696 lines
18 KiB
Markdown
696 lines
18 KiB
Markdown
---
|
||
title: 安装 Playwright 浏览器
|
||
source:
|
||
author: shenwei
|
||
published:
|
||
created: 2025-11-11
|
||
description: 可自动化、可扩展、AI增强的电商数据采集与处理系统
|
||
tags: []
|
||
link:
|
||
---
|
||
|
||
|
||
|
||
|
||
非常好的思路——你想要的是一个**可自动化、可扩展、AI增强的数据采集与处理系统**，基于 Docker + Ubuntu + n8n 搭建。下面我给你一个详细的设计与工具选择建议，从爬取到分析的整体架构。
|
||
|
||
---
|
||
|
||
## 🧩 一、系统整体架构建议
|
||
|
||
你的目标系统可以分为三个层次：
|
||
|
||
|层次|组件|说明|
|
||
|---|---|---|
|
||
|**数据采集层（爬虫）**|Scrapy / Playwright / Selenium / Apify|从各大电商网站采集结构化信息（标题、描述、图片、视频等）|
|
||
|**数据处理层（自动化管道）**|n8n + LLM API (e.g., OpenAI, Ollama, LM Studio)|对采集数据进行清洗、分类、摘要、翻译、属性提取等AI处理|
|
||
|**存储与展示层**|PostgreSQL / SQLite + MinIO / NAS + Grafana / Metabase|存储文本、图片和视频元数据，并可视化结果|
|
||
|
||
---
|
||
|
||
## 🕷️ 二、爬虫工具推荐与对比
|
||
|
||
|工具|适用场景|优点|缺点|
|
||
|---|---|---|---|
|
||
|**Scrapy**|静态页面、电商产品信息|轻量高效、插件生态丰富、可Docker化部署|对JS渲染页面支持弱，需要配合Splash或Playwright|
|
||
|**Playwright (Python/Node.js)**|动态渲染页面、滚动加载、视频图片加载|可模拟浏览器、支持无头模式、可靠性高|相对重，适合单站点深度采集|
|
||
|**Apify (Open Source SDK)**|通用网页爬取+API接口+调度|已内置防封禁策略、支持Docker|学习曲线略陡、对纯本地化部署需定制|
|
||
|**Colly (Go语言)**|高性能爬虫服务、轻量API爬取|性能强、可编译为二进制Docker镜像|JS支持弱，不适合电商复杂页面|
|
||
|**Crawlee (Node.js)**|Apify的开源核心框架，支持Playwright/Selenium|与n8n、LangChain容易集成|需要JS/TS开发基础|
|
||
|
||
**推荐组合：**
|
||
|
||
> ✅ **Scrapy + Playwright（或Crawlee + Playwright）**
|
||
|
||
- Scrapy 负责结构化抓取、分页调度、下载媒体；
|
||
|
||
- Playwright 负责加载动态页面；
|
||
|
||
- 两者可通过 Docker Compose 容器化；
|
||
|
||
- 输出 JSON 或 CSV 文件供 n8n 消费。
|
||
|
||
|
||
---
|
||
|
||
## ⚙️ 三、Docker 架构示例
|
||
|
||
|
||
Scrapy 爬完的数据（JSON/CSV）存放在 `/data`，由 n8n 定时读取和处理。
|
||
|
||
---
|
||
|
||
## 🤖 四、n8n 中的AI集成与自动化
|
||
|
||
在 n8n 中可以通过 workflow 实现整个管线自动化：
|
||
|
||
**工作流示例：**
|
||
|
||
1. **Cron Trigger** → 定时启动
|
||
|
||
2. **Execute Command Node** → 执行 `docker exec scraper scrapy crawl amazon`
|
||
|
||
3. **Read File Node** → 读取爬取的 JSON 文件
|
||
|
||
4. **OpenAI / Ollama Node** → 调用 LLM 对产品描述提取属性（品牌、型号、价格区间、关键词等）
|
||
|
||
5. **Database Node** → 写入 PostgreSQL / SQLite
|
||
|
||
6. **Webhook / Email Node** → 生成报表或通知
|
||
|
||
|
||
你也可以用 n8n 的 “HTTP Request Node” 抓取 API 页面（适合公开电商API如阿里、Shopee）。
|
||
|
||
---
|
||
|
||
## 🧠 五、AI 处理建议
|
||
|
||
结合 LLM 做如下任务：
|
||
|
||
- **内容摘要与分类**（提炼产品特点）
|
||
|
||
- **多语言翻译**
|
||
|
||
- **特征提取**（品牌、价格、类别）
|
||
|
||
- **异常检测**（自动识别异常价格或缺图产品）
|
||
|
||
- **结构化JSON输出**，便于数据库存储
|
||
|
||
|
||
可以本地使用 **Ollama** (Mistral, Llama3) 模型，通过 n8n 的 HTTP Request 调用本地 `http://localhost:11434/api/generate`。
|
||
|
||
---
|
||
|
||
## 🗄️ 六、数据存储与管理建议
|
||
|
||
|数据类型|存储建议|工具|
|
||
|---|---|---|
|
||
|文本/结构化数据|PostgreSQL / SQLite|适合AI分析、可视化|
|
||
|图片/视频|MinIO / 本地NAS|模拟S3存储方式|
|
||
|Dashboard|Grafana / Metabase|生成电商趋势与分析报表|
|
||
|
||
---
|
||
|
||
## 🛡️ 七、防封与安全建议
|
||
|
||
- 使用 **User-Agent轮换**（Scrapy 内置 `scrapy-user-agents`）
|
||
|
||
- **代理池**（BrightData / ScraperAPI / 免费代理）
|
||
|
||
- 对同一网站设置 **下载延迟 + 随机化访问**
|
||
|
||
- **分布式调度** 可用 `Scrapyd` 或 `Archetype`（Scrapy集群）
|
||
|
||
- 不采集个人隐私信息或侵权内容（仅公开页面信息）
|
||
|
||
|
||
---
|
||
|
||
## 🚀 八、后续可扩展方向
|
||
|
||
- 增加 **FastAPI 服务层**：暴露 REST API 给前端或 BI 工具；
|
||
|
||
- 使用 **LangChain** 结合 Vector DB（e.g. Qdrant / Milvus）存储产品语义信息；
|
||
|
||
- n8n 可触发自动生成市场趋势报告（AI Summarizer）；
|
||
|
||
- 可视化电商类目趋势（Grafana + Postgres）。
|
||
|
||
|
||
---
|
||
|
||
## ✅ 推荐起步路径
|
||
|
||
1. **在 Ubuntu 上安装 Docker + Docker Compose**
|
||
|
||
2. 启动基础环境：`scrapy + playwright + n8n`
|
||
|
||
3. 选择 1–2 个电商站点（如 Amazon / JD / Taobao）
|
||
|
||
4. 构建 Scrapy 爬虫模板
|
||
|
||
5. 用 n8n 处理数据并测试 AI 工作流
|
||
|
||
6. 逐步扩展至全自动管线
|
||
|
||
|
||
---
|
||
|
||
太好了 ✅ 我将为你生成一个**可直接部署的爬虫自动化系统骨架**，基于以下目标设计：
|
||
|
||
> **目标：** 在 Ubuntu 上使用 Docker Compose 部署一个自动化电商爬虫系统，采集电商网站产品信息 → 存入本地文件 → 由 n8n 自动读取并利用 AI 处理。
|
||
|
||
下面是完整结构与代码示例。
|
||
|
||
---
|
||
|
||
## 🧩 一、项目目录结构
|
||
|
||
```bash
|
||
ecommerce-crawler/
|
||
├── docker-compose.yml
|
||
├── scrapy/
|
||
│ ├── Dockerfile
|
||
│ ├── requirements.txt
|
||
│ ├── scrapy.cfg
|
||
│ └── ecommerce/
|
||
│ ├── __init__.py
|
||
│ ├── items.py
|
||
│ ├── middlewares.py
|
||
│ ├── pipelines.py
|
||
│ ├── settings.py
|
||
│ └── spiders/
|
||
│ └── amazon_spider.py
|
||
├── n8n_data/ # n8n 数据持久化
|
||
├── data/ # 爬取结果输出目录
|
||
└── README.md
|
||
```
|
||
|
||
---
|
||
|
||
## 🧱 二、`docker-compose.yml`
|
||
|
||
```yaml
|
||
version: '3.8'
|
||
services:
|
||
# Scrapy 爬虫容器
|
||
scraper:
|
||
build: ./scrapy
|
||
container_name: ecommerce-scraper
|
||
working_dir: /app/scrapy
|
||
volumes:
|
||
- ./data:/app/data
|
||
depends_on:
|
||
- playwright
|
||
environment:
|
||
- PLAYWRIGHT_BROWSERS_PATH=/ms-playwright
|
||
networks:
|
||
- crawler-net
|
||
|
||
|
||
# Playwright 浏览器支持容器
|
||
playwright:
|
||
image: mcr.microsoft.com/playwright/python:v1.48.0-jammy
|
||
shm_size: 2gb
|
||
networks:
|
||
- crawler-net
|
||
|
||
# n8n 自动化平台
|
||
#n8n:
|
||
# image: n8nio/n8n:latest
|
||
# container_name: n8n
|
||
# ports:
|
||
# - 5678:5678
|
||
# environment:
|
||
# - N8N_BASIC_AUTH_ACTIVE=true
|
||
# - N8N_BASIC_AUTH_USER=admin
|
||
# - N8N_BASIC_AUTH_PASSWORD=changeme
|
||
# - N8N_PATH=/workflows
|
||
# volumes:
|
||
# - ./n8n_data:/home/node/.n8n
|
||
# - ./data:/data
|
||
# networks:
|
||
# - crawler-net
|
||
|
||
networks:
|
||
crawler-net:
|
||
```
|
||
|
||
---
|
||
|
||
## 🐍 三、Scrapy 部分
|
||
|
||
### `scrapy/Dockerfile`
|
||
|
||
```dockerfile
|
||
FROM mcr.microsoft.com/playwright/python:v1.48.0-jammy
|
||
|
||
WORKDIR /app
|
||
COPY requirements.txt .
|
||
RUN pip install --no-cache-dir -r requirements.txt
|
||
COPY . .
|
||
|
||
# 安装 Playwright 浏览器
|
||
RUN playwright install
|
||
|
||
WORKDIR /app
|
||
CMD ["scrapy", "crawl", "amazon"]
|
||
|
||
```
|
||
|
||
---
|
||
|
||
### `scrapy/requirements.txt`
|
||
|
||
```txt
|
||
scrapy==2.13.3
|
||
playwright==1.48.0
|
||
scrapy-playwright==0.0.44
|
||
```
|
||
|
||
> 说明：`scrapy-playwright` 插件可直接让 Scrapy 调用 Playwright 渲染动态页面，非常适合电商网站。
|
||
|
||
---
|
||
|
||
### `scrapy/scrapy.cfg`
|
||
|
||
```ini
|
||
[settings]
|
||
default = settings
|
||
|
||
[deploy]
|
||
# 如果你将来要用 scrapyd 部署，可以在这里定义目标（可忽略）
|
||
# url = http://localhost:6800/
|
||
# project = crawler
|
||
|
||
```
|
||
|
||
---
|
||
|
||
### `scrapy/ecommerce/settings.py`
|
||
|
||
```python
|
||
BOT_NAME = "scrapy"
|
||
|
||
SPIDER_MODULES = ["spiders"] # 指向当前目录下的 spiders
|
||
NEWSPIDER_MODULE = "spiders" # 新建 spider 时默认放在这里
|
||
|
||
ROBOTSTXT_OBEY = False
|
||
DOWNLOAD_DELAY = 2
|
||
|
||
DOWNLOAD_HANDLERS = {
|
||
"http": "scrapy_playwright.handler.ScrapyPlaywrightDownloadHandler",
|
||
"https": "scrapy_playwright.handler.ScrapyPlaywrightDownloadHandler",
|
||
}
|
||
|
||
TWISTED_REACTOR = "twisted.internet.asyncioreactor.AsyncioSelectorReactor"
|
||
|
||
PLAYWRIGHT_LAUNCH_OPTIONS = {
|
||
"headless": True,
|
||
"args": ["--no-sandbox", "--disable-setuid-sandbox"],
|
||
}
|
||
PLAYWRIGHT_BROWSER_TYPE = "chromium"
|
||
|
||
FEEDS = {
|
||
"/app/data/amazon.json": {"format": "json", "overwrite": True},
|
||
}
|
||
|
||
```
|
||
|
||
---
|
||
|
||
### `scrapy/ecommerce/items.py`
|
||
|
||
```python
|
||
import scrapy
|
||
|
||
class ProductItem(scrapy.Item):
|
||
title = scrapy.Field()
|
||
price = scrapy.Field()
|
||
rating = scrapy.Field()
|
||
image_urls = scrapy.Field()
|
||
images = scrapy.Field()
|
||
product_url = scrapy.Field()
|
||
```
|
||
|
||
---
|
||
|
||
### `scrapy/ecommerce/pipelines.py`
|
||
|
||
```python
|
||
import json
|
||
|
||
class JsonWriterPipeline:
|
||
def open_spider(self, spider):
|
||
self.file = open("data/products.json", "w", encoding="utf-8")
|
||
self.file.write("[\n")
|
||
|
||
def close_spider(self, spider):
|
||
self.file.write("]\n")
|
||
self.file.close()
|
||
|
||
def process_item(self, item, spider):
|
||
line = json.dumps(dict(item), ensure_ascii=False) + ",\n"
|
||
self.file.write(line)
|
||
return item
|
||
```
|
||
|
||
---
|
||
|
||
### `scrapy/ecommerce/spiders/amazon_spider.py`
|
||
|
||
> ⚠️ 仅作演示用途，使用通用搜索页采集，不涉及登录或侵权内容。
|
||
|
||
```python
|
||
import scrapy
|
||
from scrapy_playwright.page import PageMethod
|
||
|
||
class QuotesSpider(scrapy.Spider):
|
||
name = "amazon"
|
||
start_urls = ["https://quotes.toscrape.com/js/"]
|
||
|
||
custom_settings = {
|
||
"PLAYWRIGHT_LAUNCH_OPTIONS": {"headless": True},
|
||
"PLAYWRIGHT_CONTEXT_ARGS": {"viewport": {"width": 1280, "height": 720}},
|
||
}
|
||
|
||
def start_requests(self):
|
||
for url in self.start_urls:
|
||
yield scrapy.Request(
|
||
url,
|
||
meta={
|
||
"playwright": True,
|
||
"playwright_page_methods": [
|
||
PageMethod("wait_for_selector", "div.quote")
|
||
],
|
||
},
|
||
)
|
||
|
||
async def parse(self, response):
|
||
quotes = response.css("div.quote")
|
||
for quote in quotes:
|
||
text = quote.css("span.text::text").get()
|
||
author = quote.css("small.author::text").get()
|
||
yield {"text": text, "author": author}
|
||
|
||
# 翻页
|
||
next_page = response.css("li.next a::attr(href)").get()
|
||
if next_page:
|
||
next_url = response.urljoin(next_page)
|
||
yield scrapy.Request(
|
||
next_url,
|
||
meta={
|
||
"playwright": True,
|
||
"playwright_page_methods": [
|
||
PageMethod("wait_for_selector", "div.quote")
|
||
],
|
||
},
|
||
callback=self.parse
|
||
)
|
||
```
|
||
|
||
### `scrapy/ecommerce/spiders/__init__.py`
|
||
|
||
```python
|
||
#empty
|
||
```
|
||
|
||
|
||
|
||
---
|
||
|
||
## ⚙️ 四、运行方式
|
||
|
||
```bash
|
||
# 1. 构建并启动所有容器
|
||
docker compose up --build -d
|
||
|
||
# 2. 进入 Scrapy 容器执行爬虫
|
||
docker exec -it ecommerce-scraper scrapy crawl amazon
|
||
|
||
# 3. 查看数据
|
||
cat data/products.json
|
||
```
|
||
|
||
### Docker 命令
|
||
|
||
|
||
``` bash
|
||
docker compose down -v
|
||
docker compose up --build -d
|
||
|
||
docker exec -it ecommerce-scraper bash
|
||
|
||
#进入交互模式
|
||
docker run -it --entrypoint bash ecommerce-scraper
|
||
```
|
||
---
|
||
|
||
## 🤖 五、n8n 自动化流程（示例）
|
||
|
||
你可以在 n8n Web UI（[http://localhost:5678）导入如下逻辑：](http://localhost:5678%EF%BC%89%E5%AF%BC%E5%85%A5%E5%A6%82%E4%B8%8B%E9%80%BB%E8%BE%91%EF%BC%9A/)
|
||
|
||
**Workflow 示例逻辑：**
|
||
|
||
1. **Cron Node**：每天凌晨触发
|
||
|
||
2. **Execute Command Node**：执行命令
|
||
`docker exec ecommerce-scraper scrapy crawl amazon`
|
||
|
||
3. **Read File Node**：读取 `/data/products.json`
|
||
|
||
4. **OpenAI Node**（或 HTTP Request 调本地 Ollama）：
|
||
|
||
- Prompt 示例：
|
||
|
||
```text
|
||
从以下JSON中提取每个产品的简短摘要（不超过30字）并分类。
|
||
```
|
||
|
||
5. **Write to Database Node**（或 Google Sheet / CSV）
|
||
保存提炼结果
|
||
|
||
6. **Telegram / Email Node**
|
||
发送日报通知
|
||
|
||
|
||
---
|
||
|
||
## 💡 六、后续增强建议
|
||
|
||
|功能|工具|说明|
|
||
|---|---|---|
|
||
|图像下载/视频下载|Scrapy 自带 `ImagesPipeline` 或 `requests`|下载并存储本地或 MinIO|
|
||
|多站点并行爬取|Scrapy 多 Spider + n8n 并行调度|自动轮询不同站点|
|
||
|数据可视化|Metabase / Grafana|接 PostgreSQL 实时展示产品趋势|
|
||
|向量搜索|Qdrant / Weaviate|用于AI语义检索|
|
||
|本地AI处理|Ollama + Llama3|替代外部API|
|
||
|
||
---
|
||
|
||
是否希望我帮你再生成一个 **完整可导入的 n8n workflow JSON 文件模板**（包含 Cron、执行 Scrapy、读取 JSON、调用 AI、存入数据库）？
|
||
这样你只需导入到 n8n UI 就能立即自动化整个流程。
|
||
|
||
非常好 👍
|
||
下面是一个**完整可导入的 n8n Workflow JSON 模板**，它能自动：
|
||
|
||
1️⃣ 定时运行 Scrapy 爬虫（通过 Docker 命令）
|
||
2️⃣ 读取爬取的数据（`products.json`）
|
||
3️⃣ 调用 AI（可选择 OpenAI 或本地 Ollama）进行摘要与分类
|
||
4️⃣ 存入数据库或文件
|
||
5️⃣ 发送报告通知
|
||
|
||
---
|
||
|
||
## 🧩 一、Workflow 功能概览
|
||
|
||
|步骤|节点类型|功能|
|
||
|---|---|---|
|
||
|1|**Cron Trigger**|每天凌晨 2:00 自动触发|
|
||
|2|**Execute Command**|执行 `docker exec ecommerce-scraper scrapy crawl amazon`|
|
||
|3|**Read Binary File**|读取 `/data/products.json`|
|
||
|4|**OpenAI (或 HTTP Request)**|提炼摘要与分类（可切换 Ollama）|
|
||
|5|**Write Binary File**|输出 `data/products_summary.json`|
|
||
|6|**Email (或 Telegram)**|发送日报通知|
|
||
|
||
---
|
||
|
||
## 📦 二、Workflow JSON 模板（可直接导入）
|
||
|
||
将以下 JSON 内容保存为
|
||
👉 `workflow_ecommerce_automation.json`
|
||
然后在 n8n Web UI → **Import from file** 导入。
|
||
|
||
```json
|
||
{
|
||
"name": "Ecommerce Crawler + AI Summary",
|
||
"nodes": [
|
||
{
|
||
"parameters": {
|
||
"triggerTimes": {
|
||
"item": [
|
||
{
|
||
"mode": "everyDay",
|
||
"hour": 2
|
||
}
|
||
]
|
||
}
|
||
},
|
||
"id": "1",
|
||
"name": "Cron Trigger",
|
||
"type": "n8n-nodes-base.cron",
|
||
"typeVersion": 1,
|
||
"position": [250, 250]
|
||
},
|
||
{
|
||
"parameters": {
|
||
"command": "docker exec ecommerce-scraper scrapy crawl amazon"
|
||
},
|
||
"id": "2",
|
||
"name": "Run Scrapy Crawler",
|
||
"type": "n8n-nodes-base.executeCommand",
|
||
"typeVersion": 1,
|
||
"position": [500, 250]
|
||
},
|
||
{
|
||
"parameters": {
|
||
"path": "/data/products.json",
|
||
"options": {}
|
||
},
|
||
"id": "3",
|
||
"name": "Read Products JSON",
|
||
"type": "n8n-nodes-base.readBinaryFile",
|
||
"typeVersion": 1,
|
||
"position": [750, 250]
|
||
},
|
||
{
|
||
"parameters": {
|
||
"functionCode": "const data = JSON.parse(Buffer.from(items[0].binary.data.data, 'base64').toString());\nreturn data.map(p => ({ json: p }));"
|
||
},
|
||
"id": "4",
|
||
"name": "Parse JSON",
|
||
"type": "n8n-nodes-base.function",
|
||
"typeVersion": 1,
|
||
"position": [1000, 250]
|
||
},
|
||
{
|
||
"parameters": {
|
||
"model": "gpt-4-turbo",
|
||
"prompt": "你是一个电商产品分析助手。请从以下产品信息中提取每个产品的简短摘要（不超过30字）并归类到相应产品类别。\n\n输入数据：{{$json[\"title\"]}}，价格：{{$json[\"price\"]}}，评分：{{$json[\"rating\"]}}。\n\n输出格式：{\"title\":\"...\",\"summary\":\"...\",\"category\":\"...\"}"
|
||
},
|
||
"id": "5",
|
||
"name": "AI Summarize & Categorize",
|
||
"type": "n8n-nodes-base.openAi",
|
||
"typeVersion": 2,
|
||
"position": [1250, 250],
|
||
"credentials": {
|
||
"openAIApi": {
|
||
"id": "YOUR-OPENAI-CREDENTIAL-ID",
|
||
"name": "OpenAI API"
|
||
}
|
||
}
|
||
},
|
||
{
|
||
"parameters": {
|
||
"fileName": "/data/products_summary.json",
|
||
"options": {},
|
||
"dataPropertyName": "json"
|
||
},
|
||
"id": "6",
|
||
"name": "Write Summary JSON",
|
||
"type": "n8n-nodes-base.writeBinaryFile",
|
||
"typeVersion": 1,
|
||
"position": [1500, 250]
|
||
},
|
||
{
|
||
"parameters": {
|
||
"fromEmail": "noreply@yourdomain.com",
|
||
"toEmail": "your@email.com",
|
||
"subject": "Daily Product Summary Report",
|
||
"text": "今日电商产品摘要已生成，请查看 /data/products_summary.json 文件。"
|
||
},
|
||
"id": "7",
|
||
"name": "Send Email Notification",
|
||
"type": "n8n-nodes-base.emailSend",
|
||
"typeVersion": 1,
|
||
"position": [1750, 250]
|
||
}
|
||
],
|
||
"connections": {
|
||
"Cron Trigger": { "main": [[{ "node": "Run Scrapy Crawler", "type": "main", "index": 0 }]] },
|
||
"Run Scrapy Crawler": { "main": [[{ "node": "Read Products JSON", "type": "main", "index": 0 }]] },
|
||
"Read Products JSON": { "main": [[{ "node": "Parse JSON", "type": "main", "index": 0 }]] },
|
||
"Parse JSON": { "main": [[{ "node": "AI Summarize & Categorize", "type": "main", "index": 0 }]] },
|
||
"AI Summarize & Categorize": { "main": [[{ "node": "Write Summary JSON", "type": "main", "index": 0 }]] },
|
||
"Write Summary JSON": { "main": [[{ "node": "Send Email Notification", "type": "main", "index": 0 }]] }
|
||
},
|
||
"active": true
|
||
}
|
||
```
|
||
|
||
---
|
||
|
||
## 🧠 三、使用说明
|
||
|
||
1. **导入流程**
|
||
|
||
- 访问 `http://<你的服务器IP>:5678`
|
||
|
||
- 点击右上角「Import from File」→ 选择上面的 JSON 文件
|
||
|
||
2. **配置 OpenAI 凭证**
|
||
|
||
- n8n → Credentials → 新建 OpenAI API Key
|
||
|
||
- 将 `YOUR-OPENAI-CREDENTIAL-ID` 替换为对应 ID
|
||
|
||
- 若使用本地 Ollama，则：
|
||
|
||
- 替换节点 `AI Summarize & Categorize` 为 “HTTP Request”
|
||
|
||
- URL 填入 `http://ollama:11434/api/generate`
|
||
|
||
- POST body 示例：
|
||
|
||
```json
|
||
{
|
||
"model": "llama3",
|
||
"prompt": "分析以下产品信息：{{$json[\"title\"]}} ..."
|
||
}
|
||
```
|
||
|
||
3. **设置邮箱凭证**
|
||
|
||
- n8n → Credentials → Email Send (SMTP)
|
||
|
||
- 支持 Gmail, Outlook, AWS SES 等
|
||
|
||
4. **测试运行**
|
||
|
||
- 手动点击 ▶️ 运行
|
||
|
||
- 查看 `/data/products_summary.json`
|
||
|
||
- 邮箱将收到日报通知
|
||
|
||
|
||
---
|
||
|
||
## 🚀 四、可选扩展
|
||
|
||
|目标|实现方式|
|
||
|---|---|
|
||
|改为多站点爬取|添加更多 spider（如 `jd_spider.py`, `ebay_spider.py`）并在 n8n 串联运行|
|
||
|增加数据库写入|在 n8n 中加入 “PostgreSQL Node” 写入产品与AI摘要|
|
||
|生成图表报表|把数据库接入 Grafana / Metabase|
|
||
|向量检索增强|在 n8n 中调用 Qdrant 插入语义向量|
|
||
|
||
---
|
||
|
||
是否希望我在下一步帮你生成一个适配 **Ollama 本地模型（离线AI处理）** 的 workflow 版本？
|
||
它将不依赖外部API，完全在本地自动运行分析。
|
||
|
||
|
||
|