Skip to content

Commit 7d4a3d0

Browse files
committed
test: test
1 parent 0e285b8 commit 7d4a3d0

File tree

6 files changed

+44
-349
lines changed

6 files changed

+44
-349
lines changed

src/ai/index.ts

Lines changed: 0 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -115,8 +115,6 @@ function createXCrawlOpenAI(
115115
? (JSON.parse(content ?? '{}') as any)
116116
: content
117117

118-
console.log(result)
119-
120118
return result
121119
}
122120

test/start/.gitignore

Lines changed: 1 addition & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1 @@
1+
.env

test/start/envConfig.ts

Lines changed: 8 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,8 @@
1+
import path from 'node:path'
2+
import dotenv from 'dotenv'
3+
4+
const pathResolve = (dirPath: string) => path.resolve(__dirname, dirPath)
5+
6+
dotenv.config({ path: pathResolve('.env') })
7+
8+
export const { BASE_URL, API_KEY } = process.env

test/start/index.ts

Lines changed: 28 additions & 29 deletions
Original file line numberDiff line numberDiff line change
@@ -1,34 +1,33 @@
11
import path from 'node:path'
2-
import xCrawl from 'x-crawl'
2+
import xCrawl, { createXCrawlOpenAI } from 'x-crawl'
3+
4+
import { BASE_URL, API_KEY } from './envConfig'
35

46
const pathResolve = (dirPath: string) => path.resolve(__dirname, dirPath)
57

6-
const testXCrawl = xCrawl()
8+
const xCrawlOpenAIApp = createXCrawlOpenAI({
9+
clientOptions: { baseURL: BASE_URL, apiKey: API_KEY }
10+
})
11+
12+
const HTMLContent = `
13+
<div class="box">
14+
<div class="list-item">男装大衣</div>
15+
<div class="list-item">女装大衣</div>
16+
<div class="scroll-list">
17+
<div class="list-item">男装卫衣</div>
18+
<div class="list-item">女装卫衣</div>
19+
<div class="list-item">男装带帽卫衣</div>
20+
</div>
21+
<div class="scroll-list" id="short">
22+
<div class="list-item">男装纯棉短袖</div>
23+
<div class="list-item">男装纯棉短袖</div>
24+
<div class="list-item">男装冰丝短袖</div>
25+
<div class="list-item">男装圆领短袖</div>
26+
</div>
27+
<div class="list-item">男装裤子</div>
28+
</div>
29+
`
730

8-
testXCrawl
9-
.crawlFile({
10-
targets: [
11-
{
12-
url: 'https://raw.githubusercontent.com/coder-hxl/airbnb-upload/master/area/4408.jpg',
13-
fileName: '4408',
14-
priority: 1
15-
},
16-
{
17-
url: 'https://raw.githubusercontent.com/coder-hxl/airbnb-upload/master/area/4401.jpg',
18-
fileName: '4401',
19-
priority: 3
20-
},
21-
{
22-
url: 'https://raw.githubusercontent.com/coder-hxl/airbnb-upload/master/area/4406.jpg',
23-
fileName: '4406',
24-
priority: 2
25-
}
26-
],
27-
proxy: { urls: ['http://localhost:14892'] },
28-
storeDirs: pathResolve('./upload')
29-
})
30-
.then((res) => {
31-
res.forEach((item) => {
32-
console.log(item.id, item.data?.data.fileName)
33-
})
34-
})
31+
xCrawlOpenAIApp.parseElements(HTMLContent, `获取男装, 并去重`).then((res) => {
32+
console.log(res)
33+
})

test/start/package.json

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -10,6 +10,6 @@
1010
"author": "",
1111
"license": "ISC",
1212
"devDependencies": {
13-
"sharp": "^0.32.0"
13+
"dotenv": "^16.4.5"
1414
}
1515
}

0 commit comments

Comments
 (0)