Complete Getting Started Guide – Liris

Discover how to automate your code production and generate custom datasets with Liris

Video Demonstration

Watch this video to see Liris in action and understand the platform’s capabilities

3-Step Start

1 Configure Projects

Upload your project or access existing configurations. Liris analyzes the structure and builds the relationship graph.

2 Define Objective (Prompt)

Write a natural language prompt describing the code to generate (Liris Vibe Coding) or the dataset to create (Liris Data Science).

3 Launch Automation

Liris uses the project context (code or dataset) to generate a relevant response and displays it in the snippet.

Project Configuration (Vibe Coding & Data Science)

Basic Project Setup

  1. Install the Liris extension for VS Code and log in.
  2. Configure your API keys for AI platforms (OpenAI, Anthropic, etc.).
  3. Upload your project. Liris analyzes the project structure (folders, files, functions/classes) to create its knowledge graph.

Knowledge Graph & Relations

This section allows you to model your project’s architecture in a graph database to ensure that the generated code/dataset is contextually accurate.

How it works

  • Context Saving: The context of each project is stored and can be updated.
  • Optimized Ontology: Creation of an ontology adapted to use cases, allowing navigation and visualization of complex relationships.

Relations between nodes

The graph captures semantic and structural relationships between all code components, including:

Class inheritance, extension, class instances, calls/imports between files, usage (`use`), etc.

Each file and each function/class becomes a descriptive node, and the edges (relationships) represent how these elements interact.

Liris Vibe Coding- Code Automation

Objective

Liris Vibe Coding automates code production by generating project-specific code from natural language prompts.

The system adapts to your project’s context by analyzing its structure and the relationships between components.

How it works

  1. Project structure analysis
  2. Creation of a knowledge graph of relationships between components
  3. Generation of code adapted to the project context
  4. Direct integration into VS Code
Operational details

Liris Data Science – Dataset Generation

Objective

Liris Data Science allows you to generate custom datasets for your data analysis and machine learning projects.

Configure context typologies and combine them to create datasets tailored to your needs.

Key Steps

  1. Configuration of context typologies
  2. Definition of combinations for batches
  3. Taxonomic structuring (optional)
  4. Generation and export of the dataset
Step-by-step guide

Detailed Step-by-Step Processes

Liris Dev Step Order

  1. Project Upload: Upload the project (the code) to be analyzed by Liris.
  2. Analysis and Configuration: Liris analyzes the structure and creates the relationship graph.
  3. Visualization: Visualize semantic relationships in the knowledge graph.
  4. Write your prompt: Define the code objective in natural language.
  5. Launch automation: Liris generates the code using the graph context and integrates the result into the snippet.

Liris Data Science Step Order

  1. Access Configuration menu: Go to the Dataset Configuration tab and create your context typologies.
  2. Define hierarchical structure: (If needed) Create the taxonomic structure (cluster > root label > parent > child).
  3. Configure combinations: Define the context combinations for batch generation.
  4. Visualize distribution: Observe the combination distribution with the distribution chart.
  5. Switch to Generation tab: Write your prompt and choose output settings.
  6. Launch generation: Download your custom dataset.