Suggest query reformulation

Update animation and get current frame

Serialize the tree to JSON format

Convert complexity to hierarchical level

Toggle visibility

Tokenize text into words

Calculate total degree of all nodes in a community

Truncate a sentence to fit within max_length

Truncate summary to fit within length limits

Build a turnkey config for the given workspace, optionally enabling Ollama.

Update existing node incrementally

Update node first/last seen timestamps

Validate an API key

Validate a response after generation

Validate a JWT token

Vector similarity search

Checks rate limit and waits if necessary before allowing request Automatically resets the counter when entering a new second. If the rate limit is re

Get the dimension with the lowest score

Set the pipeline approach (optional) Options: "semantic", "algorithmic", "hybrid"

Create metadata with chapter information

Set the chunk overlap (optional)

Override chunk size and overlap (overlap defaults to 20 % of size).

Create with custom confidence threshold

Add entities to the chunk

Enable or disable incremental mode

Set keep_alive for KV cache persistence (e.g. "1h", "30m").

Create metadata with keywords

Set minimum chunk size

Set the similarity threshold (optional)

Set temperature for LLM generation (default: 0.1)

Enable/disable temporal consistency requirement

Set temporal validity range for this entity

Pre-populate with vectors for testing

Access the entry, updating access time and count, and return a clone of the value

Adaptive ranking and deduplication based on query analysis

Retrieve relevant communities using adaptive query routing Automatically selects the appropriate hierarchical level based on query complexity and ret

Add new content incrementally

Add dictionary entries (bulk)

Add a document to the system asynchronously

Add new edge incrementally

Add a custom text normalizer

Add source attribution to content Appends a formatted list of source citations to the response content if citation mode is enabled, limiting to the t

Add query metrics to history

Advanced graph traversal for complex queries

Aggregate neighbors using attention mechanism

Aggregate one layer of neighborhood information

Aggregate neighbors using LSTM (order-dependent)

Aggregate neighbor features

Get all concepts (initial + expanded)

Analyze dependencies between subqueries based on their types

Analyze graph statistics

Analyze query to determine optimal retrieval strategy

Generate an answer to a question using async pipeline

Apply content-based normalization (most aggressive)

Apply pattern filtering to entities based on configured patterns

Apply quality filtering to extracted entities

Apply relevance feedback to boost concept scores

Apply Reciprocal Rank Fusion (RRF) to merge multiple result sets

Apply a single extraction rule

Apply query templates

Apply a single update (placeholder - to be implemented by graph manager)

Get the vector data as a slice

Assemble context from search results

Assemble context asynchronously

Perform async retrieval

Convert atomic facts to graph elements (entities and relationships) # Arguments `facts` - Vector of atomic facts to convert `chunk_id` - ID of the s

Get average operation duration

Get average time saved per hit

Batch add multiple vectors in parallel with proper synchronization

Parallel batch generation with chunking for very large datasets

Process multiple queries in batch

Benchmark a single query

Calculate betweenness centrality (simplified)

BM25-based search

Boost results based on enriched metadata

Build the cache configuration without validation

Build the tree bottom-up by merging nodes at each level

Build co-occurrence relationships between concepts