Fit Waveform Data with DAP¶

Use this guide when a Waveform already has a source curve, and you want to fit the curve and plot the fitted curve on top. For basic Waveform creation, plotting, and curve styling, start with Control a Waveform from the IPython Client.

Prerequisites¶

• BEC is running with a Dock Area.
• The samx and bpm4i devices are available in dev.
• A Waveform source curve exists, or you are ready to create one.

1. Create a source curve¶

Create a Waveform and plot bpm4i against the samx motor position:

wf = gui.bec.new(gui.available_widgets.Waveform)
wf.plot(device_x=dev.samx, device_y=dev.bpm4i)

Run a scan to produce data:

scans.line_scan(dev.samx, -5, 5, steps=25, exp_time=0.1, relative=False)

The default source curve label is bpm4i-bpm4i, built from the y-axis device and signal.

2. Add a DAP curve to an existing curve¶

Attach a Gaussian Data Analysis Pipeline (DAP) to the existing bpm4i-bpm4i curve:

wf.add_dap_curve(device_label="bpm4i-bpm4i", dap_name="GaussianModel")

device_label is the existing source curve to fit. dap_name is the DAP model to attach to that source curve.

3. Initialize the curve with a fitting model¶

If you know in advance that your plot should also include a fit, add it in the initial plot call:

wf.plot(device_x=dev.samx, device_y=dev.bpm4i, dap="GaussianModel")

Do not call plot again with the same device pair after the source curve already exists. That creates a duplicate source curve and raises an error. Use wf.add_dap_curve(...) to add a DAP model to an existing curve.

4. Use multiple DAP models¶

dap_name can also be a list of model names when you want a composite model:

wf.add_dap_curve(
    device_label="bpm4i-bpm4i",
    dap_name=["GaussianModel", "LinearModel"],
)

You can pass parameter overrides with dap_parameters. For composite models, pass either a list aligned with the model list or a dictionary keyed by model name:

wf.add_dap_curve(
    device_label="bpm4i-bpm4i",
    dap_name=["GaussianModel", "LinearModel"],
    dap_parameters={
        "GaussianModel": {
            "center": {"value": 0.0, "vary": True},
            "sigma": {"value": 1.0, "vary": False, "min": 0.0},
        }
    },
)

Use the same model names as the LMFit built-in models classes, such as GaussianModel, LinearModel, or ConstantModel.

5. Inspect DAP output¶

Inspect the DAP result summary:

wf.get_dap_summary()

Inspect the fitted model parameters:

wf.get_dap_params()

DAP curves are also Waveform curve items. You can retrieve and style them in the same way as other curves:

fit_curve = wf.get_curve("bpm4i-bpm4i-GaussianModel")
fit_curve.set_color("orange")
fit_curve.set_pen_width(3)

GUI equivalent¶

You can configure the same DAP setup from the Waveform curve settings dialog:

1. Open the Waveform curve settings dialog from the plot toolbar .
2. Set the x-axis mode to device.
3. Set the x-axis device to samx.
4. Add a y-axis curve for bpm4i.
5. Add the GaussianModel DAP model to the curve.
6. Confirm the dialog and run the scan.