anything-world 0.1.8

Anything World API wrapper library and CLI

Anything World Python Client

Python library and CLI for Anything World API.

To use this package you'll need to obtain a valid Anything World API key. Please, create an account in our website and then visit your profile page to get access to your API key.

Before sending your 3D models to be animated, please make sure to follow our guidelines on preparing your 3D model for success.

To understand better the underlying web requests called by this Python package, please check our API documentation.

For more information of what changed between versions, please check the CHANGELOG file.

Installing

pip install anything-world

Using

First of all, create a new .env file with the right API key and URLs:

AW_API_KEY=<YOUR API KEY>
AW_API_URL=https://api.anything.world
AW_POLLING_URL=https://api.anything.world/user-processed-model

From Python

This library provides both synchronous and asynchronous implementations. The sync implementation is based on requests and tries to keep dependencies at minimum (i.e. no need to install asyncio nor aiohttp). However, if you use-case requires async calls, please follow the next section for its implementation.

Sync

from anything_world.sync_api import AWClient
from dotenv import load_dotenv
load_dotenv()

# Create a client to be able to query Anything World's API
client = AWClient()

# Search for 3D models of `cats`:
response = client.find('cats')

# Upload files from ./examples/cat folder to be animated
response = client.animate(
    files_dir='./examples/cat',
    model_name='some_cat',
    model_type='cat',
    auto_rotate=True,
    is_symmetric=True
)

# If the `model_type` is not given, the AI pipeline will try to find it automatically!
# response = client.animate(
#     files_dir='./examples/cat',
#     model_name='some_cat',
#     auto_rotate=True,
#     is_symmetric=True
# )

# Response has the model_id of the 3D model that our AI pipeline is currently animating
model_id = response["model_id"]

# Runs a long-polling loop, starting it only after 10 seconds and after that,
# checking every 5 secs if the API is done animating the model
model_data = client.get_animated_model(
    model_id=model_id,
    waiting_time=5,
    warmup_time=10,
    verbose=True
)

# Check if our AI pipeline is done animating the model
is_finished = client.is_animation_done(model_id)
assert is_finished == True

# Gets all model data
model_data = client.get_model(model_id)

From CLI

All AWClient methods are exposed as commands of the anything CLI tool. Just call the anything tool and it will display a manual page of the available commands:

anything

You can do exactly the same we did in Python before through the CLI:

anything find "cat"
anything animate ./examples/cat "some cat" --is_symmetric --auto_rotate
anything get_animated_model <MODEL_ID> --verbose
anything get_model <MODEL_ID>

To know more about the parameters of a specific command, just run the command with no arguments to get a help message:

anything find

For more information about a command, please run the command with the --help suffix:

anything find --help

Developing

Installing from source

git clone [e-mail]:anythingworld/anything-world-python.git
cd anything-world-python
python3 -m venv venv
source venv/bin/activate
pip install -e .

Testing

pytest -v -s

Releasing a new pip package

First bump the version in pyproject.toml and then:

rm -rf dist/*
python3 -m build
python3 -m twine upload dist/* --username __token__ --password <password>

License

This Python module has a MIT-style license, as found in the LICENSE file.

The cat model in examples/cat folder is from Google Poly (CC-BY 4.0) license.

The ASCII logo was generated by https://patorjk.com/software/taag/ using the original font fuzzy.flf by Juan Car ([e-mail]).