Clean up content by removing placeholders and normalizing whitespace Removes unfilled placeholders, normalizes whitespace, and ensures proper sentenc

Get cleanup interval as Duration

Clear all vectors

Clear all dirty markers

Clear dirty status for an edge

Clear all entities and relationships while preserving documents and chunks This is useful for rebuilding the graph from scratch without reloading doc

Calculate closeness centrality

Cluster similar concepts across documents

Get clustering coefficient # Returns Average clustering coefficient

Get the color for this status type

CombSUM fusion (simple addition of normalized scores)

Combine node features with aggregated neighbor features

Combine semantic and keyword results using the configured fusion method

Community-based search for exploratory queries

Compare baseline and improved results

Comprehensive search that combines multiple retrieval strategies (legacy)

Compute overall similarity between two entities

Compute aggregate summary

Concatenate results (simple strategy)

Get total number of concepts

Process multiple queries concurrently

Get a mutable reference to the current configuration Allows direct manipulation of the config for advanced use cases

Calculate containment similarity

Check if this section contains the given offset

Check for common SQL injection patterns (basic check)

Get corpus statistics

Calculate cosine similarity between two vectors

Helper: Compute cosine similarity between two embeddings

Calculate cosine similarity between two vectors

Calculate cost savings based on average LLM call cost

Create configurations for each approach

Create configurations for each approach

Create baseline configuration (no enhancements)

Create chunks from sentences using breakpoints

Create content signature for deduplication

Create a human-readable display key

Create a global entity from an entity cluster

Create improved configuration (all enhancements enabled)

Create leaf nodes from text chunks with parallel processing

Create QA prompt from context

Create a sample benchmark dataset

Create test documents with varying content

Create test documents with varying content

Create a new workspace

Cross-strategy fusion to create hybrid results

Perform cross-strategy fusion of results

Add custom metadata

Set custom queries for warming (used with Custom strategy)

Decompose a complex query into simpler sub-queries

Decompose query with fallback handling

Deduplicate entities and merge their mentions

Deduplicate entities by name and type

Phase 2.1: Symbolic Anchoring

Mode defaults