Perform Text Embedding

This handler (com.rierino.handler.langchain4j.EmbeddingEventHandler) provides ability to convert text into embedding for advanced search functionality

Handler Parameters

For Onnx models available on a target system:

Parameter	Definition	Example	Default
model.store.system	Name of the system to get Onnx model details from	model_fs	A default huggingface repo
model.path	File path for the model to use	/model.onnx	A default ONNX model
tokenizer.path	File path for the tokenizer to use	/tokenizer.json	A default tokenizer
model.local.dir	Local directory to store downloaded model files	/app/models	System temp dir

For non-Onnx models serviced through custom embedding classes:

Parameter	Definition	Example	Default
embedding.class	Java class for using a custom embedding provider	dev.langchain4j.model.openai.OpenAiEmbeddingModel	-
embeddingMethods	Json object configuration of methods to call on embedding class	{"apiKey": "xxx"}	-

This handler requires the following dependency added to deployment contents as well as any custom embedding class dependencies:

implementation (group:'com.rierino.custom', name: 'langchain4j', version:"${rierinoVersion}")

Actions

EmbedText

Transforms text into an embedding. Event metadata fields applicable for this action are as follows:

Fields (table)

Field	Definition	Example	Default
Input Element	Json path for the input in request event payload	input	-
Output Element	Json path for the output in response event payload	output	-

Fields (JSON Schema)

With event metadata parameters as:

Parameters (table)

Parameter	Definition	Example	Default
Input Pattern	JMESpath pattern to apply on input element for getting text and metadata fields	{text: "", metadata: ""}	-
Extra Action	Optional extra action to perform after generating embedding ("add" to local index, "search" in local index)	add	-
Max Results	Maximum results to return if "search" extra action is used	5	-
Min Score	Minimum similarity score required if "search" extra action is used	0.5	-

Parameters (JSON Schema)

Input element can be a text value, or an object in {chat, message} format, where "chat" field includes ID of an ongoing chat for an assistant type use case with memory.