Publish technique metadata in Bliss
===================================

Use the scan metadata mechanism in `BLISS <https://bliss.gitlab-pages.esrf.fr/bliss/>`_ to add
metadata automatically to any scan and dataset based on the content of the scan. For example any scan
that has an "energy" counter is an "XAS" scan:

.. code:: python

    from bliss import current_session
    from bliss.scanning.scan_meta import get_user_scan_meta, META_TIMING
    from esrf_ontologies import technique

    def add_xas_technique(scan):
        channels = scan.scan_info.get("channels", dict())
        if "energy" in channels:
            technique_metadata = technique.get_technique_metadata("XAS")

            # For ICAT:
            technique_metadata.fill_dataset_metadata(current_session.scan_saving.dataset)

            # For HDF5:
            return technique_metadata.get_scan_metadata()

    scan_meta = get_user_scan_meta()
    scan_meta.add_categories({technique.BLISS_SCANINFO_CATEGORY})
    root_category = getattr(scan_meta, technique.BLISS_SCANINFO_CATEGORY)
    root_category.timing = META_TIMING.START
    root_category.set("add_xas_technique", add_xas_technique)

Note that `get_technique_metadata` accepts one or more technique names or acronyms.

When the beamline has technique specific commands instead of the generic BLISS scan command
it can be done like this

.. code:: python

    from bliss import setup_globals
    from bliss import current_session
    from esrf_ontologies import technique

    def xas_scan(start, stop, intervals, count_time, scan_info=None):
        technique_metadata = technique.get_technique_metadata("XAS")

        # For ICAT:
        technique_metadata.fill_dataset_metadata(current_session.scan_saving.dataset)

        # For HDF5:
        if scan_info is None:
            scan_info = technique_metadata.get_scan_info()
        else:
            technique_metadata.fill_scan_info(scan_info)

        setup_globals.ascan(setup_globals.energy, start, stop, intervals, count_time, scan_info=scan_info)

Adding the same technique to a dataset several times is not a problem; a unique list of techniques
is maintained, preventing duplicates.