Code Search

Stellarion provides two powerful ways to search your codebase: semantic search that understands meaning, and pattern search for precise matching.

Semantic Search

Find code by describing what it does in natural language. Stellarion uses AI to understand your intent and find relevant code.

Example Queries

Find error handling code

Where is user authentication implemented?

Find functions that validate email addresses

Show me database connection logic

How It Works

1. Your project is indexed to create embeddings (vector representations of code)
2. Your query is converted to the same embedding format
3. Stellarion finds code with similar meaning, not just matching keywords

Best Practices

• Be descriptive about functionality
• Include context when helpful
• Use domain terms from your codebase

Pattern Search

Search for exact text patterns using regular expressions. Ideal for finding specific syntax, identifiers, or code patterns.

Example Queries

Find all TODO comments

Search for console.log statements

Find functions starting with "handle"

Search for deprecated API usage

Filtering Options

Narrow your search by:

• File types: Search only .ts files, .py files, etc.
• Directories: Limit to src/, exclude node_modules/
• Context lines: Include surrounding code for context

Element Extraction

Get specific functions, classes, or methods by name:

Get the implementation of the UserService class

Show me the validateInput function

Extract the handleSubmit method from FormComponent

What You Get

• Full source code
• Function signature and parameters
• Documentation comments
• Complexity metrics
• File location

Choosing the Right Search

Indexing for Better Results

For large codebases, index your project to enable semantic search:

Index this project for semantic search

Indexing:

• Creates embeddings for all code
• Enables natural language queries
• Improves search speed
• Should be refreshed after major changes