Weekly update

Changelog

New SDK methods:

• uiclient.documents.correct_image_orientation
• uiclient.documents.template.documentai methods -> Predefines templates to get started more easily

Image Operations

Image_parameters allow to tune the operations performed on images by the API:

class ImageOperations(TypedDict, total=False):
    correct_image_orientation: bool # Whether to correct the image orientation
    dpi : int | Literal['auto'] # The DPI of the image
    image_to_text: Literal["ocr", "llm_description"] # Whether to convert the image to text
    browser_canvas: Literal['A3', 'A4', 'A5'] # The canvas of the browser (default = A4) - `A3`: 11.7in x 16.54in, `A4`: 8.27in x 11.7in, `A5`: 5.83in x 8.27in

New modalities

• native+text
• Handling of MIME HTML .mhtml files

WIP - Stay tuned

• Datasets management dashboard.
• Datasets management SDK
• Prompt Optimization SDK
• Emails webhook API