可视化解析树结构

时间:2015-11-02 08:05:43

标签: r nlp opennlp

我想将 openNLP 中的解析(POS标记)显示为树结构可视化。下面我提供了来自 openNLP 的解析树,但我无法将其绘制为Python's parsing共有的可视树。

install.packages(
    "http://datacube.wu.ac.at/src/contrib/openNLPmodels.en_1.5-1.tar.gz",  
    repos=NULL, 
    type="source"
)

library(NLP)
library(openNLP)

x <- 'Scroll bar does not work the best either.'
s <- as.String(x)

## Annotators
sent_token_annotator <- Maxent_Sent_Token_Annotator()
word_token_annotator <- Maxent_Word_Token_Annotator()
parse_annotator <- Parse_Annotator()

a2 <- annotate(s, list(sent_token_annotator, word_token_annotator))
p <- parse_annotator(s, a2)
ptext <- sapply(p$features, `[[`, "parse")
ptext
Tree_parse(ptext)

## > ptext
## [1] "(TOP (S (NP (NNP Scroll) (NN bar)) (VP (VBZ does) (RB not) (VP (VB work) (NP (DT the) (JJS best)) (ADVP (RB either))))(. .)))"
## > Tree_parse(ptext)
## (TOP
##   (S
##     (NP (NNP Scroll) (NN bar))
##     (VP (VBZ does) (RB not) (VP (VB work) (NP (DT the) (JJS best)) (ADVP (RB either))))
##     (. .)))

树结构看起来应该类似于:

enter image description here

有没有办法显示这个树形象?

我发现了this related tree viz问题,用于绘制可能有用的数字表达式,但我无法概括为句子解析可视化。

1 个答案:

答案 0 :(得分:10)

这是igraph版本。此函数将Parse_annotator的结果作为输入,因此示例中为ptextNLP::Tree_parse已经创建了一个漂亮的树结构,因此这里的想法是递归遍历它并创建一个边缘列表以插入igraph。边缘列表只是头部 - 尾部值的2列矩阵。

为了使igraph在适当的节点之间创建边缘,它们需要具有唯一标识符。我通过在使用regmatches<-之前将一系列整数(使用Tree_parse)附加到文本中的单词来完成此操作。

内部函数edgemaker遍历树,填充edgelist。可以选择将叶子与其余节点分开着色,但是如果您传递选项vertex.label.color,它将为它们着色所有相同的颜色。

## Make a graph from Tree_parse result
parse2graph <- function(ptext, leaf.color='chartreuse4', label.color='blue4',
                        title=NULL, cex.main=.9, ...) {
    stopifnot(require(NLP) && require(igraph))

    ## Replace words with unique versions
    ms <- gregexpr("[^() ]+", ptext)                                      # just ignoring spaces and brackets?
    words <- regmatches(ptext, ms)[[1]]                                   # just words
    regmatches(ptext, ms) <- list(paste0(words, seq.int(length(words))))  # add id to words

    ## Going to construct an edgelist and pass that to igraph
    ## allocate here since we know the size (number of nodes - 1) and -1 more to exclude 'TOP'
    edgelist <- matrix('', nrow=length(words)-2, ncol=2)

    ## Function to fill in edgelist in place
    edgemaker <- (function() {
        i <- 0                                       # row counter
        g <- function(node) {                        # the recursive function
            if (inherits(node, "Tree")) {            # only recurse subtrees
                if ((val <- node$value) != 'TOP1') { # skip 'TOP' node (added '1' above)
                    for (child in node$children) {
                        childval <- if(inherits(child, "Tree")) child$value else child
                        i <<- i+1
                        edgelist[i,1:2] <<- c(val, childval)
                    }
                }
                invisible(lapply(node$children, g))
            }
        }
    })()

    ## Create the edgelist from the parse tree
    edgemaker(Tree_parse(ptext))

    ## Make the graph, add options for coloring leaves separately
    g <- graph_from_edgelist(edgelist)
    vertex_attr(g, 'label.color') <- label.color  # non-leaf colors
    vertex_attr(g, 'label.color', V(g)[!degree(g, mode='out')]) <- leaf.color
    V(g)$label <- sub("\\d+", '', V(g)$name)      # remove the numbers for labels
    plot(g, layout=layout.reingold.tilford, ...)
    if (!missing(title)) title(title, cex.main=cex.main)
}

因此,使用您的示例,字符串x及其带注释的版本ptext,看起来像

x <- 'Scroll bar does not work the best either.'
ptext
# [1] "(TOP (S (NP (NNP Scroll) (NN bar)) (VP (VBZ does) (RB not) (VP (VB work) (NP (DT the) (JJS best)) (ADVP (RB either))))(. .)))"

通过调用

创建图表
library(igraph)
library(NLP)

parse2graph(ptext,  # plus optional graphing parameters
            title = sprintf("'%s'", x), margin=-0.05,
            vertex.color=NA, vertex.frame.color=NA,
            vertex.label.font=2, vertex.label.cex=1.5, asp=0.5,
            edge.width=1.5, edge.color='black', edge.arrow.size=0)

enter image description here