Class EmbeddingStoreIngestor.Builder

java.lang.Object

dev.langchain4j.store.embedding.EmbeddingStoreIngestor.Builder

Enclosing class:

EmbeddingStoreIngestor

public static class EmbeddingStoreIngestor.Builder
extends Object

EmbeddingStoreIngestor builder.

Constructor Summary

Constructors

Constructor	Description
Builder()	Creates a new EmbeddingStoreIngestor builder.

Method Summary

Modifier and Type	Method	Description
EmbeddingStoreIngestor	build()	Builds the EmbeddingStoreIngestor.
EmbeddingStoreIngestor.Builder	documentSplitter(DocumentSplitter documentSplitter)	Sets the document splitter.
EmbeddingStoreIngestor.Builder	documentTransformer(DocumentTransformer documentTransformer)	Sets the document transformer.
EmbeddingStoreIngestor.Builder	embeddingInputType(EmbeddingInputType embeddingInputType)	Embeds the text segments with the given EmbeddingInputType (typically EmbeddingInputType.DOCUMENT), so providers that encode queries and documents differently can produce a document-optimized embedding.
EmbeddingStoreIngestor.Builder	embeddingModel(EmbeddingModel embeddingModel)	Sets the embedding model.
EmbeddingStoreIngestor.Builder	embeddingStore(EmbeddingStore<TextSegment> embeddingStore)	Sets the embedding store.
EmbeddingStoreIngestor.Builder	textSegmentTransformer(TextSegmentTransformer textSegmentTransformer)	Sets the text segment transformer.

Methods inherited from class Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Details

Builder

public Builder()

Creates a new EmbeddingStoreIngestor builder.

Method Details

embeddingInputType

public EmbeddingStoreIngestor.Builder embeddingInputType(EmbeddingInputType embeddingInputType)

Embeds the text segments with the given EmbeddingInputType (typically EmbeddingInputType.DOCUMENT), so providers that encode queries and documents differently can produce a document-optimized embedding.

When left null (the default), no input type is sent. If set, the chosen EmbeddingModel must support the input type parameter, otherwise ingestion fails fast.

Parameters:

embeddingInputType - the input type to embed documents with.

Returns:

this

documentTransformer

public EmbeddingStoreIngestor.Builder documentTransformer(DocumentTransformer documentTransformer)

Sets the document transformer. Optional.

Parameters:

documentTransformer - the document transformer.

Returns:

this

documentSplitter

public EmbeddingStoreIngestor.Builder documentSplitter(DocumentSplitter documentSplitter)

Sets the document splitter. Optional. If none is specified, it tries to load one through SPI (see DocumentSplitterFactory).
DocumentSplitters.recursive() from main (langchain4j) module is a good starting point.

Parameters:

documentSplitter - the document splitter.

Returns:

this

textSegmentTransformer

public EmbeddingStoreIngestor.Builder textSegmentTransformer(TextSegmentTransformer textSegmentTransformer)

Sets the text segment transformer. Optional.

Parameters:

textSegmentTransformer - the text segment transformer.

Returns:

this

embeddingModel

public EmbeddingStoreIngestor.Builder embeddingModel(EmbeddingModel embeddingModel)

Sets the embedding model. Mandatory. If none is specified, it tries to load one through SPI (see EmbeddingModelFactory).

Parameters:

embeddingModel - the embedding model.

Returns:

this

embeddingStore

public EmbeddingStoreIngestor.Builder embeddingStore(EmbeddingStore<TextSegment> embeddingStore)

Sets the embedding store. Mandatory.

Parameters:

embeddingStore - the embedding store.

Returns:

this

build

public EmbeddingStoreIngestor build()

Builds the EmbeddingStoreIngestor.

Returns:

the EmbeddingStoreIngestor.