001/*
002 *  Copyright (c) 2023-2026, Agents-Flex (fuhai999@gmail.com).
003 *  <p>
004 *  Licensed under the Apache License, Version 2.0 (the "License");
005 *  you may not use this file except in compliance with the License.
006 *  You may obtain a copy of the License at
007 *  <p>
008 *  http://www.apache.org/licenses/LICENSE-2.0
009 *  <p>
010 *  Unless required by applicable law or agreed to in writing, software
011 *  distributed under the License is distributed on an "AS IS" BASIS,
012 *  WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
013 *  See the License for the specific language governing permissions and
014 *  limitations under the License.
015 */
016package com.agentsflex.core.file2text.extractor.impl;
017
018
019import com.agentsflex.core.file2text.extractor.FileExtractor;
020import com.agentsflex.core.file2text.source.DocumentSource;
021import org.apache.poi.xwpf.usermodel.*;
022
023import java.io.IOException;
024import java.io.InputStream;
025import java.util.Collections;
026import java.util.HashSet;
027import java.util.List;
028import java.util.Set;
029import java.util.stream.Collectors;
030
031/**
032 * DOCX 文档提取器（.docx, .dotx）
033 * 支持段落、表格、列表文本提取
034 */
035public class DocxExtractor implements FileExtractor {
036
037    private static final Set<String> KNOWN_MIME_TYPES;
038    private static final String MIME_PREFIX = "application/vnd.openxmlformats-officedocument.wordprocessingml";
039    private static final Set<String> SUPPORTED_EXTENSIONS;
040
041    static {
042        // 精确 MIME（可选）
043        Set<String> mimeTypes = new HashSet<>();
044        mimeTypes.add("application/vnd.openxmlformats-officedocument.wordprocessingml.document");
045        KNOWN_MIME_TYPES = Collections.unmodifiableSet(mimeTypes);
046
047        // 支持的扩展名
048        Set<String> extensions = new HashSet<>();
049        extensions.add("docx");
050        extensions.add("dotx");
051        SUPPORTED_EXTENSIONS = Collections.unmodifiableSet(extensions);
052    }
053
054    @Override
055    public boolean supports(DocumentSource source) {
056        String mimeType = source.getMimeType();
057        String fileName = source.getFileName();
058
059        // 1. MIME 精确匹配
060        if (mimeType != null && KNOWN_MIME_TYPES.contains(mimeType)) {
061            return true;
062        }
063
064        // 2. MIME 前缀匹配
065        if (mimeType != null && mimeType.startsWith(MIME_PREFIX)) {
066            return true;
067        }
068
069        // 3. 扩展名匹配
070        if (fileName != null) {
071            String ext = getExtension(fileName);
072            if (ext != null && SUPPORTED_EXTENSIONS.contains(ext.toLowerCase())) {
073                return true;
074            }
075        }
076
077        return false;
078    }
079
080    @Override
081    public String extractText(DocumentSource source) throws IOException {
082        StringBuilder text = new StringBuilder();
083
084        try (InputStream is = source.openStream();
085             XWPFDocument document = new XWPFDocument(is)) {
086
087            // 提取段落
088            for (XWPFParagraph paragraph : document.getParagraphs()) {
089                String paraText = getParagraphText(paragraph);
090                if (paraText != null && !paraText.trim().isEmpty()) {
091                    text.append(paraText).append("\n");
092                }
093            }
094
095            // 提取表格
096            for (XWPFTable table : document.getTables()) {
097                text.append("\n[Table Start]\n");
098                for (XWPFTableRow row : table.getRows()) {
099                    List<String> cellTexts = row.getTableCells().stream()
100                        .map(this::getCellText)
101                        .map(String::trim)
102                        .collect(Collectors.toList());
103                    text.append(cellTexts).append("\n");
104                }
105                text.append("[Table End]\n\n");
106            }
107
108        } catch (Exception e) {
109            throw new IOException("Failed to extract DOCX: " + e.getMessage(), e);
110        }
111
112        return text.toString().trim();
113    }
114
115    private String getParagraphText(XWPFParagraph paragraph) {
116        StringBuilder text = new StringBuilder();
117        for (XWPFRun run : paragraph.getRuns()) {
118            String runText = run.text();
119            if (runText != null) {
120                text.append(runText);
121            }
122        }
123        return text.length() > 0 ? text.toString() : null;
124    }
125
126    private String getCellText(XWPFTableCell cell) {
127        String simpleText = cell.getText();
128        if (simpleText != null && !simpleText.isEmpty()) {
129            return simpleText;
130        }
131        StringBuilder text = new StringBuilder();
132        for (XWPFParagraph p : cell.getParagraphs()) {
133            String pt = getParagraphText(p);
134            if (pt != null) {
135                text.append(pt).append(" ");
136            }
137        }
138        return text.toString().trim();
139    }
140
141    @Override
142    public int getOrder() {
143        return 10;
144    }
145
146    private String getExtension(String fileName) {
147        if (fileName == null || !fileName.contains(".")) return null;
148        int lastDot = fileName.lastIndexOf('.');
149        return fileName.substring(lastDot + 1);
150    }
151}