Package org.apache.tika.inference

Class ChunkSerializer

java.lang.Object

org.apache.tika.inference.ChunkSerializer

public final class ChunkSerializer
extends Object

Serializes and deserializes a list of Chunk objects to/from JSON. Vectors are stored as base64-encoded little-endian float32 via VectorSerializer. Locators are nested under a "locators" object with optional text, paginated, spatial, and temporal arrays.

Method Summary

Modifier and Type	Method	Description
static List<Chunk>	fromJson(String json)	Deserialize a JSON array string back to a list of chunks.
static void	mergeInto(Metadata metadata, List<Chunk> newChunks)	Reads any existing chunks from the metadata field, appends the new chunks, and writes the merged list back.
static String	toJson(List<Chunk> chunks)	Serialize chunks to a JSON array string.

Methods inherited from class java.lang.Object

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

Method Details

toJson

public static String toJson(List<Chunk> chunks)
                     throws IOException

Serialize chunks to a JSON array string.

Throws:

IOException

mergeInto

public static void mergeInto(Metadata metadata,
 List<Chunk> newChunks)
                      throws IOException

Reads any existing chunks from the metadata field, appends the new chunks, and writes the merged list back. This allows multiple components (text chunker, image embedder, etc.) to contribute to the same chunks array.

Parameters:

metadata - the metadata to read from and write to

newChunks - chunks to append

Throws:

IOException

fromJson

public static List<Chunk> fromJson(String json)
                            throws IOException

Deserialize a JSON array string back to a list of chunks.

Throws:

IOException