API Reference

This is the api REFERENCE FOR THE tiomagic Python package, which allows you to run the Tio Magic Animation Toolkit

After following setup instructions, at the top of your file add from tiomagic import tm

Configuration

tm.configure(provider=str, gpu=Optional[str], timeout=Optional[str], scaledown_window=Optional[str])

Establishes which provider you will be using. This is required to run any calls.

• provider: To view a list of existing providers, features, and models, run tm.list_implementations

Note that gpu, timeout, and scaledown_window are options for the Modal provider.

• gpu: To view gpu options, go to the Modal gpu docs.
• timeout: Execution timeout for Modal calls. All are set to a default of 1800 (30 minutes)
• scaledown_window: Keeps Modal container alive for x minutes after the last request. All are set to a default of 900 (15 minutes)

Features

text_to_video

tm.text_to_video(model=str, required_args=Dict[str: Any], **optional_args=Dict[str: Any])

Runs text to video calls on a model with given arguments.

• model: To view a list of existing models on a provider tm.get_models(feature="text_to_video", provider="...")
• required_args: A dictionary with arguments needed to run the specified model. Refer to Parameter Documentationto see the accepted required_args of a particular model, or run tm.get_schema(feature=str, model=str).
  • For text to video, a prompt (str) is required.
• optional_args: A dictionary with arguments that can be added to the generation call. Refer to Parameter Documentation to see the accepted optional_args of a particular model, or run tm.get_schema(feature=str, model=str).

image_to_video

tm.image_to_video(model=str, required_args=Dict[str: Any], **optional_args=Dict[str: Any])

Runs image to video calls on a model with given arguments.

• model: To view a list of existing models on a provider tm.get_models(feature="image_to_video", provider="...")
• required_args: A dictionary with arguments needed to run the specified model. Refer to Parameter Documentationto see the accepted required_args of a particular model, or run tm.get_schema(feature=str, model=str).
  • For image to video, a prompt (str) and an image (local or URL) are required.
• optional_args: A dictionary with arguments that can be added to the generation call. Refer to Parameter Documentation to see the accepted optional_args of a particular model, or run tm.get_schema(feature=str, model=str).

interpolate

tm.interpolate(model=str, required_args=Dict[str: Any], **optional_args=Dict[str: Any])

Runs interpolate (in-between) calls on a model with given arguments.

• model: To view a list of existing models on a provider tm.get_models(feature="interpolate", provider="...")
• required_args: A dictionary with arguments needed to run the specified model. Refer to Parameter Documentationto see the accepted required_args of a particular model, or run tm.get_schema(feature=str, model=str).
  • For interpolate, a prompt (str), first_frame (local or URL), and last_frame (local or URL) are required.
• optional_args: A dictionary with arguments that can be added to the generation call. Refer to Parameter Documentation to see the accepted optional_args of a particular model, or run tm.get_schema(feature=str, model=str).

pose_guidance

tm.pose_guidance(model=str, required_args=Dict[str: Any], **optional_args=Dict[str: Any])

Runs pose guidance calls on a model with given arguments.

• model: To view a list of existing models on a provider tm.get_models(feature="pose_guidance", provider="...")
• required_args: A dictionary with arguments needed to run the specified model. Refer to Parameter Documentationto see the accepted required_args of a particular model, or run tm.get_schema(feature=str, model=str).
  • For pose guidance, a prompt (str), image (local or URL), and either a guiding video (a video to guide the pose of the output video) or a pose video (a pose skeleton video) is required.
• optional_args: A dictionary with arguments that can be added to the generation call. Refer to Parameter Documentation to see the accepted optional_args of a particular model, or run tm.get_schema(feature=str, model=str).

check_generation_status

tm.check_generation_status(job_id=str)

Checks and updates the current status of a video generation job. You can find the job_id in generation_log.json. THis method can be called periodically to monitor long-running generation tasks. If you are running on Modal provider, you can also go onto your Modal dashboard to track generation.

If a generation is completed, calling check_generation_status will download the resulting video into output_videos directory.

At the moment, this will only run on the Modal provider. Local providers are all synchronous and will download the resulting video once generation is complete

cancel_job

tm.cancel_job(job_id=str)

Attempts to cancel a running job with the provider. The success of cancellation depends on the provider’s capabilities and the current state of the job. The result can be found in generation_log.json.

At the moment, this will only run on the Modal provider. Local providers are all synchronous.

get_providers

tm.get_providers

List all providers available.

get_models

tm.get_models(feature: str, provider: str)

List all models available for a provider and feature.

get_schema

tm.get_schema(feature: str, model: str)

List schema for particular implementation.

list_implementations

tm.list_implementations()

List all implementatoins available. Models listed and sorted by provider and feature.