feat: Add AI tool building capabilities (#7336)

Github issue / Community forum post (link here to close automatically):
https://community.n8n.io/t/langchain-memory-chat/23733

---------

Signed-off-by: Oleg Ivaniv <me@olegivaniv.com>
Co-authored-by: Oleg Ivaniv <me@olegivaniv.com>
Co-authored-by: Val <68596159+valya@users.noreply.github.com>
Co-authored-by: Alex Grozav <alex@grozav.com>
Co-authored-by: कारतोफ्फेलस्क्रिप्ट™ <aditya@netroy.in>
Co-authored-by: Deborah <deborah@starfallprojects.co.uk>
Co-authored-by: Jesper Bylund <mail@jesperbylund.com>
Co-authored-by: Jon <jonathan.bennetts@gmail.com>
Co-authored-by: Michael Kret <88898367+michael-radency@users.noreply.github.com>
Co-authored-by: Giulio Andreini <andreini@netseven.it>
Co-authored-by: Mason Geloso <Mason.geloso@gmail.com>
Co-authored-by: Mason Geloso <hone@Masons-Mac-mini.local>
Co-authored-by: Mutasem Aldmour <mutasem@n8n.io>

packages/@n8n/nodes-langchain/utils/N8nJsonLoader.ts

@@ -0,0 +1,98 @@
+import {
+	type IExecuteFunctions,
+	type INodeExecutionData,
+	NodeConnectionType,
+	NodeOperationError,
+} from 'n8n-workflow';
+
+import type { CharacterTextSplitter } from 'langchain/text_splitter';
+import type { Document } from 'langchain/document';
+import { JSONLoader } from 'langchain/document_loaders/fs/json';
+import { TextLoader } from 'langchain/document_loaders/fs/text';
+import { getMetadataFiltersValues } from './helpers';
+
+export class N8nJsonLoader {
+	private context: IExecuteFunctions;
+
+	private optionsPrefix: string;
+
+	constructor(context: IExecuteFunctions, optionsPrefix = '') {
+		this.context = context;
+		this.optionsPrefix = optionsPrefix;
+	}
+
+	async processAll(items?: INodeExecutionData[]): Promise<Document[]> {
+		const docs: Document[] = [];
+
+		if (!items) return [];
+
+		for (let itemIndex = 0; itemIndex < items.length; itemIndex++) {
+			const processedDocuments = await this.processItem(items[itemIndex], itemIndex);
+
+			docs.push(...processedDocuments);
+		}
+
+		return docs;
+	}
+
+	async processItem(item: INodeExecutionData, itemIndex: number): Promise<Document[]> {
+		const mode = this.context.getNodeParameter('jsonMode', itemIndex, 'allInputData') as
+			| 'allInputData'
+			| 'expressionData';
+
+		const pointers = this.context.getNodeParameter(
+			`${this.optionsPrefix}pointers`,
+			itemIndex,
+			'',
+		) as string;
+		const pointersArray = pointers.split(',').map((pointer) => pointer.trim());
+
+		const textSplitter = (await this.context.getInputConnectionData(
+			NodeConnectionType.AiTextSplitter,
+			0,
+		)) as CharacterTextSplitter | undefined;
+		const metadata = getMetadataFiltersValues(this.context, itemIndex) ?? [];
+
+		if (!item) return [];
+
+		let documentLoader: JSONLoader | TextLoader | null = null;
+
+		if (mode === 'allInputData') {
+			const itemString = JSON.stringify(item.json);
+			const itemBlob = new Blob([itemString], { type: 'application/json' });
+			documentLoader = new JSONLoader(itemBlob, pointersArray);
+		}
+
+		if (mode === 'expressionData') {
+			const dataString = this.context.getNodeParameter('jsonData', itemIndex) as string | object;
+			if (typeof dataString === 'object') {
+				const itemBlob = new Blob([JSON.stringify(dataString)], { type: 'application/json' });
+				documentLoader = new JSONLoader(itemBlob, pointersArray);
+			}
+
+			if (typeof dataString === 'string') {
+				const itemBlob = new Blob([dataString], { type: 'text/plain' });
+				documentLoader = new TextLoader(itemBlob);
+			}
+		}
+
+		if (documentLoader === null) {
+			// This should never happen
+			throw new NodeOperationError(this.context.getNode(), 'Document loader is not initialized');
+		}
+
+		const docs = textSplitter
+			? await documentLoader.loadAndSplit(textSplitter)
+			: await documentLoader.load();
+
+		if (metadata) {
+			docs.forEach((doc) => {
+				doc.metadata = {
+					...doc.metadata,
+					...metadata,
+				};
+			});
+		}
+		return docs;
+	}
+}