Quickstart¶
From a fresh laptop to your first uploaded measurement and first query. You need a CloudSprite account (accept your invite first) and Python 3.11 or newer.
1. Install the SDK¶
The library is a private beta — install the wheel from the URL in your invite, not PyPI:
pip install "cloudsprite[rf] @ <wheel URL from your invite>"
[rf] adds Touchstone / scikit-rf parsing; drop it if you only work with CSV.
2. Create an API key¶
In the web app, open account settings → API keys (cloudsprite.io/settings/api-keys) and create a key. Copy the secret once — CloudSprite will not show it again. Details: API keys and login.
3. Configure the SDK¶
cloudsprite init
init prompts for the key, stores it in ~/.cloudsprite/config with owner-only permissions, and can discover your default org and team. In CI or headless jobs, set CLOUDSPRITE_API_KEY instead.
4. Connect and pick your scope¶
import cloudsprite as cs
cs.connect() # raises cs.AuthError on a bad or expired key
cs.set_org("acme") # only needed for multi-org keys
cs.set_team("rf-lab")
cs.set_project("cable-qual")
5. Upload a measurement¶
project = cs.projects["cable-qual"]
ds = project.create_dataset(title="sweep-25c", description="First sweep at 25 °C")
ds.upload("sweep_25c.s2p")
Parsing is asynchronous — traces such as S21 appear on the dataset shortly after upload. Set parameters in the same sitting (lot, temp, sn, instrument); the filename is not a catalog.
6. First query¶
s21 = ds.traces["S21"]
x, y = s21.x, s21.y # network I/O happens here
lot15 = project.datasets.filter(cs.params.lot == "15")
Where next¶
- Getting started — the first hour in the web app
- Python SDK how-to — the navigate / compute / publish loop
- Upload measurement files — formats and the app upload path
- Assistant and Claude plugin — ask the assistant from your AI client