v2.0.0

Added

• Require crwlr/crawler v2 and crwlr/crwl-extension-utils min v2.4.

v1.4.0

Added

• Screenshot::take() in addition to Screenshot::loadAndTake(), allowing to take a screenshot of an already opened page, in a separate step. This way you can even add this step after an Http::crawl() step and get screenshots of all the crawled pages.

Fixed

• Change calls to HttpLoader::browserHelper() to HttpLoader::browser() and require crwlr/crawler with constraint ^1.9.3 to make sure the method exists.

v1.3.0

Added

• Merge things from the crwlr/crwl-ext-browser package to this one, because they are too tightly coupled. The other package will be abandoned.
• Add timeout config param to ScreenshotBuilder, therefore also require crwlr/crawler v1.9.0 (or greater), with the new functionality to configure timeouts for the headless browser.

Fixed

• Prepare for crwlr/crawler v2.0.

v1.2.1

Fixed

• Remove input validation in screenshot step, so it automatically uses the validation method of the HttpBase step, so it also allows to use the useInputKeyAs... methods.

v1.2.0

Added

• Option to wait a certain amount of time after loading a page, before taking the screenshot (Screenshot::waitAfterPageLoaded()).

v1.1.0

Added

• Get all colors, not only the ones making up more than 0.5 percent of the image. But also add a method onlyAbovePercentageOfImage() to the GetColors step, to manually set a custom threshold.

Fixed

• Improve memory usage of getting colors from an image.

v1.0.0

Changed

• Change the output of the Screenshot step, from an array ['response' => RespondedRequest, 'screenshotPath' => string] to a RespondedRequestWithScreenshot object, that has a screenshotPath property. The problem with the previous solution was: when using the response cache, the step failed, because it gets a cached response from the loader that was not actually loaded in the headless browser. When the step afterwards tries to take a screenshot from the page that is still open in the browser, it just fails because there is no open page. Now, with the new RespondedRequestWithScreenshot object, the screenshotPath is also saved in the cached response.

v0.1.2

Fixed

• Upgrade to crwlr/crawler v1.5.3 and remove the separate HeadlessBrowserLoader and HeadlessBrowserCrawler. The steps shall simply use the normal HttpLoader and automatically switch to use the headless browser for loading and switch back afterwards if the loader was configured to use the HTTP client.

v0.1.1

Fixed

• Set required crwlr/crawler version to ^1.5.

v0.1.0

Added

• Initial version containing a HeadlessBrowserLoader, a HeadlessBrowserCrawler (like HttpCrawler) and two steps: Screenshot and GetColors.