AI自动化神器n8n，保姆级教程来了！ - 知识铺
Datawhale干货
作者：王晓亮，Datawhale成员
目录
作为一个程序员，我一直期望可以追踪 GitHub 的 Trending 项目，但是每次都要去 GitHub 上手动查看并挨个浏览 GitHub 仓库内容，非常麻烦。
在接触到 n8n 之后，我萌生一个想法，利用 n8n 自动化进行 GitHub Trending 项目的追踪， 并生成邮件通知，甚至自动化发布到我的博客站点中。通过周末两天时间完成该工作流的搭建，并成功运行。以下是 n8n 相关介绍及工作流搭建过程，分享给大家。
n8n 初识：一个自动化超级英雄
n8n 是 nodemation 的简称，是一个开源的、高度可扩展的工作流自动化工具。GitHub 仓库为 https://github.com/n8n-io/n8n ，当前 Star 数已经达到 108K。它允许你通过一个直观的可视化界面，将不同的应用、服务和数据连接起来，创建复杂的自动化流程，从而提升效率、减少重复性工作。
核心特点：
-
可视化工作流编辑器：n8n 提供一个直观的拖拽式界面，让你可以轻松地构建和管理复杂的自动化流程，即使是没有任何编程背景的用户也能快速上手。
-
丰富的节点库：拥有庞大的预构建节点库，支持连接几乎所有主流的应用和服务，包括各种 API、数据库（如 Supabase）、SaaS 平台（如 GitHub、Slack）、以及自定义 Webhooks 等。这使得数据流转和任务协同变得异常简单。
-
高度灵活性：支持各种复杂的逻辑控制，如条件分支（if/else）、循环（loop）、并行执行等， 你可以根据业务需求构建出高度定制化的自动化流程。
-
自托管/云服务选项：n8n 提供极大的部署灵活性。你可以选择将其私有部署在自己的服务器上， 完全掌控数据安全和隐私；同时，n8n 也提供云服务选项，方便快速启动和管理。
-
低代码/无代码特性：n8n 巧妙地平衡了易用性和功能性。对于非技术人员，它提供无代码的拖拽操作，实现快速自动化；而对于开发者，它允许在工作流中嵌入自定义代码（如 JavaScript），实现更高级、更个性化的功能扩展。
相较于市面上其他自动化工具，n8n 的核心优势在于其开源的本质和卓越的可扩展性。开源意味着透明、社区驱动的快速发展，并且可以根据你的具体需求进行深度定制。而其强大的节点系统和灵活的编程能力， 使得它不仅能满足日常的简单自动化需求，更能应对复杂的企业级工作流挑战，真正成为“自动化超级英雄”。
部署 n8n 环境：开启自动化之旅
n8n 提供多种自托管部署方式，且官方提供参考仓库 https://github.com/n8n-io/n8n-hosting 。
这里我使用 docker compose 部署，同时为了后期 OAuth 相关认证支持，还需要准备一个子域名用作 n8n 服务的域名。需提前将域名解析指向 n8n 服务的公网地址。
1. 在服务创建目录 mkdir n8n-compose && cd n8n-compose
2. 参考 官方示例（https://github.com/n8n-io/n8n-hosting/tree/main/docker-compose/withPostgresAndWorker）创建文件 docker-compose.yml 与 init-data.sh
3. 准备环境变量配置文件.env（更改下列配置项为实际值）
<span><span><span><span data-pm-slice="0 0 []"><span leaf="">POSTGRES_USER</span></span><span leaf="">=</span></span><span leaf="">changeUser</span><span leaf=""><br></span></span></span><span><span><span><span leaf="">POSTGRES_PASSWORD</span></span><span><span leaf="">=</span></span><span leaf="">changePassword</span><span leaf=""><br></span></span></span><span><span><span><span leaf="">POSTGRES_DB</span></span><span><span leaf="">=</span></span><span leaf="">n8n</span><span leaf=""><br></span></span></span><span><span><span leaf=""><br></span></span></span><span><span><span><span leaf="">POSTGRES_NON_ROOT_USER</span></span><span><span leaf="">=</span></span><span leaf="">changeUser</span><span leaf=""><br></span></span></span><span><span><span><span leaf="">POSTGRES_NON_ROOT_PASSWORD</span></span><span><span leaf="">=</span></span><span leaf="">changePassword</span><span leaf=""><br></span></span></span><span><span><span leaf=""><br></span></span></span><span><span><span><span leaf="">ENCRYPTION_KEY</span></span><span><span leaf="">=</span></span><span leaf="">changeEncryptionKey</span><span leaf=""><br></span></span></span><span><span><span leaf=""><br></span></span></span><span><span><span><span leaf="">N8N_EDITOR_BASE_URL</span></span><span><span leaf="">=</span></span><span leaf="">https://n8n.example.com</span><span leaf=""><br></span></span></span><span><span><span><span leaf="">WEBHOOK_URL</span></span><span><span leaf="">=</span></span><span leaf="">https://n8n.example.com</span></span></span>
4. 在容器配置文件中加上环境变量配置
<span><span><span><span leaf="">diff --git a/docker-compose/withPostgresAndWorker/docker-compose.yml b/docker-compose/withPostgresAndWorker/docker-compose.yml</span><span leaf=""><br></span></span></span></span><span><span><span><span leaf="">index b5b2de6..2d8cdfc 100644</span><span leaf=""><br></span></span></span></span><span><span><span></span><span><span leaf="">--- a/docker-compose/withPostgresAndWorker/docker-compose.yml</span><span leaf=""><br></span></span></span></span><span><span><span></span><span><span leaf="">+++ b/docker-compose/withPostgresAndWorker/docker-compose.yml</span><span leaf=""><br></span></span></span></span><span><span><span></span><span><span leaf="">@@ -19,6 +19,8 @@ x-shared: &shared</span><span leaf=""><br></span></span></span></span><span><span><span></span><span leaf=""> - QUEUE_BULL_REDIS_HOST=redis</span><span leaf=""><br></span></span></span><span><span><span leaf=""> - QUEUE_HEALTH_CHECK_ACTIVE=true</span><span leaf=""><br></span></span></span><span><span><span leaf=""> - N8N_ENCRYPTION_KEY=${ENCRYPTION_KEY}</span><span leaf=""><br></span></span></span><span><span><span><span leaf="">+ - N8N_EDITOR_BASE_URL=${N8N_EDITOR_BASE_URL}</span><span leaf=""><br></span></span></span></span><span><span><span><span leaf="">+ - WEBHOOK_URL=${WEBHOOK_URL}</span><span leaf=""><br></span></span></span></span><span><span><span></span><span leaf=""> links:</span><span leaf=""><br></span></span></span><span><span><span leaf=""> - postgres</span><span leaf=""><br></span></span></span><span><span><span leaf=""> - redis</span></span></span>
5. 启动容器 docker-compose up -d
6. 检查容器状态 docker-compose ps，可以得到如下输出
<span leaf=""><span>NAME</span> COMMAND SERVICE STATUS PORTS</span>
7. 配置反向代理服务，如 Caddyserver 可以添加如下配置至 /etc/caddy/Caddyfile
<span><span><span><span data-pm-slice="0 0 []"><span leaf="">n8n.example.com </span></span><span leaf="">{</span></span><span leaf=""><br></span></span></span><span><span><span><span leaf=""> reverse_proxy </span></span><span><span leaf="">localhost</span></span><span><span leaf="">:</span></span><span><span leaf="">5678</span></span><span leaf=""><br></span></span></span><span><span><span><span leaf="">}</span></span><span leaf=""><br></span></span></span>
经过如上配置，访问
https://n8n.example.com 就可以访问 n8n 的 UI 了。第一次访问需要配置管理员账号密码。
实现 GitHub Trending 每日追踪
接下来到了工作流搭建过程。先规划整体工作流运转流程：
-
定时获取 GitHub Trending 项目列表并存储
-
基于存储的项目列表，选择合适的项目进行 git 仓库信息获取，主要为项目 README 内容等
-
使用 LLM 对获取的信息进行文本总结，输出 Markdown 格式内容
-
将输出内容进行邮件发送或者存储至博客系统中
数据存储我使用 Supabase 的数据库，博客系统为我个人博客主站，使用 hugo 搭建并由 GitHub Actions 自动部署。整个过程中使用的核心 node 有
-
HTTP Request（https://docs.n8n.io/integrations/builtin/core-nodes/n8n-nodes-base.httprequest/）
-
GitHub（https://docs.n8n.io/integrations/builtin/app-nodes/n8n-nodes-base.github/）
-
Supabase（https://docs.n8n.io/integrations/builtin/app-nodes/n8n-nodes-base.supabase/）
-
Code（https://docs.n8n.io/integrations/builtin/core-nodes/n8n-nodes-base.code/）
-
Basic LLM Chain（https://docs.n8n.io/integrations/builtin/cluster-nodes/root-nodes/n8n-nodes-langchain.chainllm/）
获取 GitHub Trending 项目列表与信息总结、内容发布有不同的时间周期，我将其分为两个不同的工作流。
工作流 1 - 获取 GitHub Trending 项目列表并存储
由于 GitHub Trending 项目列表是动态的，且 GitHub 官方不提供 API。为此我写了一个 Python 脚本获取 GitHub Trending 项目列表，并包装成本地服务提供 API 供工作流使用，当然绝大部分代码由 LLM 生成。该服务已经开源，地址为https://github.com/tomowang/github-trending 。
为了让 n8n 节点可以访问该服务，我们需要将其运行在 n8n docker compose 对应的 network 中， 使用如下命令启动服务（其中n8n-compose_default为 n8n 所在的 docker-compose 的网络名称）：
<span><span><span data-pm-slice="0 0 []"><span leaf="">docker</span></span><span leaf=""> run -d </span><span><span leaf="">\</span></span></span></span><span><span><span></span><span leaf=""> -p 18000:8000 </span><span><span leaf="">\</span></span></span></span><span><span><span></span><span leaf=""> --name github-trending </span><span><span leaf="">\</span></span></span></span><span><span><span></span><span leaf=""> --restart always </span><span><span leaf="">\</span></span></span></span><span><span><span></span><span leaf=""> --network n8n-compose_default </span><span><span leaf="">\</span></span></span></span><span><span><span></span><span leaf=""> ghcr.io/tomowang/github-trending:latest</span></span></span>
在 n8n 中使用 HTTP Request 节点访问本地服务，配置如下（注意其中的 URL 部分，使用 docker 网络中的 service 作为域名）：
后续 Code 节点填充额外的时间、当前列表排名等信息，最后数据使用 Supabase 节点进行存储。Supabase 表结构如下：
<span><span><span><span leaf="">create </span></span><span></span><span><span leaf="">table </span></span><span></span><span><span leaf="">public</span></span><span><span leaf="">.</span></span><span><span leaf="">trending_repos </span></span><span></span><span><span leaf="">(</span></span><span><span leaf=""><br></span></span></span></span><span><span><span></span><span><span leaf=""> id </span></span><span></span><span><span leaf="">bigint </span></span><span></span><span><span leaf="">generated </span></span><span></span><span><span leaf="">by </span></span><span></span><span><span leaf="">default </span></span><span></span><span><span leaf="">as </span></span><span></span><span><span leaf="">identity </span></span><span></span><span><span leaf="">not </span></span><span></span><span><span leaf="">null</span></span><span><span leaf="">,</span></span><span><span leaf=""><br></span></span></span></span><span><span><span></span><span><span leaf=""> name </span></span><span></span><span><span leaf="">character </span></span><span></span><span><span leaf="">varying </span></span><span></span><span><span leaf="">not </span></span><span></span><span><span leaf="">null </span></span><span></span><span><span leaf="">default </span></span><span></span><span><span leaf="">''</span></span><span><span leaf="">::</span></span><span><span leaf="">character </span></span><span></span><span><span leaf="">varying</span></span><span><span leaf="">,</span></span><span><span leaf=""><br></span></span></span></span><span><span><span></span><span><span leaf=""> description </span></span><span></span><span><span leaf="">text </span></span><span></span><span><span leaf="">null </span></span><span></span><span><span leaf="">default </span></span><span></span><span><span leaf="">''</span></span><span><span leaf="">::</span></span><span><span leaf="">text</span></span><span><span leaf="">,</span></span><span><span leaf=""><br></span></span></span></span><span><span><span></span><span><span leaf=""> language </span></span><span></span><span><span leaf="">character </span></span><span></span><span><span leaf="">varying </span></span><span></span><span><span leaf="">null </span></span><span></span><span><span leaf="">default </span></span><span></span><span><span leaf="">''</span></span><span><span leaf="">::</span></span><span><span leaf="">character</span></span><span></span><span><span leaf="">varying</span></span><span><span leaf="">,</span></span><span><span leaf=""><br></span></span></span></span><span><span><span></span><span><span leaf=""> stars </span></span><span></span><span><span leaf="">integer </span></span><span></span><span><span leaf="">null</span></span><span><span leaf="">,</span></span><span><span leaf=""><br></span></span></span></span><span><span><span></span><span><span leaf=""> forks </span></span><span></span><span><span leaf="">integer </span></span><span></span><span><span leaf="">null</span></span><span><span leaf="">,</span></span><span><span leaf=""><br></span></span></span></span><span><span><span></span><span><span leaf=""> today_stars </span></span><span></span><span><span leaf="">integer </span></span><span></span><span><span leaf="">null</span></span><span><span leaf="">,</span></span><span><span leaf=""><br></span></span></span></span><span><span><span></span><span><span leaf=""> built_by </span></span><span></span><span><span leaf="">text</span></span><span><span leaf="">[] </span></span><span></span><span><span leaf="">null</span></span><span><span leaf="">,</span></span><span><span leaf=""><br></span></span></span></span><span><span><span></span><span><span leaf=""> date </span></span><span></span><span><span leaf="">timestamp </span></span><span></span><span><span leaf="">without </span></span><span></span><span><span leaf="">time </span></span><span></span><span><span leaf="">zone </span></span><span></span><span><span leaf="">null</span></span><span><span leaf="">,</span></span><span><span leaf=""><br></span></span></span></span><span><span><span></span><span><span leaf=""> rank </span></span><span></span><span><span leaf="">smallint </span></span><span></span><span><span leaf="">null</span></span><span><span leaf="">,</span></span><span><span leaf=""><br></span></span></span></span><span><span><span></span><span><span leaf=""> constraint </span></span><span></span><span><span leaf="">trending_repos_pkey </span></span><span></span><span><span leaf="">primary</span></span><span></span><span><span leaf="">key</span></span><span></span><span><span leaf="">(</span></span><span><span leaf="">id</span></span><span><span leaf="">)</span></span><span><span leaf=""><br></span></span></span></span><span><span><span></span><span><span leaf="">)</span></span><span></span><span><span leaf="">TABLESPACE</span></span><span></span><span><span leaf="">pg_default</span></span><span><span leaf="">;</span></span></span></span>
使用 Supabase 需要我们添加 Supabase 的认证信息
-
获取 Supabase Project URL 信息：Project Settings -> Data API -> Project URL
-
获取 Supabase Project API Key 信息：Project Settings -> API Keys -> service_role
详细参考官方文档：
https://docs.n8n.io/integrations/builtin/credentials/supabase/ 。
运行工作流，我们可以在 Supabase 控制台看到数据
至此该工作流就创建完成了，我们可以将工作流状态设置为 active，通过定时触发器，每小时获取获取一次数据并进行存储，作为后续工作流的数据源。
工作流 2 - GitHub Trending 项目 AI 总结与自动化发布
有了每小时的 GitHub Trending 项目列表数据，我们就可以开始进行项目信息总结了。我们需要规划下整个工作流运转流程，有几个需求需要提前梳理：
-
由于我们每小时抓取一次 GitHub Trending 项目列表，因此我们需要一些策略来挑选我们最终用 AI 进行总结项目
-
对于已经总结以及推送过的项目，需要避免重复推送
-
项目的信息需要从 GitHub 获取，特别是项目的 README 信息
关于第二点，存储已经推送过的项目，我们同样可以使用Supabase的数据库进行存储，表结构如下：
<span><span><span></span><span><span data-pm-slice="0 0 []"><span leaf="">create </span></span><span></span><span leaf="">table </span></span><span></span><span><span leaf="">public</span></span><span><span leaf="">.</span></span><span><span leaf="">pushed_repos </span></span><span></span><span><span leaf="">(</span></span><span><span leaf=""><br></span></span></span></span><span><span><span></span><span><span leaf=""> id </span></span><span></span><span><span leaf="">integer </span></span><span></span><span><span leaf="">generated </span></span><span></span><span><span leaf="">by </span></span><span></span><span><span leaf="">default </span></span><span></span><span><span leaf="">as </span></span><span></span><span><span leaf="">identity </span></span><span></span><span><span leaf="">not </span></span><span></span><span><span leaf="">null</span></span><span><span leaf="">,</span></span><span><span leaf=""><br></span></span></span></span><span><span><span></span><span><span leaf=""> pushed_at </span></span><span></span><span><span leaf="">timestamp </span></span><span></span><span><span leaf="">with </span></span><span></span><span><span leaf="">time </span></span><span></span><span><span leaf="">zone </span></span><span></span><span><span leaf="">not </span></span><span></span><span><span leaf="">null </span></span><span></span><span><span leaf="">default </span></span><span></span><span><span leaf="">now</span></span><span><span leaf="">(),</span></span><span><span leaf=""><br></span></span></span></span><span><span><span></span><span><span leaf=""> name </span></span><span></span><span><span leaf="">character </span></span><span></span><span><span leaf="">varying </span></span><span></span><span><span leaf="">not </span></span><span></span><span><span leaf="">null</span></span><span><span leaf="">,</span></span><span><span leaf=""><br></span></span></span></span><span><span><span></span><span><span leaf=""> constraint </span></span><span></span><span><span leaf="">pushed_repos_pkey </span></span><span></span><span><span leaf="">primary </span></span><span></span><span><span leaf="">key</span></span><span></span><span><span leaf="">(</span></span><span><span leaf="">id</span></span><span><span leaf="">)</span></span><span><span leaf=""><br></span></span></span></span><span><span><span></span><span><span leaf="">) </span></span><span></span><span><span leaf="">TABLESPACE</span></span><span></span><span><span leaf="">pg_default</span></span><span><span leaf="">;</span></span></span></span>
所以整体流程大概如下：
-
从之前存储的 Supabase 中获取项目列表
-
剔除其中已经推送的项目
-
通过一定策略挑选需要推送的项目
-
获取项目信息，主要为 README
-
使用大语言模型针，结合项目信息以及 README 信息，进行总结
-
使用 Email 节点推送总结
-
将总结通过 GitHub 节点发布到博客系统
-
将推送的项目信息存储到 Supabase 中
完成后的 n8n 工作流如下：
接下来我们对其中的一些关键节点及流程进行说明。
Code Node & Built-in methods and variables
首先是我们的第一个代码节点Filter out pushed repos
n8n 支持两种语言的代码节点，JavaScript & Python，其中 Python 使用Pyodide 提供支持。在示例中，我们使用 Python 节点。在 n8n 中，节点默认对工作流中传递的数据进行 循环调用（https://docs.n8n.io/flow-logic/looping/）。但是我们期望针对获取到的 GitHub 项目信息进行统一处理，所以在示例中的代码节点我们配置其运行模式为Run Once for All Items。
官方文档提示：The Code node takes longer to process Python than JavaScript. This is due to the extra compilation steps. 所以如果追求极致的性能，可以尽量使用 JavaScript 代码节点
在代码中，我们需要使用内置的方法及变量来获取上游节点传递的数据。JavaScript 以及 Python 使用不同的命名方式进行变量获取。
| JavaScript
|
Python
|
描述
$input.item |
_input.item |
当前节点关联的输入数据
|
| $input.all() | _input.all() |
当前节点关联的所有输入
|
| $json | _json | $input.item.json
或者_input.item.json 的快速访问
|
| $("<node-name>").all() | _("<node-name>").all() |
获取特定节点的全部输入
|
当我们对数据进行处理后，需要返回数据给下游节点进行后续处理。在 n8n 中，节点中的数据传输使用 对象数组（https://docs.n8n.io/data/data-structure/） ，其结构如下
<span><span><span><span leaf="">[</span></span><span leaf=""><br></span></span></span><span><span><span><span leaf=""> {</span></span><span leaf=""><br></span></span></span><span><span><span><span leaf=""> // For most data:</span><span leaf=""><br></span></span></span></span><span><span><span></span><span><span leaf=""> // Wrap each item in another object, with the key 'json'</span><span leaf=""><br></span></span></span></span><span><span><span></span><span><span leaf=""> "json"</span></span><span><span leaf="">:</span></span><span><span leaf="">{</span></span><span leaf=""><br></span></span></span><span><span><span><span leaf=""> // Example data</span><span leaf=""><br></span></span></span></span><span><span><span></span><span><span leaf=""> "apple"</span></span><span><span leaf="">:</span></span><span><span leaf="">"beets"</span></span><span><span leaf="">,</span></span><span leaf=""><br></span></span></span><span><span><span><span leaf=""> "carrot"</span></span><span><span leaf="">:</span></span><span><span leaf="">{</span></span><span leaf=""><br></span></span></span><span><span><span><span leaf=""> "dill"</span></span><span><span leaf="">:</span></span><span><span leaf="">1</span></span><span leaf=""><br></span></span></span><span><span><span><span leaf=""> }</span></span><span leaf=""><br></span></span></span><span><span><span><span leaf=""> },</span></span><span leaf=""><br></span></span></span><span><span><span><span leaf=""> // For binary data:</span><span leaf=""><br></span></span></span></span><span><span><span></span><span><span leaf=""> // Wrap each item in another object, with the key 'binary'</span><span leaf=""><br></span></span></span></span><span><span><span></span><span><span leaf=""> "binary"</span></span><span><span leaf="">:</span></span><span><span leaf="">{</span></span><span leaf=""><br></span></span></span><span><span><span><span leaf=""> // Example data</span><span leaf=""><br></span></span></span></span><span><span><span></span><span><span leaf=""> "apple-picture"</span></span><span><span leaf="">:</span></span><span><span leaf="">{</span></span><span leaf=""><br></span></span></span><span><span><span><span leaf=""> "data"</span></span><span><span leaf="">:</span></span><span><span leaf="">"...."</span></span><span><span leaf="">,</span></span><span><span leaf="">// Base64 encoded binary data (required)</span><span leaf=""><br></span></span></span></span><span><span><span></span><span><span leaf=""> "mimeType"</span></span><span><span leaf="">:</span></span><span><span leaf="">"image/png"</span></span><span><span leaf="">,</span></span><span><span leaf="">// Best practice to set if possible (optional)</span><span leaf=""><br></span></span></span></span><span><span><span></span><span><span leaf=""> "fileExtension"</span></span><span><span leaf="">:</span></span><span><span leaf="">"png"</span></span><span><span leaf="">,</span></span><span><span leaf="">// Best practice to set if possible (optional)</span><span leaf=""><br></span></span></span></span><span><span><span></span><span><span leaf=""> "fileName"</span></span><span><span leaf="">:</span></span><span><span leaf="">"example.png"</span></span><span><span leaf="">// Best practice to set if possible (optional)</span><span leaf=""><br></span></span></span></span><span><span><span></span><span><span leaf=""> }</span></span><span leaf=""><br></span></span></span><span><span><span><span leaf=""> }</span></span><span leaf=""><br></span></span></span><span><span><span><span leaf=""> }</span></span><span leaf=""><br></span></span></span><span><span><span><span leaf="">]</span></span></span></span>
所有在示例的代码节点中，最后返回过滤后的item.json数组，完整的代码如下：
<span><span><span><span leaf=""># Loop over input items and add a new field called 'myNewField' to the JSON of each one</span></span><span leaf=""><br></span></span></span><span><span><span><span leaf="">trending_repos</span></span><span><span leaf="">=</span></span><span><span leaf="">_</span></span><span><span leaf="">(</span></span><span><span leaf="">'Last 24h trending repo'</span></span><span><span leaf="">)</span></span><span><span leaf="">.</span></span><span><span leaf="">all</span></span><span><span leaf="">()</span></span><span leaf=""><br></span></span></span><span><span><span><span leaf="">pushed_repos</span></span><span><span leaf="">=</span></span><span><span leaf="">[</span></span><span><span leaf="">item</span></span><span><span leaf="">.</span></span><span><span leaf="">json</span></span><span><span leaf="">[</span></span><span><span leaf="">'name'</span></span><span><span leaf="">]</span></span><span><span leaf="">for</span></span><span><span leaf="">item</span></span><span><span leaf="">in</span></span><span><span leaf="">_</span></span><span><span leaf="">(</span></span><span><span leaf="">'Last 30d push history'</span></span><span><span leaf="">)</span></span><span><span leaf="">.</span></span><span><span leaf="">all</span></span><span><span leaf="">()]</span></span><span leaf=""><br></span></span></span><span><span><span><span leaf="">print</span></span><span><span leaf="">(</span></span><span><span leaf="">"pushed_repos: "</span></span><span><span leaf="">,</span></span><span><span leaf="">pushed_repos</span></span><span><span leaf="">)</span></span><span leaf=""><br></span></span></span><span><span><span leaf=""><br></span></span></span><span><span><span><span leaf="">return</span></span><span><span leaf="">[</span></span><span leaf=""><br></span></span></span><span><span><span><span leaf=""> item</span></span><span><span leaf="">.</span></span><span><span leaf="">json</span></span><span><span leaf="">for</span></span><span><span leaf="">item</span></span><span><span leaf="">in</span></span><span><span leaf="">trending_repos</span></span><span><span leaf="">if</span></span><span><span leaf="">item</span></span><span><span leaf="">.</span></span><span><span leaf="">json</span></span><span><span leaf="">[</span></span><span><span leaf="">'name'</span></span><span><span leaf="">]</span></span><span><span leaf="">not</span></span><span><span leaf="">in</span></span><span><span leaf="">pushed_repos</span></span><span leaf=""><br></span></span></span><span><span><span><span leaf="">]</span></span></span></span>
另外，如果需要对代码进行调试，可以使用console.log 及 print()，相关的调试结果会输出在浏览器的调试控制台中。
1. HTTP Request Node
由于原生的 GitHub 节点不支持获取项目的 README 信息，我们使用 HTTP Request 节点来进行自定义 API 调用，这同样也适用于其他提供 HTTP 接口但是没有节点支持的系统。
GitHub 仓库的 README 信息可以通过GET /repos/{owner}/{repo}/readme API 获取，具体 API 文档可以参考：https://docs.github.com/en/rest/repos/contents?apiVersion=2022-11-28#get-a-repository-readme。
在示例中，我们使用 GET 方法，并设置 Accept 为application/vnd.github.raw+json，以获取原始的 README 内容。
对于请求的 URL，我们使用{{ $json.name }}来获取上游节点传递的项目名称信息（在这种情况下只支持 JavaScript 代码，并使用{{ }}运行 表达式（https://docs.n8n.io/code/expressions/ ）。
对于请求的认证 Authentication 部分，n8n 建议在可用时使用预定义凭证类型选项。这里我们通过预定义凭证类型 GitHub API 来进行认证。
2. LLM Chain Node
n8n 提供丰富的 LLM 相关的节点，隶属于集群节点 Cluster Nodes（https://docs.n8n.io/integrations/builtin/cluster-nodes/）， 集群节点是一起工作的节点组，包含一个根节点以及多个子节点。
在示例中，我们使用Basic LLM Chain节点对项目信息进行总结，其包含一个根节点以及一个大模型子节点。
对于提示词，我们使用预设模式。在提示词中，通过表达式获取之前节点获取的 GitHub 项目信息以及 README 信息，提示词中相关片段如下：
<span><span><span leaf="">以下为Github仓库名称 `{{ $('Calc top repos').item.json.name }}`。</span><span leaf=""><br></span></span></span><span><span><span leaf="">仓库信息：</span><span leaf=""><br></span></span></span><span><span><span leaf="">```json</span><span leaf=""><br></span></span></span><span><span><span leaf="">{{ JSON.stringify($('Calc top repos').item.json, null, 2) }}</span><span leaf=""><br></span></span></span><span><span><span leaf="">```</span><span leaf=""><br></span></span></span><span><span><span leaf=""><br></span></span></span><span><span><span leaf="">仓库原始README信息：</span><span leaf=""><br></span></span></span><span><span><span leaf="">```markdown</span><span leaf=""><br></span></span></span><span><span><span leaf="">{{ $json.data }}</span><span leaf=""><br></span></span></span><span><span><span leaf="">```</span></span></span>
Basic LLM Chain 需要绑定一个大语言模型，n8n 提供了多个大语言模型节点，如 OpenAI、Hugging Face 等。
示例中我们使用 Google Gemini Chat Model，选择该类型节点后，添加对应的鉴权信息（API Key 可以通过 https://aistudio.google.com/app/apikey 创建），并选择模型版本， 如 gemini-2.5-flash-preview-05-20，这样 Basic LLM Chain 节点就具有了大模型能力，能够根据提示词信息进行文本生成工作。
展望
基于上述的工作流，我们能够实现 GitHub 项目信息自动抓取、过滤、总结，并自动发送邮件以及发布至博客中。通过 Execution 面板，我们可以观测工作流的运转过程，并看到每个节点实际的输入输出数据，甚至预估的 Token 数。
最终自动发布的博客地址为 https://tomo.dev/aigc/ ，自动发送的邮件如下：
该工作流展示了 n8n 强大的集成能力，以及灵活的节点配置方式。通过类似的方法，我们可以拓展更多数据源获取、处理以及大语言模型解读等场景，实现更多自动化工作。
最后附上工作流文件，方便大家参考。
https://tomo.dev/posts/n8n-workflow-for-daily-github-trending-auto-posting/
**一起“**点****赞”三连↓
- 原文作者：知识铺
- 原文链接：https://index.zshipu.com/ai/post/202510/AI%E8%87%AA%E5%8A%A8%E5%8C%96%E7%A5%9E%E5%99%A8n8n%E4%BF%9D%E5%A7%86%E7%BA%A7%E6%95%99%E7%A8%8B%E6%9D%A5%E4%BA%86/
- 版权声明：本作品采用知识共享署名-非商业性使用-禁止演绎 4.0 国际许可协议进行许可，非商业转载请注明出处（作者，原文链接），商业转载请联系作者获得授权。
- 免责声明：本页面内容均来源于站内编辑发布，部分信息来源互联网，并不意味着本站赞同其观点或者证实其内容的真实性，如涉及版权等问题，请立即联系客服进行更改或删除，保证您的合法权益。转载请注明来源，欢迎对文章中的引用来源进行考证，欢迎指出任何有错误或不够清晰的表达。也可以邮件至 sblig@126.com
