测试执行时间

默认情况下，该模块会检查你的机器或CI流水线中是否有本地安装的Tesseract。如果没有本地安装，它将自动使用NodeJS版本。这可能会导致一些性能降低，因为图像处理将由Node.js完成。NodeJS并不是进行重量级处理的最佳系统。

但是....，有一些方法可以优化执行时间。让我们看看下面的测试脚本

import { browser } from "@wdio/globals";

describe("Search", () => {
    it("be able to search for a value", async () => {
        await browser.url("https://webbrowser.io");
        await browser.ocrClickOnText({
            text: "Search",
        });
        await browser.ocrSetValue({
            text: "docs",
            value: "specfileretries",
        });
        await browser.ocrWaitForTextDisplayed({
            text: "specFileRetries",
        });
    });
});

当你第一次执行这个测试时，你可能会看到以下结果，完成测试需要5.9秒。

npm run wdio -- --logLevel=silent

> ocr-demo@1.0.0 wdio
> wdio run ./wdio.conf.ts --logLevel=silent


Execution of 1 workers started at 2024-05-26T04:52:53.405Z

[0-0] RUNNING in chrome - file:///test/specs/test.e2e.ts
[0-0] Estimating resolution as 182
[0-0] Estimating resolution as 124
[0-0] Estimating resolution as 126
[0-0] PASSED in chrome - file:///test/specs/test.e2e.ts

 "spec" Reporter:
------------------------------------------------------------------
[chrome 125.0.6422.78 mac #0-0] Running: chrome (v125.0.6422.78) on mac
[chrome 125.0.6422.78 mac #0-0] Session ID: d281dcdc43962b95835aea8f64cab6c7
[chrome 125.0.6422.78 mac #0-0]
[chrome 125.0.6422.78 mac #0-0] » /test/specs/test.e2e.ts
[chrome 125.0.6422.78 mac #0-0] Search
[chrome 125.0.6422.78 mac #0-0]    ✓ be able to search for a value
[chrome 125.0.6422.78 mac #0-0]
[chrome 125.0.6422.78 mac #0-0] 1 passing (5.9s)


Spec Files:      1 passed, 1 total (100% completed) in 00:00:08

裁剪屏幕的搜索区域

你可以通过提供一个裁剪区域来执行OCR，从而优化执行时间。

如果你将脚本改为这样：

import { browser } from "@wdio/globals";

describe("Search", () => {
    it("be able to search for a value", async () => {
        await browser.url("https://webdriver.io");
        await driver.ocrClickOnText({
            haystack: $(".DocSearch"),
            text: "Search",
        });
        await driver.ocrSetValue({
            haystack: $(".DocSearch-Form"),
            text: "docs",
            value: "specfileretries",
        });
        await driver.ocrWaitForTextDisplayed({
            haystack: $(".DocSearch-Dropdown"),
            text: "specFileRetries",
        });
    });
});

然后你会看到不同的执行时间。

npm run wdio -- --logLevel=silent

> ocr-demo@1.0.0 wdio
> wdio run ./wdio.conf.ts --logLevel=silent


Execution of 1 workers started at 2024-05-26T04:56:55.326Z

[0-0] RUNNING in chrome - file:///test/specs/test.e2e.ts
[0-0] Estimating resolution as 182
[0-0] Estimating resolution as 124
[0-0] Estimating resolution as 124
[0-0] PASSED in chrome - file:///test/specs/test.e2e.ts

 "spec" Reporter:
------------------------------------------------------------------
[chrome 125.0.6422.78 mac #0-0] Running: chrome (v125.0.6422.78) on mac
[chrome 125.0.6422.78 mac #0-0] Session ID: c6cb1843535bda3ee3af07920ce232b8
[chrome 125.0.6422.78 mac #0-0]
[chrome 125.0.6422.78 mac #0-0] » /test/specs/test.e2e.ts
[chrome 125.0.6422.78 mac #0-0] Search
[chrome 125.0.6422.78 mac #0-0]    ✓ be able to search for a value
[chrome 125.0.6422.78 mac #0-0]
[chrome 125.0.6422.78 mac #0-0] 1 passing (4.8s)


Spec Files:      1 passed, 1 total (100% completed) in 00:00:08

裁剪图像

这将本地执行时间从5.9秒减少到4.8秒。这是近**19%**的减少。想象一下对于包含更多数据的大型脚本，这可能带来的性能提升。

使用本地安装的Tesseract

如果你在本地机器或流水线中有Tesseract的本地安装（关于在本地系统上安装Tesseract的更多信息可以在这里找到），你可以将执行时间加速到不到一分钟。下面是使用本地安装的Tesseract运行相同脚本的执行时间。

npm run wdio -- --logLevel=silent

> ocr-demo@1.0.0 wdio
> wdio run ./wdio.conf.ts --logLevel=silent


Execution of 1 workers started at 2024-05-26T04:59:11.620Z

[0-0] RUNNING in chrome - file:///test/specs/test.e2e.ts
[0-0] PASSED in chrome - file:///test/specs/test.e2e.ts

 "spec" Reporter:
------------------------------------------------------------------
[chrome 125.0.6422.78 mac #0-0] Running: chrome (v125.0.6422.78) on mac
[chrome 125.0.6422.78 mac #0-0] Session ID: 87f8c1e949e15a383b902e4d59b1f738
[chrome 125.0.6422.78 mac #0-0]
[chrome 125.0.6422.78 mac #0-0] » /test/specs/test.e2e.ts
[chrome 125.0.6422.78 mac #0-0] Search
[chrome 125.0.6422.78 mac #0-0]    ✓ be able to search for a value
[chrome 125.0.6422.78 mac #0-0]
[chrome 125.0.6422.78 mac #0-0] 1 passing (3.9s)


Spec Files:      1 passed, 1 total (100% completed) in 00:00:06

本地安装

这将本地执行时间从5.9秒减少到3.9秒。这是近**34%**的减少。想象一下对于包含更多数据的大型脚本，这可能带来的性能提升。