323 lines
9.6 KiB
JSON
323 lines
9.6 KiB
JSON
{
|
||
"name": "Website Craw IA v4",
|
||
"nodes": [
|
||
{
|
||
"parameters": {},
|
||
"type": "n8n-nodes-base.manualTrigger",
|
||
"typeVersion": 1,
|
||
"position": [
|
||
64,
|
||
0
|
||
],
|
||
"id": "e4051790-d93e-4048-9d8e-64de8b644f57",
|
||
"name": "When clicking ‘Execute workflow’"
|
||
},
|
||
{
|
||
"parameters": {},
|
||
"type": "n8n-nodes-base.wait",
|
||
"typeVersion": 1.1,
|
||
"position": [
|
||
880,
|
||
0
|
||
],
|
||
"id": "3ec97de7-159c-46e6-af26-267d52f14d57",
|
||
"name": "Wait",
|
||
"webhookId": "1ef26c6f-55e3-4f51-bab5-6bb1e4b85fb8"
|
||
},
|
||
{
|
||
"parameters": {
|
||
"mode": "raw",
|
||
"jsonOutput": "={\n\"result\": {{ $('Get LLM Crawl').item.json.result.extracted_content }}\n}",
|
||
"options": {}
|
||
},
|
||
"type": "n8n-nodes-base.set",
|
||
"typeVersion": 3.4,
|
||
"position": [
|
||
1632,
|
||
-16
|
||
],
|
||
"id": "5d269e95-2e32-456b-8d07-54b279a0f543",
|
||
"name": "LLM Result"
|
||
},
|
||
{
|
||
"parameters": {
|
||
"url": "=https://evo-crwal4ai.5lbnb4.easypanel.host/task/{{ $json.task_id }}",
|
||
"authentication": "genericCredentialType",
|
||
"genericAuthType": "httpHeaderAuth",
|
||
"options": {}
|
||
},
|
||
"type": "n8n-nodes-base.httpRequest",
|
||
"typeVersion": 4.2,
|
||
"position": [
|
||
1056,
|
||
0
|
||
],
|
||
"id": "e1d5e1dc-27ca-4617-bc2d-eafa3292db17",
|
||
"name": "Get LLM Crawl",
|
||
"credentials": {
|
||
"httpBasicAuth": {
|
||
"id": "kSyyvvNaOBuA00C4",
|
||
"name": "crawl4aiserver"
|
||
},
|
||
"httpHeaderAuth": {
|
||
"id": "VLM9roYC9zcTGJsa",
|
||
"name": "panel"
|
||
}
|
||
}
|
||
},
|
||
{
|
||
"parameters": {
|
||
"mode": "raw",
|
||
"jsonOutput": "={\n \"task_id\": \"{{ $('Markdown Crawl Request').item.json.task_id }}\", \"status\": \"{{ $json.status }}\"\n}",
|
||
"options": {}
|
||
},
|
||
"type": "n8n-nodes-base.set",
|
||
"typeVersion": 3.4,
|
||
"position": [
|
||
1264,
|
||
0
|
||
],
|
||
"id": "7b604366-9af1-417a-b724-f039e574cc19",
|
||
"name": "LLM TID Field"
|
||
},
|
||
{
|
||
"parameters": {
|
||
"conditions": {
|
||
"options": {
|
||
"caseSensitive": true,
|
||
"leftValue": "",
|
||
"typeValidation": "strict",
|
||
"version": 2
|
||
},
|
||
"conditions": [
|
||
{
|
||
"id": "aad7f793-d74c-4f00-8d29-11b4782b43c8",
|
||
"leftValue": "={{ $json.status }}",
|
||
"rightValue": "pending",
|
||
"operator": {
|
||
"type": "string",
|
||
"operation": "notEquals"
|
||
}
|
||
},
|
||
{
|
||
"id": "c1ca239e-8ba0-447f-bc6e-bdbc83273c9f",
|
||
"leftValue": "={{ $json.status }}",
|
||
"rightValue": "processing",
|
||
"operator": {
|
||
"type": "string",
|
||
"operation": "notEquals"
|
||
}
|
||
}
|
||
],
|
||
"combinator": "and"
|
||
},
|
||
"options": {}
|
||
},
|
||
"type": "n8n-nodes-base.if",
|
||
"typeVersion": 2.2,
|
||
"position": [
|
||
1472,
|
||
0
|
||
],
|
||
"id": "4fef42a2-3682-4ca4-b2c9-4205471aa1ce",
|
||
"name": "LLM if In Process"
|
||
},
|
||
{
|
||
"parameters": {
|
||
"method": "POST",
|
||
"url": "https://evo-crwal4ai.5lbnb4.easypanel.host/crawl",
|
||
"authentication": "genericCredentialType",
|
||
"genericAuthType": "httpHeaderAuth",
|
||
"sendBody": true,
|
||
"specifyBody": "json",
|
||
"jsonBody": "={\n \"urls\": \"{{ $json.loc }}\",\n \"crawler_params\": {\n \"headless\": true\n },\n \"extraction_config\": {\n \"type\": \"llm\",\n \"params\": {\n \"llm_config\": {\n \"provider\": \"openai/gpt4o-mini\"\n },\n \"schema\": {\n \"title\": \"ProductData\",\n \"type\": \"object\",\n \"properties\": {\n \"title\": { \"type\": \"string\", \"description\": \"Product name (page title).\" },\n \"sku\": { \"type\": \"string\", \"description\": \"SKU/Referencia: use Product.sku/mpn or 'Referencia:' text on page.\" },\n \"price\": { \"type\": \"string\", \"description\": \"Numeric price only, e.g. 12.29\" },\n \"currency\": { \"type\": \"string\", \"description\": \"Price currency code/symbol if available (e.g. USD, B/., $).\" },\n \"category\": { \"type\": \"string\", \"description\": \"Breadcrumb category; prefer the node 'leches' if present, else the most specific category.\" },\n \"description\": { \"type\": \"string\", \"description\": \"Long textual description.\" },\n \"image\": { \"type\": \"string\", \"description\": \"Main image URL.\" },\n \"url\": { \"type\": \"string\", \"description\": \"Canonical product URL.\" }\n },\n \"required\": [\"title\", \"price\", \"url\"]\n },\n \"extraction_type\": \"schema\",\n \"instruction\": \"Extract data primarily from <script type='application/ld+json'> with @type='Product' and from BreadcrumbList. Map fields as follows: title=Product.name; sku=Product.sku or Product.mpn; price=Product.offers.price (return only the number, no symbols); currency=Product.offers.priceCurrency if present; description=Product.description; image=first of Product.image; url=Product.url or the page URL. For category, read BreadcrumbList item names and return 'leches' if present, otherwise the most specific (last) breadcrumb. If any field is missing in JSON-LD, fallback to DOM: title from H1; sku from the text after 'Referencia:'; price from visible price near the buy button; description from the 'Descripción' section; image from the main gallery image. Normalize whitespace. Do not invent values.\"\n }\n },\n \"cache_mode\": \"bypass\"\n}\n",
|
||
"options": {}
|
||
},
|
||
"type": "n8n-nodes-base.httpRequest",
|
||
"typeVersion": 4.2,
|
||
"position": [
|
||
672,
|
||
0
|
||
],
|
||
"id": "dd70ec7c-340d-419b-a1b1-b5b6957d7319",
|
||
"name": "Markdown Crawl Request",
|
||
"credentials": {
|
||
"httpBasicAuth": {
|
||
"id": "kSyyvvNaOBuA00C4",
|
||
"name": "crawl4aiserver"
|
||
},
|
||
"httpHeaderAuth": {
|
||
"id": "VLM9roYC9zcTGJsa",
|
||
"name": "panel"
|
||
}
|
||
}
|
||
},
|
||
{
|
||
"parameters": {
|
||
"documentId": {
|
||
"__rl": true,
|
||
"value": "1LatvxzjsFgWUcDgqIFubGbsy6yNeeSwZ7w9yg_tb8xY",
|
||
"mode": "list",
|
||
"cachedResultName": "UrlsTest",
|
||
"cachedResultUrl": "https://docs.google.com/spreadsheets/d/1LatvxzjsFgWUcDgqIFubGbsy6yNeeSwZ7w9yg_tb8xY/edit?usp=drivesdk"
|
||
},
|
||
"sheetName": {
|
||
"__rl": true,
|
||
"value": "gid=0",
|
||
"mode": "list",
|
||
"cachedResultName": "Hoja 1",
|
||
"cachedResultUrl": "https://docs.google.com/spreadsheets/d/1LatvxzjsFgWUcDgqIFubGbsy6yNeeSwZ7w9yg_tb8xY/edit#gid=0"
|
||
},
|
||
"options": {}
|
||
},
|
||
"type": "n8n-nodes-base.googleSheets",
|
||
"typeVersion": 4.7,
|
||
"position": [
|
||
192,
|
||
0
|
||
],
|
||
"id": "154ea175-3008-4066-9980-db7014b35c5b",
|
||
"name": "Get row(s) in sheet",
|
||
"credentials": {
|
||
"googleSheetsOAuth2Api": {
|
||
"id": "SE121yXzuB6zr94l",
|
||
"name": "Google Sheets account"
|
||
}
|
||
}
|
||
},
|
||
{
|
||
"parameters": {
|
||
"options": {}
|
||
},
|
||
"type": "n8n-nodes-base.splitInBatches",
|
||
"typeVersion": 3,
|
||
"position": [
|
||
400,
|
||
0
|
||
],
|
||
"id": "b3d6614a-8b26-4a86-b147-b60856976883",
|
||
"name": "Loop Over Items"
|
||
}
|
||
],
|
||
"pinData": {},
|
||
"connections": {
|
||
"When clicking ‘Execute workflow’": {
|
||
"main": [
|
||
[
|
||
{
|
||
"node": "Get row(s) in sheet",
|
||
"type": "main",
|
||
"index": 0
|
||
}
|
||
]
|
||
]
|
||
},
|
||
"Wait": {
|
||
"main": [
|
||
[
|
||
{
|
||
"node": "Get LLM Crawl",
|
||
"type": "main",
|
||
"index": 0
|
||
}
|
||
]
|
||
]
|
||
},
|
||
"Get LLM Crawl": {
|
||
"main": [
|
||
[
|
||
{
|
||
"node": "LLM TID Field",
|
||
"type": "main",
|
||
"index": 0
|
||
}
|
||
]
|
||
]
|
||
},
|
||
"LLM TID Field": {
|
||
"main": [
|
||
[
|
||
{
|
||
"node": "LLM if In Process",
|
||
"type": "main",
|
||
"index": 0
|
||
}
|
||
]
|
||
]
|
||
},
|
||
"LLM if In Process": {
|
||
"main": [
|
||
[
|
||
{
|
||
"node": "LLM Result",
|
||
"type": "main",
|
||
"index": 0
|
||
}
|
||
],
|
||
[
|
||
{
|
||
"node": "Wait",
|
||
"type": "main",
|
||
"index": 0
|
||
}
|
||
]
|
||
]
|
||
},
|
||
"Markdown Crawl Request": {
|
||
"main": [
|
||
[
|
||
{
|
||
"node": "Wait",
|
||
"type": "main",
|
||
"index": 0
|
||
}
|
||
]
|
||
]
|
||
},
|
||
"Get row(s) in sheet": {
|
||
"main": [
|
||
[
|
||
{
|
||
"node": "Loop Over Items",
|
||
"type": "main",
|
||
"index": 0
|
||
}
|
||
]
|
||
]
|
||
},
|
||
"Loop Over Items": {
|
||
"main": [
|
||
[],
|
||
[
|
||
{
|
||
"node": "Markdown Crawl Request",
|
||
"type": "main",
|
||
"index": 0
|
||
}
|
||
]
|
||
]
|
||
},
|
||
"LLM Result": {
|
||
"main": [
|
||
[
|
||
{
|
||
"node": "Loop Over Items",
|
||
"type": "main",
|
||
"index": 0
|
||
}
|
||
]
|
||
]
|
||
}
|
||
},
|
||
"active": false,
|
||
"settings": {
|
||
"executionOrder": "v1"
|
||
},
|
||
"versionId": "89103fc7-3505-45a9-98b3-c82b01d10eb3",
|
||
"meta": {
|
||
"templateCredsSetupCompleted": true,
|
||
"instanceId": "57b253ac4a90f3c1149cb4540b0ccf8677c4c106dc133e4a240f1f14ec79a39c"
|
||
},
|
||
"id": "r3mvCmZFRPHrWOGS",
|
||
"tags": []
|
||
} |