Crate tantivy

source ·
Expand description

§tantivy

Tantivy is a search engine library. Think Lucene, but in Rust.

// First we need to define a schema ...

// `TEXT` means the field should be tokenized and indexed,
// along with its term frequency and term positions.
//
// `STORED` means that the field will also be saved
// in a compressed, row-oriented key-value store.
// This store is useful to reconstruct the
// documents that were selected during the search phase.
let mut schema_builder = Schema::builder();
let title = schema_builder.add_text_field("title", TEXT | STORED);
let body = schema_builder.add_text_field("body", TEXT);
let schema = schema_builder.build();

// Indexing documents

let index = Index::create_in_dir(index_path, schema.clone())?;

// Here we use a buffer of 100MB that will be split
// between indexing threads.
let mut index_writer: IndexWriter = index.writer(100_000_000)?;

// Let's index one documents!
index_writer.add_document(doc!(
    title => "The Old Man and the Sea",
    body => "He was an old man who fished alone in a skiff in \
            the Gulf Stream and he had gone eighty-four days \
            now without taking a fish."
))?;

// We need to call .commit() explicitly to force the
// index_writer to finish processing the documents in the queue,
// flush the current index to the disk, and advertise
// the existence of new documents.
index_writer.commit()?;

// # Searching

let reader = index.reader()?;

let searcher = reader.searcher();

let query_parser = QueryParser::for_index(&index, vec![title, body]);

// QueryParser may fail if the query is not in the right
// format. For user facing applications, this can be a problem.
// A ticket has been opened regarding this problem.
let query = query_parser.parse_query("sea whale")?;

// Perform search.
// `topdocs` contains the 10 most relevant doc ids, sorted by decreasing scores...
let top_docs: Vec<(Score, DocAddress)> =
    searcher.search(&query, &TopDocs::with_limit(10))?;

for (_score, doc_address) in top_docs {
    // Retrieve the actual content of documents given its `doc_address`.
    let retrieved_doc = searcher.doc::<TantivyDocument>(doc_address)?;
    println!("{}", retrieved_doc.to_json(&schema));
}

A good place for you to get started is to check out the example code ( literate programming / source code)

§Tantivy Architecture Overview

Tantivy is inspired by Lucene, the Architecture is very similar.

§Core Concepts

  • Index: A collection of segments. The top level entry point for tantivy users to search and index data.

  • Segment: At the heart of Tantivy’s indexing structure is the Segment. It contains documents and indices and is the atomic unit of indexing and search.

  • Schema: A schema is a set of fields in an index. Each field has a specific data type and set of attributes.

  • IndexWriter: Responsible creating and merging segments. It executes the indexing pipeline including tokenization, creating indices, and storing the index in the Directory.

  • Searching: Searcher searches the segments with anything that implements Query and merges the results. The list of supported queries. Custom Queries are supported by implementing the Query trait.

  • Directory: Abstraction over the storage where the index data is stored.

  • Tokenizer: Breaks down text into individual tokens. Users can implement or use provided tokenizers.

§Architecture Flow

  1. Document Addition: Users create documents according to the defined schema. The documents fields are tokenized, processed, and added to the current segment. See Document for the structure and usage.

  2. Segment Creation: Once the memory limit threshold is reached or a commit is called, the segment is written to the Directory. Documents are searchable after commit.

  3. Merging: To optimize space and search speed, segments might be merged. This operation is performed in the background. Customize the merge behaviour via IndexWriter::set_merge_policy.

Re-exports§

Modules§

  • aggregation
    Aggregations
  • collector
    Collectors
  • directory
    WORM (Write Once Read Many) directory abstraction.
  • error
    Definition of Tantivy’s errors and results.
  • fastfield
    Column oriented field storage for tantivy.
  • fieldnorm
    The fieldnorm represents the length associated with a given Field of a given document.
  • index
    Index Module
  • indexer
    Indexing and merging data.
  • merge_policy
    Defines tantivy’s merging strategy
  • positions
    Tantivy can (if instructed to do so in the schema) store the term positions in a given field. This position is expressed as token ordinal. For instance, In “The beauty and the beast”, the term “the” appears in position 0 and position 3. This information is useful to run phrase queries.
  • postings
    Postings module (also called inverted index)
  • query
    Module containing the different query implementations.
  • schema
    Schema definition for tantivy’s indices.
  • snippet
    SnippetGenerator Generates a text snippet for a given document, and some highlighted parts inside it. Imagine you doing a text search in a document and want to show a preview of where in the document the search terms occur, along with some surrounding text to give context, and the search terms highlighted.
  • space_usage
    Representations for the space usage of various parts of a Tantivy index.
  • store
    Compressed/slow/row-oriented storage for documents.
  • termdict
    The term dictionary main role is to associate the sorted Terms to a TermInfo struct that contains some meta-information about the term.
  • tokenizer
    Tokenizer are in charge of chopping text into a stream of tokens ready for indexing.

Macros§

  • doc
    doc! is a shortcut that helps building Document objects.
  • fail_point
    Enable fail_point if feature is enabled.

Structs§

Enums§

  • DateTimePrecision
    Precision with which datetimes are truncated when stored in fast fields. This setting is only relevant for fast fields. In the docstore, datetimes are always saved with nanosecond precision.
  • Executor
    Search executor whether search request are single thread or multithread.
  • ReloadPolicy
    Defines when a new version of the index should be reloaded.

Constants§

  • COLLECT_BLOCK_BUFFER_LEN
    The collect_block method on SegmentCollector uses a buffer of this size. Passed results to collect_block will not exceed this size and will be exactly this size as long as we can fill the buffer.
  • TERMINATED
    Sentinel value returned when a DocSet has been entirely consumed.

Traits§

  • DocSet
    Represents an iterable set of sorted doc ids.
  • HasLen
    Has length trait
  • Warmer
    Warmer can be used to maintain segment-level state e.g. caches.

Functions§

  • f64_to_u64
    Maps a f64 to u64
  • i64_to_u64
    Maps a i64 to u64
  • u64_to_f64
    Reverse the mapping given by f64_to_u64().
  • u64_to_i64
    Reverse the mapping given by i64_to_u64().
  • version
    Expose the current version of tantivy as found in Cargo.toml during compilation. eg. “0.11.0” as well as the compression scheme used in the docstore.
  • version_string
    Exposes the complete version of tantivy as found in Cargo.toml during compilation as a string. eg. “tantivy v0.11.0, index_format v1, store_compression: lz4”.

Type Aliases§

  • DocId
    A u32 identifying a document within a segment. Documents have their DocId assigned incrementally, as they are added in the segment.
  • Opstamp
    A u64 assigned to every operation incrementally
  • Result
    Tantivy result.
  • Score
    A Score that represents the relevance of the document to the query
  • SegmentOrdinal
    A SegmentOrdinal identifies a segment, within a Searcher or Merger.