Update wavelengthIndex description for processed data

[sreekanthkura7]

Clarified requirements for wavelengthIndex in processed data.

Closes #174

[HanBnrd]

Thanks @sreekanthkura7, good catch! Looks good to me!

[Horschig]

I guess this change would require a respective adjustment in the validator, right? Is @sstucker still the designated validator maintainer?

[jcrdubois]

is this the final convention that has been decided on? Some readers have made their own convention, e.g. Cedalion uses -1 https://github.com/ibs-lab/cedalion/blob/main/src/cedalion/io/snirf.py#L1003

[HanBnrd]

I guess this change would require a respective adjustment in the validator, right? Is @sstucker still the designated validator maintainer?

The validator has a bit of backlog on the specs, we could sort this as part of #183

is this the final convention that has been decided on? Some readers have made their own convention, e.g. Cedalion uses -1 https://github.com/ibs-lab/cedalion/blob/main/src/cedalion/io/snirf.py#L1003

@sreekanthkura7 could we do -1 instead of 0 eventually?

[Horschig]

I do not mind if 0 or -1, but it'd be good to proceed.

[HanBnrd]

Suggested change

	Index of the "nominal" wavelength (in `probe.wavelengths`). For processed data channels (dataType=99999) that are wavelength-independent or derived from multiple wavelengths, this field MUST be set to `0`.
	Index of the "nominal" wavelength (in `probe.wavelengths`). For processed data channels (dataType=99999) that are wavelength-independent or derived from multiple wavelengths, this field MUST be set to `-1`.

[HanBnrd]

Suggested change

	Index of the "nominal" wavelength (in `probe.wavelengths`) for each channel. A 1-D array with length equal to the size of the second dimension of `/nirs(i)/data(j)/dataTimeSeries`. For processed data channels (`dataType=99999`) that are wavelength-independent or derived from multiple wavelengths, the corresponding value MUST be `0`.
	Index of the "nominal" wavelength (in `probe.wavelengths`) for each channel. A 1-D array with length equal to the size of the second dimension of `/nirs(i)/data(j)/dataTimeSeries`. For processed data channels (`dataType=99999`) that are wavelength-independent or derived from multiple wavelengths, the corresponding value MUST be `-1`.

[samuelpowell]

Change is breaking, requires entries in CHANGELOG, etc.?

[Edouard2laire]

Is there a reason why we want to enforce a specific value here? We could make the change non-breaking by saying that if the data are processed and wavelength independent; that the value can be absent or any value.