mstar.api_server.openai.serving_images

/v1/images/generations (text-to-image) and /v1/images/edits (image-to-image) handlers.

Functions

create_image_edit(api, model_name, adapter, ...)
create_images(api, model_name, adapter, req)
image_count(req)	The request's n, 1 when omitted.
max_images_per_request()	Largest n one image request may ask for: MSTAR_MAX_IMAGES_PER_REQUEST, default 10 (OpenAI's range for the endpoint), at least 1.

async mstar.api_server.openai.serving_images.create_image_edit(api, model_name, adapter, *, prompt, image_bytes, image_filename, model_kwargs, raw_request=None)

async mstar.api_server.openai.serving_images.create_images(api, model_name, adapter, req, raw_request=None)

mstar.api_server.openai.serving_images.image_count(req)

The request’s n, 1 when omitted. Outside 1..max_images_per_request() it raises ValueError (a 400), before anything is submitted.

Return type:

int

mstar.api_server.openai.serving_images.max_images_per_request()

Largest n one image request may ask for: MSTAR_MAX_IMAGES_PER_REQUEST, default 10 (OpenAI’s range for the endpoint), at least 1. The Rust frontend parses it the same way: an ASCII integer, surrounding whitespace allowed, anything else the default.

Return type:

int