Count number of tokens for the given source text :param source_text: Text to count tokens in :return: Number of tokens

Check if the given token count fits into one prompt :param token_count: Token count to check :return: True if this token coun

Initialize the LLM

Query LLM with the provided input. :param user: The actual query. :param system: "System" query. Some LLMs support ins

Parse response and find the first occurrence of text separated with `code_delimiter` :param response: Response from the model :par

This is a hacky attempt to parse ChatGPT response that contain both header an implementation. Sometimes the result would be something like:

Load target LLM. Please note: this method is separate for unit tests to be able to override it :return: Newly loaded LLM

Load translation history that is relevant to the given bit of translated code. :param code: Code that will be translated