Do not allow an empty instance of `DiffractionObject` - init requires `xarray`, `yarray`, `xtype`

[bobleesj]

Problem

Continuing our discussion from #224 - We do not want User to have an empty DiffractionObject instance.

Proposed solution

In our init method, we could modfy

from:

class DiffractionObject:
    def input_data(
        self,
        xarray,
        yarray,
        xtype,
        metadata={},
        scat_quantity=None,
        name=None,
        wavelength=None,
    ):

to:

class DiffractionObject:
    def __init__(
        self,
        xarray,
        yarray,
        xtype,
        wavelength,
        scat_quantity="",
        name="",
        metadata={},
    ):

        self._id = uuid.uuid4()
        self.input_data(xarray, yarray, xtype, wavelength, scat_quantity, name, metadata)

Here, User needs to provide xarray, yarray, xtype and wavelength. Here, xtype and wavelneght are required for the self._set_xarrays(xarray, xtype) method within input_data where we utilize self.wavelength and self.xtype as shown below:

    def _set_xarrays(self, xarray, xtype):
        self._all_arrays = np.empty(shape=(len(xarray), 4))
        if xtype.lower() in QQUANTITIES:
            self._all_arrays[:, 1] = xarray
            self._all_arrays[:, 2] = q_to_tth(xarray, self.wavelength)
            self._all_arrays[:, 3] = q_to_d(xarray)
    ....

@sbillinge I have a draft PR here: #228 for this implementation.

[sbillinge]

I think that's about right, and I think we want input_arrays to be private.

We don't want wavelength required though. Set default to none. I think we handle that in thighs arrays. Also, pretty soon Caden will need tof neutron data in the DO and then angle is fixed and lamba is in the array....

Actually, kind of interestingly, because he has lots of detectors, angle and wavelength could both be arrays. Let's leave that for a nice future project for Caden, but for now make wavelength optional.

[bobleesj]

Noted:

• Make input_arrays private
• Wavelength optional, default to None

We don't want wavelength required though. Set default to none. I think we handle that in thighs arrays. Also, pretty soon Caden will need tof neutron data in the DO and then angle is fixed and lamba is in the array....

Actually, kind of interestingly, because he has lots of detectors, angle and wavelength could both be arrays. Let's leave that for a nice future project for Caden, but for now make wavelength optional.

@cadenmyers13 Hi Caden, keeping you in the loop since you've expressed an interest before. We are still refactoring and trying to have 3.6.0 release for diffpy.utils. But keeping you in the loop on how diffpy.utils can be used for all kinds of diffraction data.