Join paths with the appropriate separator for the platform. Args: *paths: The paths to join. Returns: str: The joined p

Remove and return content item at specified index position

Merge a list of fonts into a single font.

Get a logger with the specified name and level. Args: name (str): The name of the logger. level (int): The logging level (de

Unified interface for chat and tool calls

Apply image captions to the presentation.

Calculate the normalized edit distance between two strings. Args: text1 (str): The first string. text2 (str): The second str

Parse a Presentation from a file. Args: file_path (str): The path to the presentation file. config (Config):

Remove content item from document

Save the presentation to a file. Args: file_path (str): The path to save the presentation to. layout_only (b

Extract JSON from a text response. Args: response (str): The response text. Returns: Dict[str, Any]: The extracted JSON

Get the logger for the current context

Test the connection to the LLM. Returns: bool: True if connection is successful, False otherwise.

Convert the text box to HTML. Args: style_args (StyleArg): The style arguments for HTML conversion. Returns:

Build the fill in a shape. Args: shape (BaseShape): The shape to apply fill to. fill_xml (Optional[str]): The

Find the an element in a slide. Args: slide (SlidePage): The slide element_id (int): The ID of the element. Returns:

Iterate over all shapes in the group. Yields: ShapeElement: Each shape in the group.

Load captions data

Get the inline style for the shape element. Args: style_args (StyleArg): The style arguments for HTML conversion.

Loop interface, return the message or the outcome filepath of the agent.

Parse the markdown content to extract image path and alt text. Format expected: 

Represent the slide page in text. Args: show_image (bool): Whether to show image captions. Returns:

Load todos from CSV file.

Process environment variables in config

Tool calling interface

Create a new logger

Find media by caption or path

Get document overview with sections and subsections

Insert content item after the specified index position

Check if a file path is an image based on its extension. Args: file (str): The file path to check. Returns: bool: True

Set a new logger for the current async context

Filter shapes in the slide by type. Args: shape_type (Type[ShapeElement]): The type of shapes to filter. sha

Process the response based on return options. Args: response (str): The raw response from the model. message

Post-process the response from the agent.

Execute chat or tool call using the specified client

Asynchronously edit the slide.

Save todos to CSV file.

Build a slide in the presentation.

Async version: Perform content schema extraction for the presentation.

Apply dictionary values to an object. Args: dict_obj (Dict[str, Any]): The dictionary with values to apply. obj (Any): The o

Format messages for OpenAI server call. Args: content (str): The prompt content. images (str or list[str]):

Get the documentation for a list of API functions. Args: funcs (list[callable]): A list of functions to document.

Load document JSON

Convert a html table to the image Args: html (str): html text containing a table output_path (str): Output image path, defaults to '

Extract JSON from a text response. Args: response (str): The response text. Returns: Dict|List: The extracted JSON.

Load slide induction data

Get the index position of a content item

Iterate over all media items in the document

Async version: Perform layout induction for the presentation.

Log OpenAI exceptions

Retry a failed turn with feedback and traceback.

Merge all runs in a paragraph into a single run. Args: paragraph (_Paragraph): The paragraph to merge runs in. Returns:

Set the reference presentation and extracted presentation information. Args: presentation (Presentation): The presentati

Set the run directory and create necessary subdirectories. Args: rundir (str): The run directory.

Convert a font dictionary to a CSS style string. Args: font (Dict[str, Any]): The font dictionary. Returns:

Convert the AsyncLLM to a synchronous LLM.

Add functional layouts to the outline.

Collect the history of code execution, API calls and agent steps. Returns: dict: The collected history data.

Download and extract result

Find consecutive ranges in a list of numbers.

Asynchronously generate content for the slide.

Initialize agent roles and their models

Poll parsing result, returns download URL

Process escape characters in args, url, env