File size: 26,243 Bytes
beb4f15
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
405af39
beb4f15
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
"""Core data models for configuration, batches, frame outputs, and fiber geometry."""

from __future__ import annotations

import json
from dataclasses import dataclass, field, asdict
from copy import copy
from typing import Any, Dict, Iterable, List, Optional
import os
import imageio.v3 as iio
import numpy as np
import logging
from tifffile import TiffFile
logger = logging.getLogger(__name__)

@dataclass
class AppConfig:
    # --- Model & Hardware ---
    """Store configuration values for app."""
    model_path: str = ""
    use_gpu: bool = True
    scale_label: str = "1x"
    
    # --- Centerline Params ---
    max_angle_deg: float = 30.0
    ridge_sigma: float = 1.0
    ridge_low: float = 0.01
    ridge_high: float = 0.4
    min_fragment_centerline_length_px: float = 10.0
    measurement_centerline_length_trim_fraction: float = 0.05
    grad_quality_trim_fraction: float = 0.0
    strict_centerlines: bool = False
    keep_segments_without_measurement_centerlines: bool = False
    
    # --- Segmentation Params ---
    segmentation_mode: str = "watershed"
    merge_expansion_factor: float = 0.7
    segmentation_expansion_factor: float = 1.0
    watershed_grad_threshold: float = 0.3
    # Retained for DT-based void/pore analysis; watershed uses the gradient threshold above.
    seg_threshold: int = 1
    min_segment_foreground_mean: float = 0.5
    
    # --- Analysis & Physical Units ---
    sample_spacing: float = 5.0

    # --- DT gradient quality view ---
    # Signed error breakpoints for s = |grad(dt)| - 1:
    # [p0, p1, p2, p3] with p0 < p1 <= 0 <= p2 < p3.
    grad_quality_points: list[float] = field(
        default_factory=lambda: [-0.35, -0.08, 0.08, 0.35]
    )

    # --- Save/export defaults (single image, folder batch, and stack exports) ---
    save_options: dict = field(
        default_factory=lambda: {
            "dt": True,
            "foreground": True,
            "labels": True,
            "overlay": True,
            "diameters": True,
            "orientations": True,
            "outline": False,
            "tables": True,
            "image_format": "png",
            "collision": "suffix",
            "write_manifest": True,
        }
    )

    def save_to_json(self, filepath: str):
        """Persist to json using the configured export conventions.
        
        Args:
            filepath (str): Filesystem location used by this operation."""
        with open(filepath, 'w') as f:
            json.dump(asdict(self), f, indent=4)

    @classmethod
    def from_dict(cls, data: Dict[str, Any]) -> "AppConfig":
        """Normalize a config mapping into AppConfig."""
        if not isinstance(data, dict):
            raise ValueError("Config JSON root must be an object.")
        data = dict(data)
        valid_keys = set(cls.__dataclass_fields__)
        unknown_keys = sorted(set(data.keys()) - valid_keys)
        if unknown_keys:
            raise ValueError(f"Unknown config key(s): {', '.join(unknown_keys)}")
        return cls(**data)

    @classmethod
    def load_from_json(cls, filepath: str) -> "AppConfig":
        try:
            with open(filepath, "r", encoding="utf-8") as f:
                data = json.load(f)
            return cls.from_dict(data)
        except FileNotFoundError:
            logger.info("Config file not found: %s. Using defaults.", filepath)
        return cls()


@dataclass
class SampleProperties:
    """Per-loaded-sample physical calibration used by analysis and UI."""

    pixel_size_x: float = 1.0
    pixel_size_y: float = 1.0
    voxel_size_z: float = 1.0
    unit: str = "px"
    source: str = "default"
    source_path: str = ""

    @staticmethod
    def _as_positive(value: Any, default: float = 1.0) -> float:
        if isinstance(value, (tuple, list)) and len(value) == 2:
            try:
                value = float(value[0]) / float(value[1])
            except (TypeError, ValueError, ZeroDivisionError):
                return float(default)
        try:
            out = float(value)
        except (TypeError, ValueError):
            return float(default)
        if out <= 0 or not np.isfinite(out):
            return float(default)
        return out

    @classmethod
    def defaults(cls) -> "SampleProperties":
        return cls(
            pixel_size_x=1.0,
            pixel_size_y=1.0,
            voxel_size_z=1.0,
            unit="px",
            source="default",
            source_path="",
        )

    @classmethod
    def from_image_metadata(
        cls,
        metadata: Optional[Dict[str, Any]],
        *,
        source_path: str = "",
    ) -> "SampleProperties":
        meta = metadata or {}
        px_raw = meta.get("pixel_width")
        py_raw = meta.get("pixel_height")
        pz_raw = meta.get("voxel_depth")
        unit_raw = meta.get("unit")

        has_px = px_raw is not None and cls._as_positive(px_raw, default=-1.0) > 0
        has_py = py_raw is not None and cls._as_positive(py_raw, default=-1.0) > 0
        has_pz = pz_raw is not None and cls._as_positive(pz_raw, default=-1.0) > 0

        px = cls._as_positive(px_raw) if has_px else None
        py = cls._as_positive(py_raw) if has_py else None
        pz = cls._as_positive(pz_raw) if has_pz else 1.0

        if px is None and py is None:
            px = 1.0
            py = 1.0
        elif px is None:
            px = py
        elif py is None:
            py = px

        unit_txt = str(unit_raw).strip() if unit_raw is not None else ""
        unit = unit_txt if unit_txt else "px"

        has_metadata = bool(has_px or has_py or has_pz or unit_txt)
        return cls(
            pixel_size_x=float(px),
            pixel_size_y=float(py),
            voxel_size_z=float(pz),
            unit=unit,
            source="metadata" if has_metadata else "default",
            source_path=str(source_path or meta.get("source_path") or ""),
        )

    def effective_pixel_size_xy(self) -> float:
        x = self._as_positive(self.pixel_size_x)
        y = self._as_positive(self.pixel_size_y)
        return 0.5 * (x + y)

    def spacing_zyx(self) -> tuple[float, float, float]:
        return (
            self._as_positive(self.voxel_size_z),
            self._as_positive(self.pixel_size_y),
            self._as_positive(self.pixel_size_x),
        )

@dataclass
class ImageBatch:
    """Container for an ordered batch of frames plus lightweight metadata."""
    frames: List[np.ndarray]
    metadata: Dict[str, Any] = field(default_factory=dict)

    def __len__(self) -> int:
        """Report how many items are currently tracked by this container.
        
        Returns:
            int: Number of items currently available in the container."""
        return len(self.frames)

    @property
    def kind(self) -> str:
        """Returns the kind of loaded info: 'Image', 'Folder', or 'Stack'."""
        return self.metadata.get("source_type", "Image")

    @property
    def name(self) -> str:
        """Returns the base name of the file or folder."""
        path = self.metadata.get("source_path", "")
        if not path:
            return "Unknown"
        return os.path.basename(os.path.normpath(path))

    @staticmethod
    def _extract_voxel_sizes(path: str) -> Dict[str, Any]:
        """Read physical calibration from TIFF tags and common metadata keys."""
        sizes = {"pixel_width": None, "pixel_height": None, "voxel_depth": None, "unit": None}

        def _positive(value: Any) -> Optional[float]:
            if value is None:
                return None
            if isinstance(value, (tuple, list)) and len(value) == 2:
                try:
                    value = float(value[0]) / float(value[1])
                except (TypeError, ValueError, ZeroDivisionError):
                    return None
            try:
                out = float(value)
            except (TypeError, ValueError):
                return None
            if out <= 0 or not np.isfinite(out):
                return None
            return out

        def _unit_scale_to_um(unit_value: Any) -> Optional[float]:
            if unit_value is None:
                return None
            token = str(getattr(unit_value, "name", unit_value)).strip().lower()
            if token in {"inch", "in", "inches"}:
                return 25400.0
            if token in {"centimeter", "centimetre", "cm"}:
                return 10000.0
            if token in {"millimeter", "millimetre", "mm"}:
                return 1000.0
            if token in {"resunit.inch", "resunit.centimeter", "resunit.centimetre"}:
                if "inch" in token:
                    return 25400.0
                return 10000.0
            try:
                token_num = int(unit_value)
            except (TypeError, ValueError):
                return None
            if token_num == 2:
                return 25400.0
            if token_num == 3:
                return 10000.0
            return None

        try:
            meta = iio.immeta(path)
        except Exception:
            meta = {}

        sizes["voxel_depth"] = (
            _positive(meta.get("spacing"))
            or _positive(meta.get("PhysicalSizeZ"))
        )
        sizes["unit"] = str(
            meta.get("unit")
            or meta.get("PhysicalSizeXUnit")
            or meta.get("PhysicalSizeYUnit")
            or meta.get("PhysicalSizeZUnit")
            or ""
        ).strip() or None

        sizes["pixel_width"] = (
            _positive(meta.get("pixel_width"))
            or _positive(meta.get("PhysicalSizeX"))
        )
        sizes["pixel_height"] = (
            _positive(meta.get("pixel_height"))
            or _positive(meta.get("PhysicalSizeY"))
        )

        if (
            (sizes["pixel_width"] is None or sizes["pixel_height"] is None)
            and str(path).lower().endswith((".tif", ".tiff"))
        ):
            with TiffFile(path) as tif:
                tags = tif.pages[0].tags
                x_res = _positive(tags["XResolution"].value if "XResolution" in tags else None)
                y_res = _positive(tags["YResolution"].value if "YResolution" in tags else None)
                res_unit = tags["ResolutionUnit"].value if "ResolutionUnit" in tags else None
                unit_scale = _unit_scale_to_um(res_unit)

                if sizes["pixel_width"] is None and x_res is not None:
                    sizes["pixel_width"] = (1.0 / x_res) * (unit_scale or 1.0)
                if sizes["pixel_height"] is None and y_res is not None:
                    sizes["pixel_height"] = (1.0 / y_res) * (unit_scale or 1.0)
                if sizes["unit"] is None and unit_scale is not None:
                    sizes["unit"] = "um"

        # Minimal fallback for non-TIFF metadata formats.
        if sizes["pixel_width"] is None or sizes["pixel_height"] is None:
            res = meta.get("resolution")
            rx = None
            ry = None
            if isinstance(res, (tuple, list)) and len(res) >= 2:
                rx = _positive(res[0])
                ry = _positive(res[1])
            elif isinstance(res, dict):
                rx = _positive(res.get("x") or res.get("X"))
                ry = _positive(res.get("y") or res.get("Y"))

            unit_scale = _unit_scale_to_um(
                meta.get("resolution_unit")
                or meta.get("ResolutionUnit")
                or meta.get("resolutionunit")
            )
            if sizes["pixel_width"] is None and rx is not None:
                sizes["pixel_width"] = (1.0 / rx) * (unit_scale or 1.0)
            if sizes["pixel_height"] is None and ry is not None:
                sizes["pixel_height"] = (1.0 / ry) * (unit_scale or 1.0)
            if sizes["unit"] is None and unit_scale is not None:
                sizes["unit"] = "um"

        return sizes

    @staticmethod
    def _normalize_stack(data: np.ndarray) -> List[np.ndarray]:
        if data.ndim == 2:
            return [data]
        if data.ndim == 3:
            if data.shape[2] in (3, 4):
                return [data]
            return [data[z] for z in range(data.shape[0])]
        if data.ndim == 4:
            return [data[z] for z in range(data.shape[0])]
        raise ValueError("Unsupported image dimensions for batch loading.")

    @classmethod
    def from_path(cls, path: str) -> "ImageBatch":
        data = iio.imread(path)
        frames = cls._normalize_stack(data)
        source_type = "Image" if len(frames) == 1 else "Stack"
        filenames = [os.path.basename(path)] * len(frames)
        
        # --- Extract physical sizes ---
        voxel_sizes = cls._extract_voxel_sizes(path)
        
        metadata = {
            "source_type": source_type,
            "source_path": path,
            "filenames": filenames,
            "z_indices": list(range(len(frames))),
            **voxel_sizes  # Merge the extracted sizes into the metadata dictionary
        }
        return cls(frames=frames, metadata=metadata)

    @classmethod
    def from_folder(
        cls, folder: str, extensions: Optional[Iterable[str]] = None
    ) -> "ImageBatch":
        entries = sorted(os.listdir(folder))
        ext_set = {ext.lower() for ext in extensions} if extensions else None
        frames: List[np.ndarray] = []
        filenames: List[str] = []
        
        first_valid_path = None

        for name in entries:
            path = os.path.join(folder, name)
            if not os.path.isfile(path):
                continue
            if ext_set is not None:
                if os.path.splitext(name)[1].lower() not in ext_set:
                    continue
            try:
                data = iio.imread(path)
                if first_valid_path is None:
                    first_valid_path = path # Save the first image to extract metadata
            except Exception:
                continue
            
            image_frames = cls._normalize_stack(data)
            for image_frame in image_frames:
                frames.append(image_frame)
                filenames.append(name)
                    
        # --- Extract physical sizes from the first image in the folder ---
        voxel_sizes = {}
        if first_valid_path:
            voxel_sizes = cls._extract_voxel_sizes(first_valid_path)
            
        metadata = {
            "source_type": "Folder",
            "source_path": folder,
            "filenames": filenames,
            "z_indices": list(range(len(frames))),
            **voxel_sizes
        }
        return cls(frames=frames, metadata=metadata)

    @classmethod
    def from_source(
        cls, path: str, extensions: Optional[Iterable[str]] = None
    ) -> "ImageBatch":
        if os.path.isdir(path):
            return cls.from_folder(path, extensions=extensions)
        if os.path.isfile(path):
            return cls.from_path(path)
        raise FileNotFoundError(f"Source path does not exist: {path}")
    
@dataclass
class BatchResult:
    """Capture outputs and metadata produced by batch."""
    frame_results: List[Frame]

    def frame_at(self, index: int) -> Frame:
        return self.frame_results[index]

    def save_all(self, output_dir: str, export_config: Dict[str, Any] | None = None) -> Dict[str, Any]:
        """Persist all using the configured export conventions.
        
        Args:
            output_dir (str): Filesystem location used by this operation.
            export_config (Dict[str, Any] | None): Configuration values controlling runtime behavior.
        
        Returns:
            Dict[str, Any]: Mapping of computed keys and values."""
        from .io_export import BatchResultExporter

        exporter = BatchResultExporter(export_config=export_config)
        return exporter.save_all(self, output_dir, export_config=export_config)

@dataclass
class Frame:
    """Flat, stage-aware container for a single frame's processing state."""

    # Identity
    z_index: int = 0
    frame_id: str = ""

    # Stage flags
    has_prediction: bool = False
    has_segmentation: bool = False
    has_measurements: bool = False
    prediction_sig: Optional[str] = None
    segmentation_sig: Optional[str] = None
    measurement_sig: Optional[str] = None
    error_message: Optional[str] = None

    # Source and prediction buffers
    orig_img: Optional[np.ndarray] = None
    orig_img_scaled: Optional[np.ndarray] = None
    scale_factor: float = 1.0
    dt: Optional[np.ndarray] = None
    dt_u8: Optional[np.ndarray] = None
    dt_normalized: Optional[np.ndarray] = None
    foreground_prob: Optional[np.ndarray] = None
    lapl_img: Optional[np.ndarray] = None
    gx: Optional[np.ndarray] = None
    gy: Optional[np.ndarray] = None

    # Segmentation products
    labels_2d: Optional[np.ndarray] = None
    # Fragments that own a label; never modified after segmentation, so snapshots share them.
    fragments: Optional[List[FiberFragment]] = None

    # Measurement outputs
    samples_table: Optional[Any] = None
    fiber_summary: Optional[Dict[str, Any]] = None
    measurement_samples: Optional[Any] = None
    drawing_samples: Optional[Any] = None
    thickness_units: Optional[np.ndarray] = None
    porosity: Optional[float] = None
    pore_diams_units: Optional[np.ndarray] = None
    void_mask: Optional[np.ndarray] = None
    pore_dt: Optional[np.ndarray] = None
    pores: Optional[Any] = None

    # Images built by `render.render`, keyed by view; emptied whenever any stage is cleared.
    views: Dict[str, Optional[np.ndarray]] = field(default_factory=dict, repr=False)

    @classmethod
    def empty(
        cls,
        *,
        frame_id: str = "",
        z_index: int = 0,
        orig_img: Optional[np.ndarray] = None,
    ) -> "Frame":
        return cls(frame_id=str(frame_id), z_index=int(z_index), orig_img=orig_img)

    def clear_prediction(self) -> None:
        self.has_prediction = False
        self.dt = None
        self.dt_u8 = None
        self.dt_normalized = None
        self.foreground_prob = None
        self.lapl_img = None
        self.gx = None
        self.gy = None
        self.orig_img_scaled = None
        self.scale_factor = 1.0
        self.prediction_sig = None
        self.error_message = None
        self.clear_segmentation()

    def clear_segmentation(self) -> None:
        self.has_segmentation = False
        self.labels_2d = None
        self.fragments = None
        self.segmentation_sig = None
        self.clear_measurements()

    def clear_measurements(self) -> None:
        self.has_measurements = False
        self.samples_table = None
        self.fiber_summary = None
        self.measurement_samples = None
        self.drawing_samples = None
        self.thickness_units = None
        self.porosity = None
        self.pore_diams_units = None
        self.void_mask = None
        self.pore_dt = None
        self.pores = None
        self.measurement_sig = None
        self.views = {}


    def set_prediction_outputs(
        self,
        *,
        dt: Optional[np.ndarray],
        gx: Optional[np.ndarray],
        gy: Optional[np.ndarray],
        dt_u8: Optional[np.ndarray],
        dt_normalized: Optional[np.ndarray],
        foreground_prob: Optional[np.ndarray],
        lapl_img: Optional[np.ndarray],
        orig_img_scaled: Optional[np.ndarray],
        scale_factor: float,
    ) -> None:
        self.dt = dt
        self.gx = gx
        self.gy = gy
        self.dt_u8 = dt_u8
        self.dt_normalized = dt_normalized
        self.foreground_prob = foreground_prob
        self.lapl_img = lapl_img
        self.orig_img_scaled = orig_img_scaled
        self.scale_factor = float(scale_factor)
        self.has_prediction = self.dt is not None
        self.error_message = None
        self.clear_segmentation()

    def set_segmentation_outputs(
        self,
        *,
        labels_2d: np.ndarray,
        fragments: List[FiberFragment],
    ) -> None:
        self.labels_2d = labels_2d
        self.fragments = fragments
        self.has_segmentation = True
        self.error_message = None
        self.clear_measurements()

    def set_measurement_outputs(
        self,
        *,
        samples_table: Optional[Any],
        fiber_summary: Optional[Dict[str, Any]],
        measurement_samples: Optional[Any],
        drawing_samples: Optional[Any],
        thickness_units: Optional[np.ndarray],
        porosity: Optional[float],
        pore_diams_units: Optional[np.ndarray],
        void_mask: Optional[np.ndarray],
        pore_dt: Optional[np.ndarray],
        pores: Optional[Any],
    ) -> None:
        """Update measurement outputs and synchronize dependent runtime state."""
        self.samples_table = samples_table
        self.fiber_summary = fiber_summary
        self.measurement_samples = measurement_samples
        self.drawing_samples = drawing_samples
        self.thickness_units = thickness_units
        self.porosity = porosity
        self.pore_diams_units = pore_diams_units
        self.void_mask = void_mask
        self.pore_dt = pore_dt
        self.pores = pores
        self.has_measurements = True
        self.error_message = None

    def fail(self, message: str) -> None:
        self.error_message = str(message)

    def snapshot(
        self,
        *,
        frame_id: Optional[str] = None,
        z_index: Optional[int] = None,
    ) -> "Frame":
        clone: Frame = copy(self)
        def _clone_array(value: Any) -> Any:
            if isinstance(value, np.ndarray):
                return value.copy()
            return value

        clone.dt = _clone_array(clone.dt)
        clone.dt_u8 = _clone_array(clone.dt_u8)
        clone.dt_normalized = _clone_array(clone.dt_normalized)
        clone.foreground_prob = _clone_array(clone.foreground_prob)
        clone.lapl_img = _clone_array(clone.lapl_img)
        clone.gx = _clone_array(clone.gx)
        clone.gy = _clone_array(clone.gy)

        clone.labels_2d = _clone_array(clone.labels_2d)

        clone.samples_table = _clone_array(clone.samples_table)
        clone.thickness_units = _clone_array(clone.thickness_units)
        clone.pore_diams_units = _clone_array(clone.pore_diams_units)
        clone.void_mask = _clone_array(clone.void_mask)
        clone.pore_dt = _clone_array(clone.pore_dt)

        # Cached images are never modified in place, so the clone may share them.
        clone.views = dict(clone.views)
        if clone.fiber_summary is not None:
            clone.fiber_summary = dict(clone.fiber_summary)
        if frame_id is not None:
            clone.frame_id = str(frame_id)
        if z_index is not None:
            clone.z_index = int(z_index)
        return clone


@dataclass
class FiberFragment:
    """Unified 2D fragment carrying both segmentation and measurement centerlines."""

    fragment_id: int
    segment_centerlines_xy: list[np.ndarray]
    segment_radii: Optional[list[np.ndarray]] = None
    parent_fiber_id: int = -1
    measurement_centerlines_xy: list[np.ndarray] = field(default_factory=list)
    total_measurement_centerline_length_px: float = 0.0
    grad_quality_score: float = float("nan")

    @staticmethod
    def _normalize_centerline(points_raw: np.ndarray, label: str) -> np.ndarray:
        points_xy = np.asarray(points_raw, dtype=np.float32)
        if points_xy.ndim != 2 or points_xy.shape[1] != 2:
            raise ValueError(f"{label} must be an array of shape (N, 2).")
        if points_xy.shape[0] < 2:
            raise ValueError(f"{label} must contain at least two points.")
        return points_xy

    @staticmethod
    def _normalize_radii(radii_raw: Optional[np.ndarray], n_points: int, label: str) -> np.ndarray:
        if radii_raw is None:
            return np.zeros(n_points, dtype=np.float32)
        radii = np.asarray(radii_raw, dtype=np.float32).reshape(-1)
        if radii.shape[0] != n_points:
            raise ValueError(f"{label} length must match centerline length.")
        return radii

    @staticmethod
    def _centerline_length(points_xy: np.ndarray) -> float:
        if points_xy.shape[0] < 2:
            return 0.0
        return float(np.linalg.norm(np.diff(points_xy, axis=0), axis=1).sum())

    def recompute_total_measurement_centerline_length_px(self) -> float:
        self.total_measurement_centerline_length_px = float(
            sum(self._centerline_length(points_xy) for points_xy in self.measurement_centerlines_xy)
        )
        return self.total_measurement_centerline_length_px

    def __post_init__(self) -> None:
        if not self.segment_centerlines_xy:
            raise ValueError("FiberFragment must contain at least one segmentation centerline.")

        centerlines_xy = [
            self._normalize_centerline(points_raw, "Segmentation centerline")
            for points_raw in self.segment_centerlines_xy
        ]
        if not self.segment_radii:
            centerlines_radii = [
                np.zeros(points_xy.shape[0], dtype=np.float32) for points_xy in centerlines_xy
            ]
        else:
            if len(self.segment_radii) != len(centerlines_xy):
                raise ValueError("segment_radii length must match segment_centerlines_xy length.")
            centerlines_radii = [
                self._normalize_radii(radii_raw, points_xy.shape[0], "Segmentation radii")
                for points_xy, radii_raw in zip(centerlines_xy, self.segment_radii)
            ]

        pairs = list(zip(centerlines_xy, centerlines_radii))
        pairs.sort(key=lambda pair: self._centerline_length(pair[0]), reverse=True)
        self.segment_centerlines_xy = [points_xy for points_xy, _ in pairs]
        self.segment_radii = [radii for _, radii in pairs]

        self.measurement_centerlines_xy = [
            self._normalize_centerline(points_raw, "Measurement centerline")
            for points_raw in (self.measurement_centerlines_xy or [])
        ]
        self.recompute_total_measurement_centerline_length_px()

        self.fragment_id = int(self.fragment_id)
        self.parent_fiber_id = int(self.parent_fiber_id)

    @property
    def has_measurement_centerline(self) -> bool:
        return bool(self.measurement_centerlines_xy)