Code Search

Semantic code search using ChunkHound indexes and DuckDB.

Highlights

• Natural language queries for code search
• Vector search via DuckDB vss extension
• Filter by language, chunk type, or exclude patterns
• Batch search with multiple queries

Functions

Function	Description
code.search(query, ...)	Search code by meaning
code.search_batch(queries, ...)	Batch search with multiple queries
code.status(path, db)	Check index status

Key Parameters

code.search()

Parameter	Type	Description
query	str	Natural language query (e.g., "authentication logic")
limit	int	Max results (default: 10)
language	str	Filter by language (e.g., "python")
chunk_type	str	Filter by type: "function", "class", "method", "comment"
expand	int	Context lines to include around each match
exclude	str	Pipe-separated patterns to exclude (e.g., "test|mock")
path	str	Path to project root (default: cwd)
db	str	Path to database file relative to project root

code.search_batch()

Parameter	Type	Description
queries	str	Pipe-separated queries (e.g., "auth|login|session")
limit	int	Max results per query (default: 10)
language	str	Filter by language
chunk_type	str	Filter by type
expand	int	Context lines around matches
exclude	str	Patterns to exclude
path	str	Path to project root
db	str	Path to database file relative to project root

Snippets

Snippet	Description	Example
$c_search	Basic semantic search	$c_search query="auth logic"
$c_q	Batch search	$c_q q="auth\|login\|session"
$c_type	Type-filtered search	$c_type query="validation" type="function"

Configuration

In onetool.yaml:

tools:
  code:
    base_url: https://openrouter.ai/api/v1
    model: text-embedding-3-small
    db_path: .chunkhound/chunks.db
    dimensions: 1536
    content_limit: 500        # Max chars without expand (100-10000)
    content_limit_expanded: 2000  # Max chars with expand (500-20000)

Requires

Python packages:

pip install duckdb openai

Secrets:

• OPENAI_API_KEY in secrets.yaml (for embeddings)

Index:

just index           # Index current project
just index path=.    # Explicit path

Examples

# Basic semantic search
code.search(query="authentication logic")

# Filter by language and type
code.search(query="database queries", language="python", chunk_type="function")

# Get expanded context (20 lines around match)
code.search(query="error handling", expand=20)

# Exclude test files
code.search(query="validation", exclude="test|mock|fixture")

# Batch search (multiple queries, merged results)
code.search_batch(queries="auth logic|token validation|session handling")

# Batch search excluding tests
code.search_batch(queries="error handling|validation", exclude="test")

# Check index status
code.status()