654 lines
14 KiB
JSON
654 lines
14 KiB
JSON
{
|
|
"id": "14",
|
|
"name": "extract_swifts",
|
|
"nodes": [
|
|
{
|
|
"name": "On clicking 'execute'",
|
|
"type": "n8n-nodes-base.manualTrigger",
|
|
"position": [
|
|
-140,
|
|
820
|
|
],
|
|
"parameters": {},
|
|
"typeVersion": 1
|
|
},
|
|
{
|
|
"name": "HTTP Request",
|
|
"type": "n8n-nodes-base.httpRequest",
|
|
"position": [
|
|
320,
|
|
820
|
|
],
|
|
"parameters": {
|
|
"url": "https://www.theswiftcodes.com/browse-by-country/",
|
|
"options": {},
|
|
"responseFormat": "string"
|
|
},
|
|
"typeVersion": 1
|
|
},
|
|
{
|
|
"name": "HTML Extract",
|
|
"type": "n8n-nodes-base.htmlExtract",
|
|
"position": [
|
|
510,
|
|
820
|
|
],
|
|
"parameters": {
|
|
"options": {},
|
|
"extractionValues": {
|
|
"values": [
|
|
{
|
|
"key": "countries",
|
|
"attribute": "href",
|
|
"cssSelector": "ol > li > a",
|
|
"returnArray": true,
|
|
"returnValue": "attribute"
|
|
}
|
|
]
|
|
}
|
|
},
|
|
"typeVersion": 1
|
|
},
|
|
{
|
|
"name": "SplitInBatches",
|
|
"type": "n8n-nodes-base.splitInBatches",
|
|
"position": [
|
|
910,
|
|
820
|
|
],
|
|
"parameters": {
|
|
"options": {
|
|
"reset": false
|
|
},
|
|
"batchSize": 1
|
|
},
|
|
"typeVersion": 1
|
|
},
|
|
{
|
|
"name": "HTTP Request1",
|
|
"type": "n8n-nodes-base.httpRequest",
|
|
"position": [
|
|
2250,
|
|
740
|
|
],
|
|
"parameters": {
|
|
"url": "={{$node[\"Set\"].json[\"url\"]}}",
|
|
"options": {},
|
|
"responseFormat": "file"
|
|
},
|
|
"typeVersion": 1
|
|
},
|
|
{
|
|
"name": "HTML Extract1",
|
|
"type": "n8n-nodes-base.htmlExtract",
|
|
"position": [
|
|
2750,
|
|
590
|
|
],
|
|
"parameters": {
|
|
"options": {},
|
|
"sourceData": "binary",
|
|
"extractionValues": {
|
|
"values": [
|
|
{
|
|
"key": "next_button",
|
|
"attribute": "href",
|
|
"cssSelector": "span.next > a",
|
|
"returnValue": "attribute"
|
|
},
|
|
{
|
|
"key": "names",
|
|
"cssSelector": "td.table-name",
|
|
"returnArray": true
|
|
},
|
|
{
|
|
"key": "swifts",
|
|
"cssSelector": "td.table-swift",
|
|
"returnArray": true
|
|
},
|
|
{
|
|
"key": "cities",
|
|
"cssSelector": "td.table-city",
|
|
"returnArray": true
|
|
},
|
|
{
|
|
"key": "branches",
|
|
"cssSelector": "td.table-branch",
|
|
"returnArray": true
|
|
}
|
|
]
|
|
}
|
|
},
|
|
"typeVersion": 1
|
|
},
|
|
{
|
|
"name": "MongoDB1",
|
|
"type": "n8n-nodes-base.mongoDb",
|
|
"position": [
|
|
3280,
|
|
590
|
|
],
|
|
"parameters": {
|
|
"fields": "iso_code,country,page,name,branch,city,swift_code,createdAt,updatedAt",
|
|
"options": {
|
|
"dateFields": "createdAt,updatedAt"
|
|
},
|
|
"operation": "insert",
|
|
"collection": "swifts.meetup"
|
|
},
|
|
"credentials": {
|
|
"mongoDb": "db-mongo"
|
|
},
|
|
"typeVersion": 1
|
|
},
|
|
{
|
|
"name": "uProc",
|
|
"type": "n8n-nodes-base.uproc",
|
|
"position": [
|
|
1100,
|
|
820
|
|
],
|
|
"parameters": {
|
|
"tool": "getCountryNormalized",
|
|
"group": "geographic",
|
|
"country": "={{$node[\"SplitInBatches\"].json[\"country\"].replace(/[\\/0-9]/g, \"\")}}",
|
|
"additionalOptions": {}
|
|
},
|
|
"credentials": {
|
|
"uprocApi": "uproc-miquel"
|
|
},
|
|
"typeVersion": 1
|
|
},
|
|
{
|
|
"name": "Prepare Documents",
|
|
"type": "n8n-nodes-base.function",
|
|
"position": [
|
|
2930,
|
|
590
|
|
],
|
|
"parameters": {
|
|
"functionCode": "var newItems = [];\n\nfor (i = 0; i < items[0].json.swifts.length; i++) {\n var item = {\n iso_code: $node['uProc'].json.message.code,\n country: $node['SplitInBatches'].json.country.replace(/[-\\/0-9]/g, \"\"),\n page: $node['Set Page to Scrape'].json.page,\n name: items[0].json.names[i],\n city: items[0].json.cities[i],\n branch: items[0].json.branches[i],\n swift_code: items[0].json.swifts[i],\n createdAt: new Date(),\n updatedAt: new Date()\n }\n newItems.push({json: item});\n}\n\nreturn newItems;\n\n"
|
|
},
|
|
"typeVersion": 1
|
|
},
|
|
{
|
|
"name": "More Countries",
|
|
"type": "n8n-nodes-base.if",
|
|
"position": [
|
|
2810,
|
|
1100
|
|
],
|
|
"parameters": {
|
|
"conditions": {
|
|
"string": [
|
|
{
|
|
"value1": "={{$node[\"SplitInBatches\"].context[\"noItemsLeft\"] + \"\"}}",
|
|
"value2": "true"
|
|
}
|
|
]
|
|
}
|
|
},
|
|
"typeVersion": 1
|
|
},
|
|
{
|
|
"name": "Set Page to Scrape",
|
|
"type": "n8n-nodes-base.functionItem",
|
|
"position": [
|
|
1290,
|
|
680
|
|
],
|
|
"parameters": {
|
|
"functionCode": "const staticData = getWorkflowStaticData('global');\n\nitem.page = \"\";\nif (staticData.page && staticData.page.length) {\n item.page = staticData.page;\n} else {\n item.page = $node['SplitInBatches'].json.country;\n}\nreturn item;\n"
|
|
},
|
|
"typeVersion": 1
|
|
},
|
|
{
|
|
"name": "More Pages",
|
|
"type": "n8n-nodes-base.if",
|
|
"position": [
|
|
3070,
|
|
1020
|
|
],
|
|
"parameters": {
|
|
"conditions": {
|
|
"string": [
|
|
{
|
|
"value1": "={{$json[\"more_pages\"] + \"\"}}",
|
|
"value2": "true"
|
|
}
|
|
]
|
|
}
|
|
},
|
|
"typeVersion": 1
|
|
},
|
|
{
|
|
"name": "Set More Pages",
|
|
"type": "n8n-nodes-base.function",
|
|
"position": [
|
|
3470,
|
|
590
|
|
],
|
|
"parameters": {
|
|
"functionCode": "var next_page = $node['HTML Extract1'].json.next_button && $node['HTML Extract1'].json.next_button.length ? $node['HTML Extract1'].json.next_button : \"\";\nvar more_pages = next_page.length > 0;\nconst staticData = getWorkflowStaticData('global');\n\n//all current items are after date: needs pagination\nif (more_pages) {\n staticData.page = next_page;\n} else {\n //don't check more items in previous pages;\n delete staticData.page;\n}\n\nreturn [\n {\n json: {\n more_pages: more_pages\n }\n }\n];\n"
|
|
},
|
|
"typeVersion": 1
|
|
},
|
|
{
|
|
"name": "Set",
|
|
"type": "n8n-nodes-base.set",
|
|
"position": [
|
|
1440,
|
|
680
|
|
],
|
|
"parameters": {
|
|
"values": {
|
|
"string": [
|
|
{
|
|
"name": "url",
|
|
"value": "=https://www.theswiftcodes.com{{$node[\"Set Page to Scrape\"].json[\"page\"]}}"
|
|
}
|
|
]
|
|
},
|
|
"options": {}
|
|
},
|
|
"typeVersion": 1
|
|
},
|
|
{
|
|
"name": "Generate filename",
|
|
"type": "n8n-nodes-base.functionItem",
|
|
"position": [
|
|
1600,
|
|
610
|
|
],
|
|
"parameters": {
|
|
"functionCode": "var generateNameFromUrl = function(url){\n return url.replace(/[^a-z0-9]/gi, \"_\");\n}\n\nitem.file = generateNameFromUrl(item.url) + \".html\"\nreturn item;"
|
|
},
|
|
"typeVersion": 1
|
|
},
|
|
{
|
|
"name": "Read Binary File",
|
|
"type": "n8n-nodes-base.readBinaryFile",
|
|
"position": [
|
|
1770,
|
|
610
|
|
],
|
|
"parameters": {
|
|
"filePath": "=/home/node/.cache/scrapper/{{$json[\"file\"]}}"
|
|
},
|
|
"typeVersion": 1,
|
|
"continueOnFail": true,
|
|
"alwaysOutputData": true
|
|
},
|
|
{
|
|
"name": "File exists?",
|
|
"type": "n8n-nodes-base.if",
|
|
"position": [
|
|
1950,
|
|
610
|
|
],
|
|
"parameters": {
|
|
"conditions": {
|
|
"string": [
|
|
{
|
|
"value1": "={{$node[\"Read Binary File\"].binary.data.mimeType}}",
|
|
"value2": "text/html"
|
|
}
|
|
]
|
|
}
|
|
},
|
|
"typeVersion": 1
|
|
},
|
|
{
|
|
"name": "Write Binary File",
|
|
"type": "n8n-nodes-base.writeBinaryFile",
|
|
"position": [
|
|
2400,
|
|
740
|
|
],
|
|
"parameters": {
|
|
"fileName": "=/home/node/.cache/scrapper/{{$node[\"Generate filename\"].json[\"file\"]}}",
|
|
"dataPropertyName": "=data"
|
|
},
|
|
"typeVersion": 1
|
|
},
|
|
{
|
|
"name": "Read Binary File1",
|
|
"type": "n8n-nodes-base.readBinaryFile",
|
|
"position": [
|
|
2570,
|
|
590
|
|
],
|
|
"parameters": {
|
|
"filePath": "=/home/node/.cache/scrapper/{{$json[\"file\"]}}"
|
|
},
|
|
"typeVersion": 1,
|
|
"continueOnFail": true,
|
|
"alwaysOutputData": true
|
|
},
|
|
{
|
|
"name": "Wait",
|
|
"type": "n8n-nodes-base.function",
|
|
"position": [
|
|
2090,
|
|
740
|
|
],
|
|
"parameters": {
|
|
"functionCode": "const waitTimeSeconds = 1;\n\nreturn new Promise((resolve) => {\n setTimeout(() => {\n resolve([]);\n }, waitTimeSeconds * 1000);\n});\n"
|
|
},
|
|
"typeVersion": 1,
|
|
"continueOnFail": true,
|
|
"alwaysOutputData": true
|
|
},
|
|
{
|
|
"name": "Prepare countries",
|
|
"type": "n8n-nodes-base.function",
|
|
"position": [
|
|
700,
|
|
820
|
|
],
|
|
"parameters": {
|
|
"functionCode": "return items[0].json.countries.map(function(country) {\n return {\n json: {country: country}\n }\n});"
|
|
},
|
|
"typeVersion": 1
|
|
},
|
|
{
|
|
"name": "Create Directory",
|
|
"type": "n8n-nodes-base.executeCommand",
|
|
"position": [
|
|
70,
|
|
820
|
|
],
|
|
"parameters": {
|
|
"command": "mkdir -p /home/node/.cache/scrapper/"
|
|
},
|
|
"typeVersion": 1,
|
|
"continueOnFail": true
|
|
},
|
|
{
|
|
"name": "MongoDB",
|
|
"type": "n8n-nodes-base.mongoDb",
|
|
"disabled": true,
|
|
"position": [
|
|
3100,
|
|
520
|
|
],
|
|
"parameters": {
|
|
"query": "={\"swift_code\": \"{{$json[\"swift_code\"]}}\"}",
|
|
"options": {},
|
|
"collection": "swifts.meetup"
|
|
},
|
|
"credentials": {
|
|
"mongoDb": "db-mongo"
|
|
},
|
|
"executeOnce": false,
|
|
"typeVersion": 1,
|
|
"alwaysOutputData": true
|
|
}
|
|
],
|
|
"active": false,
|
|
"settings": {},
|
|
"connections": {
|
|
"Set": {
|
|
"main": [
|
|
[
|
|
{
|
|
"node": "Generate filename",
|
|
"type": "main",
|
|
"index": 0
|
|
}
|
|
]
|
|
]
|
|
},
|
|
"Wait": {
|
|
"main": [
|
|
[
|
|
{
|
|
"node": "HTTP Request1",
|
|
"type": "main",
|
|
"index": 0
|
|
}
|
|
]
|
|
]
|
|
},
|
|
"uProc": {
|
|
"main": [
|
|
[
|
|
{
|
|
"node": "Set Page to Scrape",
|
|
"type": "main",
|
|
"index": 0
|
|
}
|
|
]
|
|
]
|
|
},
|
|
"MongoDB": {
|
|
"main": [
|
|
[]
|
|
]
|
|
},
|
|
"MongoDB1": {
|
|
"main": [
|
|
[
|
|
{
|
|
"node": "Set More Pages",
|
|
"type": "main",
|
|
"index": 0
|
|
}
|
|
]
|
|
]
|
|
},
|
|
"More Pages": {
|
|
"main": [
|
|
[
|
|
{
|
|
"node": "Set Page to Scrape",
|
|
"type": "main",
|
|
"index": 0
|
|
}
|
|
],
|
|
[
|
|
{
|
|
"node": "More Countries",
|
|
"type": "main",
|
|
"index": 0
|
|
}
|
|
]
|
|
]
|
|
},
|
|
"File exists?": {
|
|
"main": [
|
|
[
|
|
{
|
|
"node": "Read Binary File1",
|
|
"type": "main",
|
|
"index": 0
|
|
}
|
|
],
|
|
[
|
|
{
|
|
"node": "Wait",
|
|
"type": "main",
|
|
"index": 0
|
|
}
|
|
]
|
|
]
|
|
},
|
|
"HTML Extract": {
|
|
"main": [
|
|
[
|
|
{
|
|
"node": "Prepare countries",
|
|
"type": "main",
|
|
"index": 0
|
|
}
|
|
]
|
|
]
|
|
},
|
|
"HTTP Request": {
|
|
"main": [
|
|
[
|
|
{
|
|
"node": "HTML Extract",
|
|
"type": "main",
|
|
"index": 0
|
|
}
|
|
]
|
|
]
|
|
},
|
|
"HTML Extract1": {
|
|
"main": [
|
|
[
|
|
{
|
|
"node": "Prepare Documents",
|
|
"type": "main",
|
|
"index": 0
|
|
}
|
|
]
|
|
]
|
|
},
|
|
"HTTP Request1": {
|
|
"main": [
|
|
[
|
|
{
|
|
"node": "Write Binary File",
|
|
"type": "main",
|
|
"index": 0
|
|
}
|
|
]
|
|
]
|
|
},
|
|
"More Countries": {
|
|
"main": [
|
|
[],
|
|
[
|
|
{
|
|
"node": "SplitInBatches",
|
|
"type": "main",
|
|
"index": 0
|
|
}
|
|
]
|
|
]
|
|
},
|
|
"Set More Pages": {
|
|
"main": [
|
|
[
|
|
{
|
|
"node": "More Pages",
|
|
"type": "main",
|
|
"index": 0
|
|
}
|
|
]
|
|
]
|
|
},
|
|
"SplitInBatches": {
|
|
"main": [
|
|
[
|
|
{
|
|
"node": "uProc",
|
|
"type": "main",
|
|
"index": 0
|
|
}
|
|
]
|
|
]
|
|
},
|
|
"Create Directory": {
|
|
"main": [
|
|
[
|
|
{
|
|
"node": "HTTP Request",
|
|
"type": "main",
|
|
"index": 0
|
|
}
|
|
]
|
|
]
|
|
},
|
|
"Read Binary File": {
|
|
"main": [
|
|
[
|
|
{
|
|
"node": "File exists?",
|
|
"type": "main",
|
|
"index": 0
|
|
}
|
|
]
|
|
]
|
|
},
|
|
"Generate filename": {
|
|
"main": [
|
|
[
|
|
{
|
|
"node": "Read Binary File",
|
|
"type": "main",
|
|
"index": 0
|
|
}
|
|
]
|
|
]
|
|
},
|
|
"Prepare Documents": {
|
|
"main": [
|
|
[
|
|
{
|
|
"node": "MongoDB1",
|
|
"type": "main",
|
|
"index": 0
|
|
}
|
|
]
|
|
]
|
|
},
|
|
"Prepare countries": {
|
|
"main": [
|
|
[
|
|
{
|
|
"node": "SplitInBatches",
|
|
"type": "main",
|
|
"index": 0
|
|
}
|
|
]
|
|
]
|
|
},
|
|
"Read Binary File1": {
|
|
"main": [
|
|
[
|
|
{
|
|
"node": "HTML Extract1",
|
|
"type": "main",
|
|
"index": 0
|
|
}
|
|
]
|
|
]
|
|
},
|
|
"Write Binary File": {
|
|
"main": [
|
|
[
|
|
{
|
|
"node": "Read Binary File1",
|
|
"type": "main",
|
|
"index": 0
|
|
}
|
|
]
|
|
]
|
|
},
|
|
"Set Page to Scrape": {
|
|
"main": [
|
|
[
|
|
{
|
|
"node": "Set",
|
|
"type": "main",
|
|
"index": 0
|
|
}
|
|
]
|
|
]
|
|
},
|
|
"On clicking 'execute'": {
|
|
"main": [
|
|
[
|
|
{
|
|
"node": "Create Directory",
|
|
"type": "main",
|
|
"index": 0
|
|
}
|
|
]
|
|
]
|
|
}
|
|
}
|
|
} |