Insert and retrieve documents
工作流概述
这是一个包含25个节点的复杂工作流，主要用于自动化处理各种任务。
工作流源代码
{
"id": "Hjyv9FkH5Oh6Yxw4",
"meta": {
"instanceId": "2c4c1e23e7b067270c08aab616bada21d0c384d16f212b23cf1143c6baa09219"
},
"name": "Insert and retrieve documents",
"tags": [
{
"id": "msnDWKHQmwMDxWQH",
"name": "Milvus",
"createdAt": "2025-04-16T12:48:14.539Z",
"updatedAt": "2025-04-16T12:48:14.539Z"
},
{
"id": "tnCpo8hq8uKrdASK",
"name": "AI",
"createdAt": "2025-04-16T12:47:57.976Z",
"updatedAt": "2025-04-16T12:47:57.976Z"
}
],
"nodes": [
{
"id": "52044ccd-4e0d-4353-b612-cf8db1b55331",
"name": "When clicking \"Execute Workflow\"",
"type": "n8n-nodes-base.manualTrigger",
"position": [
-500,
-100
],
"parameters": {},
"typeVersion": 1
},
{
"id": "b6993775-d21b-4ae8-a59c-43aef2b7002b",
"name": "Fetch Essay List",
"type": "n8n-nodes-base.httpRequest",
"position": [
-220,
-100
],
"parameters": {
"url": "http://www.paulgraham.com/articles.html",
"options": {}
},
"typeVersion": 4.2
},
{
"id": "cbaeb236-5c93-4b34-a06b-ff0e5de8525f",
"name": "Extract essay names",
"type": "n8n-nodes-base.html",
"position": [
-20,
-100
],
"parameters": {
"options": {},
"operation": "extractHtmlContent",
"extractionValues": {
"values": [
{
"key": "essay",
"attribute": "href",
"cssSelector": "table table a",
"returnArray": true,
"returnValue": "attribute"
}
]
}
},
"typeVersion": 1.2
},
{
"id": "d92b6692-4a02-4519-b113-8a9172c71de9",
"name": "Split out into items",
"type": "n8n-nodes-base.splitOut",
"position": [
180,
-100
],
"parameters": {
"options": {},
"fieldToSplitOut": "essay"
},
"typeVersion": 1
},
{
"id": "d16ba71b-10fc-454f-8bfc-a6826280a4e7",
"name": "Fetch essay texts",
"type": "n8n-nodes-base.httpRequest",
"position": [
580,
-100
],
"parameters": {
"url": "=http://www.paulgraham.com/{{ $json.essay }}",
"options": {}
},
"typeVersion": 4.2
},
{
"id": "c4fa74ea-6af5-410c-bf5c-9d8d3decf31b",
"name": "Limit to first 3",
"type": "n8n-nodes-base.limit",
"position": [
380,
-100
],
"parameters": {
"maxItems": 3
},
"typeVersion": 1
},
{
"id": "3da8495b-62df-475d-b99d-e0f3c64266e3",
"name": "Extract Text Only",
"type": "n8n-nodes-base.html",
"position": [
900,
-100
],
"parameters": {
"options": {},
"operation": "extractHtmlContent",
"extractionValues": {
"values": [
{
"key": "data",
"cssSelector": "body",
"skipSelectors": "img,nav"
}
]
}
},
"typeVersion": 1.2
},
{
"id": "4a9b5d5d-fc94-40b7-af0c-13d992bc1eb9",
"name": "Sticky Note3",
"type": "n8n-nodes-base.stickyNote",
"position": [
-300,
-220
],
"parameters": {
"width": 1071.752021563343,
"height": 285.66037735849045,
"content": "## Scrape latest Paul Graham essays"
},
"typeVersion": 1
},
{
"id": "b8a7a288-186f-4444-b0de-33ed90009c0a",
"name": "Sticky Note5",
"type": "n8n-nodes-base.stickyNote",
"position": [
820,
-220
],
"parameters": {
"width": 625,
"height": 607,
"content": "## Load into Milvus vector store"
},
"typeVersion": 1
},
{
"id": "c9e7b166-cc65-47e2-a437-9c00017b492a",
"name": "Recursive Character Text Splitter1",
"type": "@n8n/n8n-nodes-langchain.textSplitterRecursiveCharacterTextSplitter",
"position": [
1240,
240
],
"parameters": {
"options": {},
"chunkSize": 6000
},
"typeVersion": 1
},
{
"id": "e1a75f27-7c8c-4d0d-9b0f-33fe9ec96fc6",
"name": "Generate response",
"type": "n8n-nodes-base.set",
"position": [
1240,
560
],
"parameters": {
"options": {},
"assignments": {
"assignments": [
{
"id": "11396286-0378-4c3a-86e1-c9ef51afbfc7",
"name": "text",
"type": "string",
"value": "={{ $json.answer }} {{ $if(!$json.citations.isEmpty(), \"\n\" + $json.citations.join(\"\"), '') }}"
}
]
}
},
"typeVersion": 3.4
},
{
"id": "8b3497ad-5bc8-44b3-bdf4-3a028fe265ce",
"name": "Compose citations",
"type": "n8n-nodes-base.set",
"position": [
1040,
560
],
"parameters": {
"options": {},
"assignments": {
"assignments": [
{
"id": "ace6185e-8b3d-4f89-ae36-dfe0c391a0a9",
"name": "citations",
"type": "array",
"value": "={{ $json.citations.map(i => '[' + $('Get top chunks matching query').all()[$json.citations].json.document.metadata.file_name + ', lines ' + $('Get top chunks matching query').all()[$json.citations].json.document.metadata['loc.lines.from'] + '-' + $('Get top chunks matching query').all()[$json.citations].json.document.metadata['loc.lines.to'] + ']') }}"
}
]
}
},
"typeVersion": 3.4
},
{
"id": "0452cf15-145c-49dd-8803-4c8b8a7adbea",
"name": "Answer the query based on chunks",
"type": "@n8n/n8n-nodes-langchain.informationExtractor",
"position": [
680,
560
],
"parameters": {
"text": "={{ $json.context }}
Question: {{ $('When chat message received').first().json.chatInput }}
Helpful Answer:",
"options": {
"systemPromptTemplate": "=Use the following pieces of context to answer the question at the end. If you don't know the answer, just say that you don't know, don't try to make up an answer. Important: In your response, also include the the indexes of the chunks you used to generate the answer."
},
"schemaType": "manual",
"inputSchema": "{
\"type\": \"object\",
\"required\": [\"answer\", \"citations\"],
\"properties\": {
\"answer\": {
\"type\": \"string\"
},
\"citations\": {
\"type\": \"array\",
\"items\": {
\"type\": \"number\"
}
}
}
}"
},
"typeVersion": 1
},
{
"id": "d385ac35-6f94-4101-99de-5ce1991f40c4",
"name": "Prepare chunks",
"type": "n8n-nodes-base.code",
"position": [
480,
560
],
"parameters": {
"jsCode": "let out = \"\"
for (const i in $input.all()) {
let itemText = \"--- CHUNK \" + i + \" ---\n\"
itemText += $input.all()[i].json.document.pageContent + \"\n\"
itemText += \"\n\"
out += itemText
}
return {
'context': out
};"
},
"typeVersion": 2
},
{
"id": "379837f2-4f96-43ff-8e87-722cbe6d652f",
"name": "Set max chunks to send to model",
"type": "n8n-nodes-base.set",
"position": [
-300,
560
],
"parameters": {
"options": {},
"assignments": {
"assignments": [
{
"id": "33f4addf-72f3-4618-a6ba-5b762257d723",
"name": "chunks",
"type": "number",
"value": 4
}
]
},
"includeOtherFields": true
},
"typeVersion": 3.4
},
{
"id": "9bc391bb-df47-41df-b170-9df47a6b5e87",
"name": "Embeddings OpenAI2",
"type": "@n8n/n8n-nodes-langchain.embeddingsOpenAi",
"position": [
-100,
780
],
"parameters": {
"model": "text-embedding-ada-002",
"options": {}
},
"credentials": {
"openAiApi": {
"id": "hH2PTDH4fbS7fdPv",
"name": "OpenAi account"
}
},
"typeVersion": 1.2
},
{
"id": "efb030f4-445b-4ba0-b5c9-95e4e5893664",
"name": "When chat message received",
"type": "@n8n/n8n-nodes-langchain.chatTrigger",
"position": [
-540,
560
],
"webhookId": "cd2703a7-f912-46fe-8787-3fb83ea116ab",
"parameters": {
"options": {}
},
"typeVersion": 1.1
},
{
"id": "c74943be-0008-4d4c-9dea-598a648a97a2",
"name": "Sticky Note1",
"type": "n8n-nodes-base.stickyNote",
"position": [
-380,
440
],
"parameters": {
"color": 7,
"width": 1594,
"height": 529,
"content": ""
},
"typeVersion": 1
},
{
"id": "2e27f3d8-e8a2-4647-80dd-f2643b224cb5",
"name": "Milvus Vector Store in retrieval",
"type": "@n8n/n8n-nodes-langchain.vectorStoreMilvus",
"position": [
0,
560
],
"parameters": {
"mode": "load",
"topK": 2,
"prompt": "answer the question",
"milvusCollection": {
"__rl": true,
"mode": "list",
"value": "my_collection",
"cachedResultName": "my_collection"
}
},
"credentials": {
"milvusApi": {
"id": "8tMHHoLiWXIAXa7S",
"name": "Milvus account"
}
},
"typeVersion": 1.1
},
{
"id": "a3cf7e0e-f681-4880-9ccf-5c42d5457c0f",
"name": "Milvus Vector Store",
"type": "@n8n/n8n-nodes-langchain.vectorStoreMilvus",
"position": [
1120,
-100
],
"parameters": {
"mode": "insert",
"options": {
"clearCollection": true
},
"milvusCollection": {
"__rl": true,
"mode": "list",
"value": "my_collection",
"cachedResultName": "my_collection"
}
},
"credentials": {
"milvusApi": {
"id": "8tMHHoLiWXIAXa7S",
"name": "Milvus account"
}
},
"typeVersion": 1.1
},
{
"id": "4c4cc5a5-e880-466f-a298-4af53a2acbec",
"name": "Sticky Note",
"type": "n8n-nodes-base.stickyNote",
"position": [
-700,
-260
],
"parameters": {
"width": 280,
"height": 180,
"content": "## Step 1
1. Set up a Milvus server based on [this guide](https://milvus.io/docs/install_standalone-docker-compose.md). And then create a collection named `my_collection`.
2. Click this workflow to load scrape and load Paul Graham essays to Milvus collection.
"
},
"typeVersion": 1
},
{
"id": "18f42da4-42ea-4eb0-9c43-ef8bd31ab7ff",
"name": "Sticky Note2",
"type": "n8n-nodes-base.stickyNote",
"position": [
-680,
460
],
"parameters": {
"height": 120,
"content": "## Step 2
Chat and get citations in response"
},
"typeVersion": 1
},
{
"id": "0af427ed-d901-4192-9fdc-986a63fd585b",
"name": "Embeddings OpenAI",
"type": "@n8n/n8n-nodes-langchain.embeddingsOpenAi",
"position": [
1020,
140
],
"parameters": {
"options": {}
},
"credentials": {
"openAiApi": {
"id": "hH2PTDH4fbS7fdPv",
"name": "OpenAi account"
}
},
"typeVersion": 1.2
},
{
"id": "3603852a-bf12-4289-9733-dcd29d12a4f6",
"name": "Default Data Loader",
"type": "@n8n/n8n-nodes-langchain.documentDefaultDataLoader",
"position": [
1160,
120
],
"parameters": {
"options": {},
"jsonData": "={{ $('Extract Text Only').item.json.data }}",
"jsonMode": "expressionData"
},
"typeVersion": 1
},
{
"id": "b49eb3ae-82cb-4d87-8f22-0789b3a14d83",
"name": "OpenAI Chat Model",
"type": "@n8n/n8n-nodes-langchain.lmChatOpenAi",
"position": [
680,
780
],
"parameters": {
"model": {
"__rl": true,
"mode": "list",
"value": "gpt-4o-mini"
},
"options": {}
},
"credentials": {
"openAiApi": {
"id": "hH2PTDH4fbS7fdPv",
"name": "OpenAi account"
}
},
"typeVersion": 1.2
}
],
"active": false,
"pinData": {},
"settings": {
"executionOrder": "v1"
},
"versionId": "5dc48a1d-aaf0-4052-9666-28f9e76d198c",
"connections": {
"Prepare chunks": {
"main": [
[
{
"node": "Answer the query based on chunks",
"type": "main",
"index": 0
}
]
]
},
"Fetch Essay List": {
"main": [
[
{
"node": "Extract essay names",
"type": "main",
"index": 0
}
]
]
},
"Limit to first 3": {
"main": [
[
{
"node": "Fetch essay texts",
"type": "main",
"index": 0
}
]
]
},
"Compose citations": {
"main": [
[
{
"node": "Generate response",
"type": "main",
"index": 0
}
]
]
},
"Embeddings OpenAI": {
"ai_embedding": [
[
{
"node": "Milvus Vector Store",
"type": "ai_embedding",
"index": 0
}
]
]
},
"Extract Text Only": {
"main": [
[
{
"node": "Milvus Vector Store",
"type": "main",
"index": 0
}
]
]
},
"Fetch essay texts": {
"main": [
[
{
"node": "Extract Text Only",
"type": "main",
"index": 0
}
]
]
},
"OpenAI Chat Model": {
"ai_languageModel": [
[
{
"node": "Answer the query based on chunks",
"type": "ai_languageModel",
"index": 0
}
]
]
},
"Embeddings OpenAI2": {
"ai_embedding": [
[
{
"node": "Milvus Vector Store in retrieval",
"type": "ai_embedding",
"index": 0
}
]
]
},
"Default Data Loader": {
"ai_document": [
[
{
"node": "Milvus Vector Store",
"type": "ai_document",
"index": 0
}
]
]
},
"Extract essay names": {
"main": [
[
{
"node": "Split out into items",
"type": "main",
"index": 0
}
]
]
},
"Split out into items": {
"main": [
[
{
"node": "Limit to first 3",
"type": "main",
"index": 0
}
]
]
},
"Set max chunks to send to model": {
"main": [
[
{
"node": "Milvus Vector Store in retrieval",
"type": "main",
"index": 0
}
]
]
},
"Answer the query based on chunks": {
"main": [
[
{
"node": "Compose citations",
"type": "main",
"index": 0
}
]
]
},
"Milvus Vector Store in retrieval": {
"main": [
[
{
"node": "Prepare chunks",
"type": "main",
"index": 0
}
]
]
},
"When clicking \"Execute Workflow\"": {
"main": [
[
{
"node": "Fetch Essay List",
"type": "main",
"index": 0
}
]
]
},
"Recursive Character Text Splitter1": {
"ai_textSplitter": [
[
{
"node": "Default Data Loader",
"type": "ai_textSplitter",
"index": 0
}
]
]
}
}
}功能特点
- 自动检测新邮件
- AI智能内容分析
- 自定义分类规则
- 批量处理能力
- 详细的处理日志
技术分析
节点类型及作用
- Manualtrigger
- Httprequest
- Html
- Splitout
- Limit
复杂度评估
配置难度：
维护难度：
扩展性：
实施指南
前置条件
- 有效的Gmail账户
- n8n平台访问权限
- Google API凭证
- AI分类服务订阅
配置步骤
- 在n8n中导入工作流JSON文件
- 配置Gmail节点的认证信息
- 设置AI分类器的API密钥
- 自定义分类规则和标签映射
- 测试工作流执行
- 配置定时触发器（可选）
关键参数
| 参数名称 | 默认值 | 说明 |
|---|---|---|
| maxEmails | 50 | 单次处理的最大邮件数量 |
| confidenceThreshold | 0.8 | 分类置信度阈值 |
| autoLabel | true | 是否自动添加标签 |
最佳实践
优化建议
- 定期更新AI分类模型以提高准确性
- 根据邮件量调整处理批次大小
- 设置合理的分类置信度阈值
- 定期清理过期的分类规则
安全注意事项
- 妥善保管API密钥和认证信息
- 限制工作流的访问权限
- 定期审查处理日志
- 启用双因素认证保护Gmail账户
性能优化
- 使用增量处理减少重复工作
- 缓存频繁访问的数据
- 并行处理多个邮件分类任务
- 监控系统资源使用情况
故障排除
常见问题
邮件未被正确分类
检查AI分类器的置信度阈值设置，适当降低阈值或更新训练数据。
Gmail认证失败
确认Google API凭证有效且具有正确的权限范围，重新进行OAuth授权。
调试技巧
- 启用详细日志记录查看每个步骤的执行情况
- 使用测试邮件验证分类逻辑
- 检查网络连接和API服务状态
- 逐步执行工作流定位问题节点
错误处理
工作流包含以下错误处理机制：
- 网络超时自动重试（最多3次）
- API错误记录和告警
- 处理失败邮件的隔离机制
- 异常情况下的回滚操作