnine is an Atlassian partner and a portfolio company of Atlassian Ventures. Get the app via the Atlassian marketplace.
Tabnine is the AI that you control:
Tabnine supports all major IDEs and the most popular languages, libraries, and frameworks.
Tabnine delivers best in class cloud models and a high-velocity local model that is customized based on your code in Bitbucket. We also provide built-in reporting to help developers get up to speed faster and write more consistent code.
Start by checking out the branch from Bitbucket from the Jira issue you’ve been assigned and open the project in your IDE. For these examples, we’re using Visual Studio Code and IntelliJ.
Note: If you don’t see the Tabnine Chat window, click the “Tabnine AI” icon on the activity bar.
Tabnine provides the “/workspace” command to query your code in the current workspace. If you need to locate a function, you can ask using a natural language prompt with this command in the chat.
When exploring new code, it can be helpful to have an explanation of what a particular function does. You can highlight blocks of code and use the /explain-code command in the chat to have Tabnine return an explanation. You may also ask what the highlight code does using a natural language prompt.
Tabnine can generate code from a natural language prompt through the chat. Once the code has been generated it is time to review it. If you need to adjust, Tabnine is there to help as well with midline and newline suggestions.
Next, ask Tabnine to help with refactoring the code that is not performing.
Since we have several developers using the Bitbucket repo, I want to document the details on the code that I have written. I can do this using the chat with a natural language prompt or by using the /document-code command. Once the documents have been generated I can review and choose to insert them into the code.
One last step is writing unit tests. The best time to write the tests are while I am in the code so I jump over and highlight the functions and entering natural language in the chat or by using the /generate-test-for-code. Keep the tests in this file or copy over elsewhere in my Bitbucket repos.
Tabnine provides many context aware and advanced chat features that currently work with Bitbucket. Using natural language prompts and built in commands, we’re able to quickly understand and generate code, documentation, and tests leaving you time to focus on the parts that matter most.
Get started for free today, or contact us to connect with an expert to learn more.
Ash Moosa
9 comments