Skip to main content

Quick Overview

Videos: See Artemis in Action

With 10 steps, we will show you how to conduct an initial audit of your codebase, and generate and evaluate code recommendations for suboptimal code snippets.

Step 1: Log in​

Log-in to Artemis, either via the web platform, or within your local deployment. (This will take you to the Dashboard page.)

Step 2: Start a new project​

Click on New Project on your dashboard.

New Project

Step 3: Import your codebase as a new project​

Fill in the following details in the form under New Project:

  1. Name: Name for the project (REQUIRED)
  2. Description: Description (optional)
  3. Source: Select the Source, either Git repository or a .zip on a local folder
  4. URL + Branch OR Zip file: If you are importing code from a Git repository, give the URL and the branch. If you are uploading code from a local folder, upload the folder as a.zip file.
  5. Language: Specify the language of the codebase.

Click Create Project.

Step 4: Create a new analysis​

Go to the Projects tab and locate your project. Click the Analyse button on the right corner to specify what parts of the codebase you want to analyse.

Analyse button

This will open the New Analysis page.

Select which files you want to include/exclude in the analysis, and the types of code snippets you want extracted specifically.

New analysis

Step 5: Create scores for snippets​

Go to the Snippets tab within your code project. Click on the "balance" icon next to Scores to start scoring your code snippets. Select which LLMs you want to use for the scoring and click Score:

Snippet scoring

Step 6: Evaluate scores​

Go to the Snippets tab of your code project. Check the scores under the Scores column to understand the quality of each code snippet.

Step 7: Create recommendations​

Go to the Snippets tab of your code project. Click the "plus" sign next to the Recommendations column. Select which LLMs you would like to use to generate the new recommendations.

Go to the Snippets tab in your code project. Under recommendations, you will see a list of alternative code snippets generated for each extracted code snippet. Click on each one of the entries to view them in detail, and use the code chat to clarify questions about the snippets.

Step 9: Evaluate recommendations​

Go to the Snippets tab in your code project. Click on the "balance" icon next to Recommendation Scores. Click which LLMs you would like to use for the scores and click Score.

Once the scores are generated, use them to evaluate the quality of the snippets generated.

Step 10: Validate recommendations​

Go to the Snippets tab in your code project. Click on the "play" icon next to the Validations column and select which tests you would like to use to validate recommended code snippets. Click Run. Use the validation outcomes to evaluate the quality of the code snippets.