抓取“元素周期表”的维基页面和所有链接

发布于 2024-10-06 13:56:44 字数 664 浏览 0 评论 0原文

我希望抓取以下维基文章： http://en.wikipedia.org/wiki/Periodic_table

因此，我的 R 代码的输出将是一个包含以下列的表格：

化学元素简称
化学元素全名化学
元素 wiki 页面的 URL

（显然，每个化学元素都有一行）

我正在尝试使用 XML 包获取页面内的值，但似乎停留在开头，所以我希望有一个有关如何执行此操作的示例（和/或相关示例的链接）

library(XML)
base_url<-"http://en.wikipedia.org/wiki/Periodic_table"
base_html<-getURLContent(base_url)[[1]]
parsed_html <- htmlTreeParse(base_html, useInternalNodes = TRUE)
xmlChildren(parsed_html)
getNodeSet(parsed_html, "//html", c(x = base_url))
[[1]]
attr(,"class")
[1] "XMLNodeSet"

原文

I wish to scrape the following wiki article: http://en.wikipedia.org/wiki/Periodic_table

So that the output of my R code will be a table with the following columns:

Chemical elements short name
Chemical elements full name
The URL to the chemical elements wiki page

(and with a row for each chemical element, obviously)

I am trying to get to the values inside the page using the XML package, but seems to be stuck in the beginning, so I'd appreciate an example on how to do it (and/or links to relevant examples)

library(XML)
base_url<-"http://en.wikipedia.org/wiki/Periodic_table"
base_html<-getURLContent(base_url)[[1]]
parsed_html <- htmlTreeParse(base_html, useInternalNodes = TRUE)
xmlChildren(parsed_html)
getNodeSet(parsed_html, "//html", c(x = base_url))
[[1]]
attr(,"class")
[1] "XMLNodeSet"

分享到QQ

分享到微博

如果你对这篇内容有疑问，欢迎到本站社区发帖提问参与讨论，获取更多帮助，或者扫码二维码加入 Web 技术交流群。

发布评论

需要登录才能够评论，你可以免费注册一个本站的账号。

千笙结 2024-10-13 13:56:46

试试这个：

library(XML)

URL <- "http://en.wikipedia.org/wiki/Periodic_table"
root <- htmlTreeParse(URL, useInternalNodes = TRUE)

# extract attributes and value of all 'a' tags within 3rd table
f <- function(x) c(xmlAttrs(x), xmlValue(x))
m1 <- xpathApply(root, "//table[3]//a", f)
m2 <- suppressWarnings(do.call(rbind, m1))

# extract rows that correspond to chemical symbols
ix <- grep("^[[:upper:]][[:lower:]]{0,2}", m2[, "class"])

m3 <- m2[ix, 1:3]
colnames(m3) <- c("URL", "Name", "Symbol")
m3[,1] <- sub("^", "http://en.wikipedia.org", m3[,1])
m3[,2] <- sub(" .*", "", m3[,2])

一些输出：

> dim(m3)
[1] 118   3
> head(m3)
     URL                                      Name        Symbol
[1,] "http://en.wikipedia.org/wiki/Hydrogen"  "Hydrogen"  "H"   
[2,] "http://en.wikipedia.org/wiki/Helium"    "Helium"    "He"  
[3,] "http://en.wikipedia.org/wiki/Lithium"   "Lithium"   "Li"  
[4,] "http://en.wikipedia.org/wiki/Beryllium" "Beryllium" "Be"  
[5,] "http://en.wikipedia.org/wiki/Boron"     "Boron"     "B"   
[6,] "http://en.wikipedia.org/wiki/Carbon"    "Carbon"    "C"

我们可以通过从 Jeffrey 的 xpath 表达式（因为它几乎获取顶部的元素）开始进一步增强 xpath 表达式并为其添加精确的限定条件来使其更加紧凑。在这种情况下，可以使用 xpathSApply 来消除对 do.call 或 plyr 包的需要。我们解决零碎问题的最后一点与之前相同。这会产生一个矩阵而不是一个数据框，这似乎更可取，因为内容完全是字符。

library(XML)

URL <- "http://en.wikipedia.org/wiki/Periodic_table"
root <- htmlTreeParse(URL, useInternalNodes = TRUE)

# extract attributes and value of all a tags within 3rd table
f <- function(x) c(xmlAttrs(x), xmlValue(x))
M <- t(xpathSApply(root, "//table[3]/tr/td/a[.!='']", f))[1:118,]

# nicer column names, fix up URLs, fix up Mercury.
colnames(M) <- c("URL", "Name", "Symbol")
M[,1] <- sub("^", "http://en.wikipedia.org", M[,1])
M[,2] <- sub(" .*", "", M[,2])

View(M)

Try this:

library(XML)

URL <- "http://en.wikipedia.org/wiki/Periodic_table"
root <- htmlTreeParse(URL, useInternalNodes = TRUE)

# extract attributes and value of all 'a' tags within 3rd table
f <- function(x) c(xmlAttrs(x), xmlValue(x))
m1 <- xpathApply(root, "//table[3]//a", f)
m2 <- suppressWarnings(do.call(rbind, m1))

# extract rows that correspond to chemical symbols
ix <- grep("^[[:upper:]][[:lower:]]{0,2}", m2[, "class"])

m3 <- m2[ix, 1:3]
colnames(m3) <- c("URL", "Name", "Symbol")
m3[,1] <- sub("^", "http://en.wikipedia.org", m3[,1])
m3[,2] <- sub(" .*", "", m3[,2])

A bit of the output:

> dim(m3)
[1] 118   3
> head(m3)
     URL                                      Name        Symbol
[1,] "http://en.wikipedia.org/wiki/Hydrogen"  "Hydrogen"  "H"   
[2,] "http://en.wikipedia.org/wiki/Helium"    "Helium"    "He"  
[3,] "http://en.wikipedia.org/wiki/Lithium"   "Lithium"   "Li"  
[4,] "http://en.wikipedia.org/wiki/Beryllium" "Beryllium" "Be"  
[5,] "http://en.wikipedia.org/wiki/Boron"     "Boron"     "B"   
[6,] "http://en.wikipedia.org/wiki/Carbon"    "Carbon"    "C"

We can make this more compact by enhancing the xpath expression further starting with Jeffrey's xpath expression (since it nearly gets the elements at top) and adding a qualification to it which exactly does. In that case xpathSApply can be used to eliminate the need for do.call or the plyr package. The last bit where we fix up odds and ends is the same as before. This produces a matrix rather than a data frame which seems preferable since the content is entirely character.

library(XML)

URL <- "http://en.wikipedia.org/wiki/Periodic_table"
root <- htmlTreeParse(URL, useInternalNodes = TRUE)

# extract attributes and value of all a tags within 3rd table
f <- function(x) c(xmlAttrs(x), xmlValue(x))
M <- t(xpathSApply(root, "//table[3]/tr/td/a[.!='']", f))[1:118,]

# nicer column names, fix up URLs, fix up Mercury.
colnames(M) <- c("URL", "Name", "Symbol")
M[,1] <- sub("^", "http://en.wikipedia.org", M[,1])
M[,2] <- sub(" .*", "", M[,2])

View(M)

回复收藏 0 原文

枕梦 2024-10-13 13:56:46

Tal——我以为这会很容易。我打算向您指出 readHTMLTable()，这是 XML 包中我最喜欢的函数。哎呀，它的帮助页面甚至显示了抓取维基百科页面的示例！

但可惜，这不是您想要的：

library(XML)
url = 'http://en.wikipedia.org/wiki/Periodic_table'
tables = readHTMLTable(html)

# ... look through the list to find the one you want...

table = tables[3]
table
Tal——我以为这会很容易。我打算向您指出 readHTMLTable()，这是 XML 包中我最喜欢的函数。哎呀，它的帮助页面甚至显示了抓取维基百科页面的示例！
但可惜，这不是您想要的：
NULL`
         Group #    1    2    3     4     5     6     7     8     9    10    11    12     13     14     15     16     17     18
1         Period      <NA> <NA>  <NA>  <NA>  <NA>  <NA>  <NA>  <NA>  <NA>  <NA>  <NA>   <NA>   <NA>   <NA>   <NA>   <NA>   <NA>
2              1   1H       2He  <NA>  <NA>  <NA>  <NA>  <NA>  <NA>  <NA>  <NA>  <NA>   <NA>   <NA>   <NA>   <NA>   <NA>   <NA>
3              2  3Li  4Be         5B    6C    7N    8O    9F  10Ne  <NA>  <NA>  <NA>   <NA>   <NA>   <NA>   <NA>   <NA>   <NA>
4              3 11Na 12Mg       13Al  14Si   15P   16S  17Cl  18Ar  <NA>  <NA>  <NA>   <NA>   <NA>   <NA>   <NA>   <NA>   <NA>
5              4  19K 20Ca 21Sc  22Ti   23V  24Cr  25Mn  26Fe  27Co  28Ni  29Cu  30Zn   31Ga   32Ge   33As   34Se   35Br   36Kr
6              5 37Rb 38Sr  39Y  40Zr  41Nb  42Mo  43Tc  44Ru  45Rh  46Pd  47Ag  48Cd   49In   50Sn   51Sb   52Te    53I   54Xe
7              6 55Cs 56Ba    *  72Hf  73Ta   74W  75Re  76Os  77Ir  78Pt  79Au  80Hg   81Tl   82Pb   83Bi   84Po   85At   86Rn
8              7 87Fr 88Ra   ** 104Rf 105Db 106Sg 107Bh 108Hs 109Mt 110Ds 111Rg 112Cn 113Uut 114Uuq 115Uup 116Uuh 117Uus 118Uuo
9                <NA> <NA> <NA>  <NA>  <NA>  <NA>  <NA>  <NA>  <NA>  <NA>  <NA>  <NA>   <NA>   <NA>   <NA>   <NA>   <NA>   <NA>
10 * Lanthanoids 57La 58Ce 59Pr  60Nd  61Pm  62Sm  63Eu  64Gd  65Tb  66Dy  67Ho  68Er   69Tm   70Yb   71Lu          <NA>   <NA>
11  ** Actinoids 89Ac 90Th 91Pa   92U  93Np  94Pu  95Am  96Cm  97Bk  98Cf  99Es 100Fm  101Md  102No  103Lr          <NA>   <NA>

名称消失了，原子序数变成了符号。

回到绘图板...

我的 DOM 行走能力不是很强，所以这不太漂亮。它获取表格单元格中的每个链接，仅保留具有“标题”属性的链接（即符号所在的位置），并将您想要的内容粘贴到 data.frame 中。它也获取页面上的所有其他此类链接，但我们很幸运，元素是前 118 个此类链接：

library(XML)
library(plyr) 

url = 'http://en.wikipedia.org/wiki/Periodic_table'

# don't forget to parse the HTML, doh!

doc = htmlParse(url)

# get every link in a table cell:

links = getNodeSet(doc, '//table/tr/td/a')

# make a data.frame for each node with non-blank text, link, and 'title' attribute:

df = ldply(links, function(x) {
            text = xmlValue(x)
            if (text=='') text=NULL

            symbol = xmlGetAttr(x, 'title')
            link = xmlGetAttr(x, 'href')
            if (!is.null(text) & !is.null(symbol) & !is.null(link))
                data.frame(symbol, text, link)
        } )

# only keep the actual elements -- we're lucky they're first!

df = head(df, 118)

head(df)
     symbol text            link
1  Hydrogen    H  /wiki/Hydrogen
2    Helium   He    /wiki/Helium
3   Lithium   Li   /wiki/Lithium
4 Beryllium   Be /wiki/Beryllium
5     Boron    B     /wiki/Boron
6    Carbon    C    /wiki/Carbon

Tal -- I thought this was going to be easy. I was going to point you to readHTMLTable(), my favorite function in the XML package. Heck, its help page even shows an example of scraping a Wikipedia page!

But alas, this is not what you want:

library(XML)
url = 'http://en.wikipedia.org/wiki/Periodic_table'
tables = readHTMLTable(html)

# ... look through the list to find the one you want...

table = tables[3]
table
Tal -- I thought this was going to be easy. I was going to point you to readHTMLTable(), my favorite function in the XML package. Heck, its help page even shows an example of scraping a Wikipedia page!
But alas, this is not what you want:
NULL`
         Group #    1    2    3     4     5     6     7     8     9    10    11    12     13     14     15     16     17     18
1         Period      <NA> <NA>  <NA>  <NA>  <NA>  <NA>  <NA>  <NA>  <NA>  <NA>  <NA>   <NA>   <NA>   <NA>   <NA>   <NA>   <NA>
2              1   1H       2He  <NA>  <NA>  <NA>  <NA>  <NA>  <NA>  <NA>  <NA>  <NA>   <NA>   <NA>   <NA>   <NA>   <NA>   <NA>
3              2  3Li  4Be         5B    6C    7N    8O    9F  10Ne  <NA>  <NA>  <NA>   <NA>   <NA>   <NA>   <NA>   <NA>   <NA>
4              3 11Na 12Mg       13Al  14Si   15P   16S  17Cl  18Ar  <NA>  <NA>  <NA>   <NA>   <NA>   <NA>   <NA>   <NA>   <NA>
5              4  19K 20Ca 21Sc  22Ti   23V  24Cr  25Mn  26Fe  27Co  28Ni  29Cu  30Zn   31Ga   32Ge   33As   34Se   35Br   36Kr
6              5 37Rb 38Sr  39Y  40Zr  41Nb  42Mo  43Tc  44Ru  45Rh  46Pd  47Ag  48Cd   49In   50Sn   51Sb   52Te    53I   54Xe
7              6 55Cs 56Ba    *  72Hf  73Ta   74W  75Re  76Os  77Ir  78Pt  79Au  80Hg   81Tl   82Pb   83Bi   84Po   85At   86Rn
8              7 87Fr 88Ra   ** 104Rf 105Db 106Sg 107Bh 108Hs 109Mt 110Ds 111Rg 112Cn 113Uut 114Uuq 115Uup 116Uuh 117Uus 118Uuo
9                <NA> <NA> <NA>  <NA>  <NA>  <NA>  <NA>  <NA>  <NA>  <NA>  <NA>  <NA>   <NA>   <NA>   <NA>   <NA>   <NA>   <NA>
10 * Lanthanoids 57La 58Ce 59Pr  60Nd  61Pm  62Sm  63Eu  64Gd  65Tb  66Dy  67Ho  68Er   69Tm   70Yb   71Lu          <NA>   <NA>
11  ** Actinoids 89Ac 90Th 91Pa   92U  93Np  94Pu  95Am  96Cm  97Bk  98Cf  99Es 100Fm  101Md  102No  103Lr          <NA>   <NA>

The names are gone and the atomic number runs into the symbol.

So back to the drawing board...

My DOM walking-fu is not very strong, so this isn't pretty. It gets every link in a table cell, only keeps those with a "title" attribute (that's where the symbol is), and sticks what you want in a data.frame. It gets every other such link on the page, too, but we're lucky and the elements are the first 118 such links:

library(XML)
library(plyr) 

url = 'http://en.wikipedia.org/wiki/Periodic_table'

# don't forget to parse the HTML, doh!

doc = htmlParse(url)

# get every link in a table cell:

links = getNodeSet(doc, '//table/tr/td/a')

# make a data.frame for each node with non-blank text, link, and 'title' attribute:

df = ldply(links, function(x) {
            text = xmlValue(x)
            if (text=='') text=NULL

            symbol = xmlGetAttr(x, 'title')
            link = xmlGetAttr(x, 'href')
            if (!is.null(text) & !is.null(symbol) & !is.null(link))
                data.frame(symbol, text, link)
        } )

# only keep the actual elements -- we're lucky they're first!

df = head(df, 118)

head(df)
     symbol text            link
1  Hydrogen    H  /wiki/Hydrogen
2    Helium   He    /wiki/Helium
3   Lithium   Li   /wiki/Lithium
4 Beryllium   Be /wiki/Beryllium
5     Boron    B     /wiki/Boron
6    Carbon    C    /wiki/Carbon

回复收藏 0 原文

命硬 2024-10-13 13:56:46

您必须抓取维基百科吗？您可以改为针对 Wikidata 运行此 SPARQL 查询 (结果)：

SELECT
  ?elementLabel
  ?symbol
  ?article
WHERE
{
  ?element wdt:P31 wd:Q11344;
           wdt:P1086 ?n;
           wdt:P246 ?symbol.
  OPTIONAL {
    ?article schema:about ?element;
             schema:inLanguage "en";
             schema:isPartOf <https://en.wikipedia.org/>.
  }
  FILTER (?n >= 1 && ?n <= 118).
  SERVICE wikibase:label { bd:serviceParam wikibase:language "en" . }
}
ORDER BY ?n

很抱歉，如果这不能直接回答您的问题，但这应该可以帮助人们以干净的方式获取相同的信息。

Do you have to scrape Wikipedia? You can run this SPARQL query against Wikidata instead (results):

SELECT
  ?elementLabel
  ?symbol
  ?article
WHERE
{
  ?element wdt:P31 wd:Q11344;
           wdt:P1086 ?n;
           wdt:P246 ?symbol.
  OPTIONAL {
    ?article schema:about ?element;
             schema:inLanguage "en";
             schema:isPartOf <https://en.wikipedia.org/>.
  }
  FILTER (?n >= 1 && ?n <= 118).
  SERVICE wikibase:label { bd:serviceParam wikibase:language "en" . }
}
ORDER BY ?n

Sorry if this doesn't answer your question directly but this should help people looking to scrape the same information but in a clean manner.

回复收藏 0 原文

~没有更多了~