我目前正在抓取本网站的数据。 www.allmusic.com
通过这种方式，我特意从每个艺术家页面中抓取音乐风格。
rm(list= ls())
library(xml2)
library(rvest)
library(magrittr)
library(stringr)
library(plyr)
library(dplyr)
library(tidyr)
library(knitr)
library(XML)
library(data.table)
setwd("~/AlbumScrape")
load("popalbums2004-2014.rdata")
pages <- popalbums2004-2014.rdata %>% slice(1:100) %>%
mutate(URL = paste0("http://www.allmusic.com/album/",album,"/credits"))
credits_tmp = NULL
albumAttrs_tmp = NULL
album_genre_tmp = NULL
album_style_tmp = NULL
album_mood_tmp = NULL
album_theme_tmp = NULL
pb = txtProgressBar(min = 0, max = nrow(pages), initial = 0, char = "*", style = 3)
system.time(
for(i in seq(nrow(pages))) {
return_value <- try({
test403 <- read_html(paste0("http://www.allmusic.com/album /",pages$album[i],"/credits"))
})
if(class(return_value) == "try-error"){
print("failed");
} else {
print("succeeded")
src_url <- pages$URL[i] %>% read_html()
src_artist <- pages$src_artist[i] %>%
as.character()
album <- pages$album[i] %>%
as.character()
credits_artistURL <- try_default(
expr =
src_url %>%
html_nodes(".artist a") %>%
html_attr("href") %>%
str_trim() %>%
gsub("http://www.allmusic.com/album/","",.),
default = "", quiet = T)
credit_artistRole <- try_default(
expr =
src_url %>%
html_nodes(".credit") %>%
html_text() %>%
str_trim(),
default = "",
quiet = TRUE)
album_duration <-
try_default(
expr =
src_url %>%
html_nodes(".duration span") %>%
html_text() %>%
str_trim(),
default = "",
quiet = TRUE)
album_releaseDate <-
try_default(
expr =
src_url %>%
html_nodes(".release-date span") %>%
html_text() %>%
str_trim(),
default = "",
quiet = TRUE)
album_recordingDate <-
try_default(
expr =
src_url %>%
html_nodes(".recording-date div") %>%
html_text() %>%
str_trim(),
default = "",
quiet = TRUE)
album_genre <-
try_default(
expr =
src_url %>%
html_nodes(".genre a") %>%
html_attr("href") %>%
str_trim() %>%
gsub("http://www.allmusic.com/genre/","",.),
default = "",
quiet = TRUE)
album_style <-
try_default(
expr =
src_url %>%
html_nodes(".styles a") %>%
html_attr("href") %>%
str_trim() %>%
gsub("http://www.allmusic.com/style/","",.),
default = "",
quiet = TRUE)
album_mood <-
try_default(
expr =
src_url %>%
html_nodes(".mood a") %>%
html_attr("href") %>%
str_trim() %>%
gsub("http://www.allmusic.com/mood/","",.),
default = "",
quiet = TRUE)
album_theme <-
try_default(
src_url %>%
html_nodes(".theme a") %>%
html_attr("href") %>%
str_trim() %>%
gsub("http://www.allmusic.com/theme/","",.),
default = "",
quiet = TRUE)
credits_res_tmp <- try_default(
expr =
data.frame(credits_artistURL = credits_artistURL,
credit_artistRole = credit_artistRole,
src_artist = src_artist, album = album),
default =
data.frame(credits_artistURL = "ERROR",
credit_artistRole = "ERROR",
src_artist = src_artist, album = album), quiet = T) %>%
filter(., credits_artistURL != "#")
data_credits_tmp = list(credits_tmp, credits_res_tmp)
credits_tmp <- rbindlist(data_credits_tmp)
albumAttrs_res_tmp <- try_default(
expr =
data.frame(album_duration = album_duration,
album_releaseDate = album_releaseDate,
src_artist = src_artist, album = album),
default =
data.frame(album_duration = "ERROR",
album_releaseDate = "ERROR",
src_artist = src_artist, album = album), quiet = T)
data_albumAttrs_tmp = list(albumAttrs_tmp, albumAttrs_res_tmp)
albumAttrs_tmp <- rbindlist (data_albumAttrs_tmp)
album_genre_res_tmp <- try_default(
expr =
data.frame(album_genre = album_genre,
src_artist = src_artist, album = album),
default =
data.frame(album_genre = "ERROR",
src_artist = src_artist, album = album), quiet = T)
data_album_genre_tmp = list(album_style_tmp, album_genre_res_tmp)
album_genre_tmp <- rbindlist(data_album_genre_tmp)
album_style_res_tmp <- try_default(
expr =
data.frame(album_style = album_style,
src_artist = src_artist, album = album),
default =
data.frame(album_style = "ERROR",
src_artist = src_artist, album = album), quiet = T)
data_album_style_tmp = list(album_style_tmp, album_style_res_tmp)
album_style_tmp <- rbindlist(data_album_style_tmp)
album_mood_res_tmp <- try_default(
expr =
data.frame(album_mood = album_mood,
src_artist = src_artist, album = album),
default =
data.frame(album_mood = "ERROR",
src_artist = src_artist, album = album), quiet = T)
data_album_mood_tmp = list(album_mood_tmp, album_mood_res_tmp)
album_mood_tmp <- rbindlist (data_album_mood_tmp)
album_theme_res_tmp <- try_default(
expr =
data.frame(album_theme = album_theme,
src_artist = src_artist, album = album),
default =
data.frame(album_theme = "ERROR",
src_artist = src_artist, album = album), quiet = T)
data_album_theme_tmp = list(album_theme_tmp, album_theme_res_tmp)
album_theme_tmp <- rbindlist(data_album_theme_tmp)
Sys.sleep(runif(1, 0.01, 0.05))
if(i %% 10 == 0) {setTxtProgressBar(pb, i)}
}}
)
这给了我一些结果，但是对于一些艺术家，我在样式字段中收到了一些错误。
我创建了一个单独的脚本，模仿第一个脚本，但只加载了一个只有获得错误的艺术家的保存数据框。
脚本就是这样。
rm(list= ls())
library(xml2)
library(rvest)
library(magrittr)
library(stringr)
library(plyr)
library(dplyr)
library(tidyr)
library(knitr)
library(XML)
library(data.table)
setwd("~/AlbumScrape")
load("style_errors.Rdata")
pages <- style_errors %>% slice(1:100) %>%
mutate(URL = paste0("http://www.allmusic.com/album/",album,"/credits"))
credits_tmp = NULL
albumAttrs_tmp = NULL
album_genre_tmp = NULL
album_style_tmp = NULL
album_mood_tmp = NULL
album_theme_tmp = NULL
pb = txtProgressBar(min = 0, max = nrow(pages), initial = 0, char = "*", style = 3)
system.time(
for(i in seq(nrow(pages))) {
return_value <- try({
test403 <- read_html(paste0("http://www.allmusic.com/album/",pages$album[i],"/credits"))
})
if(class(return_value) == "try-error"){
print("failed");
} else {
print("succeeded")
src_url <- pages$URL[i] %>% read_html()
src_artist <- pages$src_artist[i] %>%
as.character()
album <- pages$album[i] %>%
as.character()
album_style <-
try_default(
expr =
src_url %>%
html_nodes(".styles a") %>%
html_attr("href") %>%
str_trim() %>%
gsub("http://www.allmusic.com/style/","",.),
default = "",
quiet = TRUE)
album_style_res_tmp <- try_default(
expr =
data.frame(album_style = album_style,
src_artist = src_artist, album = album),
default =
data.frame(album_style = "ERROR",
src_artist = src_artist, album = album), quiet = T)
data_album_style_tmp = list(album_style_tmp, album_style_res_tmp)
album_style_tmp <- rbindlist(data_album_style_tmp)
Sys.sleep(runif(1, 0.01, 0.05))
if(i %% 10 == 0) {setTxtProgressBar(pb, i)}
}}
)
然而，在使用它之后，我仍然遇到了错误。
1:In if (class(return_value) == "try-error") { ... :
the condition has length > 1 and only the first element will be used
2: In if (class(return_value) == "try-error") { ... : etc etc
我认为这个问题来自于某些网页的论据，例如http://www.allmusic.com/album/ultimate-darkness-mw0000481576/credits，其中有多个字段，因此在我的逻辑中创建了超过1个单位长度。
我不确定如何更改我的代码以包含这些因素，因此我可以收集样式字段。
答案 0 :(得分：0)
返回对象有多个类。当您使用if ()时，括号中的表达式需要计算为一个逻辑长度（或可以强制为一个的长度）。如果班级的长度> 1然后你得到一个长度> gt的逻辑向量; 1，这会导致您看到的警告。例如：
foo <- 1:10
class(foo)
class(foo) <- c("myclass", class(foo))
class(foo)
class(foo) == "myclass"
> class(foo)
[1] "integer"
> class(foo) <- c("myclass", class(foo))
> class(foo)
[1] "myclass" "integer"
> class(foo) == "myclass"
[1] TRUE FALSE
> if(class(foo) == "myclass") {
+ print("Yay!")
+ }
[1] "Yay!"
Warning message:
In if (class(foo) == "myclass") { :
the condition has length > 1 and only the first element will be used
测试类成员身份的正确方法是通过继承类。这是使用inherits()函数完成的：
> if(inherits(foo, "myclass")) {
+ print("Yay!")
+ }
[1] "Yay!"