Merge pull request rapidsai#47 from jwyles/bfs_first

BFS imported and python bindings added.

CMakeLists.txt

@@ -182,6 +182,7 @@ endif (NVGRAPH_INCLUDE AND NVGRAPH_LIBRARY)
 include_directories(
     "${CMAKE_CURRENT_SOURCE_DIR}/include" 
     "${CMAKE_CURRENT_SOURCE_DIR}/src"
+    "${CMAKE_CURRENT_SOURCE_DIR}/external/cub"
     "${CMAKE_CUDA_TOOLKIT_INCLUDE_DIRECTORIES}" 
     "${CUDF_INCLUDE}"
     "${NVGRAPH_INCLUDE}"
@@ -204,6 +205,7 @@ add_library(cugraph SHARED
     src/grmat.cu
     src/cugraph.cu
     src/pagerank.cu
+    src/bfs.cu
     src/nvgraph_gdf.cu
     ${CMAKE_CURRENT_BINARY_DIR}/gunrock/gunrock/util/test_utils.cu
     ${CMAKE_CURRENT_BINARY_DIR}/gunrock/gunrock/util/error_utils.cu

include/functions.h

@@ -60,3 +60,5 @@ gdf_error gdf_delete_transpose(gdf_graph *graph);
 gdf_error gdf_pagerank(gdf_graph *graph, gdf_column *pagerank, float alpha, float tolerance, int max_iter, bool has_guess);
 
 gdf_error gdf_grmat_gen (const char* argv, size_t &vertices, size_t &edges, gdf_column* src, gdf_column* dest, gdf_column* val);
+
+gdf_error gdf_bfs(gdf_graph *graph, gdf_column *distances, gdf_column *predecessors, int start_node, bool directed);

python/bfs/bfs_wrapper.pyx

@@ -0,0 +1,59 @@
+from c_bfs cimport *
+from libcpp cimport bool
+from libc.stdint cimport uintptr_t
+from libc.stdlib cimport calloc, malloc, free
+import cudf
+from librmm_cffi import librmm as rmm
+#from pygdf import Column
+import numpy as np
+
+cpdef bfs(G, start, directed=True):
+    """
+    Find the distances and predecessors for a breadth first traversal of a graph.
+    
+    Parameters
+    ----------
+    G : cugraph.graph
+        cuGraph graph descriptor, should contain the connectivity information as an
+        adjacency list.
+    start : Integer
+        The index of the graph vertex from which the traversal begins
+    directed : bool
+        Indicates whether the graph in question is a directed graph, or whether
+        each edge has a corresponding reverse edge. (Allows optimizations if the
+        graph is undirected)
+    
+    Returns
+    -------
+    df : cudf.DataFrame
+        df['vertex'][i] gives the vertex id of the i'th vertex
+        df['distance'][i] gives the path distance for the i'th vertex from the starting vertex
+        df['predecessor'][i] gives for the i'th vertex the vertex it was reached from in the traversal
+        
+    Examples
+    --------
+    >>> M = ReadMtxFile(graph_file)
+    >>> sources = cudf.Series(M.row)
+    >>> destinations = cudf.Series(M.col)
+    >>> G = cuGraph.Graph()
+    >>> G.add_edge_list(sources,destinations,none)
+    >>> dist, pred = cuGraph.bfs(G, 0, false)
+    """
+
+    cdef uintptr_t graph = G.graph_ptr
+    cdef gdf_graph* g = <gdf_graph*>graph
+    num_verts = g.adjList.offsets.size - 1
+
+    df = cudf.DataFrame()
+    df['vertex'] = cudf.Series(np.zeros(num_verts, dtype=np.int32))
+    cdef uintptr_t vertex_ptr = create_column(df['vertex'])
+    df['distance'] = cudf.Series(np.zeros(num_verts, dtype=np.int32))
+    cdef uintptr_t distances_ptr = create_column(df['distance'])
+    df['predecessor'] = cudf.Series(np.zeros(num_verts, dtype=np.int32))
+    cdef uintptr_t predecessors_ptr = create_column(df['predecessor'])
+
+    err = g.adjList.get_vertex_identifiers(<gdf_column*>vertex_ptr)
+    cudf.bindings.cudf_cpp.check_gdf_error(err)
+
+    gdf_bfs(<gdf_graph*>g, <gdf_column*>distances_ptr, <gdf_column*>predecessors_ptr, <int>start, <bool>directed)
+    return df

python/bfs/c_bfs.pxd

@@ -0,0 +1,6 @@
+from c_graph cimport *
+from libcpp cimport bool
+
+cdef extern from "cugraph.h":
+
+    cdef gdf_error gdf_bfs(gdf_graph *graph, gdf_column *distances, gdf_column *predecessors, int start_node, bool directed)

python/bfs/test_bfs.py

@@ -0,0 +1,69 @@
+import cugraph
+import cudf
+import time
+from scipy.io import mmread
+import pytest
+import numpy as np
+
+def ReadMtxFile(mmFile):
+    print('Reading ' + str(mmFile) + '...')
+    return mmread(mmFile).asfptype()
+
+
+def cugraph_Call(M, start_vertex):
+
+    # Device data
+    M = M.tocsr()
+    sources = cudf.Series(M.indptr)
+    destinations = cudf.Series(M.indices)
+    values = cudf.Series(M.data)
+
+    G = cugraph.Graph()
+    G.add_adj_list(sources, destinations, values)
+
+    t1 = time.time()
+    df = cugraph.bfs(G, start_vertex)
+    t2 = time.time() - t1
+    print('Time : '+str(t2))
+
+    # Return distances as np.array()
+    return np.array(df['distance'])
+
+
+def base_Call(M, start_vertex):
+    intMax = 2147483647
+    M = M.tocsr()
+    offsets = M.indptr
+    indices = M.indices
+    num_verts = len(offsets) - 1
+    dist = np.zeros(num_verts, dtype=np.int32)
+
+    for i in range(num_verts):
+        dist[i] = intMax
+    import queue
+    q = queue.Queue()
+    q.put(start_vertex)
+    dist[start_vertex] = 0
+    while(not q.empty()):
+        u = q.get()
+        for iCol in range(offsets[u],offsets[u + 1]):
+            v = indices[iCol]
+            if (dist[v] == intMax):
+                dist[v] = dist[u] + 1
+                q.put(v)
+    return dist
+
+datasets = ['/datasets/networks/dolphins.mtx',
+            '/datasets/networks/karate.mtx',
+            '/datasets/golden_data/graphs/dblp.mtx']
+
+@pytest.mark.parametrize('graph_file', datasets)
+def test_bfs(graph_file):
+
+    M = ReadMtxFile(graph_file)
+    base_dist = base_Call(M, 0)
+    dist = cugraph_Call(M, 0)
+
+    assert len(base_dist) == len(dist)
+    for i in range(len(dist)):
+        assert base_dist[i] == dist[i]

python/cugraph.pyx

@@ -5,3 +5,4 @@ include "sssp/sssp_wrapper.pyx"
 include "jaccard/jaccard_wrapper.pyx"
 include "grmat/grmat_wrapper.pyx"
 include "louvain/louvain_wrapper.pyx"
+include "bfs/bfs_wrapper.pyx"

setup.py

@@ -73,7 +73,6 @@ def locate_nvgraph():
 
 # temporary fix. cudf 0.5 will have a cudf.get_include()
 cudf_include = os.path.normpath(sys.prefix) + '/include'
-
 cython_files = ['python/cugraph.pyx']
 
 extensions = [