elasticlunr.js

Elasticlunr.js is a lightweight alternative to Lunr.js designed for browser and offline search. Key advantages include:

• Query-Time Boosting: Adjust field weights during the search process rather than at index time.
• Improved Performance: Faster computation by removing the need to compute document and query vectors (unlike Lunr.js).
• Smaller Index Size: Does not store a TokenCorpus, resulting in index files approximately half the size of Lunr.js. Users can further reduce size by disabling original document storage.
• Advanced Scoring: Uses a scoring mechanism similar to Elasticsearch/Lucene, combining Boolean, TF/IDF, and Vector Space models for reliable ranking.
• Field Search: Granular control over which fields are indexed and searched.

Install the package via npm:

npm install elasticlunr

Then require it in your project:

var elasticlunr = require('elasticlunr');

var index = elasticlunr(function () {
    this.addField('title');
    this.addField('body');
});

index.addDoc({
    "id": 1,
    "title": "Oracle released its latest database",
    "body": "..."
});

const results = index.search("Oracle");

var elasticlunr = require('elasticlunr');

var index = elasticlunr(function () {
    this.addField('title')
    this.addField('body')
});

var doc1 = {
    "id": 1,
    "title": "Oracle released its latest database Oracle 12g",
    "body": "Yestaday Oracle has released its new database Oracle 12g, this would make more money for this company and lead to a nice profit report of annual year."
}

index.addDoc(doc1);
index.search("Oracle database profit");

To avoid rebuilding the index every time, you can build it once, serialize it to JSON, and load it later.

Saving an index

Use JSON.stringify(index) to serialize the index. This invokes the elasticlunr.Index.prototype.toJSON method internally.

Loading an index

Use elasticlunr.Index.load(jsonObject) to restore a previously saved index.

// Saving (Node.js example)
fs.writeFile('./index.json', JSON.stringify(idx), function (err) {
  if (err) throw err;
  console.log('done');
});

// Loading
var indexDump = JSON.parse(fileContent);
var index = elasticlunr.Index.load(indexDump);

// Saving
fs.writeFile('./example/example_index.json', JSON.stringify(idx), function (err) {
  if (err) throw err;
  console.log('done');
});

// Loading
var indexDump = JSON.parse(indexDump)
window.idx = elasticlunr.Index.load(indexDump)

Elasticlunr uses a QueryRepository to parse query objects into specific Query class instances.

When you provide a query object like { match: { field: 'value' } }, the QueryRepository identifies the match key, looks up the corresponding parser, and returns a MatchQuery instance. This allows for a nested, JSON-like syntax to build complex search logic.

Commonly used query types registered in the repository include:

• match
• terms
• bool
• not
• match_all

Once the index is configured, you can manage documents using the following methods:

• addDoc(doc): Adds a JSON document to the index. Fields not configured in the index setup will not be indexed.
• removeDoc(doc): Removes a document from the index. This removes each token of that document's fields from the inverted index.
• updateDoc(doc): Updates an existing document in the index.

Note: If you used saveDocument(false) during setup, updating and removing documents may be difficult as the original data is not stored.

To perform a basic search, pass a query string to the search method. The method returns an array of result objects, sorted in descending order by their similarity score. Each object contains:

• ref: The document reference (e.g., the ID).
• score: The similarity measurement.

index.search("Oracle database profit");

To create a search index, call elasticlunr() and configure the fields you want to index and the reference field. If you do not specify fields, no fields will be searchable. By default, the reference field is 'id'. You can use this.setRef('fieldName') to change this.

You can also call this.saveDocument(false) during configuration to prevent the index from storing the original JSON documents, which significantly reduces the index size (useful for offline search), but makes updating or deleting documents more difficult.

var index = elasticlunr(function () {
    this.addField('title');
    this.addField('body');
    this.setRef('id');
    this.saveDocument(false); // Optional: reduces index size
});

Elasticlunr.js includes ~120 default English stop words. You can clear them or add your own.

• Remove default stop words: Use elasticlunr.clearStopWords().
• Add customized stop words: Pass an array of strings to elasticlunr.addStopWords(list).

// Remove defaults
elasticlunr.clearStopWords();

// Add custom words
var customized_stop_words = ['an', 'hello', 'xyzabc'];
elasticlunr.addStopWords(customized_stop_words);