Clean up resources by closing sessions and clearing tool lists.

Clean up all servers properly.

Clean up all servers properly.

Close Graphiti connection.

Close database connection pool.

Close graph client.

Connect to the MCP server using the configuration file. Raises: ConfigurationError: If the configuration file is

Count the total number of unique documents in the database. Returns: Number of unique documents (based on unique

Count tokens in text using tiktoken. Args: text: The text to count tokens in model: The model name for encoding selection

Create appropriate chunker based on configuration. Args: config: Chunking configuration Returns: Chunker instance

Create appropriate chunker based on configuration. Args: config: Chunking configuration Returns: Chunker instance

Async context manager for creating and managing a Claude agent. Args: working_directory: Working directory for file operations.

Create a contextual enricher instance. Args: model: LLM model to use Returns: ContextualEnricher instance

Creates the output directory if it doesn't exist

Create embedding generator with optional caching. Args: model: Embedding model to use use_cache: Whether to use caching

Create embedding generator with optional caching. Args: model: Embedding model to use use_cache: Whether to use caching

Create graph builder instance.

Create a new session. Args: user_id: Optional user identifier metadata: Optional session metadata timeout_minute

Deactivates all streams and marks all replies as inactive for a given session. This function uses the provided session ID to deactivate exis

Generate embeddings for multiple texts in small batches. Args: texts: List of texts to embed batch_s

Generate embeddings for document chunks. Args: chunks: List of document chunks progress_callback: Op

Generate embeddings for document chunks. Args: chunks: List of document chunks progress_callback: Op

Generate embeddings for document chunks. Args: chunks: List of document chunks progress_callback: Op

Generate embeddings for document chunks. Args: chunks: List of document chunks progress_callback: Op

Generate an embedding for a single text with retry logic. Args: text: The text to embed max_retries:

Enrich multiple chunks with contextual information (with concurrency control). Args: chunks: List of chunk texts to enri

Ensure a directory exists, create it if it doesn't

Extract entities from chunks and add to metadata. Args: chunks: List of document chunks extract_comp

Extract tool calls from Pydantic AI result. Args: result: Pydantic AI result object Returns: List of ToolCall o

Fetch articles related to the query using the Brave API. Args: query (str): The search query. session_id (str): The session

Generate a unique completion ID in OpenAI format.

Generate a unique completion ID in OpenAI format.

Generate embedding for text using OpenAI. Args: text: Text to embed Returns: Embedding vector

Generate embedding for a single text. Args: text: Text to embed Returns: Embedding

Generate embedding for a single text. Args: text: Text to embed Returns: Embedding

Generate embedding for a single text. Args: text: Text to embed Returns: Embedding

Generate embedding for a single text. Args: text: Text to embed Returns: Embedding

Generate three engaging Twitter drafts based on article content. Args: articles (list): A list of articles with titles, URLs, and de

Get a list of all unique document sources. Returns: List of unique source URLs/identifiers

Returns the available LLM providers based on the environment variables The model will be available if all its variables are set

Get a list of all available sources in the knowledge base. Returns: List of source identifiers

Retrieve a list of available tools from the MCP server. Simplify the schema for each tool to make it compatible with the OpenAI API.

Fetch betting odds for a game

Returns the buttons configuration for ideation from question (prompt)

Reduce the list of buttons submitted to a single boolean value to determine if the form was submitted

Get recent conversation context. Args: session_id: Session ID max_messages: Maximum number of messages to retrieve

Returns a formatted date and time

Get all chunks for a document. Args: document_id: Document UUID Returns: List of chunks ordered by chunk index

Generates an embedding vector for a given text using the Gemini model. This function uses the `google.generativeai` library to generate an embedd

Get OpenAI client for embeddings. Returns: Configured OpenAI client for embeddings

Get OpenAI client for embeddings. Returns: Configured OpenAI client for embeddings

Get OpenAI client for embeddings. Returns: Configured OpenAI client for embeddings

Get embedding model name. Returns: Embedding model name

Get embedding model name. Returns: Embedding model name

Get embedding model name. Returns: Embedding model name

Returns the form session state key

Returns the form session state key

Returns the form session state key

Fetch games for a specific date

Returns the ideation form configuration

Get the last retrieved conversation. If "last_retrieved_conversation" entry is found and the id matches, return the buffered conversa

Get LLM model configuration based on environment variables. Supports any OpenAI-compatible API provider. Args: model_choice:

Get LLM model configuration for OpenAI. Returns: Configured OpenAI model

Returns the LLM model object

Return the messages array for the LLM call in the format expected by the OpenAI API. The messages array is a list of dictiona

Get metadata for a TXT file. Args: file_path: Path to the TXT file Returns: Dic

Returns the index of the option in the list

Get an existing collection or create a new one if it doesn't exist. Args: client: ChromaDB client collection_name: Name of th

Get an existing collection or create a new one if it doesn't exist. Args: client: ChromaDB client collection_name: Name of th

Get existing session or create new one.

Fetch stats for multiple players.

Returns the prompt enhancement button config

Makes a GET request with retries using multiple API keys. This function iterates through a list of API keys, attempting to make a GET request

Returns the response as prompt button config

Returns the selected feature

Get messages for a session. Args: session_id: Session UUID limit: Maximum number of messages to return Returns:

Get current standings.

Fetch current injuries for a team

Fetch all players for a team.

Fetch team statistics for the current season.

Returns the title of the app

Returns the title from the question

Returns the title from the question

Get user choice to continue or quit.

Get the working directory for a user. Args: user_id: Telegram user ID Returns: Working directory path (falls back to WO

Fetches transcript of the video, if available.

Handle slash commands for adding MCP servers and listing available functions. Args: query (str): The command query.

Perform a HuggingFace query Args: api_url (str): HuggingFace API URL payload (dict): HuggingFace payload

Perform hybrid search (vector + keyword). Args: input_data: Search parameters Returns: List of matching chunks

Umport data from a uploaded JSON file into the database

Initialize graph client.

Create connection pool.

Create connection pool.

Create connection pool.

Initialize database connection pool.