在R语言中创建支持多参数的文本缩略自定义函数的实现方法

前置依赖 #

处理字符串分割需要用到stringr包，运行前先加载：

# 未安装先执行安装
# install.packages("stringr")
library(stringr)

封装完成的可调用函数 #

text_abbrev <- function(text, 
                        custom_stopwords = c("the", "really", "truly", "very"),
                        min_abbrev_len = 5,
                        extract_first_sent = TRUE) {
  # 1. 提取每个段落首句（按需开启）
  if (extract_first_sent) {
    text <- unlist(lapply(text, function(x) str_split(x, "\\.", simplify = T)[1]))
  }
  # 2. 移除指定停用词
  text <- unlist(lapply(text, function(t) {
    words <- unlist(strsplit(t, " "))
    # 过滤空字符串和停用词，统一转小写匹配避免大小写误差
    words <- words[words != "" & !tolower(words) %in% tolower(custom_stopwords)]
    paste(words, collapse = " ")
  }))
  # 3. 长单词缩写，超出指定长度的部分替换为.
  regex_pattern <- paste0("(?<=\\w{", min_abbrev_len, "})\\w+")
  text <- gsub(regex_pattern, ".", text, perl = TRUE)
  # 4. 连续空格去重、清理首尾空格
  text <- gsub("^ *|(?<= ) | *$", "", text, perl = TRUE)
  return(text)
}

参数说明 #

• text: 输入待处理文本，支持单条文本或多段落文本向量
• custom_stopwords: 自定义停用词向量，默认内置通用程度副词停用词
• min_abbrev_len: 触发缩写的单词最小长度，默认值为5，即长度超过5的单词保留前5位后用.缩写
• extract_first_sent: 逻辑值，是否开启段落首句提取功能，默认开启

使用示例 #

# 测试文本
test_input <- c(
  "The really great advantage of R language is that it has very rich text processing packages. This part of the content will be removed after extracting the first sentence.",
  "This is a second paragraph with truly supercalifragilisticexpialidocious long words. The following content will not be retained."
)
# 调用函数
result <- text_abbrev(test_input)
# 查看输出
print(result)

输出结果参考：

[1] "great advant. of R langua. is that it has rich text proces. packa."
[2] "This is a second paragr. with super. long words"

优化说明 #

相对原有零散逻辑做了两处兼容优化：

• 停用词匹配改为不区分大小写，避免输入文本大小写不统一导致的停用词漏删
• 各功能可通过参数灵活开关，不需要的模块可直接关闭无需修改函数内部代码

内容的提问来源于stack exchange，提问作者Imran Luqman