Format the error code into a JSON result. Args: code (ErrorCode): Error code to be formatted. Returns: dict:

Load knowledge graph.

Remove and return an item from the queue. If optional args block is true and timeout is None (the default), block if necessary until

Post data to hook url.

Remove and return an item from the queue. If optional args block is true and timeout is None (the default), block if necessary until

@param name: python file name @return: Logger

Log a query. Args: key (str): The key associated with the query. value (str): The value or result associated with the

Wrap http post and error handling.

Generate a sync response from local LLM. Args: prompt (str): The prompt for inference. history (list): List of previo

Processes user queries and generates appropriate responses. It involves several steps including checking for valid questions, extracti

Remove and return an item from the queue. If optional args block is true and timeout is None (the default), block if necessary until

Embed input text or image as feature, output np.ndarray with np.float32

Executes a google search with cache. If the query already exists in the cache, returns the cached result. If an exception occurs duri

Put item into the queue.

Create chunks from a list of texts.

Return the description of the error code.

Processes user queries and generates appropriate responses. It involves several steps including checking for valid questions, extracti

Get filetype depends on URI suffix.

Drop database by id.

update task into redis. :param task: HxdTask :return: bool: True or False

Create a new TextSplitter. Args: chunk_size: Maximum size of chunks to return chunk_overlap: Overlap in characters be

Choose real backend according to config.ini.

Performs a search operation in the selected repository based on the user's question. Args: llm_client: Client instance fo

Generate a response from the appropriate LLM based on the configuration. If failed, use exponential backoff. Async generator. Args:

Is input query relative with knowledge base. Return true or false, and the maxisum score

Processes a query and returns the best match from the vector store database. If the question is rejected, returns None. Args:

Return chunks most similar to query. Args: query: Multimodal query. k: Number of Documents to return. Defaults to 30.

Generates a score based on the prompt, and then compares it to threshold. Args: prompt (str): The prompt for the language

Split text into multiple components.

Extract the features required for the response pipeline based on the document.

Generate a response from the chat service. Args: prompt (str): The prompt to send to the chat service. history (list,

Initializes response and reject feature store. Only needs to be called once. Also calculates the optimal threshold based on provided

Is wx_msg a revert command.

Generate a score based on the prompt, and then compares it to threshold. Args: prompt (str): The prompt for the language model. t

Preprocesses files in a given directory. Copies each file to 'preprocess' with new name formed by joining all subdirectories with '_'.

Rerank input texts, return descending indexes, indexes[0] is the nearest chunk.

Choose real backend according to config.ini.

Build a prompt for interaction. Args: history_pair (list): List of previous interactions. instruction (str): Instruct

Generate a stream response from local LLM. Wrap transformer API to async generator Args: prompt (str): The prompt for inference.

Executes a shell command and returns its output. Args: txt (str): Command to be executed in the shell. Returns: