__init__.py 47.3 KB
Newer Older
1
#! /usr/bin/env python
2
# -*- coding: utf-8 -*-
3
#
4
5
# graph_tool -- a general graph manipulation python module
#
Tiago Peixoto's avatar
Tiago Peixoto committed
6
# Copyright (C) 2007-2012 Tiago de Paula Peixoto <tiago@skewed.de>
7
8
9
10
11
12
13
14
15
16
17
18
19
20
#
# This program is free software: you can redistribute it and/or modify
# it under the terms of the GNU General Public License as published by
# the Free Software Foundation, either version 3 of the License, or
# (at your option) any later version.
#
# This program is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with this program.  If not, see <http://www.gnu.org/licenses/>.

21
"""
22
23
``graph_tool.topology`` - Assessing graph topology
--------------------------------------------------
24
25
26
27
28
29
30

Summary
+++++++

.. autosummary::
   :nosignatures:

31
   shortest_distance
Tiago Peixoto's avatar
Tiago Peixoto committed
32
   shortest_path
Tiago Peixoto's avatar
Tiago Peixoto committed
33
   pseudo_diameter
34
   similarity
35
   isomorphism
36
37
   subgraph_isomorphism
   mark_subgraph
38
39
   max_cardinality_matching
   max_independent_vertex_set
40
41
42
43
44
45
   min_spanning_tree
   dominator_tree
   topological_sort
   transitive_closure
   label_components
   label_biconnected_components
46
   label_largest_component
47
   is_bipartite
48
   is_planar
Tiago Peixoto's avatar
Tiago Peixoto committed
49
   edge_reciprocity
50
51
52

Contents
++++++++
53

54
55
"""

56
57
from __future__ import division, absolute_import, print_function

Tiago Peixoto's avatar
Tiago Peixoto committed
58
from .. dl_import import dl_import
59
dl_import("from . import libgraph_tool_topology")
60

61
from .. import _prop, Vector_int32_t, _check_prop_writable, \
62
     _check_prop_scalar, _check_prop_vector, Graph, PropertyMap, GraphView
63
64
from .. flow import libgraph_tool_flow
import random, sys, numpy
65
__all__ = ["isomorphism", "subgraph_isomorphism", "mark_subgraph",
66
           "max_cardinality_matching", "max_independent_vertex_set",
67
           "min_spanning_tree", "dominator_tree", "topological_sort",
68
           "transitive_closure", "label_components", "label_largest_component",
69
           "label_biconnected_components", "shortest_distance", "shortest_path",
70
           "pseudo_diameter", "is_bipartite", "is_planar", "similarity", "edge_reciprocity"]
71
72
73
74
75
76
77
78
79
80


def similarity(g1, g2, label1=None, label2=None, norm=True):
    r"""Return the adjacency similarity between the two graphs.

    Parameters
    ----------
    g1 : :class:`~graph_tool.Graph`
        First graph to be compared.
    g2 : :class:`~graph_tool.Graph`
Tiago Peixoto's avatar
Tiago Peixoto committed
81
        Second graph to be compared.
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
    label1 : :class:`~graph_tool.PropertyMap` (optional, default: ``None``)
        Vertex labels for the first graph to be used in comparison. If not
        supplied, the vertex indexes are used.
    label2 : :class:`~graph_tool.PropertyMap` (optional, default: ``None``)
        Vertex labels for the second graph to be used in comparison. If not
        supplied, the vertex indexes are used.
    norm : bool (optional, default: ``True``)
        If ``True``, the returned value is normalized by the total number of
        edges.

    Returns
    -------
    similarity : float
        Adjacency similarity value.

    Notes
    -----
    The adjacency similarity is the sum of equal entries in the adjacency
    matrix, given a vertex ordering determined by the vertex labels. In other
    words it counts the number of edges which have the same source and target
    labels in both graphs.

    The algorithm runs with complexity :math:`O(E_1 + V_1 + E_2 + V_2)`.

    Examples
    --------
    >>> from numpy.random import seed
    >>> seed(42)
    >>> g = gt.random_graph(100, lambda: (3,3))
    >>> u = g.copy()
    >>> gt.similarity(u, g)
    1.0
    >>> gt.random_rewire(u);
    >>> gt.similarity(u, g)
    0.03333333333333333
    """

    if label1 is None:
        label1 = g1.vertex_index
    if label2 is None:
        label2 = g2.vertex_index
    if label1.value_type() != label2.value_type():
        raise ValueError("label property maps must be of the same type")
    s = libgraph_tool_topology.\
           similarity(g1._Graph__graph, g2._Graph__graph,
                      _prop("v", g1, label1), _prop("v", g1, label2))
    if not g1.is_directed() or not g2.is_directed():
        s /= 2
    if norm:
        s /= float(max(g1.num_edges(), g2.num_edges()))
    return s
133

Tiago Peixoto's avatar
Tiago Peixoto committed
134

135
def isomorphism(g1, g2, isomap=False):
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
    r"""Check whether two graphs are isomorphic.

    If `isomap` is True, a vertex :class:`~graph_tool.PropertyMap` with the
    isomorphism mapping is returned as well.

    Examples
    --------
    >>> from numpy.random import seed
    >>> seed(42)
    >>> g = gt.random_graph(100, lambda: (3,3))
    >>> g2 = gt.Graph(g)
    >>> gt.isomorphism(g, g2)
    True
    >>> g.add_edge(g.vertex(0), g.vertex(1))
    <...>
    >>> gt.isomorphism(g, g2)
    False

154
    """
155
156
    imap = g1.new_vertex_property("int32_t")
    iso = libgraph_tool_topology.\
157
           check_isomorphism(g1._Graph__graph, g2._Graph__graph,
Tiago Peixoto's avatar
Tiago Peixoto committed
158
                             _prop("v", g1, imap))
159
160
161
162
163
    if isomap:
        return iso, imap
    else:
        return iso

Tiago Peixoto's avatar
Tiago Peixoto committed
164

165
def subgraph_isomorphism(sub, g, max_n=0, random=False):
166
    r"""
167
168
    Obtain all subgraph isomorphisms of `sub` in `g` (or at most `max_n`
    subgraphs, if `max_n > 0`).
169

170
171
172
    If `random` = True, the vertices of `g` are indexed in random order before
    the search.

173
174
175
176
177
178
179
180
181
182
183
    It returns two lists, containing the vertex and edge property maps for `sub`
    with the isomorphism mappings. The value of the properties are the
    vertex/edge index of the corresponding vertex/edge in `g`.

    Examples
    --------
    >>> from numpy.random import seed, poisson
    >>> seed(42)
    >>> g = gt.random_graph(30, lambda: (poisson(6),poisson(6)))
    >>> sub = gt.random_graph(10, lambda: (poisson(1.8), poisson(1.9)))
    >>> vm, em = gt.subgraph_isomorphism(sub, g)
184
    >>> print(len(vm))
185
    102
186
    >>> for i in range(len(vm)):
187
188
189
190
191
192
193
194
195
196
    ...   g.set_vertex_filter(None)
    ...   g.set_edge_filter(None)
    ...   vmask, emask = gt.mark_subgraph(g, sub, vm[i], em[i])
    ...   g.set_vertex_filter(vmask)
    ...   g.set_edge_filter(emask)
    ...   assert(gt.isomorphism(g, sub))
    >>> g.set_vertex_filter(None)
    >>> g.set_edge_filter(None)
    >>> ewidth = g.copy_property(emask, value_type="double")
    >>> ewidth.a += 0.5
Tiago Peixoto's avatar
Tiago Peixoto committed
197
198
199
    >>> ewidth.a *= 2
    >>> gt.graph_draw(g, vertex_fill_color=vmask, edge_color=emask,
    ...               edge_pen_width=ewidth, output_size=(200, 200),
200
    ...               output="subgraph-iso-embed.pdf")
201
    <...>
Tiago Peixoto's avatar
Tiago Peixoto committed
202
    >>> gt.graph_draw(sub, output_size=(200, 200), output="subgraph-iso.pdf")
203
204
    <...>

Tiago Peixoto's avatar
Tiago Peixoto committed
205
206
    .. image:: subgraph-iso.*
    .. image:: subgraph-iso-embed.*
207

208

Tiago Peixoto's avatar
Tiago Peixoto committed
209
    **Left:** Subgraph searched, **Right:** One isomorphic subgraph found in main graph.
210
211
212

    Notes
    -----
213
214
215
216
    The algorithm used is described in [ullmann-algorithm-1976]. It has
    worse-case complexity of :math:`O(N_g^{N_{sub}})`, but for random graphs it
    typically has a complexity of :math:`O(N_g^\gamma)` with :math:`\gamma`
    depending sub-linearly on the size of `sub`.
217
218
219

    References
    ----------
220
    .. [ullmann-algorithm-1976] Ullmann, J. R., "An algorithm for subgraph
Tiago Peixoto's avatar
Tiago Peixoto committed
221
       isomorphism", Journal of the ACM 23 (1): 31–42, 1976, :doi:`10.1145/321921.321925`
222
    .. [subgraph-isormophism-wikipedia] http://en.wikipedia.org/wiki/Subgraph_isomorphism_problem
223
224
225
226

    """
    # vertex and edge labels disabled for the time being, until GCC is capable
    # of compiling all the variants using reasonable amounts of memory
Tiago Peixoto's avatar
Tiago Peixoto committed
227
228
    vlabels=(None, None)
    elabels=(None, None)
229
230
    vmaps = []
    emaps = []
231
    if random:
232
        seed = numpy.random.randint(0, sys.maxsize)
233
234
    else:
        seed = 42
235
236
237
238
239
240
    libgraph_tool_topology.\
           subgraph_isomorphism(sub._Graph__graph, g._Graph__graph,
                                _prop("v", sub, vlabels[0]),
                                _prop("v", g, vlabels[1]),
                                _prop("e", sub, elabels[0]),
                                _prop("e", g, elabels[1]),
241
                                vmaps, emaps, max_n, seed)
242
    for i in range(len(vmaps)):
243
244
245
246
        vmaps[i] = PropertyMap(vmaps[i], sub, "v")
        emaps[i] = PropertyMap(emaps[i], sub, "e")
    return vmaps, emaps

Tiago Peixoto's avatar
Tiago Peixoto committed
247

248
249
250
251
252
253
254
255
256
257
def mark_subgraph(g, sub, vmap, emap, vmask=None, emask=None):
    r"""
    Mark a given subgraph `sub` on the graph `g`.

    The mapping must be provided by the `vmap` and `emap` parameters,
    which map vertices/edges of `sub` to indexes of the corresponding
    vertices/edges in `g`.

    This returns a vertex and an edge property map, with value type 'bool',
    indicating whether or not a vertex/edge in `g` corresponds to the subgraph
258
    `sub`.
259
    """
260
    if vmask is None:
261
        vmask = g.new_vertex_property("bool")
262
    if emask is None:
263
264
265
266
267
268
269
270
271
272
273
274
275
276
        emask = g.new_edge_property("bool")

    vmask.a = False
    emask.a = False

    for v in sub.vertices():
        w = g.vertex(vmap[v])
        vmask[w] = True
        for ew in w.out_edges():
            for ev in v.out_edges():
                if emap[ev] == g.edge_index[ew]:
                    emask[ew] = True
                    break
    return vmask, emask
277

Tiago Peixoto's avatar
Tiago Peixoto committed
278

279
def min_spanning_tree(g, weights=None, root=None, tree_map=None):
280
281
282
283
284
285
286
287
288
289
290
    """
    Return the minimum spanning tree of a given graph.

    Parameters
    ----------
    g : :class:`~graph_tool.Graph`
        Graph to be used.
    weights : :class:`~graph_tool.PropertyMap` (optional, default: None)
        The edge weights. If provided, the minimum spanning tree will minimize
        the edge weights.
    root : :class:`~graph_tool.Vertex` (optional, default: None)
291
        Root of the minimum spanning tree. If this is provided, Prim's algorithm
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
        is used. Otherwise, Kruskal's algorithm is used.
    tree_map : :class:`~graph_tool.PropertyMap` (optional, default: None)
        If provided, the edge tree map will be written in this property map.

    Returns
    -------
    tree_map : :class:`~graph_tool.PropertyMap`
        Edge property map with mark the tree edges: 1 for tree edge, 0
        otherwise.

    Notes
    -----
    The algorithm runs with :math:`O(E\log E)` complexity, or :math:`O(E\log V)`
    if `root` is specified.

    Examples
    --------
Tiago Peixoto's avatar
Tiago Peixoto committed
309
    >>> from numpy.random import seed, random
310
    >>> seed(42)
311
312
313
    >>> g, pos = gt.triangulation(random((400, 2)) * 10, type="delaunay")
    >>> weight = g.new_edge_property("double")
    >>> for e in g.edges():
Tiago Peixoto's avatar
Tiago Peixoto committed
314
    ...    weight[e] = linalg.norm(pos[e.target()].a - pos[e.source()].a)
315
    >>> tree = gt.min_spanning_tree(g, weights=weight)
316
    >>> gt.graph_draw(g, pos=pos, output="triang_orig.pdf")
317
318
    <...>
    >>> g.set_edge_filter(tree)
319
    >>> gt.graph_draw(g, pos=pos, output="triang_min_span_tree.pdf")
320
321
322
    <...>


323
    .. image:: triang_orig.*
Tiago Peixoto's avatar
Tiago Peixoto committed
324
        :width: 400px
325
    .. image:: triang_min_span_tree.*
Tiago Peixoto's avatar
Tiago Peixoto committed
326
        :width: 400px
327
328

    *Left:* Original graph, *Right:* The minimum spanning tree.
329
330
331
332
333

    References
    ----------
    .. [kruskal-shortest-1956] J. B. Kruskal.  "On the shortest spanning subtree
       of a graph and the traveling salesman problem",  In Proceedings of the
Tiago Peixoto's avatar
Tiago Peixoto committed
334
335
       American Mathematical Society, volume 7, pages 48-50, 1956.
       :doi:`10.1090/S0002-9939-1956-0078686-7`
336
337
338
339
340
    .. [prim-shortest-1957] R. Prim.  "Shortest connection networks and some
       generalizations",  Bell System Technical Journal, 36:1389-1401, 1957.
    .. [boost-mst] http://www.boost.org/libs/graph/doc/graph_theory_review.html#sec:minimum-spanning-tree
    .. [mst-wiki] http://en.wikipedia.org/wiki/Minimum_spanning_tree
    """
341
    if tree_map is None:
342
343
344
345
        tree_map = g.new_edge_property("bool")
    if tree_map.value_type() != "bool":
        raise ValueError("edge property 'tree_map' must be of value type bool.")

346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
    try:
        g.stash_filter(directed=True)
        g.set_directed(False)
        if root is None:
            libgraph_tool_topology.\
                   get_kruskal_spanning_tree(g._Graph__graph,
                                             _prop("e", g, weights),
                                             _prop("e", g, tree_map))
        else:
            libgraph_tool_topology.\
                   get_prim_spanning_tree(g._Graph__graph, int(root),
                                          _prop("e", g, weights),
                                          _prop("e", g, tree_map))
    finally:
        g.pop_filter(directed=True)
361
    return tree_map
362

Tiago Peixoto's avatar
Tiago Peixoto committed
363

Tiago Peixoto's avatar
Tiago Peixoto committed
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
def dominator_tree(g, root, dom_map=None):
    """Return a vertex property map the dominator vertices for each vertex.

    Parameters
    ----------
    g : :class:`~graph_tool.Graph`
        Graph to be used.
    root : :class:`~graph_tool.Vertex`
        The root vertex.
    dom_map : :class:`~graph_tool.PropertyMap` (optional, default: None)
        If provided, the dominator map will be written in this property map.

    Returns
    -------
    dom_map : :class:`~graph_tool.PropertyMap`
        The dominator map. It contains for each vertex, the index of its
        dominator vertex.

    Notes
    -----
    A vertex u dominates a vertex v, if every path of directed graph from the
    entry to v must go through u.

    The algorithm runs with :math:`O((V+E)\log (V+E))` complexity.

    Examples
    --------
    >>> from numpy.random import seed
    >>> seed(42)
    >>> g = gt.random_graph(100, lambda: (2, 2))
    >>> tree = gt.min_spanning_tree(g)
    >>> g.set_edge_filter(tree)
396
    >>> root = [v for v in g.vertices() if v.in_degree() == 0]
Tiago Peixoto's avatar
Tiago Peixoto committed
397
    >>> dom = gt.dominator_tree(g, root[0])
398
    >>> print(dom.a)
Tiago Peixoto's avatar
Tiago Peixoto committed
399
400
401
    [0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 5 0 0 0 0 0 0 0 0 0 0
     0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
     0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0]
Tiago Peixoto's avatar
Tiago Peixoto committed
402
403
404

    References
    ----------
405
    .. [dominator-bgl] http://www.boost.org/libs/graph/doc/lengauer_tarjan_dominator.htm
Tiago Peixoto's avatar
Tiago Peixoto committed
406
407

    """
408
    if dom_map is None:
Tiago Peixoto's avatar
Tiago Peixoto committed
409
410
411
        dom_map = g.new_vertex_property("int32_t")
    if dom_map.value_type() != "int32_t":
        raise ValueError("vertex property 'dom_map' must be of value type" +
412
413
                         " int32_t.")
    if not g.is_directed():
Tiago Peixoto's avatar
Tiago Peixoto committed
414
        raise ValueError("dominator tree requires a directed graph.")
415
    libgraph_tool_topology.\
Tiago Peixoto's avatar
Tiago Peixoto committed
416
417
418
               dominator_tree(g._Graph__graph, int(root),
                              _prop("v", g, dom_map))
    return dom_map
419

Tiago Peixoto's avatar
Tiago Peixoto committed
420

421
def topological_sort(g):
Tiago Peixoto's avatar
Tiago Peixoto committed
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
    """
    Return the topological sort of the given graph. It is returned as an array
    of vertex indexes, in the sort order.

    Notes
    -----
    The topological sort algorithm creates a linear ordering of the vertices
    such that if edge (u,v) appears in the graph, then v comes before u in the
    ordering. The graph must be a directed acyclic graph (DAG).

    The time complexity is :math:`O(V + E)`.

    Examples
    --------
    >>> from numpy.random import seed
    >>> seed(42)
    >>> g = gt.random_graph(30, lambda: (3, 3))
    >>> tree = gt.min_spanning_tree(g)
    >>> g.set_edge_filter(tree)
    >>> sort = gt.topological_sort(g)
442
    >>> print(sort)
Tiago Peixoto's avatar
Tiago Peixoto committed
443
444
    [ 3 20  9 29 15  0 10 23  1  2 21  7  4 12 11  5 26 27  6  8 13 14 22 16 17
     28 18 19 24 25]
Tiago Peixoto's avatar
Tiago Peixoto committed
445
446
447

    References
    ----------
448
    .. [topological-boost] http://www.boost.org/libs/graph/doc/topological_sort.html
Tiago Peixoto's avatar
Tiago Peixoto committed
449
450
451
452
    .. [topological-wiki] http://en.wikipedia.org/wiki/Topological_sorting

    """

453
454
455
    topological_order = Vector_int32_t()
    libgraph_tool_topology.\
               topological_sort(g._Graph__graph, topological_order)
Tiago Peixoto's avatar
Tiago Peixoto committed
456
    return numpy.array(topological_order)
457

Tiago Peixoto's avatar
Tiago Peixoto committed
458

459
def transitive_closure(g):
Tiago Peixoto's avatar
Tiago Peixoto committed
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
    """Return the transitive closure graph of g.

    Notes
    -----
    The transitive closure of a graph G = (V,E) is a graph G* = (V,E*) such that
    E* contains an edge (u,v) if and only if G contains a path (of at least one
    edge) from u to v. The transitive_closure() function transforms the input
    graph g into the transitive closure graph tc.

    The time complexity (worst-case) is :math:`O(VE)`.

    Examples
    --------
    >>> from numpy.random import seed
    >>> seed(42)
    >>> g = gt.random_graph(30, lambda: (3, 3))
    >>> tc = gt.transitive_closure(g)

    References
    ----------
480
    .. [transitive-boost] http://www.boost.org/libs/graph/doc/transitive_closure.html
Tiago Peixoto's avatar
Tiago Peixoto committed
481
482
483
484
    .. [transitive-wiki] http://en.wikipedia.org/wiki/Transitive_closure

    """

485
486
487
488
489
490
491
    if not g.is_directed():
        raise ValueError("graph must be directed for transitive closure.")
    tg = Graph()
    libgraph_tool_topology.transitive_closure(g._Graph__graph,
                                              tg._Graph__graph)
    return tg

Tiago Peixoto's avatar
Tiago Peixoto committed
492

493
494
def label_components(g, vprop=None, directed=None):
    """
495
    Label the components to which each vertex in the graph belongs. If the
496
497
    graph is directed, it finds the strongly connected components.

498
499
500
    A property map with the component labels is returned, together with an
    histogram of component labels.

501
502
    Parameters
    ----------
503
    g : :class:`~graph_tool.Graph`
504
        Graph to be used.
505
    vprop : :class:`~graph_tool.PropertyMap` (optional, default: None)
506
507
508
509
510
511
512
513
        Vertex property to store the component labels. If none is supplied, one
        is created.
    directed : bool (optional, default:None)
        Treat graph as directed or not, independently of its actual
        directionality.

    Returns
    -------
514
    comp : :class:`~graph_tool.PropertyMap`
515
        Vertex property map with component labels.
516
517
    hist : :class:`~numpy.ndarray`
        Histogram of component labels.
518
519
520
521
522
523

    Notes
    -----
    The components are arbitrarily labeled from 0 to N-1, where N is the total
    number of components.

524
    The algorithm runs in :math:`O(V + E)` time.
525
526
527

    Examples
    --------
528
529
530
    >>> from numpy.random import seed
    >>> seed(43)
    >>> g = gt.random_graph(100, lambda: (1, 1))
531
    >>> comp, hist = gt.label_components(g)
532
    >>> print(comp.a)
Tiago Peixoto's avatar
Tiago Peixoto committed
533
534
535
    [0 0 0 1 0 2 0 0 0 0 2 0 0 0 2 1 0 2 0 1 2 0 1 0 0 1 0 2 0 2 1 0 2 0 0 0 0
     0 0 1 0 0 2 2 2 0 0 0 0 0 0 2 0 0 1 1 0 0 2 0 1 0 0 0 2 0 0 2 2 1 2 1 0 0
     2 0 0 1 2 1 2 2 0 0 0 0 0 2 0 0 0 1 1 0 0 0 1 1 2 2]
536
    >>> print(hist)
Tiago Peixoto's avatar
Tiago Peixoto committed
537
    [58 18 24]
538
539
    """

540
    if vprop is None:
541
542
543
544
545
        vprop = g.new_vertex_property("int32_t")

    _check_prop_writable(vprop, name="vprop")
    _check_prop_scalar(vprop, name="vprop")

546
547
    if directed is not None:
        g = GraphView(g, directed=directed)
548

549
550
551
552
553
554
555
    hist = libgraph_tool_topology.\
               label_components(g._Graph__graph, _prop("v", g, vprop))
    return vprop, hist


def label_largest_component(g, directed=None):
    """
556
557
    Label the largest component in the graph. If the graph is directed, then the
    largest strongly connected component is labelled.
558
559
560
561
562
563
564
565
566
567
568
569
570
571

    A property map with a boolean label is returned.

    Parameters
    ----------
    g : :class:`~graph_tool.Graph`
        Graph to be used.
    directed : bool (optional, default:None)
        Treat graph as directed or not, independently of its actual
        directionality.

    Returns
    -------
    comp : :class:`~graph_tool.PropertyMap`
572
         Boolean vertex property map which labels the largest component.
573
574
575
576
577
578
579
580
581
582
583

    Notes
    -----
    The algorithm runs in :math:`O(V + E)` time.

    Examples
    --------
    >>> from numpy.random import seed, poisson
    >>> seed(43)
    >>> g = gt.random_graph(100, lambda: poisson(1), directed=False)
    >>> l = gt.label_largest_component(g)
584
    >>> print(l.a)
Tiago Peixoto's avatar
Tiago Peixoto committed
585
586
587
    [1 0 0 0 0 0 0 0 1 1 0 1 0 0 0 0 0 0 1 0 0 1 1 1 1 0 0 0 1 0 0 0 0 0 0 0 1
     1 1 0 0 0 0 1 0 1 1 0 0 0 1 1 0 0 1 1 0 0 0 0 0 0 1 0 1 0 0 0 0 0 0 0 0 0
     0 0 0 1 1 0 1 1 0 0 0 0 0 1 1 0 1 0 1 0 1 0 0 0 0 0]
588
    >>> u = gt.GraphView(g, vfilt=l)   # extract the largest component as a graph
589
    >>> print(u.num_vertices())
Tiago Peixoto's avatar
Tiago Peixoto committed
590
    31
591
592
593
594
    """

    label = g.new_vertex_property("bool")
    c, h = label_components(g, directed=directed)
595
596
597
598
599
    vfilt, inv = g.get_vertex_filter()
    if vfilt is None:
        label.a = c.a == h.argmax()
    else:
        label.a = (c.a == h.argmax()) & (vfilt.a ^ inv)
600
    return label
601

Tiago Peixoto's avatar
Tiago Peixoto committed
602

603
def label_biconnected_components(g, eprop=None, vprop=None):
604
605
606
607
    """
    Label the edges of biconnected components, and the vertices which are
    articulation points.

608
609
610
611
    An edge property map with the component labels is returned, together a
    boolean vertex map marking the articulation points, and an histogram of
    component labels.

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
    Parameters
    ----------
    g : :class:`~graph_tool.Graph`
        Graph to be used.

    eprop : :class:`~graph_tool.PropertyMap` (optional, default: None)
        Edge property to label the biconnected components.

    vprop : :class:`~graph_tool.PropertyMap` (optional, default: None)
        Vertex property to mark the articulation points. If none is supplied,
        one is created.


    Returns
    -------
    bicomp : :class:`~graph_tool.PropertyMap`
        Edge property map with the biconnected component labels.
    articulation : :class:`~graph_tool.PropertyMap`
        Boolean vertex property map which has value 1 for each vertex which is
        an articulation point, and zero otherwise.
    nc : int
        Number of biconnected components.

    Notes
    -----

    A connected graph is biconnected if the removal of any single vertex (and
    all edges incident on that vertex) can not disconnect the graph. More
    generally, the biconnected components of a graph are the maximal subsets of
    vertices such that the removal of a vertex from a particular component will
    not disconnect the component. Unlike connected components, vertices may
    belong to multiple biconnected components: those vertices that belong to
    more than one biconnected component are called "articulation points" or,
    equivalently, "cut vertices". Articulation points are vertices whose removal
    would increase the number of connected components in the graph. Thus, a
    graph without articulation points is biconnected. Vertices can be present in
    multiple biconnected components, but each edge can only be contained in a
    single biconnected component.

    The algorithm runs in :math:`O(V + E)` time.

    Examples
    --------
    >>> from numpy.random import seed
Tiago Peixoto's avatar
Tiago Peixoto committed
656
    >>> seed(43)
657
    >>> g = gt.random_graph(100, lambda: 2, directed=False)
658
    >>> comp, art, hist = gt.label_biconnected_components(g)
659
    >>> print(comp.a)
Tiago Peixoto's avatar
Tiago Peixoto committed
660
661
662
    [0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0 0
     0 0 1 0 0 1 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 1 0 0 0 0 1
     0 0 0 0 0 0 0 1 0 1 0 0 0 0 0 0 0 0 1 0 1 0 0 0 0 0]
663
    >>> print(art.a)
664
665
666
    [0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
     0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
     0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0]
667
    >>> print(hist)
Tiago Peixoto's avatar
Tiago Peixoto committed
668
    [87 13]
669
    """
670

671
    if vprop is None:
672
        vprop = g.new_vertex_property("bool")
673
    if eprop is None:
674
675
676
677
678
679
680
        eprop = g.new_edge_property("int32_t")

    _check_prop_writable(vprop, name="vprop")
    _check_prop_scalar(vprop, name="vprop")
    _check_prop_writable(eprop, name="eprop")
    _check_prop_scalar(eprop, name="eprop")

681
682
    g = GraphView(g, directed=False)
    hist = libgraph_tool_topology.\
683
684
             label_biconnected_components(g._Graph__graph, _prop("e", g, eprop),
                                          _prop("v", g, vprop))
685
    return eprop, vprop, hist
686

Tiago Peixoto's avatar
Tiago Peixoto committed
687

688
def shortest_distance(g, source=None, weights=None, max_dist=None,
689
690
                      directed=None, dense=False, dist_map=None,
                      pred_map=False):
691
692
693
694
695
696
697
698
699
    """
    Calculate the distance of all vertices from a given source, or the all pairs
    shortest paths, if the source is not specified.

    Parameters
    ----------
    g : :class:`~graph_tool.Graph`
        Graph to be used.
    source : :class:`~graph_tool.Vertex` (optional, default: None)
700
        Source vertex of the search. If unspecified, the all pairs shortest
701
702
703
704
705
706
        distances are computed.
    weights : :class:`~graph_tool.PropertyMap` (optional, default: None)
        The edge weights. If provided, the minimum spanning tree will minimize
        the edge weights.
    max_dist : scalar value (optional, default: None)
        If specified, this limits the maximum distance of the vertices
707
        are searched. This parameter has no effect if source is None.
708
709
710
711
    directed : bool (optional, default:None)
        Treat graph as directed or not, independently of its actual
        directionality.
    dense : bool (optional, default: False)
712
713
        If true, and source is None, the Floyd-Warshall algorithm is used,
        otherwise the Johnson algorithm is used. If source is not None, this option
714
715
716
717
        has no effect.
    dist_map : :class:`~graph_tool.PropertyMap` (optional, default: None)
        Vertex property to store the distances. If none is supplied, one
        is created.
718
719
720
    pred_map : bool (optional, default: False)
        If true, a vertex property map with the predecessors is returned.
        Ignored if source=None.
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

    Returns
    -------
    dist_map : :class:`~graph_tool.PropertyMap`
        Vertex property map with the distances from source. If source is 'None',
        it will have a vector value type, with the distances to every vertex.

    Notes
    -----

    If a source is given, the distances are calculated with a breadth-first
    search (BFS) or Dijkstra's algorithm [dijkstra]_, if weights are given. If
    source is not given, the distances are calculated with Johnson's algorithm
    [johnson-apsp]_. If dense=True, the Floyd-Warshall algorithm
    [floyd-warshall-apsp]_ is used instead.

    If source is specified, the algorithm runs in :math:`O(V + E)` time, or
    :math:`O(V \log V)` if weights are given. If source is not specified, it
    runs in :math:`O(VE\log V)` time, or :math:`O(V^3)` if dense == True.

    Examples
    --------
    >>> from numpy.random import seed, poisson
    >>> seed(42)
    >>> g = gt.random_graph(100, lambda: (poisson(3), poisson(3)))
    >>> dist = gt.shortest_distance(g, source=g.vertex(0))
747
    >>> print(dist.a)
Tiago Peixoto's avatar
Tiago Peixoto committed
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
    [         0          3          6          4 2147483647          3
              4          3          4          2          3          4
              3          4          2          4          2          5
              4          4 2147483647          4 2147483647          6
              4          7          5 2147483647          3          4
              2          3          5          5          4          5
              1          5          6          1 2147483647          8
              4          2          1          5          5          6
              7          4          5          3          4          4
              5          3          3          5          4          5
              4          3          5          4          2 2147483647
              6          5          4          5          1 2147483647
              5          5          4          2          5          4
              6          3          5          3          4 2147483647
              4          4          7          4          3          5
              5          2          7          3          4          4
              4          3          4          4]
765
    >>> dist = gt.shortest_distance(g)
766
    >>> print(dist[g.vertex(0)].a)
Tiago Peixoto's avatar
Tiago Peixoto committed
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
    [         0          3          6          4 2147483647          3
              4          3          4          2          3          4
              3          4          2          4          2          5
              4          4 2147483647          4 2147483647          6
              4          7          5 2147483647          3          4
              2          3          5          5          4          5
              1          5          6          1 2147483647          8
              4          2          1          5          5          6
              7          4          5          3          4          4
              5          3          3          5          4          5
              4          3          5          4          2 2147483647
              6          5          4          5          1 2147483647
              5          5          4          2          5          4
              6          3          5          3          4 2147483647
              4          4          7          4          3          5
              5          2          7          3          4          4
              4          3          4          4]
784
785
786
787
788

    References
    ----------
    .. [bfs] Edward Moore, "The shortest path through a maze", International
       Symposium on the Theory of Switching (1959), Harvard University
Tiago Peixoto's avatar
Tiago Peixoto committed
789
790
       Press;
    .. [bfs-boost] http://www.boost.org/libs/graph/doc/breadth_first_search.html
791
792
    .. [dijkstra] E. Dijkstra, "A note on two problems in connexion with
       graphs." Numerische Mathematik, 1:269-271, 1959.
Tiago Peixoto's avatar
Tiago Peixoto committed
793
    .. [dijkstra-boost] http://www.boost.org/libs/graph/doc/dijkstra_shortest_paths.html
794
795
796
797
    .. [johnson-apsp] http://www.boost.org/libs/graph/doc/johnson_all_pairs_shortest.html
    .. [floyd-warshall-apsp] http://www.boost.org/libs/graph/doc/floyd_warshall_shortest.html
    """

798
    if weights is None:
799
800
801
802
        dist_type = 'int32_t'
    else:
        dist_type = weights.value_type()

803
804
    if dist_map is None:
        if source is not None:
805
806
807
808
809
            dist_map = g.new_vertex_property(dist_type)
        else:
            dist_map = g.new_vertex_property("vector<%s>" % dist_type)

    _check_prop_writable(dist_map, name="dist_map")
810
    if source is not None:
811
812
813
814
        _check_prop_scalar(dist_map, name="dist_map")
    else:
        _check_prop_vector(dist_map, name="dist_map")

815
    if max_dist is None:
816
817
        max_dist = 0

818
    if directed is not None:
819
820
821
822
        g.stash_filter(directed=True)
        g.set_directed(directed)

    try:
823
        if source is not None:
824
            pmap = g.copy_property(g.vertex_index, value_type="int64_t")
825
826
827
            libgraph_tool_topology.get_dists(g._Graph__graph, int(source),
                                             _prop("v", g, dist_map),
                                             _prop("e", g, weights),
828
                                             _prop("v", g, pmap),
829
830
831
832
833
834
835
                                             float(max_dist))
        else:
            libgraph_tool_topology.get_all_dists(g._Graph__graph,
                                                 _prop("v", g, dist_map),
                                                 _prop("e", g, weights), dense)

    finally:
836
        if directed is not None:
837
            g.pop_filter(directed=True)
838
    if source is not None and pred_map:
839
840
841
842
        return dist_map, pmap
    else:
        return dist_map

Tiago Peixoto's avatar
Tiago Peixoto committed
843

844
845
846
847
848
849
850
851
852
853
def shortest_path(g, source, target, weights=None, pred_map=None):
    """
    Return the shortest path from `source` to `target`.

    Parameters
    ----------
    g : :class:`~graph_tool.Graph`
        Graph to be used.
    source : :class:`~graph_tool.Vertex`
        Source vertex of the search.
Tiago Peixoto's avatar
Tiago Peixoto committed
854
    target : :class:`~graph_tool.Vertex`
855
856
        Target vertex of the search.
    weights : :class:`~graph_tool.PropertyMap` (optional, default: None)
Tiago Peixoto's avatar
Tiago Peixoto committed
857
        The edge weights.
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
    pred_map :  :class:`~graph_tool.PropertyMap` (optional, default: None)
        Vertex property map with the predecessors in the search tree. If this is
        provided, the shortest paths are not computed, and are obtained directly
        from this map.

    Returns
    -------
    vertex_list : list of :class:`~graph_tool.Vertex`
        List of vertices from `source` to `target` in the shortest path.
    edge_list : list of :class:`~graph_tool.Edge`
        List of edges from `source` to `target` in the shortest path.

    Notes
    -----

    The paths are computed with a breadth-first search (BFS) or Dijkstra's
    algorithm [dijkstra]_, if weights are given.

    The algorithm runs in :math:`O(V + E)` time, or :math:`O(V \log V)` if
    weights are given.

    Examples
    --------
    >>> from numpy.random import seed, poisson
    >>> seed(42)
    >>> g = gt.random_graph(300, lambda: (poisson(3), poisson(3)))
    >>> vlist, elist = gt.shortest_path(g, g.vertex(10), g.vertex(11))
885
    >>> print([str(v) for v in vlist])
Tiago Peixoto's avatar
Tiago Peixoto committed
886
    ['10', '222', '246', '0', '50', '257', '12', '242', '11']
887
    >>> print([str(e) for e in elist])
888
    ['(10, 222)', '(222, 246)', '(246, 0)', '(0, 50)', '(50, 257)', '(257, 12)', '(12, 242)', '(242, 11)']
889
890
891
892
893

    References
    ----------
    .. [bfs] Edward Moore, "The shortest path through a maze", International
       Symposium on the Theory of Switching (1959), Harvard University
Tiago Peixoto's avatar
Tiago Peixoto committed
894
895
       Press
    .. [bfs-boost] http://www.boost.org/libs/graph/doc/breadth_first_search.html
896
897
    .. [dijkstra] E. Dijkstra, "A note on two problems in connexion with
       graphs." Numerische Mathematik, 1:269-271, 1959.
Tiago Peixoto's avatar
Tiago Peixoto committed
898
    .. [dijkstra-boost] http://www.boost.org/libs/graph/doc/dijkstra_shortest_paths.html
899
900
    """

901
    if pred_map is None:
Tiago Peixoto's avatar
Tiago Peixoto committed
902
903
        pred_map = shortest_distance(g, source, weights=weights,
                                     pred_map=True)[1]
904

Tiago Peixoto's avatar
Tiago Peixoto committed
905
    if pred_map[target] == int(target):  # no path to source
906
907
908
909
910
        return [], []

    vlist = [target]
    elist = []

911
    if weights is not None:
912
913
914
915
916
917
918
919
920
921
922
923
924
        max_w = weights.a.max() + 1
    else:
        max_w = None

    v = target
    while v != source:
        p = g.vertex(pred_map[v])
        min_w = max_w
        pe = None
        s = None
        for e in v.in_edges() if g.is_directed() else v.out_edges():
            s = e.source() if g.is_directed() else e.target()
            if s == p:
925
                if weights is not None:
926
927
928
929
930
931
932
933
934
935
936
                    if weights[e] < min_w:
                        min_w = weights[e]
                        pe = e
                else:
                    pe = e
                    break
        elist.insert(0, pe)
        vlist.insert(0, p)
        v = p
    return vlist, elist

937

Tiago Peixoto's avatar
Tiago Peixoto committed
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
def pseudo_diameter(g, source=None, weights=None):
    """
    Compute the pseudo-diameter of the graph.

    Parameters
    ----------
    g : :class:`~graph_tool.Graph`
        Graph to be used.
    source : :class:`~graph_tool.Vertex` (optional, default: `None`)
        Source vertex of the search. If not supplied, the first vertex
        in the graph will be chosen.
    weights : :class:`~graph_tool.PropertyMap` (optional, default: `None`)
        The edge weights.

    Returns
    -------
    pseudo_diameter : int
        The pseudo-diameter of the graph.
    end_points : pair of :class:`~graph_tool.Vertex`
        The two vertices which correspond to the pseudo-diameter found.

    Notes
    -----

    The pseudo-diameter is an approximate graph diameter. It is obtained by
    starting from a vertex `source`, and finds a vertex `target` that is
    farthest away from `source`. This process is repeated by treating
    `target` as the new starting vertex, and ends when the graph distance no
    longer increases. A vertex from the last level set that has the smallest
    degree is chosen as the final starting vertex u, and a traversal is done
    to see if the graph distance can be increased. This graph distance is
    taken to be the pseudo-diameter.

    The paths are computed with a breadth-first search (BFS) or Dijkstra's
    algorithm [dijkstra]_, if weights are given.

    The algorithm runs in :math:`O(V + E)` time, or :math:`O(V \log V)` if
    weights are given.

    Examples
    --------
    >>> from numpy.random import seed, poisson
    >>> seed(42)
    >>> g = gt.random_graph(300, lambda: (poisson(3), poisson(3)))
    >>> dist, ends = gt.pseudo_diameter(g)
983
    >>> print(dist)
984
    9.0
985
    >>> print(int(ends[0]), int(ends[1]))
986
    0 255
Tiago Peixoto's avatar
Tiago Peixoto committed
987
988
989
990
991
992
993
994
995
996
997
998
999
1000

    References
    ----------
    .. [pseudo-diameter] http://en.wikipedia.org/wiki/Distance_%28graph_theory%29
    """

    if source is None:
        source = g.vertex(0)
    dist, target = 0, source
    while True:
        new_source = target
        new_target, new_dist = libgraph_tool_topology.get_diam(g._Graph__graph,
                                                               int(new_source),
                                                               _prop("e", g, weights))