Karate框架从URL直接下载图片至目标文件夹的技术问询

问题背景 #

在使用Karate进行图片下载测试时，遇到以下问题：

• 初始请求返回302重定向，可捕获图片真实URL，但直接用read()读取该URL会报错，因为Karate会将URL当作本地文件路径处理
• 直接请求图片URL返回200状态码，但无法获取到响应体中的图片内容

测试代码示例 #

Feature: Photo Download Test
Background:
    * configure continueOnStepFailure = true
    * configure followRedirects = false

  Scenario:
    # initial request to capture random photo
    * url 'https://source.unsplash.com'
    Given path 'user/c_v_r/100x100'
    When method GET
    Then status 302
    * def imageFilePath = karate.response.header('Location')
    * print 'imageFilePath-',imageFilePath

    * def photoName = imageFilePath.split('?')[0].split('/')[3]
    * print 'photoName-',photoName

    * def imageType = imageFilePath.split('?')[1].split('&')[3].split('=')[1]
    * print 'imageType-',imageType

    * def fileName = photoName + '.' + imageType
    * def contentType = 'image/' + imageType

    # image file capture attempts
    * def image = read(imageFilePath)
    * def image = read('file:'+imageFilePath)
    * def image = karate.read('file:'+imageFilePath)
    * print 'image-',image

    # 200 that contains no response value
    * url imageFilePath
    Given path ''
    When method GET
    Then status 200
    * def image = response

遇到的具体问题 #

请求图片URL返回200，但响应体无法获取图片内容：

11:17:16.494 [com.intuit.karate.Main.main()] DEBUG com.intuit.karate - response time in milliseconds: 8281
2 < 200
2 < Connection: keep-alive
2 < Content-Length: 3576
2 < last-modified: Wed, 21 Feb 2024 18:17:16 GMT
2 < cache-control: public, max-age=31536000
2 < x-imgix-id: 865a9f7e088d398a398951413abaddd4d059a862
2 < Server: Google Frontend
2 < X-Imgix-Render-Farm: 02.66056
2 < Date: Wed, 21 Feb 2024 18:17:16 GMT
2 < Age: 0
2 < Accept-Ranges: bytes
2 < Content-Type: image/jpeg
2 < Access-Control-Allow-Origin: *
2 < Timing-Allow-Origin: *
2 < Cross-Origin-Resource-Policy: cross-origin
2 < X-Content-Type-Options: nosniff
2 < X-Served-By: cache-sjc1000139-SJC, cache-dfw-kdfw8210069-DFW
2 < X-Cache: MISS, MISS

尝试直接读取重定向URL时，Karate会将其解析为本地文件路径，触发文件未找到错误：

01: read(imageFilePath)
<<<<
org.graalvm.polyglot.PolyglotException: java.io.FileNotFoundException: /localfilepath for test feature/https:/images.unsplash.com/photo-1608231975456-2f2d9fb1b49b?crop=entropy&cs=tinysrgb&fit=crop&fm=jpg&h=100&ixid=MnwxfDB8MXxyYW5kb218MHx8fHx8fHx8MTcwODU0MTgyNQ&ixlib=rb-4.0.3&q=80&utm_campaign=api-credit&utm_medium=referral&utm_source=unsplash_source&w=100 (No such file or directory)

01: read('file:'+imageFilePath)
<<<<
org.graalvm.polyglot.PolyglotException: not found: https://images.unsplash.com/photo-1597428963794-aba7182f3abf?crop=entropy&cs=tinysrgb&fit=crop&fm=jpg&h=100&ixid=MnwxfDB8MXxyYW5kb218MHx8fHx8fHx8MTcwODU0MTc4MA&ixlib=rb-4.0.3&q=80&utm_campaign=api-credit&utm_medium=referral&utm_source=unsplash_source&w=100

01: karate.read('file:'+imageFilePath)
<<<<
org.graalvm.polyglot.PolyglotException: not found: https://images.unsplash.com/photo-1597431842922-d9686a23baa6?crop=entropy&cs=tinysrgb&fit=crop&fm=jpg&h=100&ixid=MnwxfDB8MXxyYW5kb218MHx8fHx8fHx8MTcwODU0MTQ5OQ&ixlib=rb-4.0.3&q=80&utm_campaign=api-credit&utm_medium=referral&utm_source=unsplash_source&w=100

解决方案 #

使用Karate的karate.write()方法结合HTTP请求的二进制响应体，可直接将图片写入目标文件夹，后续通过本地路径读取。

修改后的测试代码 #

Feature: Photo Download Test
Background:
    * configure continueOnStepFailure = true
    * configure followRedirects = false

  Scenario:
    # 获取图片的重定向URL
    * url 'https://source.unsplash.com'
    Given path 'user/c_v_r/100x100'
    When method GET
    Then status 302
    * def imageUrl = karate.response.header('Location')
    
    # 解析文件名与存储路径
    * def photoName = imageUrl.split('\\?')[0].split('/')[3]
    * def imageType = imageUrl.split('\\?')[1].split('&')[3].split('=')[1]
    * def fileName = photoName + '.' + imageType
    * def targetPath = 'target/' + fileName

    # 请求图片并写入本地文件
    * url imageUrl
    When method GET
    Then status 200
    * karate.write(response, targetPath)
    
    # 验证文件可正常读取
    * def downloadedImage = read(targetPath)
    * print '图片已成功下载并可读取'

关键说明 #

• karate.write(response, targetPath)：Karate处理二进制响应时会自动转为字节数组，该方法可直接将字节数组写入指定路径的文件
• read(targetPath)：文件写入完成后，直接通过本地路径即可读取内容，无需额外前缀
• 若target目录不存在，可手动添加创建步骤：* java.io.File('target').mkdirs()

内容的提问来源于stack exchange，提问作者mike