<repositories>
    <repository>
        <id>com.e-iceblue</id>
        <name>e-iceblue</name>
        <url>https://repo.e-iceblue.com/nexus/content/groups/public/</url>
    </repository>
</repositories>
<dependencies>
    <dependency>
        <groupId>e-iceblue</groupId>
        <artifactId>spire.doc</artifactId>
        <version>14.4.9</version>
    </dependency>
</dependencies>

import com.spire.doc.Document;

public class WordCountExample {
    public static void main(String[] args) {
        // 加載 Word 文檔
        Document doc = new Document();
        doc.loadFromFile("SampleDocument.docx");

        // 獲取文檔中的文本內(nèi)容
        String text = doc.getText();

        // 統(tǒng)計(jì)單詞數(shù)量
        int wordCount = countWords(text);

        System.out.println("文檔中的單詞數(shù)量: " + wordCount);

        doc.close();
    }

    /**
     * 統(tǒng)計(jì)文本中的單詞數(shù)量（針對(duì)英文）
     * @param text 輸入文本
     * @return 單詞數(shù)量
     */
    private static int countWords(String text) {
        if (text == null || text.trim().isEmpty()) {
            return 0;
        }
        // 按空格、換行、標(biāo)點(diǎn)符號(hào)等分隔符進(jìn)行拆分
        String[] words = text.trim().split("[\\s\\p{Punct}]+");
        return words.length;
    }
}

import com.spire.doc.Document;
import com.spire.doc.documents.Paragraph;

public class ParagraphWordCountExample {
    public static void main(String[] args) {
        Document doc = new Document();
        doc.loadFromFile("SampleDocument.docx");

        int totalWordCount = 0;
        int paragraphIndex = 1;

        // 遍歷文檔中的所有段落
        for (Object obj : doc.getSections()) {
            com.spire.doc.Section section = (com.spire.doc.Section) obj;
            for (Object paraObj : section.getParagraphs()) {
                Paragraph paragraph = (Paragraph) paraObj;
                String text = paragraph.getText();
                int wordCount = countWords(text);
                totalWordCount += wordCount;
                System.out.println("段落 " + paragraphIndex + " 單詞數(shù): " + wordCount);
                paragraphIndex++;
            }
        }

        System.out.println("文檔總單詞數(shù): " + totalWordCount);
        doc.close();
    }

    private static int countWords(String text) {
        if (text == null || text.trim().isEmpty()) {
            return 0;
        }
        String[] words = text.trim().split("[\\s\\p{Punct}]+");
        return words.length;
    }
}

import com.spire.doc.Document;
import com.spire.doc.Table;
import com.spire.doc.documents.Paragraph;

public class FullWordCountExample {
    public static void main(String[] args) {
        Document doc = new Document();
        doc.loadFromFile("DocumentWithTable.docx");

        int totalWordCount = 0;

        // 統(tǒng)計(jì)段落中的單詞
        for (Object obj : doc.getSections()) {
            com.spire.doc.Section section = (com.spire.doc.Section) obj;
            for (Object paraObj : section.getParagraphs()) {
                Paragraph paragraph = (Paragraph) paraObj;
                totalWordCount += countWords(paragraph.getText());
            }
        }

        // 統(tǒng)計(jì)表格中的單詞
        for (Object tableObj : doc.getTables()) {
            Table table = (Table) tableObj;
            for (Object rowObj : table.getRows()) {
                com.spire.doc.TableRow row = (com.spire.doc.TableRow) rowObj;
                for (Object cellObj : row.getCells()) {
                    com.spire.doc.TableCell cell = (com.spire.doc.TableCell) cellObj;
                    for (Object paraObj : cell.getParagraphs()) {
                        Paragraph paragraph = (Paragraph) paraObj;
                        totalWordCount += countWords(paragraph.getText());
                    }
                }
            }
        }

        System.out.println("文檔總單詞數(shù)（包含表格）: " + totalWordCount);
        doc.close();
    }

    private static int countWords(String text) {
        if (text == null || text.trim().isEmpty()) {
            return 0;
        }
        String[] words = text.trim().split("[\\s\\p{Punct}]+");
        return words.length;
    }
}

import com.spire.doc.Document;
import com.spire.doc.Section;

public class SectionWordCountExample {
    public static void main(String[] args) {
        Document doc = new Document();
        doc.loadFromFile("MultiSectionDocument.docx");

        // 獲取第一個(gè)節(jié)
        Section firstSection = doc.getSections().get(0);
        
        String sectionText = firstSection.getText();
        int wordCount = countWords(sectionText);
        
        System.out.println("第一節(jié)的單詞數(shù)量: " + wordCount);
        
        doc.close();
    }

    private static int countWords(String text) {
        if (text == null || text.trim().isEmpty()) {
            return 0;
        }
        String[] words = text.trim().split("[\\s\\p{Punct}]+");
        return words.length;
    }
}

public class MixedLanguageCountExample {
    
    public static void main(String[] args) {
        Document doc = new Document();
        doc.loadFromFile("MixedLanguageDocument.docx");
        
        String text = doc.getText();
        
        int englishWordCount = countEnglishWords(text);
        int chineseCharCount = countChineseCharacters(text);
        
        System.out.println("英文單詞數(shù): " + englishWordCount);
        System.out.println("中文字符數(shù): " + chineseCharCount);
        
        doc.close();
    }
    
    private static int countEnglishWords(String text) {
        if (text == null || text.trim().isEmpty()) {
            return 0;
        }
        // 匹配英文單詞（字母序列）
        String[] words = text.split("[^a-zA-Z]+");
        int count = 0;
        for (String w : words) {
            if (!w.isEmpty()) {
                count++;
            }
        }
        return count;
    }
    
    private static int countChineseCharacters(String text) {
        if (text == null) {
            return 0;
        }
        int count = 0;
        for (char c : text.toCharArray()) {
            // 判斷是否為中文字符（Unicode 范圍：4E00-9FFF）
            if (c >= 0x4E00 && c <= 0x9FFF) {
                count++;
            }
        }
        return count;
    }
}