Refactor to read xml directly

src/esssans/isis/__init__.py

@@ -8,20 +8,23 @@
     lab_frame_transform,
     sans2d_tube_detector_pixel_shape,
 )
+from . import io, mantidio, masking
 from .common import transmission_from_background_run, transmission_from_sample_run
-from .mantidio import CalibrationFilename, Filename, PixelMask, PixelMaskFilename
-from .mantidio import providers as mantidio_providers
-from .masking import apply_pixel_masks
+from .io import PixelMaskFilename
+from .mantidio import CalibrationFilename, Filename
+from .masking import PixelMask
 
 providers = (
-    apply_pixel_masks,
-    get_detector_data,
-    get_monitor,
-    lab_frame_transform,
-    sans2d_tube_detector_pixel_shape,
-) + mantidio_providers
-
-del mantidio_providers
+    (
+        get_detector_data,
+        get_monitor,
+        lab_frame_transform,
+        sans2d_tube_detector_pixel_shape,
+    )
+    + mantidio.providers
+    + masking.providers
+    + io.providers
+)
 
 __all__ = [
     'Filename',

src/esssans/isis/io.py

@@ -0,0 +1,55 @@
+# SPDX-License-Identifier: BSD-3-Clause
+# Copyright (c) 2024 Scipp contributors (https://github.com/scipp)
+"""
+File loading functions for ISIS data, NOT using Mantid.
+"""
+from typing import NewType
+
+import scipp as sc
+
+PixelMaskFilename = NewType('PixelMaskFilename', str)
+
+
+MaskedDetectorIDs = NewType('MaskedDetectorIDs', sc.Variable)
+"""1-D variable listing all masked detector IDs."""
+
+
+def read_xml_detector_masking(filename: PixelMaskFilename) -> MaskedDetectorIDs:
+    """Read a pixel mask from an XML file.
+
+    The format is as follows, where the detids are inclusive ranges of detector IDs:
+
+    <?xml version="1.0"?>
+    <detector-masking>
+        <group>
+            <detids>1400203-1400218,1401199,1402190-1402223</detids>
+        </group>
+    </detector-masking>
+
+    Parameters
+    ----------
+    filename:
+        Path to the XML file.
+    """
+    import xml.etree.ElementTree as ET  # nosec
+
+    tree = ET.parse(filename)  # nosec
+    root = tree.getroot()
+
+    masked_detids = []
+    for group in root.findall('group'):
+        for detids in group.findall('detids'):
+            for detid in detids.text.split(','):
+                detid = detid.strip()
+                if '-' in detid:
+                    start, stop = detid.split('-')
+                    masked_detids += list(range(int(start), int(stop) + 1))
+                else:
+                    masked_detids.append(int(detid))
+
+    return MaskedDetectorIDs(
+        sc.array(dims=['detector_id'], values=masked_detids, unit=None, dtype='int32')
+    )
+
+
+providers = (read_xml_detector_masking,)

src/esssans/isis/mantidio.py

@@ -9,7 +9,7 @@
 import scipp as sc
 import scippneutron as scn
 from mantid.api import MatrixWorkspace, Workspace
-from mantid.simpleapi import CopyInstrumentParameters, Load, LoadMask
+from mantid.simpleapi import CopyInstrumentParameters, Load
 from scipp.constants import g
 
 from ..types import (
@@ -20,18 +20,9 @@
     SampleRun,
 )
 
-IDFFilename = NewType('IDFFilename', str)
-PixelMaskFilename = NewType('PixelMaskFilename', str)
-PixelMask = NewType('PixelMask', sc.Variable)
-
-PixelMaskWorkspace = NewType('PixelMaskWorkspace', MatrixWorkspace)
-
 CalibrationFilename = NewType('CalibrationFilename', str)
 CalibrationWorkspace = NewType('CalibrationWorkspace', MatrixWorkspace)
 
-DetectorIDs = NewType('DetectorIDs', sc.Variable)
-"""Detector ID corresponding to each spectrum."""
-
 
 class DataWorkspace(sciline.Scope[RunType, MatrixWorkspace], MatrixWorkspace):
     """Workspace containing data"""
@@ -42,7 +33,7 @@ def _make_detector_info(ws: MatrixWorkspace) -> sc.DataGroup:
     return sc.DataGroup(det_info.coords)
 
 
-def get_detector_ids(ws: DataWorkspace[SampleRun]) -> DetectorIDs:
+def get_detector_ids(ws: DataWorkspace[SampleRun]) -> sc.Variable:
     det_info = _make_detector_info(ws)
     dim = 'spectrum'
     da = sc.DataArray(det_info['detector'], coords={dim: det_info[dim]})
@@ -52,32 +43,7 @@ def get_detector_ids(ws: DataWorkspace[SampleRun]) -> DetectorIDs:
         sc.arange('detector', da.sizes['detector'], dtype='int32', unit=None),
     ):
         raise ValueError("Spectrum-detector mapping is not 1:1, this is not supported.")
-    return DetectorIDs(da.data.rename_dims(detector='spectrum'))
-
-
-def get_idf_filename(ws: DataWorkspace[SampleRun]) -> IDFFilename:
-    lines = repr(ws).split('\n')
-    line = [line for line in lines if 'Instrument from:' in line][0]
-    path = line.split('Instrument from:')[1].strip()
-    return IDFFilename(path)
-
-
-def from_pixel_mask_workspace(ws: PixelMaskWorkspace, detids: DetectorIDs) -> PixelMask:
-    mask = scn.from_mantid(ws)['data']
-    # The 'spectrum' coord of `mask` is the spectrum *number*, but the detector info
-    # uses the spectrum *index*, i.e., a simple 0-based index.
-    mask.coords['spectrum'] = sc.arange(
-        'spectrum', mask.sizes['spectrum'], dtype='int32', unit=None
-    )
-    index_to_mask = sc.lookup(mask, dim='spectrum', mode='previous')
-    mask_det_info = _make_detector_info(ws)
-    det_mask = sc.DataArray(
-        index_to_mask[mask_det_info['spectrum']],
-        coords={'detector': mask_det_info['detector']},
-    )
-    det_mask = sc.sort(det_mask, 'detector')
-    detid_to_mask = sc.lookup(det_mask, dim='detector', mode='previous')
-    return PixelMask(detid_to_mask[detids])
+    return da.data.rename_dims(detector='spectrum')
 
 
 def load_calibration(filename: CalibrationFilename) -> CalibrationWorkspace:
@@ -101,32 +67,20 @@ class Filename(sciline.Scope[RunType, str], str):
 
 
 def from_data_workspace(
-    ws: DataWorkspace[RunType], calibration: CalibrationWorkspace
+    ws: DataWorkspace[RunType],
+    calibration: CalibrationWorkspace,
 ) -> LoadedFileContents[RunType]:
     CopyInstrumentParameters(
         InputWorkspace=calibration, OutputWorkspace=ws, StoreInADS=False
     )
     up = ws.getInstrument().getReferenceFrame().vecPointingUp()
     dg = scn.from_mantid(ws)
     dg['data'] = dg['data'].squeeze()
+    dg['data'].coords['detector_id'] = get_detector_ids(ws)
     dg['data'].coords['gravity'] = sc.vector(value=-up) * g
     return LoadedFileContents[RunType](dg)
 
 
-def load_pixel_mask(
-    filename: PixelMaskFilename,
-    idf_path: IDFFilename,
-    ref_workspace: DataWorkspace[SampleRun],
-) -> PixelMaskWorkspace:
-    mask = LoadMask(
-        Instrument=idf_path,
-        InputFile=str(filename),
-        RefWorkspace=ref_workspace,
-        StoreInADS=False,
-    )
-    return PixelMaskWorkspace(mask)
-
-
 def load_run(filename: Filename[RunType]) -> DataWorkspace[RunType]:
     loaded = Load(Filename=str(filename), LoadMonitors=True, StoreInADS=False)
     if isinstance(loaded, Workspace):
@@ -140,11 +94,7 @@ def load_run(filename: Filename[RunType]) -> DataWorkspace[RunType]:
 
 providers = (
     from_data_workspace,
-    from_pixel_mask_workspace,
-    get_detector_ids,
-    get_idf_filename,
     load_calibration,
     load_direct_beam,
-    load_pixel_mask,
     load_run,
 )

src/esssans/isis/masking.py

@@ -1,9 +1,31 @@
 # SPDX-License-Identifier: BSD-3-Clause
 # Copyright (c) 2023 Scipp contributors (https://github.com/scipp)
+from typing import NewType
+
+import numpy as np
 import sciline
+import scipp as sc
+
+from ..types import MaskedData, RawData, RunType, SampleRun
+from .io import MaskedDetectorIDs
+
+PixelMask = NewType('PixelMask', sc.Variable)
 
-from ..types import MaskedData, RawData, RunType
-from .mantidio import PixelMask
+
+def to_pixel_mask(data: RawData[SampleRun], masked: MaskedDetectorIDs) -> PixelMask:
+    """Convert a list of masked detector IDs to a pixel mask.
+
+    Parameters
+    ----------
+    data:
+        Raw data.
+    masked:
+        The masked detector IDs.
+    """
+    ids = data.coords['detector_id']
+    mask = sc.zeros(sizes=ids.sizes, dtype='bool')
+    mask.values[np.isin(ids.values, masked.values)] = True
+    return PixelMask(mask)
 
 
 def apply_pixel_masks(
@@ -22,3 +44,9 @@ def apply_pixel_masks(
     for name, mask in masks.items():
         data.masks[name] = mask
     return MaskedData[RunType](data)
+
+
+providers = (
+    to_pixel_mask,
+    apply_pixel_masks,
+)