Python Graph.add_edges_from方法代码示例

# 需要导入模块: from igraph import Graph [as 别名]
# 或者: from igraph.Graph import add_edges_from [as 别名]
def parse_full_arnetminer_dataset(should_profile, use_igraph):
    """
      Parse the full arnetminer dataset in plaintext format. Vertex types in the graph are:

        1: Authors
        2: Papers
        3: Conferences
        4: Terms
    """

    print "Parsing nodes for graph..."
    input_file = open(os.path.join(project_root, 'data', 'Arnetminer-Full.txt'))
    beginning = input_file.tell()

    # Use either igraph or networkx
    if use_igraph:
        graph = Graph(directed=True)
    else:
        graph = MultiDiGraph()

    # Counts for statistics
    VALID_PAPERS = 8579222  # Most recent count of valid papers from Arnetminer
    papers_processed = 0

    # Caches for vertices & edges to add (batch adds together when using igraph)
    vertices_to_add = []
    edges_to_add = []
    types_of_vertices_to_add = []
    FLUSH_FREQUENCY = 100000  # 'Flush' cached vertices and edges this often

    # Add each paper to graph (adding missing associated terms, authors, and conferences)
    for title, authors, conference, terms, paper_index in __papers_from_file(input_file, should_profile):

        # Output any test paper indices found
        if title in test_papers:
            print "Found test paper '%s' by %s, index: %d" % (title, ','.join(authors), paper_index)

        if use_igraph:

            # Use string, because otherwise integer overflow in igraph
            paper_index = str(paper_index)

            # Collect vertices and edges to add
            vertices_to_add += [paper_index, conference] + authors + terms
            types_of_vertices_to_add += [2, 3] + ([1] * len(authors)) + ([4] * len(terms))
            for author in authors:
                edges_to_add += [(author, paper_index), (paper_index, author)]
            edges_to_add += [(conference, paper_index), (paper_index, conference)]
            for term in terms:
                edges_to_add += [(term, paper_index), (paper_index, term)]

            # Every so often, actually mutate the graph
            if papers_processed % FLUSH_FREQUENCY == 0:
                for vertex, vertex_type in zip(vertices_to_add, types_of_vertices_to_add):
                    graph.add_vertex(vertex, type=vertex_type)
                graph.add_edges(edges_to_add)
                vertices_to_add = []
                edges_to_add = []
                types_of_vertices_to_add = []

        else:

            # Add symmetric edges & nodes (if they don't already exist in the network)
            for author in authors:
                graph.add_edges_from([(author, paper_index), (paper_index, author)])
            graph.add_edges_from([(conference, paper_index), (paper_index, conference)])
            for term in terms:
                graph.add_edges_from([(term, paper_index), (paper_index, term)])

        # Output progress
        papers_processed += 1
        if papers_processed % 10 == 0:
            sys.stdout.write("\r Processed %d / %d papers..." % (papers_processed, VALID_PAPERS))
            sys.stdout.flush()

    # Basic statistics about cleanliness of data
    count_and_percent_of_papers = lambda a: (a, total_papers, 100 * float(a) / total_papers)
    print "\n\nTotal Papers: %d" % total_papers
    print "  Added (Successful): %d / %d (%2.2f%%)" % count_and_percent_of_papers(successful_papers)
    print "  Ignored (Bad Title): %d / %d (%2.2f%%)" % count_and_percent_of_papers(skipped_bad_title)
    print "  Skipped (Missing Conference): %d / %d (%2.2f%%)" % count_and_percent_of_papers(skipped_missing_conference)
    print "  Invalid (Unknown): %d / %d (%2.2f%%)\n\n" % count_and_percent_of_papers(invalid_papers)

    # Rewind file
    input_file.seek(beginning)

    print "Parsing citations for graph..."

    # Counts for statistics
    papers_processed = 0
    successful_citations = 0
    omitted_paper_citations = 0
    invalid_paper_citations = 0
    invalid_citations = 0

    # Cache for citations to add (batch graph actions for igraph)
    citations_to_add = []

    # Add citations to the graph
    for citing_id, citations in __citations_from_file(input_file, should_profile):
#.........这里部分代码省略.........