Store operational information locally for use outside of bluesky

[prjemian]

Users want to see a list of reflections, samples, etc. from their recent session, even after bluesky stops. Multiple sessions could (optionally) access the same list or use a different one.

This would be a local file, stored in the pwd. A session could load this file by default, start a new one, or load from somewhere else.

@strempfer : Thanks for the suggestion!

[prjemian]

@rodolakis - You described an implementation of this idea that uses EPICS PVs. It was useful for when the bluesky session needed to be restarted. (So that any session info could be restored.)

[prjemian]

One way is to generalize the support so it is easy and consistent to call. Consider this class:

from dataclasses import dataclass

from hkl.diffract import Diffractometer


@dataclass
class DiffractometerPreservation:
    """Preserve Diffractometer Configuration."""

    meter: Diffractometer

    def export_dict(self):
        me = self.meter

        d = {
            "name": me.name,
            "geometry": me.calc._geometry.name_get(),
            "class": me.__class__.__name__,
            "engine": me.calc.engine.name,
            "mode": me.calc.engine.mode,
            "class": me.__class__.__name__,
            "hklpy_version": me._hklpy_version_,
            "energy_keV": me.calc.energy,
            "wavelength_angstrom": me.calc.wavelength,
            # fmt: off
            "constraints": {
                k: {
                    nm: getattr(v, nm)
                    for nm in "low_limit high_limit value fit".split()
                }
                for k, v in me._constraints_dict.items()
            },
            "samples": {
                sname: {
                    "name": sample.name,
                    "reflections": sample.reflections_details,
                    "U": sample.reflections_details,
                    "UB": sample.reflections_details,
                }
                for sname, sample in me.calc._samples.items()
            },
            # fmt: on
            "real_axes": list(me.RealPosition._fields),
            "reciprocal_axes": list(me.PseudoPosition._fields),
        }
        return d

    def export_json(self):
        import json

        return json.dumps(self.export_dict())

    def export_yaml(self):
        import yaml

        return yaml.dump(self.export_dict())

An agent could be created: agent = DiffractometerPreservation(e4cv). The default configuration could be exported as JSON by calling agent.export_json():

{"name": "e4cv", "geometry": "E4CV", "class": "SimulatedE4CV", "engine": "hkl", "mode": "bissector", "hklpy_version": "1.0.3", "energy_keV": 8.050921974025975, "wavelength_angstrom": 1.54, "constraints": {"omega": {"low_limit": -180.0, "high_limit": 180.0, "value": 0.0, "fit": true}, "chi": {"low_limit": -180.0, "high_limit": 180.0, "value": 0.0, "fit": true}, "phi": {"low_limit": -180.0, "high_limit": 180.0, "value": 0.0, "fit": true}, "tth": {"low_limit": -180.0, "high_limit": 180.0, "value": 0.0, "fit": true}}, "samples": {"main": {"name": "main", "reflections": [], "U": [], "UB": []}}, "real_axes": ["omega", "chi", "phi", "tth"], "reciprocal_axes": ["h", "k", "l"]}

When orientation info exists for a sample (sample and reflections are added and UB matrix is computed):

{"name": "e4cv", "geometry": "E4CV", "class": "SimulatedE4CV", "engine": "hkl", "mode": "bissector", "hklpy_version": "1.0.3", "energy_keV": 8.050921974025975, "wavelength_angstrom": 1.54, "constraints": {"omega": {"low_limit": -180.0, "high_limit": 180.0, "value": 0.0, "fit": true}, "chi": {"low_limit": -180.0, "high_limit": 180.0, "value": 0.0, "fit": true}, "phi": {"low_limit": -180.0, "high_limit": 180.0, "value": 0.0, "fit": true}, "tth": {"low_limit": -180.0, "high_limit": 180.0, "value": 0.0, "fit": true}}, "samples": {"main": {"name": "main", "reflections": [], "U": [[1.0, 0.0, 0.0], [0.0, 1.0, 0.0], [0.0, 0.0, 1.0]], "UB": [[4.079990459207523, -2.4982736282101165e-16, -2.4982736282101165e-16], [0.0, 4.079990459207523, -2.4982736282101165e-16], [0.0, 0.0, 4.079990459207523]]}, "silicon": {"name": "silicon", "reflections": [{"reflection": {"h": 4.0, "k": 0.0, "l": 0.0}, "flag": 1, "wavelength": 1.54, "position": {"omega": -145.451, "chi": 0.0, "phi": 0.0, "tth": 69.0966}, "orientation_reflection": true}, {"reflection": {"h": 0.0, "k": 4.0, "l": 0.0}, "flag": 1, "wavelength": 1.54, "position": {"omega": -145.451, "chi": 0.0, "phi": 90.0, "tth": 69.0966}, "orientation_reflection": true}], "U": [[-1.2217304763832569e-05, -0.9999999999253688, 0.0], [0.0, 0.0, 1.0], [-0.9999999999253688, 1.2217304763832569e-05, 0.0]], "UB": [[-1.4134284639502065e-05, -1.1569069374686927, 7.084098436944218e-17], [0.0, 0.0, 1.156906937555034], [-1.1569069374686927, 1.4134284639572905e-05, 7.083925341879798e-17]]}}, "real_axes": ["omega", "chi", "phi", "tth"], "reciprocal_axes": ["h", "k", "l"]}

Export as YAML is demonstrated, above. Could add exporters to file, EPICS PV (CA or PVA), or other.

Importers would have to match the geometry of the diffractometer object first. Basically a reverse of the export steps. Convert from incoming format to a standard dict, then a single code that takes the standard dict and updates the diffractometer object.

[prjemian]

The export_dict() method (above) could be part of the Diffractometer() class.

[prjemian]

While hklpy should not design how to persist such information outside the package, it can provide an easy way to import & export. The export_dict() method above is an example.

[prjemian]

Not necessary for the 1.0.4 release.

[prjemian]

The dictionary stored as a run's configuration: diffractometer.read_configuration() but this dict only contains a single sample.