Add motor info to QNexuswidget

[mspito]

Needs #1087 first

[vasole]

Please rebase your work on current master that includes #1087

[vasole]

Some comments:

PyMca is used at other facilities than the ESRF and former ESRF files did not look like current ones.
I find nxpositioners = nxentry.get("instrument/positioners_start", None) very ESRF-centric and restrictive.

I would aim at using something similar to the function getPositionersGroup(h5file, path) in PyMca5.PyMcaCore.NexusTools.py That would make the use of the new feature not restricted to ESRF. For instance, A getStartingPositionersGroup in PyMca5.PyMcaCore.NexusTools.py seems to me more appropriate. It could use your hard-coded specific call and default to getPositionersGroup if nothing found. That would imply that if one positioner is an array of positions, the first position is used as starting point to allow the use of get_motor_positions.

[mspito]

I tried to address your suggestions. Is there anything else to be done?

[vasole]

I tried to address your suggestions. Is there anything else to be done?

Yes. It continues to be ESRF centric and only applicable to files generated recently.

Please take a look at NexusTools.getPositionersGroup

There you will find a more generic approach to figure out the positioners associated to a scan. In absence of the new ESRF positioners_start group, the information provided in that group is perfectly usable taking the first value of a positioner in case multiple values are present.

[vasole]

The proposed feature should be of use on old and new ESRF files as well as Sardana HDF5 files.

[woutdenolf]

What makes you think positioners_start is "new"? As far as I can remember it has always been there.

[vasole]

Prior to it there was a group inside NXinstrument named positioners and no positioners_start

[vasole]

What makes you think positioners_start is "new"? As far as I can remember it has always been there.

If it serves as reminder, silx convert is still generating the NXinstrument/positioners group.

[woutdenolf]

For reference: NXinstrument/positioners provides motor positions in the scope of entire scan. This means static motors are scalars and moving motors are arrays. NXinstrument/positioners_start and NXinstrument/positioners_end are snapshots of motor positions at the start/end of the scan and are therefore scalars.

[vasole]

For reference: NXinstrument/positioners provides motor positions in the scope of entire scan. This means static motors are scalars and moving motors are arrays. NXinstrument/positioners_start and NXinstrument/positioners_end are snapshots of motor positions at the start/end of the scan and are therefore scalars.

positioners is always there while start and end are not always there . Just convert a SPEC file with silx convert to verify it. Therefore, I want to check first for positioners_start, and if not present, to use positioners, and if not present to try the Sardana equivalents. It is a 10 minutes work. Simply I do not consider to be my job to do it.

[woutdenolf]

So what needs to be done is

• remove getStartingPositionersGroup and use getPositionersGroup in get_motor_positions instead, taking into account the datasets can be 0D or 1D.
• modify getPositionersGroup to look for positioners_start before looking for positioners.

Is that correct?

[vasole]

One cannot escape to having two different functions (or a function with keywords).

You need to keep into account that to match the SPEC behaviour, you want the motor values associated to the positions when the command was issued. If the information is not available, then one uses the group provided by getPositionersGroup

What you propose should work for this issue, but when I look for the positioners group in PyMca, I actually want the group containing the scanned ones because I want the positions during the scan and not when the command was issued.

So, I would just extend the getStartingPositionersGroup to handle the case positioners_start is not there and it gets the getPositionersGroup instead. The handling of 1D data instead of 0D would indeed be necessary. If you want to centralize the dimensions handling as getStartingPositionersAndValues or getStartingPositioners with a keyword to get the values, or whatever, I am pretty sure it can avoid duplicating code somewhere else.

[mspito]

We added the function getStartingPositionerValues which calls the function getStartingPositionersGroup which falls back to the function getPositionersGroup. Is that what you had in mind to be more generic ?

[vasole]

We added the function getStartingPositionerValues which calls the function getStartingPositionersGroup which falls back to the function getPositionersGroup. Is that what you had in mind to be more generic ?

Yes, it is. Thank you. Please, just make the new functionality is available only at NXentry level.

[woutdenolf]

We thought it would be easier if you could click anywhere in the scan and see the list of motors.

[vasole]

I understand your point of view, but I always keep in mind PyMca is used to deal with HDF5 files from many sources (and not even Nexus files). My point of view is that in many cases the motors information will not be there.

To have systematically a Motors tab, associated to any group or dataset of an HDF5 file is overkill to me.

[mspito]

Now we can access to the motor positions only at NXentry level. Is that what you expected ?

[vasole]

Thanks a lot. I just added the license header to the new file prior to the merge.