n8n-nodes-google-vertex-embeddings-extended
Version:
n8n community sub-node for Google Vertex AI Embeddings with output dimensions and configurable batch size support - resolves LangChain compatibility issues
83 lines • 4.02 kB
JavaScript
;
Object.defineProperty(exports, "__esModule", { value: true });
exports.logWrapper = logWrapper;
const n8n_workflow_1 = require("n8n-workflow");
async function callMethodAsync(parameters) {
try {
return await parameters.method.call(this, ...parameters.arguments);
}
catch (error) {
const connectedNode = parameters.executeFunctions.getNode();
throw new n8n_workflow_1.NodeOperationError(connectedNode, error);
}
}
function logAiEvent(executeFunctions, eventType) {
try {
if ('logAiEvent' in executeFunctions && typeof executeFunctions.logAiEvent === 'function') {
executeFunctions.logAiEvent({
type: eventType,
});
}
}
catch (error) {
}
}
function logWrapper(originalInstance, executeFunctions) {
console.log('VertexEmbeddingsLogWrapper: Wrapping instance of type:', originalInstance.constructor.name);
return new Proxy(originalInstance, {
get(target, prop, receiver) {
const originalValue = Reflect.get(target, prop, receiver);
if (typeof originalValue === 'function' && typeof prop === 'string') {
console.log('VertexEmbeddingsLogWrapper: Method accessed:', prop);
}
if ('embedDocuments' in target || 'embedQuery' in target) {
if (prop === 'embedDocuments' && 'embedDocuments' in target) {
return async (documents) => {
console.log('VertexEmbeddingsLogWrapper: embedDocuments intercepted, docs:', documents?.length || 0);
const connectionType = "ai_embedding";
const { index } = executeFunctions.addInputData(connectionType, [
[{ json: { documents } }],
]);
const response = (await callMethodAsync.call(target, {
executeFunctions,
connectionType,
currentNodeRunIndex: index,
method: target[prop],
arguments: [documents],
}));
console.log('VertexEmbeddingsLogWrapper: embedDocuments completed, embeddings:', response?.length || 0);
logAiEvent(executeFunctions, 'ai-document-embedded');
executeFunctions.addOutputData(connectionType, index, [
[{ json: { response } }],
]);
return response;
};
}
if (prop === 'embedQuery' && 'embedQuery' in target) {
return async (query) => {
console.log('VertexEmbeddingsLogWrapper: embedQuery intercepted, query length:', query?.length || 0);
const connectionType = "ai_embedding";
const { index } = executeFunctions.addInputData(connectionType, [
[{ json: { query } }],
]);
const response = (await callMethodAsync.call(target, {
executeFunctions,
connectionType,
currentNodeRunIndex: index,
method: target[prop],
arguments: [query],
}));
console.log('VertexEmbeddingsLogWrapper: embedQuery completed, embedding dimension:', response?.length || 0);
logAiEvent(executeFunctions, 'ai-query-embedded');
executeFunctions.addOutputData(connectionType, index, [
[{ json: { response } }],
]);
return response;
};
}
}
return originalValue;
},
});
}
//# sourceMappingURL=logWrapper.js.map