feat: Initial commit of PDF Tools project

server/models/File.js

@@ -0,0 +1,307 @@
+const mongoose = require('mongoose');
+
+const fileSchema = new mongoose.Schema({
+  fileId: {
+    type: String,
+    required: true,
+    unique: true,
+    index: true
+  },
+  userId: {
+    type: String,
+    required: false, // 允许匿名用户上传
+    index: true
+  },
+  originalName: {
+    type: String,
+    required: true,
+    trim: true
+  },
+  fileName: {
+    type: String,
+    required: true,
+    unique: true
+  },
+  fileSize: {
+    type: Number,
+    required: true,
+    min: 0
+  },
+  mimeType: {
+    type: String,
+    required: true,
+    enum: [
+      'application/pdf',
+      'application/vnd.openxmlformats-officedocument.wordprocessingml.document',
+      'text/html',
+      'text/plain',
+      'image/png',
+      'image/jpeg'
+    ]
+  },
+  filePath: {
+    type: String,
+    required: true
+  },
+  status: {
+    type: String,
+    enum: ['uploading', 'ready', 'processing', 'completed', 'failed', 'deleted'],
+    default: 'ready'
+  },
+  metadata: {
+    pageCount: Number,
+    width: Number,
+    height: Number,
+    hasImages: Boolean,
+    hasText: Boolean,
+    isEncrypted: Boolean,
+    pdfVersion: String,
+    creationDate: Date,
+    modificationDate: Date,
+    author: String,
+    title: String,
+    subject: String,
+    keywords: String
+  },
+  checksum: {
+    type: String,
+    index: true // 用于去重
+  },
+  downloadCount: {
+    type: Number,
+    default: 0
+  },
+  lastDownloaded: Date,
+  tags: [String],
+  isPublic: {
+    type: Boolean,
+    default: false
+  },
+  expiresAt: {
+    type: Date,
+    index: { expireAfterSeconds: 0 } // MongoDB TTL索引，自动删除过期文件
+  }
+}, {
+  timestamps: true
+});
+
+// 复合索引
+fileSchema.index({ userId: 1, createdAt: -1 });
+fileSchema.index({ status: 1, createdAt: -1 });
+fileSchema.index({ mimeType: 1, status: 1 });
+fileSchema.index({ fileSize: 1 });
+
+// 虚拟字段：格式化文件大小
+fileSchema.virtual('formattedFileSize').get(function() {
+  const sizes = ['Bytes', 'KB', 'MB', 'GB', 'TB'];
+  if (this.fileSize === 0) return '0 Bytes';
+  
+  const i = Math.floor(Math.log(this.fileSize) / Math.log(1024));
+  const size = (this.fileSize / Math.pow(1024, i)).toFixed(1);
+  return size + ' ' + sizes[i];
+});
+
+// 虚拟字段：文件扩展名
+fileSchema.virtual('fileExtension').get(function() {
+  const mimeToExt = {
+    'application/pdf': 'pdf',
+    'application/vnd.openxmlformats-officedocument.wordprocessingml.document': 'docx',
+    'text/html': 'html',
+    'text/plain': 'txt',
+    'image/png': 'png',
+    'image/jpeg': 'jpg'
+  };
+  return mimeToExt[this.mimeType] || 'unknown';
+});
+
+// 虚拟字段：是否已过期
+fileSchema.virtual('isExpired').get(function() {
+  if (!this.expiresAt) return false;
+  return new Date() > this.expiresAt;
+});
+
+// 实例方法：设置过期时间
+fileSchema.methods.setExpiration = function(hours = 24) {
+  this.expiresAt = new Date(Date.now() + hours * 60 * 60 * 1000);
+  return this.save();
+};
+
+// 实例方法：增加下载次数
+fileSchema.methods.incrementDownload = function() {
+  this.downloadCount += 1;
+  this.lastDownloaded = new Date();
+  return this.save();
+};
+
+// 实例方法：更新状态
+fileSchema.methods.updateStatus = function(status) {
+  this.status = status;
+  return this.save();
+};
+
+// 静态方法：根据校验和查找重复文件
+fileSchema.statics.findByChecksum = function(checksum) {
+  return this.findOne({ checksum, status: { $ne: 'deleted' } });
+};
+
+// 静态方法：获取用户文件
+fileSchema.statics.getUserFiles = function(userId, options = {}) {
+  const {
+    page = 1,
+    limit = 10,
+    status,
+    mimeType,
+    sortBy = 'createdAt',
+    sortOrder = -1
+  } = options;
+
+  const query = { userId };
+  
+  if (status) query.status = status;
+  if (mimeType) query.mimeType = mimeType;
+
+  const skip = (page - 1) * limit;
+  const sort = { [sortBy]: sortOrder };
+
+  return this.find(query)
+    .sort(sort)
+    .skip(skip)
+    .limit(limit)
+    .lean();
+};
+
+// 静态方法：清理过期文件
+fileSchema.statics.cleanupExpiredFiles = async function() {
+  const expiredFiles = await this.find({
+    expiresAt: { $lt: new Date() },
+    status: { $ne: 'deleted' }
+  });
+
+  const fs = require('fs');
+  const path = require('path');
+  let cleanedCount = 0;
+  let freedSpace = 0;
+
+  for (const file of expiredFiles) {
+    try {
+      // 删除物理文件
+      if (fs.existsSync(file.filePath)) {
+        freedSpace += file.fileSize;
+        fs.unlinkSync(file.filePath);
+      }
+      
+      // 更新数据库记录
+      await this.findByIdAndUpdate(file._id, { status: 'deleted' });
+      cleanedCount++;
+    } catch (error) {
+      console.error(`清理文件失败: ${file.fileName}`, error);
+    }
+  }
+
+  return { cleanedCount, freedSpace };
+};
+
+// 静态方法：获取文件统计
+fileSchema.statics.getFileStats = async function() {
+  const pipeline = [
+    {
+      $group: {
+        _id: '$status',
+        count: { $sum: 1 },
+        totalSize: { $sum: '$fileSize' }
+      }
+    }
+  ];
+
+  const statusStats = await this.aggregate(pipeline);
+  
+  const typeStats = await this.aggregate([
+    {
+      $group: {
+        _id: '$mimeType',
+        count: { $sum: 1 },
+        totalSize: { $sum: '$fileSize' }
+      }
+    }
+  ]);
+
+  const totalStats = await this.aggregate([
+    {
+      $group: {
+        _id: null,
+        totalFiles: { $sum: 1 },
+        totalSize: { $sum: '$fileSize' },
+        totalDownloads: { $sum: '$downloadCount' }
+      }
+    }
+  ]);
+
+  return {
+    byStatus: statusStats,
+    byType: typeStats,
+    total: totalStats[0] || { totalFiles: 0, totalSize: 0, totalDownloads: 0 }
+  };
+};
+
+// 静态方法：搜索文件
+fileSchema.statics.searchFiles = function(searchTerm, options = {}) {
+  const {
+    userId,
+    page = 1,
+    limit = 10,
+    status = { $ne: 'deleted' }
+  } = options;
+
+  const query = {
+    status,
+    $or: [
+      { originalName: { $regex: searchTerm, $options: 'i' } },
+      { tags: { $in: [new RegExp(searchTerm, 'i')] } },
+      { 'metadata.title': { $regex: searchTerm, $options: 'i' } },
+      { 'metadata.subject': { $regex: searchTerm, $options: 'i' } }
+    ]
+  };
+
+  if (userId) query.userId = userId;
+
+  const skip = (page - 1) * limit;
+
+  return this.find(query)
+    .sort({ createdAt: -1 })
+    .skip(skip)
+    .limit(limit)
+    .lean();
+};
+
+// 中间件：删除前清理物理文件
+fileSchema.pre('remove', async function(next) {
+  try {
+    const fs = require('fs');
+    if (fs.existsSync(this.filePath)) {
+      fs.unlinkSync(this.filePath);
+    }
+    next();
+  } catch (error) {
+    console.error('删除物理文件失败:', error);
+    next(error);
+  }
+});
+
+// 中间件：保存后更新用户统计
+fileSchema.post('save', async function(doc) {
+  if (doc.userId && doc.isNew) {
+    try {
+      const User = require('./User');
+      await User.findOneAndUpdate(
+        { userId: doc.userId },
+        { $inc: { 'statistics.totalFileSize': doc.fileSize } }
+      );
+    } catch (error) {
+      console.error('更新用户统计失败:', error);
+    }
+  }
+});
+
+const File = mongoose.model('File', fileSchema);
+
+module.exports = File;