Skip to content

Specify step and flyscan paths in a serializable, efficient and Pythonic way

License

Notifications You must be signed in to change notification settings

bluesky/scanspec

Repository files navigation

CI Coverage PyPI License

scanspec

Specify step and flyscan paths in a serializable, efficient and Pythonic way using combinations of:

  • Specs like Line or Spiral
  • Optionally Snaking
  • Zip, Product and Concat to compose
  • Masks with multiple Regions to restrict

Serialize the Spec rather than the expanded Path and reconstruct it on the server. It can them be iterated over like a cycler, or a stack of scan Frames can be produced and expanded Paths created to consume chunk by chunk.

Source https://github.com/bluesky/scanspec
PyPI pip install scanspec
Docker docker run ghcr.io/bluesky/scanspec:latest
Documentation https://bluesky.github.io/scanspec
Releases https://github.com/bluesky/scanspec/releases

An example ScanSpec of a 2D snaked grid flyscan inside a circle spending 0.4s at each point:

from scanspec.specs import Line, fly
from scanspec.regions import Circle

grid = Line(y, 2.1, 3.8, 12) * ~Line(x, 0.5, 1.5, 10)
spec = fly(grid, 0.4) & Circle(x, y, 1.0, 2.8, radius=0.5)

Which when plotted looks like:

plot

Scan points can be iterated through directly for convenience:

for point in spec.midpoints():
    print(point)
# ...
# {'y': 3.1818181818181817, 'x': 0.8333333333333333, 'DURATION': 0.4}
# {'y': 3.1818181818181817, 'x': 0.7222222222222222, 'DURATION': 0.4}

or a Path created from the stack of Frames and chunks of a given length consumed from it for performance:

from scanspec.core import Path

stack = spec.calculate()
len(stack[0])  # 44
stack[0].axes()  # ['y', 'x', 'DURATION']

path = Path(stack, start=5, num=30)
chunk = path.consume(10)
chunk.midpoints  # {'x': <ndarray len=10>, 'y': <ndarray len=10>, 'DURATION': <ndarray len=10>}
chunk.upper  # bounds are same dimensionality as positions

See https://bluesky.github.io/scanspec for more detailed documentation.