woutdenolf commented 4 months ago

Current NXazint Status

The docs are build daily from the main branch: NXDL --> HTML

First time contributors

Clone the project locally and install all python packages needed to build the docs

python -m venv azintenv
source azintenv/bin/activate

git clone git@github.com:nexuscontributions/nxazint.git
cd nxazint

make install

Build the docs locally

source azintenv/bin/activate
make local
firefox build/manual/build/html/index.html

Existing application definitions that are related

NXmonopd: Monochromatic Neutron and X-Ray Powder diffractometer.
NXcanSAS: Implementation of the canSAS standard to store reduced small-angle scattering data of any dimension.

Initialize NXazint contribution

[x] fork official repo https://github.com/nexusformat/definitions
[x] add at least one MAXIV member as owners to the fork
[x] initialize NXazint (#1 )
[x] setup NeXus documentation rendering for the fork here
[x] create PR of main to official NeXus repo (https://github.com/nexusformat/definitions/pull/1395)
[x] convert the MAXIV proposal to the first NXazint version (#2 )
[x] host MAXIV proposal example here

woutdenolf commented 4 months ago

Ping @meghdadyazdi Please accept the invitation and invite other MAXIV members to the nxazint team on github.

kif commented 4 months ago

FYI, pyFAI is already compatible with:

fgjorup commented 4 months ago

Feel free to add me when you get the chance @meghdadyazdi

woutdenolf commented 4 months ago

Seems like +1 length for edges is actually allowed in NXdata but got lost over time in the docs: https://github.com/nexusformat/definitions/pull/1396#issuecomment-2234015257. So good news for this MR.

zdemat commented 3 months ago

Feel free to add me here @woutdenolf , my GitHub user name is @zdemat.

zdemat commented 1 month ago

@woutdenolf just a status report on this:

we would like to refactor this contributions at MAX IV
- we likely want to split 1D and 2D case
- we want to make NXazit rather a base class that beside data keep data about azint procedure parameters and sw
- we would like to add Application classes that refer to the NXazint
we would like to discuss this more with ESRF beamline staff we ... refers here to a MAX IV working group of more than 5 people

fgjorup commented 1 month ago

Just to elaborate a bit on Zdenek's comment: We suggest a base class, for now called NXazint, and a series of application classes. I don't know whether all of those application classes needs to be strictly defined in NeXus, or if they should just be suggestions to data structuring. The idea would be to have one or more application classes as NXsubentries in one unified NXentry, which also contains the common instrument information (monochromator, source, etc.), which the subentries soft link to.

Unfortunately I could not attach the example.h5 here on github, but here is a rough summary tree:

entry/ - (NXentry)
entry/instrument/ - (NXinstrument)
    entry/instrument/azint1d/ - (NXazint)
        entry/instrument/azint1d/data - (NX_FLOAT)
        entry/instrument/azint1d/process/ - (NXprocess)
            entry/instrument/azint1d/process/date - (NX_DATE_TIME)
            entry/instrument/azint1d/process/note - (NX_CHAR)
            entry/instrument/azint1d/process/parameters/ - (NXparameters)
                entry/instrument/azint1d/process/parameters/error_model - (NX_CHAR)
                entry/instrument/azint1d/process/parameters/mask_file - (NX_CHAR)
                entry/instrument/azint1d/process/parameters/num_of_azi_bins - (NX_INT)
                entry/instrument/azint1d/process/parameters/num_of_radial_bins - (NX_INT)
                entry/instrument/azint1d/process/parameters/pixel_splitting - (NX_INT)
                entry/instrument/azint1d/process/parameters/polarization_factor - (NX_FLOAT)
                entry/instrument/azint1d/process/parameters/poni - (NX_CHAR)
                entry/instrument/azint1d/process/parameters/solid_angle_correction - (NX_BOOLEAN)
                entry/instrument/azint1d/process/parameters/unit - (NX_CHAR)
                entry/instrument/azint1d/process/parameters/wavelength - (NX_FLOAT)
            entry/instrument/azint1d/process/program - (NX_CHAR)
            entry/instrument/azint1d/process/reference - (NX_CHAR)
            entry/instrument/azint1d/process/version - (NX_CHAR)
        entry/instrument/azint1d/radial_axis - (NX_FLOAT)
    entry/instrument/azint2d/ - (NXazint)
        entry/instrument/azint2d/azimuthal_axis - (NX_FLOAT)
        entry/instrument/azint2d/data - (NX_FLOAT)
        entry/instrument/azint2d/data_errors - (NX_FLOAT)
        entry/instrument/azint2d/process/ - (NXprocess)
            entry/instrument/azint2d/process/date - (NX_DATE_TIME)
            entry/instrument/azint2d/process/note - (NX_CHAR)
            entry/instrument/azint2d/process/parameters/ - (NXparameters)
                entry/instrument/azint2d/process/parameters/error_model - (NX_CHAR)
                entry/instrument/azint2d/process/parameters/mask_file - (NX_CHAR)
                entry/instrument/azint2d/process/parameters/num_of_azi_bins - (NX_INT)
                entry/instrument/azint2d/process/parameters/num_of_radial_bins - (NX_INT)
                entry/instrument/azint2d/process/parameters/pixel_splitting - (NX_INT)
                entry/instrument/azint2d/process/parameters/polarization_factor - (NX_FLOAT)
                entry/instrument/azint2d/process/parameters/poni - (NX_CHAR)
                entry/instrument/azint2d/process/parameters/solid_angle_correction - (NX_BOOLEAN)
                entry/instrument/azint2d/process/parameters/unit - (NX_CHAR)
                entry/instrument/azint2d/process/parameters/wavelength - (NX_FLOAT)
            entry/instrument/azint2d/process/program - (NX_CHAR)
            entry/instrument/azint2d/process/reference - (NX_CHAR)
            entry/instrument/azint2d/process/version - (NX_CHAR)
        entry/instrument/azint2d/radial_axis - (NX_FLOAT)
    entry/instrument/monochromator/ - (NXmonochromator)
        entry/instrument/monochromator/energy - (NX_FLOAT)
        entry/instrument/monochromator/wavelength - (NX_FLOAT)
    entry/instrument/name - (NX_CHAR)
    entry/instrument/source/ - (NXsource)
        entry/instrument/source/name - (NX_CHAR)
        entry/instrument/source/probe - (NX_CHAR)
        entry/instrument/source/type - (NX_CHAR)
entry/integration1d/ - (NXsubentry)
    entry/integration1d/data/ - (NXdata)
    entry/integration1d/definition
    entry/integration1d/instrument/ - (NXinstrument)
entry/integration2d/ - (NXsubentry)
    entry/integration2d/data/ - (NXdata)
    entry/integration2d/definition
    entry/integration2d/instrument/ - (NXinstrument)
entry/xrdmap/ - (NXsubentry)
    entry/xrdmap/data/ - (NXdata)
        entry/xrdmap/data/I - (NX_FLOAT)
    entry/xrdmap/definition

zdemat commented 1 month ago

There is NXiqproc. It is an application class and there are more classes like this. I think the advantage of the application class is that we may be more strict in requirements on the present fields and their names.

@SAnsell can NXiqproc somehow cover the QxQy case?

woutdenolf commented 1 month ago

Note that we can now use single class inheritance in NeXus.

zdemat commented 1 month ago

That is a good point but NXiqproc is really quite an Application class and we do not want most of its fields (instrument, sample etc.) and specialized fields like qx etc. The idea of NXazint (not relevant if it is base-class or application-class) is that the AZINT program, script, pipeline can write it independently of other info, as e.g. detector data.

fgjorup commented 4 weeks ago

Summary of meeting 24/10-2024

Scrap nxazint base class and move data and process directly to the application classes
Define 5 application classes: azint1d, azint2d, azintmap1d, azintmap2d, azintQxQy. (names are subject to change)
Include definitions of allowed units and long names (degrees, 1/angstrom, angstrom, etc.)

zdemat commented 4 weeks ago

Hi,

I agree with the 3 points
note that it was effectively already done with this commit on the separate-2d branch. There is now just single NXazint class that covers only "azint1d". I think we just need to rename it and include the long names.
once the 1d class is polished, we can build the 2d class ... it may happen that they will have much in common and we can use inheritance, but I prefer to do it step by step

zdemat commented 4 weeks ago

@woutdenolf is there any CI to build & update the "web-representation" and publish it? I think the initial one at ESRF is too old and most of the activities are happening at MAX IV now. However for some of us it can be complicated to understand the source definition. We can build CI at MAX IV but we are behind firewall and it will not be visible outside.

woutdenolf commented 4 weeks ago

is there any CI to build & update the "web-representation" and publish it?

Unfortunately not. For the main branch of this repo I have a CI running every day on our gitlab (https://gitlab.esrf.fr/hdf5/nexus/nxazint) which publishes here: https://hdf5.gitlab-pages.esrf.fr/nexus/nxazint/classes/contributed_definitions/NXazint.html.

The CI of the main NeXus repo builds documentation which I'm guessing you can download somehow? https://github.com/nexusformat/definitions/actions/workflows/ci.yaml

Edit: we have it too but I don't know where to download the HTML files: https://github.com/nexuscontributions/nxazint/actions/workflows/ci.yaml.

woutdenolf commented 4 weeks ago

I think the initial one at ESRF is too old

It builds the main branch of this repo every day.

woutdenolf commented 4 weeks ago

I just rebased to main branch of this repo on the main branch of the official repo.

zdemat commented 4 weeks ago

note that the current NXazint is in the separate-2d branch. It is the current 1D case, so it should be renamed.
a CI build at ESRF once per day is I think OK for us

nexuscontributions / nxazint

Initialize NXazint contribution #3

Current NXazint Status

First time contributors

Existing application definitions that are related

Initialize NXazint contribution