Check if URL should be scraped based on content changes

Start the debouncer task.

Convert to observations for entity storage.

Remove the pre-commit hook.

Write all manual entries to memories.md with [ ] checkbox format.

Initialize BM25 embedder with configurable parameters. Args: model_name: Identifier for the BM25 model k1: Co

Initialize summarizer with OpenAI configuration.

Initialize parser with Claude projects directory.

Initialize reporter with chat parser and summarizer.

Initialize with optional project path for context.

Initialize with pre-compiled regex patterns for performance.

Validate chunk after creation.

Validate relation chunk after creation.

Validate chat chunk after creation.

Validate entity after creation.

Validate relation after creation.

Initialize mutable default values.

Store all points in batch with detailed logging.

Character-based token approximation fallback.

Delegate orphaned relation cleanup to backend

Safely clean up temporary file with exception handling.

Collect cost data from embedding results.

Create a CoreIndexer instance for Git+Meta optimized processing.

Check if entity exists with valid current content hash

Convert entity to text for embedding.

Extract function signature with parameters and return type.

Extract import relations from import statements.

Recursively find all nodes matching given types.

Get all entity names from the collection. Returns: Set of entity names currently in the collection.

Extract parameter name from catch clause.

Get existing entity IDs for file and chunk type.

Calculate SHA256 hash of file contents (follows existing pattern).

Get human-readable explanation of health status with actionable advice.

Check if the parse tree contains syntax errors.

Initialize tiktoken - Voyage uses similar tokenization to OpenAI's cl100k_base.

Inject project-specific parser configurations.

Check if a file is a test file - DISABLED.

Callback for the coalescer to process a batch of files.

Process a file change or creation with Git+Meta deduplication.

Rebuild full state file from all current files.

Convert relation to text for embedding.

Save current statistics to state file.

Delegate scroll collection to backend.

Handle shutdown signals.

Background timer that periodically checks for ready files.

Add a file change event to the debounce queue.

Add MCP server configuration for a collection.

Add metrics to the context.

Create new entity with additional observation (immutable).

Add a project to the service watch list.

Assert that an embedding has the correct shape and type.

Process a batch of entities, relations, and chunks immediately.

Check if this parser can handle the file.

Check if this parser can handle the file.