ܛ7iEdZddlZddgZejj ddZdZejj dejdZ y) z6 Algorithms for computing distance measures on trees. Ncentercentroiddirectedc<t|j}|jDchc] \}}|dk(s |}}}t|dkDr|rt }|D]r}||=|j |D]Y}||vr||xxdzcc<||x}dk(r|j |1|dk(s7t|dk7sFtjdt|}t|dkDr|rt|} | dk(r|r| dvrtjdt|Scc}}w)aReturns the center of an undirected tree graph. The center of a tree consists of nodes that minimize the maximum eccentricity. That is, these nodes minimize the maximum distance to all other nodes. This implementation currently only works for unweighted edges. If the input graph is not a tree, results are not guaranteed to be correct and while some non-trees will raise an ``nx.NotATree`` exception, not all non-trees will be discovered. Thus, this function should not be used if caller is unsure whether the input graph is a tree. Use ``nx.is_tree(G)`` to check. Parameters ---------- G : NetworkX graph A tree graph (undirected, acyclic graph). Returns ------- center : list A list of nodes forming the center of the tree. This can be one or two nodes. Raises ------ NetworkXNotImplemented If the input graph is directed. NotATree If the algorithm detects the input graph is not a tree. There is no guarantee this error will always raise if a non-tree is passed. Notes ----- This algorithm iteratively removes leaves (nodes with degree 1) from the tree until there are only 1 or 2 nodes left. The remaining nodes form the center of the tree. This algorithm's time complexity is ``O(N)`` where ``N`` is the number of nodes in the tree. Examples -------- >>> G = nx.Graph([(1, 2), (1, 3), (2, 4), (2, 5)]) >>> nx.tree.center(G) [1, 2] >>> G = nx.path_graph(5) >>> nx.tree.center(G) [2] rzinput graph is not a tree>rr) dictdegreeitemslenset neighborsaddnxNotATreelist) Gcenter_candidates_degreenoder leaves new_leavesleafneighborcddnns g/home/rose/Desktop/poly/venv/lib/python3.12/site-packages/networkx/algorithms/tree/distance_measures.pyrr s2b $AHH~'?'E'E'G W'G|tV6UV;d'GF W & '! +U D(.KK-#;;(2a724X>>D1DNN8,QY3'?#@A#E++&ABB. & '! + $%A Qv!6/kk566 ( ))1Xs DDcd|di}|g}tj||D]O\}}|d|k7r,|j}||dxx||z cc<|d|k7r,|j|d||<Qtjj t |D]\}}||xx||z cc<|S)a0Return a `dict` of the size of each subtree, for every subtree of a tree rooted at a given node. For every node in the given tree, consider the new tree that would be created by detaching it from its parent node (if any). The number of nodes in the resulting tree rooted at that node is then assigned as the value for that node in the return dictionary. Parameters ---------- G : NetworkX graph A tree. root : node A node in `G`. Returns ------- s : dict Dictionary of number of nodes in every subtree of this tree, keyed on the root node for each subtree. Examples -------- >>> _subtree_sizes(nx.path_graph(4), 0) {0: 4, 1: 3, 2: 2, 3: 1} >>> _subtree_sizes(nx.path_graph(4), 2) {2: 4, 1: 2, 0: 1, 3: 1} r)r dfs_edgespopappendutilspairwisereversed)rrootsizesstackparentchild descendants r_subtree_sizesr+Zs@1IE FEa. Bi6!J %) j 1 1 Bi6!  Ue / **8E?; v f u% < Lc"tjstjddtjj }}t |j }fd}|||}t||z j|d|dz kDr6||}}|||}t||z j|d|dz kDr6|gj|Dcgc]}||k7s ||dz k(s|c}zScc}w)aReturn the centroid of an unweighted tree. The centroid of a tree is the set of nodes such that removing any one of them would split the tree into a forest of subtrees, each with at most ``N / 2`` nodes, where ``N`` is the number of nodes in the original tree. This set may contain two nodes if removing an edge between them results in two trees of size exactly ``N / 2``. Parameters ---------- G : NetworkX graph A tree. Returns ------- c : list List of nodes in centroid of the tree. This could be one or two nodes. Raises ------ NotATree If the input graph is not a tree. NotImplementedException If the input graph is directed. NetworkXPointlessConcept If `G` has no nodes or edges. Notes ----- This algorithm's time complexity is ``O(N)`` where ``N`` is the number of nodes in the tree. In unweighted trees the centroid coincides with the barycenter, the node or nodes that minimize the sum of distances to all other nodes. However, this concept is different from that of the graph center, which is the set of nodes minimizing the maximum distance to all other nodes. Examples -------- >>> G = nx.path_graph(4) >>> nx.tree.centroid(G) [1, 2] A star-shaped tree with one long branch illustrates the difference between the centroid and the center. The center lies near the middle of the long branch, minimizing maximum distance. The centroid, however, limits the size of any resulting subtree to at most half the total nodes, forcing it to remain near the hub when enough short branches are present. >>> G = nx.star_graph(6) >>> nx.add_path(G, [6, 7, 8, 9, 10]) >>> nx.tree.centroid(G), nx.tree.center(G) ([0], [7]) See Also -------- :func:`~networkx.algorithms.distance_measures.barycenter` :func:`~networkx.algorithms.distance_measures.center` center : tree center zprovided graph is not a treeNcftfdj|DjdS)Nc3.K|] }|k7s |yw)N).0xprevs r z4centroid.._heaviest_child..s 7)1Q$YQ)s )keydefault)maxrget)r3r%rr&s` r_heaviest_childz!centroid.._heaviest_childs* 7 D) 7UYYPT  r,rr) ris_treerr"arbitrary_elementr+number_of_nodesr7r8r)rr3r% total_sizer9hcr2r&s` @rrrsD ::a=kk899rxx11!4$D 1d #E""$J t $B j5;& "a(8 9JN J2d T4 ( j5;& "a(8 9JN J 6;;t$$aT eAh*q.6P$ s, D 7 D D ) __doc__networkxr__all__r"not_implemented_forrr+ _dispatchablerr0r,rrDs}   j)I**I*X*Zj)R*Rr,