Catalog#

Every dataset in the package, generated from the dataset manifests.

nbdemo.data.catalog returns the datasets. It reads manifests only, so nothing is downloaded.

import pandas as pd

import nbdemo.data as nd

pd.DataFrame(
    {
        "id": d.id,
        "description": d.description,
        "size (MB)": round(d.size / 1e6, 2),
        "forms": ", ".join(d.forms),
        "license": d.license,
    }
    for d in nd.catalog()
).style.hide(axis="index")
id description size (MB) forms license
aero/flight_dc A 120 s simulated flight near Washington, DC, at 10 Hz: takeoff, climb and three turns, with position, attitude, speeds and engine data 0.180000 path, df Apache-2.0
med/ecg_mitbih_119 The first minute of MIT-BIH Arrhythmia Database record 119: two ECG leads at 360 Hz with beat and rhythm annotations 0.070000 path, native, df ODC-By-1.0
rf/sigmf_synthetic IQEngine's synthetic SigMF recording: about 2 s of complex samples at 480 kS/s, centred on 8.486 GHz 8.000000 path, native CC-BY-4.0

Look one up by its id when the name is only known at run time:

nd.get("aero/flight_dc")
Dataset('aero/flight_dc', 0.18 MB, local, forms=('path', 'df'))

The attribution for datasets whose license asks for one:

for dataset in nd.catalog():
    if dataset.attribution:
        print(f"{dataset.id}: {dataset.attribution}\n")
med/ecg_mitbih_119: Moody GB, Mark RG. The impact of the MIT-BIH Arrhythmia Database. IEEE Eng in Med and Biol 20(3):45-50 (2001). Goldberger A, et al. PhysioBank, PhysioToolkit, and PhysioNet. Circulation 101(23):e215-e220 (2000).

rf/sigmf_synthetic: Marc, 'synthetic', via IQEngine. Licensed under CC BY 4.0.