keyword_extractor

⚠️ Experimental Package — Not ready for production

keyword_extractor is a Dart package for extracting keywords from structured text data.

It supports:

  • ✅ Basic word-based tokenization
  • ✅ Word prefix and phrase n-gram generation
  • ✅ Field-specific keyword extraction

Features

  • Extracts keywords from Map<String, dynamic> data
  • Works with any object that provides .toMap() or .toJson()
  • Swappable tokenizer strategies:
    • DefaultTokenizer: simple word splitting
    • AdvancedTokenizer: word prefixes + phrase n-grams
  • SelectiveKeywordExtractor for targeting specific fields

Getting Started

import 'package:keyword_extractor/keyword_extractor.dart';

void main() {
  final data = {
    'title': 'Improving search accuracy with keyword extraction',
    'summary': 'This article explores simple and advanced tokenization techniques.',
  };

  final extractor = DefaultKeywordExtractor(
    tokenizer: const DefaultTokenizer(),
  );

  final keywords = extractor.extract(data);
  print(keywords);
}

Selective Field Extraction

final extractor = SelectiveKeywordExtractor(
  tokenizer: const AdvancedTokenizer(),
  fields: ['title'], // extract only from the 'title' field
);

final keywords = extractor.extract(data);
print(keywords);

Input & Output Example

Input Map:

{
  "title": "Improving search accuracy with keyword extraction",
  "summary": "This article explores simple and advanced tokenization techniques."
}

DefaultTokenizer Output:

[
  "improving",
  "search",
  "accuracy",
  "with",
  "keyword",
  "extraction",
  "this",
  "article",
  "explores",
  "simple",
  "and",
  "advanced",
  "tokenization",
  "techniques"
]

AdvancedTokenizer Output (partial):

[
  "imp",
  "impr",
  "impro",
  "improv",
  "improvi",
  "improvin",
  "improving",
  "sea",
  "sear",
  "searc",
  "search",
  "keyword extraction",
  "extraction techniques",
  "simple and advanced",
  "improving search",
  "search accuracy",
  "accuracy with keyword"
]

Tokenizers

Tokenizer Description
DefaultTokenizer Splits text on spaces and punctuation
AdvancedTokenizer Adds word prefixes and phrase n-gram tokens

Disclaimer

This package is experimental and under active development.
Do not use it in production environments. APIs may change, and edge cases may not be fully covered yet.

Roadmap

  • Stopword filtering
  • Fuzzy variant generation
  • Nested field/key support
  • Token ranking and weighting

License

MIT

Libraries

keyword_extractor