如何使用R从.docx文件提取指定纯文本（剔除摘要、图表等）

Hey there, let's break down how to pull out the Introduction, Methods, Results, and Conclusion sections from your academic .docx papers—while ditching abstracts, references, and figures/tables. I'll start with a solid R solution, then cover some no-code online options you might not have tried yet.

R Language Solution: Precise Section Extraction #

This approach relies on the officer and docxtractr packages, purpose-built for handling Word documents in R. We'll also use stringr for text filtering.

Step 1: Install & Load Required Packages #

First, get the tools set up:

install.packages(c("officer", "docxtractr", "stringr"))
library(officer)
library(docxtractr)
library(stringr)

Step 2: Load Your Document & Extract All Text #

Read in your .docx file and pull out every paragraph (we'll track their positions to target sections):

# Replace with your actual file path
doc <- read_docx("your_academic_paper.docx")

# Extract all text, preserving paragraph order
all_paragraphs <- docx_extract_all_text(doc)

Step 3: Target Sections & Filter Out Unwanted Content #

Academic papers usually follow consistent heading structures—we'll use that to pinpoint our desired sections. Adjust the keywords below to match your paper's actual headings (e.g., if your Methods section is labeled "材料与方法", update that):

# Define sections we want to keep
target_headings <- c("引言", "方法", "结果", "结论")
# Define sections we want to exclude
exclude_headings <- c("摘要", "参考文献")

# Find the starting index of each target section
start_indices <- sapply(target_headings, function(heading) {
  grep(str_c("^", heading), all_paragraphs)  # Matches headings at the start of a line
})

# Find the start of the first excluded section after Conclusion (usually References)
ref_start <- if ("参考文献" %in% exclude_headings) {
  grep("^参考文献", all_paragraphs)
} else length(all_paragraphs)

# Extract each section
introduction <- all_paragraphs[start_indices["引言"]:(start_indices["方法"] - 1)]
methods <- all_paragraphs[start_indices["方法"]:(start_indices["结果"] - 1)]
results <- all_paragraphs[start_indices["结果"]:(start_indices["结论"] - 1)]
conclusion <- all_paragraphs[start_indices["结论"]:(ref_start - 1)]

# Filter out figure/table captions (matches lines starting with "图X" or "表X")
clean_section <- function(text) {
  text[!str_detect(text, "^(图|表)\\d+")]
}

# Apply cleaning to all sections
intro_clean <- clean_section(introduction)
methods_clean <- clean_section(methods)
results_clean <- clean_section(results)
conclusion_clean <- clean_section(conclusion)

Step 4: Save the Extracted Text #

Combine the cleaned sections and save to a plain text file:

final_text <- c(intro_clean, methods_clean, results_clean, conclusion_clean)
writeLines(final_text, "extracted_academic_content.txt")

Note: If your headings include numbers (e.g., "1 引言"), update the regex in grep() to str_c("^\\d+\\s+", heading) to match numbered headings.

No-Code Online Alternatives #

If coding isn't your vibe, here are a few workarounds:

• Google Docs: Upload your .docx, then use "Find and Replace" (with regex) to delete abstracts/references. For figures/tables, select all images via Ctrl+Alt+Shift+I (Windows) or Cmd+Option+Shift+I (Mac) and delete them in bulk. Then copy the remaining text.
• LLM File Upload: Upload your .docx directly to tools like ChatGPT with a prompt like: "Extract only the Introduction, Methods, Results, and Conclusion sections from this academic paper. Remove the abstract, references, all figure/table captions, and any images." This works great for single papers.
• Adobe Acrobat Online: Convert your .docx to PDF, use the "Extract Text" tool, then paste the text into a tool like Notepad++ (online version) and use regex to isolate your target sections.

内容的提问来源于stack exchange，提问作者user9650316