## 自動化工作流程下的 LLM ## 驅動無程式碼網頁爬蟲與EDA 2025-06-19 陳奎銘 Ben @南區統計年會

無程式碼？

我只能說盡量不寫程式

<!-- .slide: data-background-iframe="media/Ben.html" -->

- # `playwright` - # AI Agent - # n8n

## `playwright` * Microsoft 出品的跨瀏覽器自動化測試框架，支援 Chromium、Firefox 與 WebKit。 * 多語言 API 與 MCP。 * 內建並行測試、網路攔截、截圖與錄影等功能，適用於端到端測試與爬蟲。

## Install `playwright` ``` pip install playwright playwright install --with-deps ```

## `playwright` Demo ``` playwright codegen --target python \ --output download_script.py \ https://data.moenv.gov.tw/dataset/detail/AQX_P_432 ``` <iframe width="560" height="315" src="https://www.youtube.com/embed/P4ytwrt8HiM?si=yqyb5m4YwKxj0uaq" title="YouTube video player" frameborder="0" allow="accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture; web-share" referrerpolicy="strict-origin-when-cross-origin" allowfullscreen></iframe>

## generate code by `playwright` 產出可以直接下載資料的程式碼 ```python[7-10|5, 12] from playwright.sync_api import Playwright, sync_playwright def run(playwright: Playwright) -> None: browser = playwright.chromium.launch(headless=False) context = browser.new_context(accept_downloads=True) page = context.new_page() page.goto("https://data.moenv.gov.tw/dataset/detail/AQX_P_432") page.get_by_role("button", name="下載篩選後資料").click() with page.expect_download() as download_info: page.get_by_label("下載篩選後資料").get_by_text("CSV").click() download = download_info.value download.save_as("AQX_P_432.csv") page.close() # --------------------- context.close() browser.close() with sync_playwright() as playwright: run(playwright) ```

## 關鍵是取得網頁內容 ```python browser = p.chromium.launch(headless=False) page = browser.new_page() page.goto(url, timeout=self.timeout, wait_until="networkidle") content = page.content() ```

## 架設 API Service - 用現成的，不用自己寫 code - docker.io/kuiming/playwright-api - https://github.com/KuiMing/n8n_agent

## AI Agent 能感知環境、規劃動作與利用記憶進行推理，並透過工具或 API 呼叫執行任務以達成預定目標的系統

## AI Agent 

## `n8n` ### API 懶人包 + 樂高 #### 自動化工作流程工具

## WHY `n8n` - **開源且免費**：可自架設於本地或雲端，無需擔心資料隱私。 - **高度彈性**：支援自訂節點與腳本，滿足各種自動化需求，可設置多種觸發器。 - **資源豐富**：擁有龐大開發者社群、豐富範例與插件，官方快速迭代回應新需求，對比封閉商業平台更易獲得資源與技術支援。

## `n8n` Nodes 

## `n8n` Credential- 地端 

## `n8n` Credential- 雲端 

### 範例：我需要取得球賽資訊 [FIFA Club World Cup](https://www.fifa.com/en/tournaments/mens/club-world-cup/usa-2025/standings)

<!-- .slide: data-background="media/n8n-playwright-1.png" -->

<!-- .slide: data-background="media/n8n-playwright-2.png" -->

<!-- .slide: data-background="media/n8n-playwright-3.png" -->

<!-- .slide: data-background="media/n8n-playwright-4.png" -->

<!-- .slide: data-background="media/n8n-playwright-5.png" -->

<!-- .slide: data-background="media/n8n-playwright-6.png" -->

<!-- .slide: data-background="media/n8n-playwright-7.png" -->

<!-- .slide: data-background="media/n8n-playwright-8.png" -->

<!-- .slide: data-background="media/n8n-playwright-9.png" -->

<!-- .slide: data-background="media/n8n-playwright-schedule.png" -->

## Take away - 若有爬蟲需求，請嘗試使用 `playwright` - 請用 AI Agent 減少自己需要動手的雜事 - 請用 n8n 讓雜事自動化

#### https://github.com/KuiMing/n8n_agent

## `Facebook` <img src=media/QR_R_Ladies_Taipei.png width=50%></img><img src=media/QR_Ben_facebook.png width=50%></img>

<!-- .slide: data-background-iframe="https://www.accupass.com/go/openfun_rladies" -->

# Thank you