Merge pull request #125 from easyscience/develop

Release: merge develop into master

Author: AndrewSazonov

.github/workflows/coverage.yaml

@@ -86,7 +86,7 @@ jobs:
           verbose: true
           token: ${{ secrets.CODECOV_TOKEN }}
 
-  # Job 2: Run integration tests with coverage and upload to Codecov
+  # Job 3: Run integration tests with coverage and upload to Codecov
   integration-tests-coverage:
     runs-on: ubuntu-latest
 

pyproject.toml

@@ -22,6 +22,7 @@ classifiers = [
 ]
 requires-python = '>=3.11,<3.14'
 dependencies = [
+  'essdiffraction', # ESS Diffraction library
   'numpy',          # Numerical computing library
   'colorama',       # Color terminal output
   'tabulate',       # Pretty-print tabular data for terminal output
@@ -190,7 +191,13 @@ fail_under = 65     # Temporarily reduce to allow gradual improvement
 
 [tool.ruff]
 # Temporarily exclude some directories until we have improved the code quality there
-exclude = ['tests', 'tmp']
+#exclude = ['tests', 'tmp']
+exclude = [
+  'tmp',
+  'tests/unit',
+  'tests/integration/fitting',
+  'tests/integration/scipp-analysis/tmp',
+]
 indent-width = 4
 line-length = 99
 # Enable new rules that are not yet stable, like DOC
@@ -264,7 +271,8 @@ ban-relative-imports = 'all'
 force-single-line = true
 
 [tool.ruff.lint.per-file-ignores]
-'*test_*.py' = ['S101'] # allow asserts in test files
+'*test_*.py' = ['S101']  # allow asserts in test files
+'conftest.py' = ['S101'] # allow asserts in test files
 # Vendored jupyter_dark_detect: keep as-is from upstream for easy updates
 # https://github.com/OpenMined/jupyter-dark-detect/tree/main/jupyter_dark_detect
 'src/easydiffraction/utils/_vendored/jupyter_dark_detect/*' = [

src/easydiffraction/analysis/calculators/cryspy.py

@@ -361,8 +361,8 @@ def _convert_experiment_to_cryspy_cif(
                     cif_lines.append(f'{engine_key_name} {attr_obj.value}')
 
         x_data = experiment.data.x
-        twotheta_min = float(x_data.min())
-        twotheta_max = float(x_data.max())
+        twotheta_min = f'{np.round(x_data.min(), 5):.5f}'  # float(x_data.min())
+        twotheta_max = f'{np.round(x_data.max(), 5):.5f}'  # float(x_data.max())
         cif_lines.append('')
         if expt_type.beam_mode.value == BeamModeEnum.CONSTANT_WAVELENGTH:
             cif_lines.append(f'_range_2theta_min {twotheta_min}')

src/easydiffraction/experiments/categories/data/bragg_pd.py

@@ -76,7 +76,7 @@ def __init__(self, **kwargs):
             description='Standard uncertainty of the measured intensity at this data point.',
             value_spec=AttributeSpec(
                 type_=DataTypes.NUMERIC,
-                default=0.0,
+                default=1.0,
                 content_validator=RangeValidator(ge=0),
             ),
             cif_handler=CifHandler(
@@ -321,7 +321,22 @@ def meas(self) -> np.ndarray:
 
     @property
     def meas_su(self) -> np.ndarray:
-        return np.fromiter((p.intensity_meas_su.value for p in self._calc_items), dtype=float)
+        # TODO: The following is a temporary workaround to handle zero
+        #  or near-zero uncertainties in the data, when dats is loaded
+        #  from CIF files. This is necessary because zero uncertainties
+        #  cause fitting algorithms to fail.
+        #  The current implementation is inefficient.
+        #  In the future, we should extend the functionality of
+        #  the NumericDescriptor to automatically replace the value
+        #  outside of the valid range (`content_validator`) with a
+        #  default value (`default`), when the value is set.
+        #  BraggPdExperiment._load_ascii_data_to_experiment() handles
+        #  this for ASCII data, but we also need to handle CIF data and
+        #  come up with a consistent approach for both data sources.
+        original = np.fromiter((p.intensity_meas_su.value for p in self._calc_items), dtype=float)
+        # Replace values smaller than 0.0001 with 1.0
+        modified = np.where(original < 0.0001, 1.0, original)
+        return modified
 
     @property
     def calc(self) -> np.ndarray:

src/easydiffraction/utils/utils.py

@@ -74,7 +74,7 @@ def _fetch_data_index() -> dict:
     _validate_url(index_url)
 
     # macOS: sha256sum index.json
-    index_hash = 'sha256:e78f5dd2f229ea83bfeb606502da602fc0b07136889877d3ab601694625dd3d7'
+    index_hash = 'sha256:9aceaf51d298992058c80903283c9a83543329a063692d49b7aaee1156e76884'
     destination_dirname = 'easydiffraction'
     destination_fname = 'data-index.json'
     cache_dir = pooch.os_cache(destination_dirname)

tests/integration/scipp-analysis/dream/conftest.py

@@ -0,0 +1,60 @@
+# SPDX-License-Identifier: BSD-3-Clause
+# Copyright (c) 2026 DMSC
+"""Shared fixtures for DREAM scipp-analysis integration tests.
+
+This module provides pytest fixtures for downloading and parsing
+reduced diffraction data from the DREAM instrument in CIF format.
+"""
+
+from pathlib import Path
+
+import gemmi
+import pytest
+from pooch import retrieve
+
+# Remote CIF file URL (regenerated nightly by scipp reduction pipeline)
+CIF_URL = 'https://pub-6c25ef91903d4301a3338bd53b370098.r2.dev/dream_reduced.cif'
+
+# Expected datablock name in the CIF file
+DATABLOCK_NAME = 'reduced_tof'
+
+
+@pytest.fixture(scope='module')
+def cif_path(
+    tmp_path_factory: pytest.TempPathFactory,
+) -> str:
+    """Download CIF file fresh each test session and return its path.
+
+    Uses tmp_path_factory to avoid pooch caching, ensuring the latest
+    version of the nightly-regenerated CIF file is always used.
+    """
+    tmp_dir = tmp_path_factory.mktemp('dream_data')
+    return retrieve(url=CIF_URL, known_hash=None, path=tmp_dir)
+
+
+@pytest.fixture(scope='module')
+def cif_content(
+    cif_path: str,
+) -> str:
+    """Read the CIF file content as text."""
+    return Path(cif_path).read_text()
+
+
+@pytest.fixture(scope='module')
+def cif_document(
+    cif_path: str,
+) -> gemmi.cif.Document:
+    """Read the CIF file with gemmi and return the document."""
+    return gemmi.cif.read(cif_path)
+
+
+@pytest.fixture(scope='module')
+def cif_block(
+    cif_document: gemmi.cif.Document,
+) -> gemmi.cif.Block:
+    """Return the 'reduced_tof' data block from the CIF document."""
+    block = cif_document.find_block(DATABLOCK_NAME)
+    assert block is not None, (
+        f'Expected CIF datablock {DATABLOCK_NAME!r} was not found in the document.'
+    )
+    return block

tests/integration/scipp-analysis/dream/test_analyze_reduced_data.py

@@ -0,0 +1,213 @@
+# SPDX-License-Identifier: BSD-3-Clause
+# Copyright (c) 2026 DMSC
+"""Tests for analyzing reduced diffraction data using easydiffraction.
+
+These tests verify the complete workflow:
+1. Define project
+2. Add sample model manually defined
+3. Modify experiment CIF file
+4. Add experiment from modified CIF file
+5. Modify default experiment configuration
+6. Select parameters to be fitted
+7. Do fitting
+"""
+
+from pathlib import Path
+
+import pytest
+
+import easydiffraction as ed
+
+# CIF experiment type tags required by easydiffraction to identify
+# the experiment configuration (powder TOF neutron diffraction)
+EXPT_TYPE_TAGS = {
+    '_expt_type.sample_form': 'powder',
+    '_expt_type.beam_mode': 'time-of-flight',
+    '_expt_type.radiation_probe': 'neutron',
+    '_expt_type.scattering_type': 'bragg',
+}
+
+
+@pytest.fixture(scope='module')
+def prepared_cif_path(
+    cif_path: str,
+    tmp_path_factory: pytest.TempPathFactory,
+) -> str:
+    """Prepare CIF file with experiment type tags for
+    easydiffraction.
+    """
+    with Path(cif_path).open() as f:
+        content = f.read()
+
+    # Add experiment type tags if missing
+    for tag, value in EXPT_TYPE_TAGS.items():
+        if tag not in content:
+            content += f'\n{tag} {value}'
+
+    # Write to temp file
+    tmp_dir = tmp_path_factory.mktemp('dream_data')
+    prepared_path = tmp_dir / 'dream_reduced_prepared.cif'
+    prepared_path.write_text(content)
+
+    return str(prepared_path)
+
+
+@pytest.fixture(scope='module')
+def project_with_data(
+    prepared_cif_path: str,
+) -> ed.Project:
+    """Create project with sample model, experiment data, and
+    configuration.
+
+    1. Define project
+    2. Add sample model manually defined
+    3. Modify experiment CIF file
+    4. Add experiment from modified CIF file
+    5. Modify default experiment configuration
+    """
+    # Step 1: Define Project
+    project = ed.Project()
+
+    # Step 2: Define Sample Model manually
+    project.sample_models.add(name='si')
+    sample_model = project.sample_models['si']
+
+    sample_model.space_group.name_h_m = 'F d -3 m'
+    sample_model.space_group.it_coordinate_system_code = '1'
+
+    sample_model.cell.length_a = 5.43146
+
+    sample_model.atom_sites.add(
+        label='Si',
+        type_symbol='Si',
+        fract_x=0.125,
+        fract_y=0.125,
+        fract_z=0.125,
+        wyckoff_letter='c',
+        b_iso=1.1,
+    )
+
+    # Step 3: Add experiment from modified CIF file
+    project.experiments.add(cif_path=prepared_cif_path)
+    experiment = project.experiments['reduced_tof']
+
+    # Step 4: Configure experiment
+    # Link phase
+    experiment.linked_phases.add(id='si', scale=0.8)
+
+    # Instrument setup
+    experiment.instrument.setup_twotheta_bank = 90.0
+    experiment.instrument.calib_d_to_tof_linear = 18630.0
+
+    # Peak profile parameters
+    experiment.peak.broad_gauss_sigma_0 = 48500.0
+    experiment.peak.broad_gauss_sigma_1 = 3000.0
+    experiment.peak.broad_gauss_sigma_2 = 0.0
+    experiment.peak.broad_mix_beta_0 = 0.05
+    experiment.peak.broad_mix_beta_1 = 0.0
+    experiment.peak.asym_alpha_0 = 0.0
+    experiment.peak.asym_alpha_1 = 0.26
+
+    # Excluded regions
+    experiment.excluded_regions.add(id='1', start=0, end=10000)
+    experiment.excluded_regions.add(id='2', start=70000, end=200000)
+
+    # Background points
+    background_points = [
+        ('2', 10000, 0.01),
+        ('3', 14000, 0.2),
+        ('4', 21000, 0.7),
+        ('5', 27500, 0.6),
+        ('6', 40000, 0.3),
+        ('7', 50000, 0.6),
+        ('8', 61000, 0.7),
+        ('9', 70000, 0.6),
+    ]
+    for id_, x, y in background_points:
+        experiment.background.add(id=id_, x=x, y=y)
+
+    return project
+
+
+@pytest.fixture(scope='module')
+def fitted_project(
+    project_with_data: ed.Project,
+) -> ed.Project:
+    """Perform fit and return project with results.
+
+    6. Select parameters to be fitted
+    7. Do fitting
+    """
+    project = project_with_data
+    sample_model = project.sample_models['si']
+    experiment = project.experiments['reduced_tof']
+
+    # Step 5: Select parameters to be fitted
+    # Set free parameters for sample model
+    sample_model.atom_sites['Si'].b_iso.free = True
+
+    # Set free parameters for experiment
+    experiment.linked_phases['si'].scale.free = True
+    experiment.instrument.calib_d_to_tof_linear.free = True
+
+    experiment.peak.broad_gauss_sigma_0.free = True
+    experiment.peak.broad_gauss_sigma_1.free = True
+    experiment.peak.broad_mix_beta_0.free = True
+
+    # Set free parameters for background
+    for point in experiment.background:
+        point.y.free = True
+
+    # Step 6: Do fitting
+    project.analysis.fit()
+
+    return project
+
+
+# Test: Data Loading
+
+
+def test_analyze_reduced_data__load_cif(
+    project_with_data: ed.Project,
+) -> None:
+    """Verify CIF data loads into project correctly."""
+    assert 'reduced_tof' in project_with_data.experiments.names
+
+
+def test_analyze_reduced_data__data_size(
+    project_with_data: ed.Project,
+) -> None:
+    """Verify loaded data has expected size."""
+    experiment = project_with_data.experiments['reduced_tof']
+    # Data should have substantial number of points
+    assert experiment.data.x.size > 100
+
+
+# Test: Configuration
+
+
+def test_analyze_reduced_data__phase_linked(
+    project_with_data: ed.Project,
+) -> None:
+    """Verify phase is correctly linked to experiment."""
+    experiment = project_with_data.experiments['reduced_tof']
+    assert 'si' in experiment.linked_phases.names
+
+
+def test_analyze_reduced_data__background_set(
+    project_with_data: ed.Project,
+) -> None:
+    """Verify background points are configured."""
+    experiment = project_with_data.experiments['reduced_tof']
+    assert len(experiment.background.names) >= 5
+
+
+# Test: Fitting
+
+
+def test_analyze_reduced_data__fit_quality(
+    fitted_project: ed.Project,
+) -> None:
+    """Verify fit quality is reasonable (chi-square value)."""
+    chi_square = fitted_project.analysis.fit_results.reduced_chi_square
+    assert chi_square == pytest.approx(16.0, abs=0.1)