Skip to main content

A community-maintained Python package for wrapping a variety of REST API calls for different AI art models.

Project description

Pycurtain

A community-maintained python package for wrapping REST APIs that interface with a variety of different AI models. Think "curtain", like wizard of oz behind the curtain. This tool lets the developer close the curtains and only focus on implementing one code base for any supported AI models.

Install

pip install pycurtain

Usage

There are tasks (image-related) and sources (AI art models) that are supported. The following is an example of how to use the package to generate an image from a model source.

import pycurtain

source = pycurtain.source.SourceType.STABLE_DIFFUSION
prompt = "A blooming flower in the middle of a desert"
img_o = pycurtain.task.image.generate(source, prompt)[0]
img_o.save('output.png')

Supported Tasks

image

  • generate (text2img, creation)
  • vary (img2img, style transfer)
  • edit (img2img, with masks)
  • interrogate (img2text, description)

text

  • summarize (tl;dr)
  • fix_grammar (punctuation, spelling, capitalization, etc.)
  • tag (subject matter, genres, subgenres, etc.)

Supported Processes

A process is a way to combine multiple tasks together. For example, you can convert a sequence of images into a video.

  • text_to_video_zoom (zoom_in and_out)
from pycurtain.process import text_to_video_zoom
from pycurtain.source import SourceImageType
text_to_video_zoom.zoom_in(prompts=["A grateful dead poster"], source=SourceImageType.STABLE_DIFFUSION, n_imgs=10, file_name="output.mp4")

AI Model Sources

Currently all but one supported AI Models require an API key generated from the REST API host (Craiyon being the exception). Pycurtain is a third-party (open-source / community driven) python wrapper SDK and is not associated with any of the AI Model providers. Each API key can be generated by signing up from the AI Model hosting site. The current implementation requires each (in-use) API key to be set in the environment variables following the \secrets\stuff.py format.

No API Key or private information is stored by Pycurtain.

Sources Offical Sites

Road Map

  • provide prelimiary wrappers for these AI model sources

  • allow users to adjust common parameters between models

  • provide utility features beyond the model wrapper

    • equirectangluar images (360 degree images)
    • simple inpainting mask creation
  • beyond images... videos, text, audio. All leveraging the available REST API accessible ML models of today.

Developing Pycurtain

To install pycurtain, along with the tools you need to develop and run test, run the following in your virtual environment:

pip install -e .[dev]

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

pycurtain-0.0.7.tar.gz (14.7 kB view details)

Uploaded Source

Built Distribution

pycurtain-0.0.7-py3-none-any.whl (17.6 kB view details)

Uploaded Python 3

File details

Details for the file pycurtain-0.0.7.tar.gz.

File metadata

  • Download URL: pycurtain-0.0.7.tar.gz
  • Upload date:
  • Size: 14.7 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/4.0.2 CPython/3.9.16

File hashes

Hashes for pycurtain-0.0.7.tar.gz
Algorithm Hash digest
SHA256 5d06ce803b1065d8f9d343149015b61e7492833711d9ddf086dbca0e0e7d6816
MD5 41f027a799effa70824715f08e8f57ec
BLAKE2b-256 1899aad590d591c6fe720204d1661ac2d15390ea0aa4f079b46be38a18ed496b

See more details on using hashes here.

File details

Details for the file pycurtain-0.0.7-py3-none-any.whl.

File metadata

  • Download URL: pycurtain-0.0.7-py3-none-any.whl
  • Upload date:
  • Size: 17.6 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/4.0.2 CPython/3.9.16

File hashes

Hashes for pycurtain-0.0.7-py3-none-any.whl
Algorithm Hash digest
SHA256 4b3b59b459859361696c74b16239f36054c9a8ab04254bbfe7daad31282f3f76
MD5 082e78b8b3494fa1f8813745c0b5d75a
BLAKE2b-256 805298f805ea3c861ee7fdfec85e044ca80ffb2c6731ba8593e7df8b5278cefb

See more details on using hashes here.

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page