如何从多级列表标题PDF提取最低级标题及对应文本并导入Excel？

Approach Recommendation #

Regex is a reliable choice for this task, especially since your headings follow a numbered hierarchical structure. It’s straightforward to match the pattern of lowest-level headings and capture their associated content. However, if the PDF conversion introduced inconsistent formatting (like line breaks in headings), a PDF-specific parser like pdfplumber might yield better results.

Regex-Based Solution (Using Converted Text File) #

This script will parse your output.txt file, identify the deepest (lowest-level) numbered headings, extract their adjacent text, and save the pairs to a CSV file.

Step 1: Python Code #

import re
import csv

# Read the converted text file
with open('output.txt', 'r', encoding='utf-8') as f:
    text = f.read()

# Regex pattern to match numbered headings (e.g., 1.2.3) and their subsequent content
pattern = r'^(\d+(\.\d+)+)\s+(.+?)\n([\s\S]*?)(?=^\d+(\.\d+)+|$)'
matches = re.findall(pattern, text, flags=re.MULTILINE)

# Process matches to get heading title, content, and hierarchy depth
entries = []
for match in matches:
    heading_num, _, heading_title, content, _ = match
    # Calculate depth (number of levels in the heading)
    depth = len(heading_num.split('.'))
    # Clean up content: remove extra newlines and whitespace
    cleaned_content = re.sub(r'\s+', ' ', content).strip()
    entries.append((depth, heading_title.strip(), cleaned_content))

if not entries:
    print("No numbered headings found in the text file.")
    exit()

# Identify the lowest level (maximum depth)
max_depth = max(entry[0] for entry in entries)

# Filter only entries from the lowest level
lowest_level_entries = [entry[1:] for entry in entries if entry[0] == max_depth]

# Write to CSV (columns will map to Excel's C and D)
with open('output.csv', 'w', newline='', encoding='utf-8') as csvfile:
    writer = csv.writer(csvfile)
    # Optional header (Excel will ignore this if you start importing at column C)
    writer.writerow(['Heading', 'Text'])
    for title, content in lowest_level_entries:
        writer.writerow([title, content])

print(f"Successfully exported {len(lowest_level_entries)} lowest-level entries to output.csv")

Key Details #

• Regex Pattern: Matches headings with numeric hierarchy (e.g., 1.2.3), captures the heading text, then grabs all content until the next heading or end of file.
• Depth Calculation: Determines the lowest level by counting the number of dots in the heading number (more dots = deeper level).
• Content Cleaning: Normalizes whitespace to ensure clean text in Excel.

Alternative: PDF Parser Approach (No Pre-Conversion Needed) #

If the text conversion introduced formatting issues, use pdfplumber to directly extract structured content from the PDF. This method relies on font size or indentation to identify headings.

Step 1: Install Dependencies #

pip install pdfplumber

Step 2: Python Code #

import pdfplumber
import csv

lowest_level_entries = []

# Adjust these values to match your PDF's formatting
HEADING_FONT_SIZE = 12  # Font size of lowest-level headings
CONTENT_FONT_SIZE = 10  # Font size of body text

with pdfplumber.open('doc.pdf') as pdf:
    for page in pdf.pages:
        # Extract words with font size metadata
        words = page.extract_words(extra_attrs=['fontsize'])
        current_heading = None
        current_content = []
        
        for word in words:
            # Check if the word belongs to a lowest-level heading
            if word['fontsize'] == HEADING_FONT_SIZE:
                # Save the previous heading-content pair if exists
                if current_heading is not None:
                    cleaned_content = ' '.join(current_content).strip()
                    lowest_level_entries.append((current_heading, cleaned_content))
                current_heading = word['text']
                current_content = []
            # Add words to content if we're in a heading's section
            elif current_heading is not None and word['fontsize'] == CONTENT_FONT_SIZE:
                current_content.append(word['text'])
        
        # Save the last entry on the page
        if current_heading is not None:
            cleaned_content = ' '.join(current_content).strip()
            lowest_level_entries.append((current_heading, cleaned_content))

# Write to CSV
with open('output.csv', 'w', newline='', encoding='utf-8') as csvfile:
    writer = csv.writer(csvfile)
    writer.writerow(['Heading', 'Text'])
    for title, content in lowest_level_entries:
        writer.writerow([title, content])

print(f"Successfully exported {len(lowest_level_entries)} entries to output.csv")

Key Details #

• Font Size Detection: Adjust HEADING_FONT_SIZE and CONTENT_FONT_SIZE to match your PDF’s actual formatting (use pdfplumber to inspect font sizes if unsure).
• Structural Preservation: Directly reads the PDF, avoiding issues from text conversion like broken headings.

Excel Import Tips #

When importing the CSV into Excel:

• Go to Data > From Text/CSV.
• Select output.csv and click Import.
• In the wizard, choose Delimited and set the delimiter to comma.
• Under Data preview, use the arrow buttons to skip columns A and B (so the first column maps to C, second to D).
• Click Load to complete the import.

内容的提问来源于stack exchange，提问作者wuddadid