two-time corr for NeXus definition description online

[ambarb]

• [x] description: two-time intensity correlation function
• [x] plain text: two_time_corr_func

some debate about format of data:

• should it be normalized in a different way?
• should we allow for both types (so baseline represented as 1 or 0)

[ambarb]

This is related to #46. The questions here should also be resolved for this, and then we can decide on both #46 and #49.

The questions: https://github.com/lightsources/BES-XPCS-Pilot/issues/46#issuecomment-898592843

[ambarb]

in discussing, we have two options:

if mutually exclusive, we need only one field. however, we add an attribute to declare if ideal baseline =0 or =1. This attribute should be required. if not always mutually exclusive, then we need 2.

So we can make mutually exclusive and someone can just "re-run the calculation" and create a new file. I think we should add an attribute to the two time called:

baseline_from_normalization baseline_reference baseline_nomralization_method The allowed value with be a constant (typically 1 or 0) depending on the methods imposed for normalization. Since the g2_partials are extracted from the two-time, then should all be consistent and (should all be kept together as fields belonging to the same subentry.

@JulReinhardt these additions are you latest commit.

[JulReinhardt]

@ambarb yes the baseline information is added to the NX defintion xml file, not yet to the creator class. Will do so in a next commit.

[JulReinhardt]

I just noted that as a last step we need to modify the loaders to reflect all changes to the NXxpcs structure. I'll add a new issue.

[JulReinhardt]

this should be done too, was probably not linked to the PR, but should be close?

[ambarb]

I think this can be closed as I see it.