R - ウェブあなたの助けとあなたの検討のため

Question

まず、多くの感謝を掻き取り、

私はあなたが1つのウェブサイトになったら、このページhttp://ieeexplore.ieee.org/document/6875970/keywordsから何かを抽出したい、私は に関連する情報をスクラップウェブに興味を持っています - IEEEキーワード - INSPEC：制御されたインデックス - INSPEC：非制御インデックス

install.packages("rvest") 
library(rvest) 
keywords<-read_html("http://ieeexplore.ieee.org/document/6875970/keywords") 

keywords %>% 
a<-html_node(lego_movie,"text/javascript") %>% 
html_text(a) 

しかし、それは動作しません！

私を助けてもらえますか？

ありがとうございます！

Answer 1

# if you get an error while importing, just use install.packages('jsonlite') and /or install.packages('stringr') 
library(jsonlite) 
library(stringr) 

# the ieee doc url we're interested in 
url <- 'http://ieeexplore.ieee.org/document/6875970/keywords' 

# read the document as text, no HTML parsing at all 
doc <- readLines(url) 

# after inspecting it, we notice it's built with angular-js, 
# and the data we need to extract is defined as a single javascript variable. 

# so, we first find the id of the line which defines 
# the javascript variable containing the tags 
idx <- which(!is.na(str_match(doc, 'global.document.metadata='))) 

# get the line ;-) 
line <- doc[idx] 

# since it's a javascript variable, we need to massage it 
# a little to be able to read it as json. 
# step 1: remove the "var global.document.metadata=" part (everything before the actual json) 
line <- str_replace(line, '^[^{]*', '') 

# step 2: remove the trailing ';' symbol 
line <- str_replace(line, ';$', '') 

# now we can parse the json data 
df <- fromJSON(line) 

# and get the information we need 
df$keywords[df$keywords$type == 'IEEE Keywords',]$kwd[[1]] 

df$keywords[df$keywords$type == 'INSPEC: Controlled Indexing',]$kwd[[1]] 

df$keywords[df$keywords$type == 'INSPEC: Non-Controlled Indexing',]$kwd[[1]] 

サンプル出力：

[1] "CTC incident datasets"        
[2] "proactive spatiotemporal resource allocation"  
[3] "predictive visual analytics"     
[4] "community policing"        
[5] "law enforcement"