Get all nodes connected to given node.

Generate message list for chat completion from mixed inputs. Args: inputs (list): List of input dictionaries with 'type' and 'content' ke

Retry get_response up to MAX_RETRIES times with error handling. Args: model (str): Model identifier messages (list): List of mess

Process multiple texts in parallel to get embeddings. Args: model (str): Model identifier texts (list): List of texts to embed

Get video files matching the specified name and config. Files can be .mp4, .mp3, .srt, .txt, depending on the base_path type. Args: b

Get chat completion response from specified model. Args: model (str): Model identifier messages (list): List of message dictionar

Print faces for given image nodes in a grid layout with 9 faces per row. Args: img_nodes (list): List of image node IDs t

Process video frames to detect, cluster and track faces. Args: video_graph: Graph object to store face embeddings and relationships

Add or update bidirectional weighted edges between two nodes. Text-to-text connections are not allowed between same type text nodes.

Add a new text node with episodic or semantic content. Args: text: Text content text_type: Type of text node

Retry get_embedding up to MAX_RETRIES times with error handling. Args: model (str): Model identifier text (str): Text to embed

Load video graph from pickle file.

Reinforce all edges connected to the given node. Args: node_id: ID of the node to reinforce delta_weight: Amo

Update weight of existing bidirectional edge.

Update an existing node. Args: node_id: ID of target node update_info: Dictionary of update information

Weaken all edges connected to the given node. Args: node_id: ID of the node to weaken delta_weight: Amount to

Calculate average cosine similarity between two lists of embeddings.

Add a new face node with initial image embedding(s). Args: img_embedding: Single embedding or list of embeddings

Add a new voice node with initial audio embedding(s). Args: audio_embedding: Single embedding or list of embeddings

Get embedding for text using specified model. Args: model (str): Model identifier text (str): Text to embed Returns:

Get normalized audio embeddings for a list of base64 audio strings Args: base64_audios (list): List of base64 en

Get video/audio information using appropriate libraries. Args: file_path (str): Path to video or audio file Returns:

Transcribe audio file using Whisper model. Args: model (str): Model identifier file_path (str): Path to audio file Returns:

Retry Whisper transcription with error handling. Args: model (str): Model identifier file_path (str): Path to audio file Ret

Normalize embedding to unit length.

Process multiple audio files in parallel using Whisper model. Args: model (str): Model identifier file_paths (list): List of audi

Clean and format JSON string by removing markdown code blocks. Args: json_str (str): Raw JSON string with potential markdown formatting

Search for face nodes using image embeddings.

Search for text nodes using text embeddings. Args: query_embeddings: Query embeddings range_nodes: Optional l

Search for voice nodes using audio embeddings.

Process video segments at specified intervals with given fps. Args: video_graph (VideoGraph): Graph object to store video information

Process video segments at specified intervals with given fps. Args: video_graph (VideoGraph): Graph object to store video information

Validate and fix Python list string. Args: invalid_list (str): Raw Python list string with potential markdown formatting Returns:

Initialize a video graph with nodes for faces, voices and text events. Args: max_img_embeddings: Maximum number of image

Process a batch of video frames to detect faces. Args: params (tuple): A tuple containing: - frames (lis