File: test_refine.py

package info (click to toggle)
dials 3.25.0%2Bdfsg3-3
  • links: PTS, VCS
  • area: main
  • in suites: sid
  • size: 20,112 kB
  • sloc: python: 134,740; cpp: 34,526; makefile: 160; sh: 142
file content (477 lines) | stat: -rw-r--r-- 16,947 bytes parent folder | download
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
"""
Test command line program dials.refine by running a job with saved data and
comparing with expected output.

This serves as a high level test that not only checks whether refinement works,
but also that the command line program is functioning and that the output models
have not changed format and so on.
"""

from __future__ import annotations

import math
import random
import shutil
import subprocess

import numpy as np
import pytest
from annlib_ext import AnnAdaptor

from dxtbx.model import Beam, Crystal, Detector, Experiment, Goniometer
from dxtbx.model.experiment_list import ExperimentList, ExperimentListFactory

from dials.algorithms.refinement.engine import Journal
from dials.array_family import flex
from dials.command_line.refine import _find_disjoint_sets


def test_i04_weak_data(dials_data, tmp_path):
    data_dir = dials_data("refinement_test_data", pathlib=True)
    experiments_path = data_dir / "i04-weak.json"
    pickle_path = data_dir / "i04-weak.pickle"

    # set some old defaults
    cmd = (
        shutil.which("dials.refine"),
        "close_to_spindle_cutoff=0.05",
        "reflections_per_degree=100",
        "outlier.separate_blocks=False",
        "scan_varying=False",
        "reflections.outlier.nproc=1",
        experiments_path,
        pickle_path,
    )
    result = subprocess.run(cmd, cwd=tmp_path, capture_output=True)
    assert not result.returncode and not result.stderr
    # load results
    reg_exp = ExperimentListFactory.from_json_file(
        data_dir / "i04-weak-regression.json", check_format=False
    )[0]
    ref_exp = ExperimentListFactory.from_json_file(
        tmp_path / "refined.expt", check_format=False
    )[0]

    # test refined models against expected
    assert reg_exp.crystal == ref_exp.crystal
    assert reg_exp.detector == ref_exp.detector
    assert reg_exp.beam == ref_exp.beam

    # test cell parameter esds
    assert ref_exp.crystal.get_cell_parameter_sd() == pytest.approx(
        (0.0009903, 0.0009903, 0.0021227, 0.0, 0.0, 0.0), abs=1e-6
    )
    assert ref_exp.crystal.get_cell_volume_sd() == pytest.approx(23.8051262, abs=1e-6)


def test_scan_varying_refinement_rmsds(dials_data, tmp_path):
    """Run scan-varying refinement, comparing RMSD table with expected values."""

    data_dir = dials_data("refinement_test_data", pathlib=True)
    experiments_path = data_dir / "from-xds.json"
    pickle_path = data_dir / "from-xds-all.pickle"

    # scan-static refinement first to get refined.expt as start point
    result = subprocess.run(
        (
            shutil.which("dials.refine"),
            experiments_path,
            pickle_path,
            "scan_varying=False",
            "reflections_per_degree=50",
            "outlier.algorithm=null",
            "close_to_spindle_cutoff=0.05",
        ),
        cwd=tmp_path,
        capture_output=True,
    )
    assert not result.returncode and not result.stderr
    # NB Requesting corrgram.pdf and history.json exercises
    # https://github.com/dials/dials/issues/1923
    result = subprocess.run(
        (
            shutil.which("dials.refine"),
            "refined.expt",
            pickle_path,
            "scan_varying=true",
            "output.history=history.json",
            "correlation_plot.filename=corrgram.pdf",
            "reflections_per_degree=50",
            "outlier.algorithm=null",
            "close_to_spindle_cutoff=0.05",
            "crystal.orientation.smoother.interval_width_degrees=36.0",
            "crystal.unit_cell.smoother.interval_width_degrees=36.0",
            "set_scan_varying_errors=True",
            "reflections.outlier.nproc=1",
        ),
        cwd=tmp_path,
        capture_output=True,
    )
    assert not result.returncode and not result.stderr

    # load and check results
    history = Journal.from_json_file(tmp_path / "history.json")

    expected_rmsds = [
        (0.088488398, 0.114583571, 0.001460382),
        (0.080489334, 0.086406517, 0.001284069),
        (0.078835086, 0.086052630, 0.001195882),
        (0.077476911, 0.086194611, 0.001161143),
        (0.076755840, 0.086090630, 0.001157239),
        (0.076586376, 0.085939462, 0.001155641),
        (0.076603722, 0.085878953, 0.001155065),
        (0.076611382, 0.085862959, 0.001154863),
        (0.076608732, 0.085856935, 0.001154384),
        (0.076605731, 0.085852271, 0.001153858),
        (0.076604576, 0.085852318, 0.001153643),
        (0.076603981, 0.085854175, 0.001153594),
    ]
    for a, b in zip(history["rmsd"], expected_rmsds):
        assert a == pytest.approx(b, abs=1e-6)

    # check that the used_in_refinement flag got set correctly
    rt = flex.reflection_table.from_file(tmp_path / "refined.refl")
    uir = rt.get_flags(rt.flags.used_in_refinement)
    assert uir.count(True) == history["num_reflections"][-1]


def test_scan_varying_with_automated_outlier_rejection_block_width_interval_width(
    dials_data, tmp_path
):
    """Strict check for scan-varying refinement using automated outlier rejection
    block width and interval width setting"""

    data_dir = dials_data("refinement_test_data", pathlib=True)
    experiments_path = data_dir / "from-xds.json"
    pickle_path = data_dir / "from-xds-all.pickle"

    result = subprocess.run(
        (
            shutil.which("dials.refine"),
            experiments_path,
            pickle_path,
            "scan_varying=true",
            "max_iterations=5",
            "output.history=history.json",
            "crystal.orientation.smoother.interval_width_degrees=auto",
            "crystal.unit_cell.smoother.interval_width_degrees=auto",
            "reflections.outlier.nproc=1",
        ),
        cwd=tmp_path,
        capture_output=True,
    )
    assert not result.returncode and not result.stderr

    # load and check results
    history = Journal.from_json_file(tmp_path / "history.json")

    expected_rmsds = [
        [0.619507829, 0.351326044, 0.006955399],
        [0.174024575, 0.113486044, 0.004704006],
        [0.098351363, 0.084052519, 0.002660408],
        [0.069202909, 0.072796782, 0.001451734],
        [0.064305277, 0.071560831, 0.001165639],
        [0.062955462, 0.071315612, 0.001074453],
    ]
    for a, b in zip(history["rmsd"], expected_rmsds):
        assert a == pytest.approx(b, abs=1e-6)

    # check the refined unit cell
    ref_exp = ExperimentListFactory.from_json_file(
        tmp_path / "refined.expt", check_format=False
    )[0]
    unit_cell = ref_exp.crystal.get_unit_cell().parameters()
    assert unit_cell == pytest.approx(
        [42.27482, 42.27482, 39.66893, 90.00000, 90.00000, 90.00000], abs=1e-3
    )

    refined_refl = flex.reflection_table.from_file(tmp_path / "refined.refl")
    # re-predict reflections using the refined experiments
    predicted = flex.reflection_table.from_predictions_multi([ref_exp])

    matched, reference, unmatched = predicted.match_with_reference(refined_refl)
    # assert most refined reflections are matched with predictions
    assert reference.size() > (0.997 * refined_refl.size())

    # second check with nearest neighbour matching that the predictions match up
    ann = AnnAdaptor(data=predicted["xyzcal.px"].as_double(), dim=3, k=1)
    ann.query(refined_refl["xyzcal.px"].as_double())
    assert (ann.distances < 0.5).count(True) > (0.998 * refined_refl.size())


@pytest.mark.parametrize("fix", ["cell", "orientation"])
def test_scan_varying_with_fixed_crystal(fix, dials_data, tmp_path):
    location = dials_data("multi_crystal_proteinase_k", pathlib=True)
    refls = location / "reflections_1.pickle"
    expts = location / "experiments_1.json"

    result = subprocess.run(
        [shutil.which("dials.refine"), expts, refls, f"crystal.fix={fix}"],
        cwd=tmp_path,
        capture_output=True,
    )
    assert not result.returncode and not result.stderr


def test_scan_varying_missing_segments_multi_crystal(dials_data, tmp_path):
    # https://github.com/dials/dials/issues/1053
    location = dials_data("i19_1_pdteet_index", pathlib=True)
    refls = location / "indexed.refl"
    expts = location / "indexed.expt"

    # first remove some reflections to keep dials.refine sharp
    data = flex.reflection_table.from_file(refls)

    # prune only indexed reflections

    data = data.select(data.get_flags(data.flags.indexed))
    z = data["xyzobs.px.value"].parts()[2]

    # take overlapping subsets of reflections from two lattices which do not
    # fill the entire scan

    sel = ((data["id"] == 0) & (z > 200)) | ((data["id"] == 1) & (z < 650))
    trimmed = data.select(sel)

    trimmed_filename = tmp_path / "indexed_trim.refl"
    trimmed.as_file(trimmed_filename)

    result = subprocess.run(
        [shutil.which("dials.refine"), expts, trimmed_filename],
        cwd=tmp_path,
        capture_output=True,
    )
    assert not result.returncode and not result.stderr

    el = ExperimentListFactory.from_json_file(
        tmp_path / "refined.expt", check_format=False
    )

    # verify scans start and finish correctly
    image_ranges = [e.scan.get_image_range() for e in el]

    assert image_ranges[0][1] == 850
    assert image_ranges[1][0] == 1


@pytest.mark.parametrize("gcb", ["None", "3000"])
def test_scan_varying_multi_scan_one_crystal(gcb, dials_data, tmp_path):
    # https://github.com/dials/dials/issues/994
    location = dials_data("l_cysteine_dials_output", pathlib=True)
    refls = location / "indexed.refl"
    expts = location / "indexed.expt"

    # Set options for quick rather than careful refinement
    result = subprocess.run(
        [
            shutil.which("dials.refine"),
            expts,
            refls,
            "output.history=history.json",
            "outlier.algorithm=tukey",
            "max_iterations=3",
            "unit_cell.smoother.interval_width_degrees=56",
            "orientation.smoother.interval_width_degrees=56",
            "gradient_calculation_blocksize=" + gcb,
        ],
        cwd=tmp_path,
        capture_output=True,
    )
    assert not result.returncode and not result.stderr

    el = ExperimentListFactory.from_json_file(
        tmp_path / "refined.expt", check_format=False
    )

    # Crystal has been copied into each experiment for scan-varying refinement
    assert len(el.crystals()) == 4

    # load and check results
    history = Journal.from_json_file(tmp_path / "history.json")

    expected_rmsds = [
        (0.1401658782847504, 0.2225931584837884, 0.002349912655443814),
        (0.12060230585178289, 0.1585977879739876, 0.002114318828411418),
        (0.10970832317567975, 0.1348574975434352, 0.001955034565537597),
        (0.10373159352273859, 0.12827852889951505, 0.0017901404193256304),
    ]
    for a, b in zip(history["rmsd"], expected_rmsds):
        assert a == pytest.approx(b, abs=1e-6)


def test_find_disjoint_sets():
    # Create some experiments with a shared Beam
    beam = Beam()
    shared_beam = []
    for i in range(random.randint(1, 5)):
        crystal = Crystal((10, 0, 0, 0, 10, 0, 0, 0, 10), "P1")
        goniometer = Goniometer()
        detector = Detector()
        shared_beam.append(
            Experiment(
                beam=beam, goniometer=goniometer, crystal=crystal, detector=detector
            )
        )

    # Create some experiments with a shared Crystal
    crystal = Crystal((10, 0, 0, 0, 10, 0, 0, 0, 10), "P1")
    shared_crystal = []
    for i in range(random.randint(1, 5)):
        beam = Beam()
        goniometer = Goniometer()
        detector = Detector()
        shared_crystal.append(
            Experiment(
                beam=beam, goniometer=goniometer, crystal=crystal, detector=detector
            )
        )

    # Create some experiments with a shared Detector
    detector = Detector()
    shared_detector = []
    for i in range(random.randint(1, 5)):
        beam = Beam()
        crystal = Crystal((10, 0, 0, 0, 10, 0, 0, 0, 10), "P1")
        goniometer = Goniometer()
        shared_detector.append(
            Experiment(
                beam=beam, goniometer=goniometer, crystal=crystal, detector=detector
            )
        )

    # Create some experiments with a shared Goniometer
    goniometer = Goniometer()
    shared_goniometer = []
    for i in range(random.randint(1, 5)):
        beam = Beam()
        crystal = Crystal((10, 0, 0, 0, 10, 0, 0, 0, 10), "P1")
        detector = Detector()
        shared_goniometer.append(
            Experiment(
                beam=beam, goniometer=goniometer, crystal=crystal, detector=detector
            )
        )

    # Randomly re-order and make an experiment list
    experiments = shared_beam + shared_crystal + shared_detector + shared_goniometer
    random.shuffle(experiments)
    experiments = ExperimentList(experiments)

    disjoint_sets = _find_disjoint_sets(experiments)

    # We know there are 4 groups
    assert len(disjoint_sets) == 4

    # The length of each group must match one of the known group lengths
    check = {
        len(shared_beam),
        len(shared_crystal),
        len(shared_detector),
        len(shared_goniometer),
    }
    for group in disjoint_sets:
        assert len(group) in check


def test_refinement_of_disjoint_sets(dials_data, tmp_path):
    # Take 22 input experiments that are in 7 disjoint groups
    location = dials_data("polyhedra_narrow_wedges", pathlib=True)
    expts = [
        location / f"sweep_00{i}_experiments.json"
        for i in ["2", "3", "4", "5", "6", "7", "9"]
    ]
    refls = [
        location / f"sweep_00{i}_reflections.pickle"
        for i in ["2", "3", "4", "5", "6", "7", "9"]
    ]
    result = subprocess.run(
        [
            shutil.which("dials.combine_experiments"),
        ]
        + expts
        + refls,
        cwd=tmp_path,
        capture_output=True,
    )
    assert not result.returncode and not result.stderr

    # Refine treating the disjoint groups separately and calculate final RMSDs
    result = subprocess.run(
        [
            shutil.which("dials.refine"),
            "combined.expt",
            "combined.refl",
            "scan_varying=False",
            "output.reflections=separate.refl",
        ],
        cwd=tmp_path,
        capture_output=True,
    )
    separate_refl = flex.reflection_table.from_file(tmp_path / "separate.refl")
    separate_refl = separate_refl.select(
        separate_refl.get_flags(separate_refl.flags.used_in_refinement)
    )
    x_obs, y_obs, z_obs = separate_refl["xyzobs.mm.value"].parts()
    x_calc, y_calc, z_calc = separate_refl["xyzcal.mm"].parts()
    x_resid2 = (x_calc - x_obs) ** 2
    y_resid2 = (y_calc - y_obs) ** 2
    z_resid2 = (z_calc - z_obs) ** 2
    separate_rmsd_x = []
    separate_rmsd_y = []
    separate_rmsd_z = []
    for i in range(22):
        sel = separate_refl["id"] == i
        n = sel.count(True)
        separate_rmsd_x.append(math.sqrt(flex.sum(x_resid2.select(sel)) / n))
        separate_rmsd_y.append(math.sqrt(flex.sum(y_resid2.select(sel)) / n))
        separate_rmsd_z.append(math.sqrt(flex.sum(z_resid2.select(sel)) / n))

    # Refine all together in one joint refinement and calculate final RMSDs
    result = subprocess.run(
        [
            shutil.which("dials.refine"),
            "combined.expt",
            "combined.refl",
            "scan_varying=False",
            "separate_independent_sets=False",
            "output.reflections=joint.refl",
        ],
        cwd=tmp_path,
        capture_output=True,
    )
    joint_refl = flex.reflection_table.from_file(tmp_path / "joint.refl")
    joint_refl = joint_refl.select(
        joint_refl.get_flags(joint_refl.flags.used_in_refinement)
    )
    x_obs, y_obs, z_obs = joint_refl["xyzobs.mm.value"].parts()
    x_calc, y_calc, z_calc = joint_refl["xyzcal.mm"].parts()
    x_resid2 = (x_calc - x_obs) ** 2
    y_resid2 = (y_calc - y_obs) ** 2
    z_resid2 = (z_calc - z_obs) ** 2
    joint_rmsd_x = []
    joint_rmsd_y = []
    joint_rmsd_z = []
    for i in range(22):
        sel = joint_refl["id"] == i
        n = sel.count(True)
        joint_rmsd_x.append(math.sqrt(flex.sum(x_resid2.select(sel)) / n))
        joint_rmsd_y.append(math.sqrt(flex.sum(y_resid2.select(sel)) / n))
        joint_rmsd_z.append(math.sqrt(flex.sum(z_resid2.select(sel)) / n))

    # Check difference in RMSDs within 1/100th of a pixel and 1/100th of a degree.
    # We don't expect them to be exactly the same (https://github.com/dials/dials/pull/2336#issue-1572772140)
    # but we want the two ways to be of equivalent quality.
    assert (
        flex.max(flex.abs(flex.double(separate_rmsd_x) - flex.double(joint_rmsd_x)))
        < 0.01
    )
    assert (
        flex.max(flex.abs(flex.double(separate_rmsd_y) - flex.double(joint_rmsd_y)))
        < 0.01
    )
    assert (
        np.degrees(
            flex.max(flex.abs(flex.double(separate_rmsd_z) - flex.double(joint_rmsd_z)))
        )
        < 0.01
    )