Flutter文本处理插件flutter_tf_idf的使用

flutter_tf_idf 是一个用于在 Flutter 应用程序中计算 TF-IDF（词频-逆文档频率）并执行文本分析任务的 Dart 包。

功能

• 计算文档集合的 TF-IDF 矩阵。
• 计算文档之间的余弦相似度和距离。
• 获取特定文档的顶级词汇。
• 基于查询搜索文档。
• 获取特定词汇和文档的 TF-IDF 分数。
• 查找给定词汇的最高得分文档。

开始使用

1. 添加依赖

在 pubspec.yaml 文件中添加 flutter_tf_idf 依赖：

dependencies:
  flutter_tf_idf: ^1.0.0

2. 运行 flutter pub get

打开终端并运行以下命令以获取依赖项：

flutter pub get

3. 导入包

在需要使用该包的文件中导入：

import 'package:flutter_tf_idf/flutter_tf_idf.dart';

使用示例

下面是一个完整的示例，展示了如何使用 flutter_tf_idf 包来创建一个简单的 Flutter 应用程序。

示例代码

import 'package:flutter/material.dart';
import 'package:flutter_tf_idf/flutter_tf_idf.dart';

void main() {
  runApp(TfIdfDemoApp());
}

class TfIdfDemoApp extends StatelessWidget {
  [@override](/user/override)
  Widget build(BuildContext context) {
    return MaterialApp(
      title: 'TF-IDF Demo',
      theme: ThemeData(
        primarySwatch: Colors.blue,
      ),
      home: TfIdfDemo(),
    );
  }
}

class TfIdfDemo extends StatelessWidget {
  final TfIdf tfIdf;

  TfIdfDemo()
      : tfIdf = TfIdf([
          Document('1', 'The art of baking delicious cakes'),
          Document('2', 'Painting techniques for beginners'),
          Document('3', 'Culinary arts: baking and beyond'),
        ]);

  [@override](/user/override)
  Widget build(BuildContext context) {
    return Scaffold(
      appBar: AppBar(
        title: Text('TF-IDF Demo'),
      ),
      body: Padding(
        padding: const EdgeInsets.all(16.0),
        child: SingleChildScrollView(
          child: Column(
            crossAxisAlignment: CrossAxisAlignment.start,
            children: [
              Text(
                'Documents:',
                style: TextStyle(fontWeight: FontWeight.bold, fontSize: 18),
              ),
              ...tfIdf.documents
                  .map((doc) => Text('Document ${doc.id}: ${doc.content}')),
              SizedBox(height: 20),
              Text(
                'Top terms in document 1:',
                style: TextStyle(fontWeight: FontWeight.bold, fontSize: 16),
              ),
              Text(tfIdf.getTopTerms('1', 3).join(', ')),
              SizedBox(height: 20),
              Text(
                'Search results for "art":',
                style: TextStyle(fontWeight: FontWeight.bold, fontSize: 16),
              ),
              Text(tfIdf.searchDocuments('art', 2).join(', ')),
              SizedBox(height: 20),
              Text(
                'Cosine similarity between document 1 and document 2:',
                style: TextStyle(fontWeight: FontWeight.bold, fontSize: 16),
              ),
              Text(
                  tfIdf.calculateCosineSimilarity('1', '2').toStringAsFixed(3)),
              SizedBox(height: 20),
              Text(
                'Cosine distance between document 1 and document 2:',
                style: TextStyle(fontWeight: FontWeight.bold, fontSize: 16),
              ),
              Text(tfIdf.calculateCosineDistance('1', '2').toStringAsFixed(3)),
              SizedBox(height: 20),
              Text(
                'TF-IDF score of "art" in document 1:',
                style: TextStyle(fontWeight: FontWeight.bold, fontSize: 16),
              ),
              Text(tfIdf.getTfIdfScore('art', '1').toStringAsFixed(3)),
              SizedBox(height: 20),
              Text(
                'Highest scoring document for the term "art":',
                style: TextStyle(fontWeight: FontWeight.bold, fontSize: 16),
              ),
              Text(tfIdf.getHighestScoringDocument('art')),
              SizedBox(height: 20),
              Text(
                'TF-IDF Matrix:',
                style: TextStyle(fontWeight: FontWeight.bold, fontSize: 18),
              ),
              for (var term in tfIdf.terms)
                Text(
                    '$term: ${tfIdf.documents.map((doc) => tfIdf.getTfIdfScore(term, doc.id).toStringAsFixed(3)).join(', ')}'),
            ],
          ),
        ),
      ),
    );
  }
}