An interactive widget for Jupyter notebooks to explore the parameters of t-SNE.
pip install outerspaceAdditionally, if you use jupyter notebook:
jupyter nbextension enable --py widgetsnbextension... or for jupyter lab:
jupyter labextension install @jupyter-widgets/jupyterlab-manager
jupyter labextension install @bokeh/jupyter_bokehRun t-SNE on the digits data set (see result in the image above):
from outerspace import tsne_playground
from sklearn.datasets import load_digits
digits = load_digits()
X, y = digits.data, digits.target
tsne_playground(X, y)Show the actual digit images in a tooltip:
from outerspace import tsne_playground, array2d_to_html_img
from sklearn.datasets import load_digits
digits = load_digits()
X, y, images = digits.data, digits.target, digits.images
images = 256 - images * 16 # convert range and invert
images = images.clip(0, 255) # clip values at 255
images = images.astype('uint8') # convert to uint8
images = [array2d_to_html_img(image, resize=(32,32))
for image in images] # convert to HTML images
tsne_playground(X, y,
additional_columns=dict(images=images),
tooltips='@images{safe}') # safe = don't escape HTMLEvaluating the chemical space of a set of molecules (with molecule images as tooltip):
from outerspace import tsne_playground, pil_to_html_img
from rdkit.Chem import SDMolSupplier, Draw, AllChem
import requests
import numpy as np
url = ('https://raw.githubusercontent.com/rdkit/rdkit'
'/Release_2020_03/Docs/Book/data/solubility.test.sdf')
response = requests.get(url)
supplier = SDMolSupplier()
supplier.SetData(response.text)
ms = [m for m in supplier]
X = np.array([list(AllChem.GetMACCSKeysFingerprint(m)) for m in ms])
y = [m.GetProp('SOL_classification') for m in ms]
images = [pil_to_html_img(Draw.MolToImage(m, size=(150, 150)))
for m in ms]
tsne_playground(X, y,
additional_columns=dict(images=images),
tooltips='@images{safe}')