272 lines
6.3 KiB
JSON
272 lines
6.3 KiB
JSON
{
|
|
"name": "71 Index policy docs (reference · Lab 01 example 2)",
|
|
"nodes": [
|
|
{
|
|
"name": "Index webhook",
|
|
"type": "n8n-nodes-base.webhook",
|
|
"typeVersion": 2,
|
|
"position": [
|
|
0,
|
|
300
|
|
],
|
|
"parameters": {
|
|
"httpMethod": "POST",
|
|
"path": "my-index",
|
|
"responseMode": "responseNode",
|
|
"options": {}
|
|
}
|
|
},
|
|
{
|
|
"name": "List docs",
|
|
"type": "n8n-nodes-base.httpRequest",
|
|
"typeVersion": 4.2,
|
|
"position": [
|
|
200,
|
|
300
|
|
],
|
|
"parameters": {
|
|
"url": "http://data-api:8800/docs",
|
|
"options": {}
|
|
}
|
|
},
|
|
{
|
|
"name": "One item per doc",
|
|
"type": "n8n-nodes-base.splitOut",
|
|
"typeVersion": 1,
|
|
"position": [
|
|
400,
|
|
300
|
|
],
|
|
"parameters": {
|
|
"fieldToSplitOut": "docs",
|
|
"options": {}
|
|
}
|
|
},
|
|
{
|
|
"name": "Get doc",
|
|
"type": "n8n-nodes-base.httpRequest",
|
|
"typeVersion": 4.2,
|
|
"position": [
|
|
600,
|
|
300
|
|
],
|
|
"parameters": {
|
|
"url": "=http://data-api:8800/docs/{{ $json.name }}",
|
|
"options": {}
|
|
}
|
|
},
|
|
{
|
|
"name": "Make chunks",
|
|
"type": "n8n-nodes-base.code",
|
|
"typeVersion": 2,
|
|
"position": [
|
|
800,
|
|
300
|
|
],
|
|
"parameters": {
|
|
"jsCode": "// Split each policy document into chunks. Body \"chunking\": \"section\" (one chunk per ## section, default) or a number of characters.\nconst body = $('Index webhook').first().json.body ?? {};\nconst CHUNKING = body.chunking ?? 'section';\nconst chunks = [];\nfor (const item of $input.all()) {\n const { name, title, text } = item.json;\n const content = text.split('\\n').slice(1).join('\\n').trim();\n let pieces;\n if (CHUNKING === 'section') {\n pieces = content.split(/\\n(?=## )/).map(part => part.trim()).filter(Boolean);\n } else {\n const size = Number(CHUNKING), step = size - Math.floor(size / 5);\n pieces = [];\n for (let i = 0; i < content.length; i += step) pieces.push(content.slice(i, i + size));\n }\n pieces.forEach((piece, i) => {\n const section = (piece.match(/^## (.+)$/m) ?? [])[1] ?? '';\n chunks.push({ json: { source: name, title, chunk: i + 1, text: `${title}\\n${piece}` } });\n });\n}\nreturn chunks;"
|
|
}
|
|
},
|
|
{
|
|
"name": "Store chunks",
|
|
"type": "@n8n/n8n-nodes-langchain.vectorStoreInMemory",
|
|
"typeVersion": 1.3,
|
|
"position": [
|
|
1020,
|
|
300
|
|
],
|
|
"parameters": {
|
|
"mode": "insert",
|
|
"memoryKey": {
|
|
"__rl": true,
|
|
"mode": "list",
|
|
"value": "my_policy_docs",
|
|
"cachedResultName": "my_policy_docs"
|
|
},
|
|
"clearStore": true
|
|
}
|
|
},
|
|
{
|
|
"name": "Embeddings",
|
|
"type": "@n8n/n8n-nodes-langchain.embeddingsOpenAi",
|
|
"typeVersion": 1.2,
|
|
"position": [
|
|
960,
|
|
520
|
|
],
|
|
"parameters": {
|
|
"model": "text-embedding-3-large",
|
|
"options": {}
|
|
},
|
|
"credentials": {
|
|
"openAiApi": {
|
|
"id": "openAiCourse0001",
|
|
"name": "OpenAI (.env)"
|
|
}
|
|
}
|
|
},
|
|
{
|
|
"name": "Chunk loader",
|
|
"type": "@n8n/n8n-nodes-langchain.documentDefaultDataLoader",
|
|
"typeVersion": 1.1,
|
|
"position": [
|
|
1120,
|
|
520
|
|
],
|
|
"parameters": {
|
|
"jsonMode": "expressionData",
|
|
"jsonData": "={{ $json.text }}",
|
|
"textSplittingMode": "simple",
|
|
"options": {
|
|
"metadata": {
|
|
"metadataValues": [
|
|
{
|
|
"name": "source",
|
|
"value": "={{ $json.source }}"
|
|
},
|
|
{
|
|
"name": "title",
|
|
"value": "={{ $json.title }}"
|
|
},
|
|
{
|
|
"name": "chunk",
|
|
"value": "={{ $json.chunk }}"
|
|
}
|
|
]
|
|
}
|
|
}
|
|
}
|
|
},
|
|
{
|
|
"name": "Count",
|
|
"type": "n8n-nodes-base.code",
|
|
"typeVersion": 2,
|
|
"position": [
|
|
1260,
|
|
300
|
|
],
|
|
"parameters": {
|
|
"jsCode": "// One summary item after all chunks are stored.\nreturn [{ json: { indexed_chunks: $('Make chunks').all().length, chunking: ($('Index webhook').first().json.body ?? {}).chunking ?? 'section' } }];"
|
|
}
|
|
},
|
|
{
|
|
"name": "Send result",
|
|
"type": "n8n-nodes-base.respondToWebhook",
|
|
"typeVersion": 1.1,
|
|
"position": [
|
|
1460,
|
|
300
|
|
],
|
|
"parameters": {
|
|
"respondWith": "json",
|
|
"responseBody": "={{ $json }}",
|
|
"options": {}
|
|
}
|
|
}
|
|
],
|
|
"connections": {
|
|
"Index webhook": {
|
|
"main": [
|
|
[
|
|
{
|
|
"node": "List docs",
|
|
"type": "main",
|
|
"index": 0
|
|
}
|
|
]
|
|
]
|
|
},
|
|
"List docs": {
|
|
"main": [
|
|
[
|
|
{
|
|
"node": "One item per doc",
|
|
"type": "main",
|
|
"index": 0
|
|
}
|
|
]
|
|
]
|
|
},
|
|
"One item per doc": {
|
|
"main": [
|
|
[
|
|
{
|
|
"node": "Get doc",
|
|
"type": "main",
|
|
"index": 0
|
|
}
|
|
]
|
|
]
|
|
},
|
|
"Get doc": {
|
|
"main": [
|
|
[
|
|
{
|
|
"node": "Make chunks",
|
|
"type": "main",
|
|
"index": 0
|
|
}
|
|
]
|
|
]
|
|
},
|
|
"Make chunks": {
|
|
"main": [
|
|
[
|
|
{
|
|
"node": "Store chunks",
|
|
"type": "main",
|
|
"index": 0
|
|
}
|
|
]
|
|
]
|
|
},
|
|
"Store chunks": {
|
|
"main": [
|
|
[
|
|
{
|
|
"node": "Count",
|
|
"type": "main",
|
|
"index": 0
|
|
}
|
|
]
|
|
]
|
|
},
|
|
"Count": {
|
|
"main": [
|
|
[
|
|
{
|
|
"node": "Send result",
|
|
"type": "main",
|
|
"index": 0
|
|
}
|
|
]
|
|
]
|
|
},
|
|
"Embeddings": {
|
|
"ai_embedding": [
|
|
[
|
|
{
|
|
"node": "Store chunks",
|
|
"type": "ai_embedding",
|
|
"index": 0
|
|
}
|
|
]
|
|
]
|
|
},
|
|
"Chunk loader": {
|
|
"ai_document": [
|
|
[
|
|
{
|
|
"node": "Store chunks",
|
|
"type": "ai_document",
|
|
"index": 0
|
|
}
|
|
]
|
|
]
|
|
}
|
|
},
|
|
"settings": {
|
|
"executionOrder": "v1"
|
|
}
|
|
}
|