File size: 29,838 Bytes
3193174
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
"""
Service layer over rustworkx algorithms for graph analysis.

Provides:
- K shortest paths
- Centralities (betweenness, closeness, degree, eigenvector, PageRank)
- Community detection
- Cycle detection
- Subgraph filtering by metadata
- Integration with the router
"""

import contextlib
from collections import deque
from collections.abc import Callable
from enum import Enum
from typing import Any

import rustworkx as rx
import torch
from pydantic import BaseModel, ConfigDict, Field

__all__ = [
    "CentralityResult",
    # Enums
    "CentralityType",
    "CommunityResult",
    "CycleInfo",
    # Main service
    "GraphAlgorithms",
    "PathMetric",
    # Data classes
    "PathResult",
    "SubgraphFilter",
    # Utility functions
    "compute_all_centralities",
    "find_critical_nodes",
    "get_graph_metrics",
]


class CentralityType(str, Enum):
    """Centrality types."""

    BETWEENNESS = "betweenness"
    CLOSENESS = "closeness"
    DEGREE = "degree"
    EIGENVECTOR = "eigenvector"
    PAGERANK = "pagerank"
    KATZ = "katz"


class PathMetric(str, Enum):
    """Metric for path computation."""

    HOPS = "hops"
    WEIGHT = "weight"
    LATENCY = "latency"
    COST = "cost"
    RELIABILITY = "reliability"


class PathResult(BaseModel):
    """Description of a found path with weights and arbitrary metadata."""

    nodes: list[str]
    total_weight: float
    edge_weights: list[float]
    metadata: dict[str, Any] = Field(default_factory=dict)

    @property
    def length(self) -> int:
        """Number of edges in the path."""
        return len(self.nodes) - 1 if len(self.nodes) > 1 else 0

    def __repr__(self) -> str:
        return f"PathResult({' -> '.join(self.nodes)}, weight={self.total_weight:.3f})"


class CentralityResult(BaseModel):
    """Result of centrality computation for graph nodes."""

    centrality_type: CentralityType
    values: dict[str, float]
    normalized: bool = True

    def top_k(self, k: int = 5) -> list[tuple[str, float]]:
        """Return the top-k nodes by centrality value."""
        sorted_items = sorted(self.values.items(), key=lambda x: x[1], reverse=True)
        return sorted_items[:k]

    def get_node_rank(self, node_id: str) -> int | None:
        """Node position in the ranking (1-based) or None if absent."""
        sorted_nodes = sorted(self.values.keys(), key=lambda n: self.values[n], reverse=True)
        try:
            return sorted_nodes.index(node_id) + 1
        except ValueError:
            return None


class CommunityResult(BaseModel):
    """Result of community detection."""

    communities: list[set[str]]
    modularity: float | None = None
    algorithm: str = "unknown"

    @property
    def num_communities(self) -> int:
        """Number of detected communities."""
        return len(self.communities)

    def get_node_community(self, node_id: str) -> int | None:
        """Find the community index that the node belongs to."""
        for i, community in enumerate(self.communities):
            if node_id in community:
                return i
        return None

    def get_community_sizes(self) -> list[int]:
        """Return a list of community sizes."""
        return [len(c) for c in self.communities]


class CycleInfo(BaseModel):
    """Information about a detected cycle."""

    nodes: list[str]
    edges: list[tuple[str, str]]
    total_weight: float = 0.0

    @property
    def length(self) -> int:
        """Number of nodes in the cycle."""
        return len(self.nodes)


class SubgraphFilter(BaseModel):
    """Rules for filtering nodes and edges when building a subgraph."""

    model_config = ConfigDict(arbitrary_types_allowed=True)

    node_filter: Callable[[str, dict[str, Any]], bool] | None = None
    edge_filter: Callable[[str, str, dict[str, Any]], bool] | None = None
    include_nodes: set[str] | None = None
    exclude_nodes: set[str] | None = None
    min_weight: float | None = None
    max_weight: float | None = None
    required_attrs: list[str] | None = None

    def matches_node(self, node_id: str, attrs: dict[str, Any]) -> bool:
        """Check whether the node satisfies the given conditions."""
        if self.exclude_nodes and node_id in self.exclude_nodes:
            return False
        if self.include_nodes and node_id not in self.include_nodes:
            return False
        if self.required_attrs and not all(attr in attrs for attr in self.required_attrs):
            return False
        return not (self.node_filter and not self.node_filter(node_id, attrs))

    def matches_edge(self, src: str, tgt: str, attrs: dict[str, Any]) -> bool:
        """Check whether the edge satisfies the given conditions."""
        weight = attrs.get("weight", 1.0)
        if self.min_weight is not None and weight < self.min_weight:
            return False
        if self.max_weight is not None and weight > self.max_weight:
            return False
        return not (self.edge_filter and not self.edge_filter(src, tgt, attrs))


class GraphAlgorithms:
    """Service layer over `rustworkx` for analysing a `RoleGraph`."""

    def __init__(
        self,
        graph: Any,  # RoleGraph
        weight_attr: str = "weight",
        default_weight: float = 1.0,
    ):
        """
        Initialise the wrapper around the graph.

        Args:
            graph: A RoleGraph instance (or an object with a `graph: PyDiGraph` attribute).
            weight_attr: Key for the edge weight in the data.
            default_weight: Weight value when not specified in the edge.

        """
        self._role_graph = graph
        self._graph: rx.PyDiGraph = graph.graph
        self._weight_attr = weight_attr
        self._default_weight = default_weight

        self._node_id_to_idx: dict[str, int] = {}
        self._idx_to_node_id: dict[int, str] = {}
        self._rebuild_index_cache()

    def _rebuild_index_cache(self) -> None:
        """Rebuild the cache mapping node_id ↔ rustworkx index."""
        self._node_id_to_idx.clear()
        self._idx_to_node_id.clear()
        for idx in self._graph.node_indices():
            data = self._graph.get_node_data(idx)
            if isinstance(data, dict):
                node_id = data.get("id", str(idx))
            elif hasattr(data, "agent_id"):
                node_id = data.agent_id
            else:
                node_id = str(idx)
            self._node_id_to_idx[node_id] = idx
            self._idx_to_node_id[idx] = node_id

    def _get_node_idx(self, node_id: str) -> int:
        """Get the node index by ID."""
        if node_id not in self._node_id_to_idx:
            self._rebuild_index_cache()
        if node_id not in self._node_id_to_idx:
            msg = f"Node '{node_id}' not found in graph"
            raise ValueError(msg)
        return self._node_id_to_idx[node_id]

    def _get_node_id(self, idx: int) -> str:
        """Get the node ID by its internal graph index."""
        if idx not in self._idx_to_node_id:
            self._rebuild_index_cache()
        return self._idx_to_node_id.get(idx, str(idx))

    def _get_edge_weight(
        self,
        edge_data: Any,
        metric: PathMetric = PathMetric.WEIGHT,
    ) -> float:
        """Get the edge weight according to the selected metric."""
        if edge_data is None:
            return self._default_weight

        if isinstance(edge_data, dict):
            if metric == PathMetric.HOPS:
                return 1.0
            if metric == PathMetric.WEIGHT:
                return edge_data.get(self._weight_attr, self._default_weight)
            if metric == PathMetric.LATENCY:
                return edge_data.get("latency", self._default_weight)
            if metric == PathMetric.COST:
                return edge_data.get("cost", self._default_weight)
            if metric == PathMetric.RELIABILITY:
                rel = edge_data.get("reliability", 1.0)
                return -torch.log(torch.tensor(max(rel, 1e-10))).item()

        return self._default_weight

    def k_shortest_paths(
        self,
        source: str,
        target: str,
        k: int = 3,
        metric: PathMetric = PathMetric.WEIGHT,
    ) -> list[PathResult]:
        """Find k shortest paths between nodes using the given metric."""
        src_idx = self._get_node_idx(source)
        tgt_idx = self._get_node_idx(target)

        def weight_fn(edge_data: Any) -> float:
            return self._get_edge_weight(edge_data, metric)

        paths = self._yen_k_shortest_paths(src_idx, tgt_idx, k, weight_fn)

        results = []
        for path_indices, total_weight in paths:
            nodes = [self._get_node_id(idx) for idx in path_indices]
            edge_weights = []
            for i in range(len(path_indices) - 1):
                edge_data = self._graph.get_edge_data(path_indices[i], path_indices[i + 1])
                edge_weights.append(self._get_edge_weight(edge_data, metric))

            results.append(
                PathResult(
                    nodes=nodes,
                    total_weight=total_weight,
                    edge_weights=edge_weights,
                    metadata={"metric": metric.value},
                )
            )

        return results

    def _find_initial_shortest_path(
        self,
        source: int,
        target: int,
        weight_fn: Callable[[Any], float],
    ) -> tuple[list[int], float] | None:
        """Find the first shortest path."""
        try:
            distances = rx.dijkstra_shortest_path_lengths(self._graph, source, weight_fn)
            if target not in distances:
                return None

            path_map = rx.dijkstra_shortest_paths(self._graph, source, target=target, weight_fn=weight_fn)
            if target not in path_map:
                return None

            return list(path_map[target]), distances[target]
        except (ValueError, KeyError, RuntimeError):
            return None

    def _remove_conflicting_edges(
        self,
        found_paths: list[tuple[list[int], float]],
        root_path: list[int],
        j: int,
    ) -> list[tuple[int, int, Any]]:
        """Remove edges that conflict with the found paths."""
        removed_edges = []
        for path, _ in found_paths:
            if len(path) > j and path[: j + 1] == root_path and j + 1 < len(path):
                try:
                    edge_data = self._graph.get_edge_data(path[j], path[j + 1])
                    self._graph.remove_edge(path[j], path[j + 1])
                    removed_edges.append((path[j], path[j + 1], edge_data))
                except (ValueError, KeyError, RuntimeError):
                    pass
        return removed_edges

    def _calculate_path_weight(self, path: list[int], weight_fn: Callable[[Any], float]) -> float:
        """Compute the total weight of a path."""
        total_weight = 0.0
        for idx in range(len(path) - 1):
            edge_data = self._graph.get_edge_data(path[idx], path[idx + 1])
            total_weight += weight_fn(edge_data) if edge_data else self._default_weight
        return total_weight

    def _find_spur_path(
        self,
        spur_node: int,
        target: int,
        root_path: list[int],
        weight_fn: Callable[[Any], float],
    ) -> tuple[list[int], float] | None:
        """Find an alternative path from the spur node."""
        try:
            spur_distances = rx.dijkstra_shortest_path_lengths(self._graph, spur_node, weight_fn)
            if target not in spur_distances:
                return None

            spur_path_map = rx.dijkstra_shortest_paths(self._graph, spur_node, target=target, weight_fn=weight_fn)
            if target not in spur_path_map:
                return None
        except (ValueError, KeyError, RuntimeError):
            return None
        else:
            spur_path = list(spur_path_map[target])
            total_path = root_path[:-1] + spur_path
            total_weight = self._calculate_path_weight(total_path, weight_fn)
            return total_path, total_weight

    def _yen_k_shortest_paths(
        self,
        source: int,
        target: int,
        k: int,
        weight_fn: Callable[[Any], float],
    ) -> list[tuple[list[int], float]]:
        """Yen's algorithm: return paths as lists of node indices and total weight."""
        import heapq

        initial_path = self._find_initial_shortest_path(source, target, weight_fn)
        if not initial_path:
            return []

        first_path, first_weight = initial_path
        found_paths = [(first_path, first_weight)]
        candidate_heap: list[tuple[float, list[int]]] = []

        for i in range(1, k):
            if i - 1 >= len(found_paths):
                break

            prev_path, _ = found_paths[i - 1]

            for j in range(len(prev_path) - 1):
                spur_node = prev_path[j]
                root_path = prev_path[: j + 1]

                removed_edges = self._remove_conflicting_edges(found_paths, root_path, j)

                try:
                    spur_result = self._find_spur_path(spur_node, target, root_path, weight_fn)
                    if spur_result:
                        total_path, total_weight = spur_result
                        if not any(p == total_path for _, p in candidate_heap) and not any(
                            p == total_path for p, _ in found_paths
                        ):
                            heapq.heappush(candidate_heap, (total_weight, total_path))
                finally:
                    # Restore removed edges in any case
                    for u, v, data in removed_edges:
                        self._graph.add_edge(u, v, data)

            if candidate_heap:
                weight, path = heapq.heappop(candidate_heap)
                found_paths.append((path, weight))

        return found_paths

    def shortest_path(
        self,
        source: str,
        target: str,
        metric: PathMetric = PathMetric.WEIGHT,
    ) -> PathResult | None:
        """Find one shortest path between two nodes."""
        paths = self.k_shortest_paths(source, target, k=1, metric=metric)
        return paths[0] if paths else None

    def all_pairs_shortest_paths(
        self,
        metric: PathMetric = PathMetric.WEIGHT,
    ) -> dict[str, dict[str, float]]:
        """Compute shortest paths between all pairs of nodes."""

        def weight_fn(edge_data: Any) -> float:
            return self._get_edge_weight(edge_data, metric)

        all_distances = rx.all_pairs_dijkstra_path_lengths(self._graph, weight_fn)

        result = {}
        for src_idx, distances in all_distances.items():
            src_id = self._get_node_id(src_idx)
            result[src_id] = {}
            for tgt_idx, dist in distances.items():
                tgt_id = self._get_node_id(tgt_idx)
                result[src_id][tgt_id] = dist

        return result

    def compute_centrality(
        self,
        centrality_type: CentralityType,
        normalized: bool = True,
        **kwargs: Any,
    ) -> CentralityResult:
        """Compute the selected centrality type for all graph nodes."""
        values: dict[int, int | float] = {}

        if centrality_type == CentralityType.BETWEENNESS:
            raw_result = rx.betweenness_centrality(self._graph, normalized=normalized)
            values = (
                dict(raw_result.items())
                if hasattr(raw_result, "items")
                else dict(enumerate(raw_result))
                if isinstance(raw_result, list)
                else raw_result
            )

        elif centrality_type == CentralityType.CLOSENESS:
            undirected = self._graph.to_undirected()
            raw_values = rx.closeness_centrality(undirected)
            values = (
                dict(enumerate(raw_values))
                if isinstance(raw_values, list)
                else dict(raw_values.items())
                if hasattr(raw_values, "items")
                else raw_values
            )

        elif centrality_type == CentralityType.DEGREE:
            for idx in self._graph.node_indices():
                in_deg = self._graph.in_degree(idx)
                out_deg = self._graph.out_degree(idx)
                values[idx] = float(in_deg + out_deg)
            if normalized and self._graph.num_nodes() > 1:
                max_deg = 2 * (self._graph.num_nodes() - 1)
                values = {k: v / max_deg for k, v in values.items()}

        elif centrality_type == CentralityType.EIGENVECTOR:
            try:
                raw = rx.eigenvector_centrality(self._graph)
                values = (
                    dict(enumerate(raw))
                    if isinstance(raw, list)
                    else dict(raw.items())
                    if hasattr(raw, "items")
                    else raw
                )
            except (ValueError, RuntimeError, AttributeError):
                raw_pr = rx.pagerank(self._graph)
                values = (
                    dict(raw_pr.items())
                    if hasattr(raw_pr, "items")
                    else dict(enumerate(raw_pr))
                    if isinstance(raw_pr, list)
                    else raw_pr
                )

        elif centrality_type == CentralityType.PAGERANK:
            alpha = kwargs.get("alpha", 0.85)
            raw_pr = rx.pagerank(self._graph, alpha=alpha)
            values = (
                dict(raw_pr.items())
                if hasattr(raw_pr, "items")
                else dict(enumerate(raw_pr))
                if isinstance(raw_pr, list)
                else raw_pr
            )

        elif centrality_type == CentralityType.KATZ:
            alpha = kwargs.get("alpha", 0.1)
            beta = kwargs.get("beta", 1.0)
            try:
                raw_katz = rx.katz_centrality(self._graph, alpha=alpha, beta=beta)
                values = (
                    dict(raw_katz.items())
                    if hasattr(raw_katz, "items")
                    else dict(enumerate(raw_katz))
                    if isinstance(raw_katz, list)
                    else raw_katz
                )
            except (ValueError, RuntimeError, AttributeError):
                raw_pr = rx.pagerank(self._graph)
                values = (
                    dict(raw_pr.items())
                    if hasattr(raw_pr, "items")
                    else dict(enumerate(raw_pr))
                    if isinstance(raw_pr, list)
                    else raw_pr
                )

        result_values = {}
        for idx, val in values.items():
            node_id = self._get_node_id(idx)
            result_values[node_id] = float(val)

        return CentralityResult(
            centrality_type=centrality_type,
            values=result_values,
            normalized=normalized,
        )

    def compute_all_centralities(self, normalized: bool = True) -> dict[CentralityType, CentralityResult]:
        """Compute all centralities and return a dict keyed by type."""
        results = {}
        for ct in CentralityType:
            with contextlib.suppress(Exception):
                results[ct] = self.compute_centrality(ct, normalized=normalized)
        return results

    def detect_communities(
        self,
        algorithm: str = "louvain",
        _resolution: float = 1.0,
    ) -> CommunityResult:
        """Detect communities using the specified algorithm (louvain/label_propagation)."""
        undirected = self._graph.to_undirected()

        communities: list[set[str]] = []
        modularity: float | None = None

        if algorithm == "louvain":
            try:
                components = rx.connected_components(undirected)
                communities = [{self._get_node_id(idx) for idx in comp} for comp in components]
            except (ValueError, RuntimeError):
                communities = [{self._get_node_id(idx) for idx in undirected.node_indices()}]

        elif algorithm == "label_propagation":
            communities = self._label_propagation(undirected)

        elif algorithm == "connected_components":
            components = rx.connected_components(undirected)
            communities = [{self._get_node_id(idx) for idx in comp} for comp in components]

        else:
            components = rx.connected_components(undirected)
            communities = [{self._get_node_id(idx) for idx in comp} for comp in components]

        return CommunityResult(
            communities=communities,
            modularity=modularity,
            algorithm=algorithm,
        )

    def _label_propagation(self, graph: rx.PyGraph) -> list[set[str]]:
        """Simple label propagation implementation for an undirected graph."""
        import random

        labels = {idx: idx for idx in graph.node_indices()}

        for _ in range(100):
            changed = False
            nodes = list(graph.node_indices())
            random.shuffle(nodes)

            for node in nodes:
                neighbors = list(graph.neighbors(node))
                if not neighbors:
                    continue

                label_counts: dict[int, int] = {}
                for neighbor in neighbors:
                    lbl = labels[neighbor]
                    label_counts[lbl] = label_counts.get(lbl, 0) + 1

                max_count = max(label_counts.values())
                best_labels = [label for label, count in label_counts.items() if count == max_count]
                # Use first label if only one, otherwise pick randomly (non-cryptographic use)
                new_label = best_labels[0] if len(best_labels) == 1 else random.choice(best_labels)

                if labels[node] != new_label:
                    labels[node] = new_label
                    changed = True

            if not changed:
                break

        label_to_nodes: dict[int, set[str]] = {}
        for node, label in labels.items():
            if label not in label_to_nodes:
                label_to_nodes[label] = set()
            label_to_nodes[label].add(self._get_node_id(node))

        return list(label_to_nodes.values())

    def detect_cycles(self, max_length: int | None = None) -> list[CycleInfo]:
        """Find simple cycles, optionally limiting the maximum length."""
        cycles = []

        try:
            simple_cycles = rx.simple_cycles(self._graph)
            for cycle_indices in simple_cycles:
                if max_length and len(cycle_indices) > max_length:
                    continue

                nodes = [self._get_node_id(idx) for idx in cycle_indices]
                edges = []
                total_weight = 0.0

                for i in range(len(cycle_indices)):
                    src = cycle_indices[i]
                    tgt = cycle_indices[(i + 1) % len(cycle_indices)]
                    edges.append((self._get_node_id(src), self._get_node_id(tgt)))

                    edge_data = self._graph.get_edge_data(src, tgt)
                    if edge_data and isinstance(edge_data, dict):
                        total_weight += edge_data.get(self._weight_attr, self._default_weight)
                    else:
                        total_weight += self._default_weight

                cycles.append(
                    CycleInfo(
                        nodes=nodes,
                        edges=edges,
                        total_weight=total_weight,
                    )
                )
        except (ValueError, RuntimeError):
            pass  # Cycle detection may fail

        return cycles

    def is_dag(self) -> bool:
        """Check whether the graph is a directed acyclic graph (DAG)."""
        return rx.is_directed_acyclic_graph(self._graph)

    def topological_sort(self) -> list[str] | None:
        """Return the topological ordering of nodes if the graph is a DAG."""
        if not self.is_dag():
            return None

        order = rx.topological_sort(self._graph)
        return [self._get_node_id(idx) for idx in order]

    def filter_subgraph(
        self,
        filter_spec: SubgraphFilter,
    ) -> "GraphAlgorithms":
        """Filter nodes/edges by rules and return a wrapper over the subgraph."""
        keep_nodes = set()
        for idx in self._graph.node_indices():
            node_id = self._get_node_id(idx)
            node_data = self._graph.get_node_data(idx)
            attrs = node_data if isinstance(node_data, dict) else {}

            if filter_spec.matches_node(node_id, attrs):
                keep_nodes.add(idx)

        new_graph = rx.PyDiGraph()
        old_to_new: dict[int, int] = {}

        for old_idx in keep_nodes:
            node_data = self._graph.get_node_data(old_idx)
            new_idx = new_graph.add_node(node_data)
            old_to_new[old_idx] = new_idx

        for edge_idx in self._graph.edge_indices():
            src, tgt = self._graph.get_edge_endpoints_by_index(edge_idx)
            if src not in keep_nodes or tgt not in keep_nodes:
                continue

            edge_data = self._graph.get_edge_data_by_index(edge_idx)
            attrs = edge_data if isinstance(edge_data, dict) else {}

            src_id = self._get_node_id(src)
            tgt_id = self._get_node_id(tgt)

            if filter_spec.matches_edge(src_id, tgt_id, attrs):
                new_graph.add_edge(old_to_new[src], old_to_new[tgt], edge_data)

        class SubgraphWrapper:
            def __init__(self, g: rx.PyDiGraph):
                self.graph = g

        return GraphAlgorithms(
            SubgraphWrapper(new_graph),
            weight_attr=self._weight_attr,
            default_weight=self._default_weight,
        )

    def get_reachable_nodes(self, source: str, max_depth: int | None = None) -> set[str]:
        """Return the set of nodes reachable from source, optionally limited by depth."""
        src_idx = self._get_node_idx(source)

        visited = set()
        queue = deque([(src_idx, 0)])

        while queue:
            node, depth = queue.popleft()
            if node in visited:
                continue
            if max_depth is not None and depth > max_depth:
                continue

            visited.add(node)

            successors_to_add = [
                (successor, depth + 1) for successor in self._graph.successor_indices(node) if successor not in visited
            ]
            queue.extend(successors_to_add)

        return {self._get_node_id(idx) for idx in visited}

    def get_predecessors(self, node: str, max_depth: int | None = None) -> set[str]:
        """Return the set of predecessors of a node, optionally limited by depth."""
        node_idx = self._get_node_idx(node)

        visited = set()
        queue = deque([(node_idx, 0)])

        while queue:
            n, depth = queue.popleft()
            if n in visited:
                continue
            if max_depth is not None and depth > max_depth:
                continue

            visited.add(n)

            predecessors_to_add = [
                (predecessor, depth + 1)
                for predecessor in self._graph.predecessor_indices(n)
                if predecessor not in visited
            ]
            queue.extend(predecessors_to_add)

        visited.discard(node_idx)
        return {self._get_node_id(idx) for idx in visited}

    def get_routing_metrics(self, source: str, target: str) -> dict[str, Any]:
        """Collect a brief summary of paths and centrality for a pair of nodes."""
        paths_list: list[dict[str, Any]] = []
        centrality_dict: dict[str, float] = {}
        is_reachable = False

        for metric in [PathMetric.WEIGHT, PathMetric.LATENCY, PathMetric.COST]:
            try:
                paths = self.k_shortest_paths(source, target, k=3, metric=metric)
                if paths:
                    is_reachable = True
                    paths_list.append(
                        {
                            "metric": metric.value,
                            "best_path": paths[0].nodes,
                            "best_weight": paths[0].total_weight,
                            "alternatives": len(paths) - 1,
                        }
                    )
            except (ValueError, RuntimeError) as e:
                from config.logging import logger

                logger.debug(f"Error: {e}")

        try:
            pr = self.compute_centrality(CentralityType.PAGERANK)
            centrality_dict["pagerank"] = pr.values.get(target, 0.0)
        except (ValueError, RuntimeError):
            pass  # Centrality computation may fail

        return {
            "source": source,
            "target": target,
            "paths": paths_list,
            "centrality": centrality_dict,
            "is_reachable": is_reachable,
        }


def compute_all_centralities(graph: Any) -> dict[str, CentralityResult]:
    """Compute all centrality types and return them by string keys."""
    alg = GraphAlgorithms(graph)
    results = alg.compute_all_centralities()
    return {ct.value: result for ct, result in results.items()}


def find_critical_nodes(graph: Any, top_k: int = 5) -> list[str]:
    """Return the nodes with the highest betweenness centrality."""
    alg = GraphAlgorithms(graph)
    bc = alg.compute_centrality(CentralityType.BETWEENNESS)
    return [node_id for node_id, _ in bc.top_k(top_k)]


def get_graph_metrics(graph: Any) -> dict[str, Any]:
    """Collect key graph metrics: size, DAG status, communities, cycles."""
    alg = GraphAlgorithms(graph)

    return {
        "num_nodes": graph.graph.num_nodes(),
        "num_edges": graph.graph.num_edges(),
        "is_dag": alg.is_dag(),
        "num_communities": alg.detect_communities().num_communities,
        "num_cycles": len(alg.detect_cycles(max_length=10)),
    }