How to use the Bleve library for full-text search in Golang

Bleve is a lightweight, embeddable full-text search library for Go that enables efficient indexing and querying without external dependencies. First, install it via go get github.com/blevesearch/bleve/v2. Then create a disk-based index using bleve.New("example.bleve", mapping), where mapping := bleve.NewIndexMapping() defines text analysis rules. Index structured data like structs or maps by calling index.Index("id", doc)—for example, a Product struct with name, description, and price. Search with queries such as bleve.NewMatchQuery("laptop"), wrapped in a bleve.NewSearchRequest(), then execute against the index to retrieve results with IDs and scores. Update documents by reindexing with the same ID; delete using index.Delete("id"). Customize indexing behavior by defining field-specific mappings, controlling analyzers, or disabling indexing on certain fields. Close the index with index.Close() and reopen later via bleve.Open("example.bleve"). While not as powerful as Elasticsearch, Bleve excels in simplicity and performance for small to medium-scale applications requiring embedded search.

Bleve is a powerful full-text search library for Go that allows you to index and query data efficiently. It’s ideal for adding search capabilities directly into your Go applications without relying on external services like Elasticsearch or Solr. Here's how to get started using Bleve in Golang.

Install Bleve

To begin, install the Bleve library using Go modules:

This command adds Bleve as a dependency to your project.

Create an Index

Bleve stores searchable data in an index. You can create an in-memory or persistent index on disk. Below is an example of creating a basic disk-based index:

mapping := bleve.NewIndexMapping()
index, err := bleve.New("example.bleve", mapping)
if err != nil {
  log.Fatal(err)
}

The NewIndexMapping() function sets up how Bleve analyzes and indexes text. The default mapping works well for general use.

Index Some Data

Once the index is created, you can add documents. A document can be a struct, map, or any JSON-marshalable type. For example:

The first argument to Index() is a unique ID for the document.

Search the Index

To perform a search, create a query and execute it against the index. Here's how to do a simple keyword search:

The MatchQuery looks for documents where any indexed field contains the term "laptop". You can also use more specific queries like NewMatchPhraseQuery or NewTermQuery for exact matches.

Update or Delete Documents

To update a document, simply call Index() again with the same ID. Bleve will replace the existing document.

To delete a document, use:

Customize Index Mapping (Optional)

If you need control over how fields are analyzed, define a custom mapping. For example, to disable indexing on a field:

You can specify which fields to index, their types, and which analyzers to use (e.g., standard, keyword, etc.).

Close and Reopen Index

When done, close the index:

To reopen later:

This loads the previously saved index from disk.

Basically, Bleve gives you a lightweight, embeddable search engine in pure Go. With just a few lines, you can index structured data and run fast, relevant text searches. It's not as feature-rich as large-scale search servers, but it's perfect for small to medium applications where simplicity and low overhead matter.

The above is the detailed content of How to use the Bleve library for full-text search in Golang. For more information, please follow other related articles on the PHP Chinese website!