root package
package root
This is the documentation for the splish library.
Package Information
The splish package contains a single class splish.Splash with methods for interacting with a ScrapingHub Splash instance.
If you haven’t hit the above link yet and are unfamilar with Splash, the TLDR is that it’s an alternative to Selenium in that it’s a full browser and executes javascript. The full rendering engine is based on Qt Webkit and Splash instances have a REST API that provides a ton of flexibility when needed and ease of use for more casual scraping tasks.
You can get it up and running locally with Docker via:
sudo docker run -p 8050:8050 -p 5023:5023 scrapinghub/splash
If you've built the source and run sbt packInstall, you can start playing
with splash on the command line via ~/local/bin/splash-main. Here's the help:
splash 1.0 Usage: splash [options] url url the URL to scrape -r, --render html request action; one of 'html', 'json' or 'har' --help prints this usage text -w, --wait <value> How long to wait (in seconds) after loading the page (to allow js onX scripts to run). Default is 2 seconds -t, --timeout <value> Overall page/connection timeout. Defaults to 30 seconds -h, --host <value> Splash instance host name or IP address (defaults to localhost) -p, --port <value> Splash instance port if not the default (8050) -u, --user <value> Splash username (if authentication is required). Default is no authentcation. -p, --pass <value> Splash password (if authentication is required). Default is no authentication. -s, --ssl Use an SSL connection to the Splash instance? (defaults to false)
The first thing we need to do is make a connection to the server
import splish.Splash val s = Splash() println(s) ## Splash(localhost,8050,null,null,false)
We can test that connection and get some other information as well:
println( "Server is up? " + s.isActive() + "\n" + "What's the server version? " + s.version()("splash") + "\n" + "How long has the server been up? " + s.performanceStatistics()("cputime").num ) ## Server is up? true ## What's the server version? "3.2" ## How long has the server been up? 68.84
The library makes use of
[uJson](http://www.lihaoyi.com/upickle/#uJson) for more complex return
types and a few methods return a requests
[Response](https://github.com/lihaoyi/requests-scala/blob/master/requests/src/requests/Model.scala#L235-L276)
object due to the result of a call to more dynamic endpoints being
un-knowable at call time (Splash allows you to use Lua to perform
complex page interaction and you can return images, plaintext, HTML or
JSON via the Lua interface).
The classic use case for Splash is to feed it a URL and get HTML back after it’s had time to process any javascript. The URL in the following example relies on javascript to add content to the page:
val html = s.renderHTML("https://rud.is/splash-js-test.html") println(html) ## <html><head> ## <title>Test</title> ## </head> ## <body onload="addElements()"> ## < p>This is a Splash test page. ## < p><span id="target">This won't be here if javascript is disabled</span> ## <script> ## function addElements() { ## document.getElementById("target").innerHTML = "This won't be here if javascript is disabled" ; ## } ## </script> ## ## ## </body></html>
Here’s what that looks like just using the requests library:
import requests._ val res = requests.get("https://rud.is/splash-js-test.html") println(res.text) ## <html> ## <head> ## <title>Test</title> ## </head> ## <body onload="addElements()"> ## < p>This is a Splash test page. ## < p><span id="target"></span> ## <script> ## function addElements() { ## document.getElementById("target").innerHTML = "This won't be here if javascript is disabled" ; ## } ## </script> ## </body> ## </html>
Most of the other Splash API endpoints have corresponding methods in the library (the image-oriented ones are on the TODO list). We can get the same page in both Splash JSON:
println(s.renderJSON("https://rud.is/splash-js-test.html", responseBody = true, html = true)) ## {"title":"Test","requestedUrl":"https://rud.is/splash-js-test.html","url":"https://rud.is/splash-js-test.html","geometry":[0,0,1024,768],"html":"\nTest \n \n \n < p>This is a Splash test page.\n < p>This won't be here if javascript is disabled\n \n \n\n"}
and HAR formats:
println(s.renderHAR("https://rud.is/splash-js-test.html?1", responseBody = true)) ## {"log":{"browser":{"version":"602.1","comment":"PyQt 5.9, Qt 5.9.1","name":"QWebKit"},"pages":[{"pageTimings":{"_onPrepareStart":286,"_onStarted":1,"onContentLoad":285,"onLoad":286},"id":"1","title":"Test","startedDateTime":"2018-08-18T22:27:57.550751Z"}],"version":"1.2","creator":{"version":"3.2","name":"Splash"},"entries":[{"pageref":"1","time":121,"timings":{"connect":-1,"blocked":-1,"send":0,"ssl":-1,"receive":1,"dns":-1,"wait":120},"request":{"url":"https://rud.is/splash-js-test.html?2","headers":[{"value":"Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/602.1 (KHTML, like Gecko) splash Version/9.0 Safari/602.1","name":"User-Agent"},{"value":"text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8","name":"Accept"}],"queryString":[{"value":"","name":"2"}],"method":"GET","httpVersion":"HTTP/1.1","headersSize":188,"cookies":[],"bodySize":-1},"cache":{},"response":{"headers":[{"value":"nginx/1.13.9","name":"Server"},{"value":"Sat, 18 Aug 2018 22:20:16 GMT","name":"Date"},{"value":"text/html","name":"Content-Type"},{"value":"Sat, 18 Aug 2018 21:49:08 GMT","name":"Last-Modified"},{"value":"chunked","name":"Transfer-Encoding"},{"value":"keep-alive","name":"Connection"},{"value":"Accept-Encoding","name":"Vary"},{"value":"W/\"5b789454-15d\"","name":"ETag"},{"value":"Sun, 19 Aug 2018 21:49:08 GMT","name":"Expires"},{"value":"max-age=84532","name":"Cache-Control"},{"value":"max-age=31536000; includeSubDomains; preload","name":"Strict-Transport-Security"},{"value":"SAMEORIGIN","name":"X-Frame-Options"},{"value":"<3","name":"X-Powered-By"},{"value":"frame-ancestors 'self', default-src * 'self' data: 'unsafe-inline' 'unsafe-eval'; report-uri 'https://hrbrmstr.report-uri.com/r/d/csp/reportOnly';","name":"Content-Security-Policy"},{"value":"default-src * 'self' data: 'unsafe-inline' 'unsafe-eval'; report-uri 'https://hrbrmstr.report-uri.com/r/d/csp/reportOnly';","name":"X-Content-Security-Policy"},{"value":"default-src * 'self' data: 'unsafe-inline' 'unsafe-eval'; report-uri 'https://hrbrmstr.report-uri.com/r/d/csp/reportOnly';","name":"X-WebKit-CSP"},{"value":"1; mode=block","name":"X-XSS-Protection"},{"value":"nosniff","name":"X-Content-Type-Options"},{"value":"gzip","name":"Content-Encoding"}],"ok":true,"redirectURL":"","httpVersion":"HTTP/1.1","bodySize":349,"cookies":[],"status":200,"content":{"encoding":"base64","mimeType":"text/html","text":"PGh0bWw+CiAgPGhlYWQ+CiAgICA8dGl0bGU+VGVzdDwvdGl0bGU+CiAgPC9oZWFkPgogIDxib2R5IG9ubG9hZD0iYWRkRWxlbWVudHMoKSI+CiAgICA8cD5UaGlzIGlzIGEgU3BsYXNoIHRlc3QgcGFnZS48L3A+CiAgICA8cD48c3BhbiBpZD0idGFyZ2V0Ij48L3NwYW4+PC9wPgogICAgPHNjcmlwdD4KICAgICAgZnVuY3Rpb24gYWRkRWxlbWVudHMoKSB7CiAgICAgICAgZG9jdW1lbnQuZ2V0RWxlbWVudEJ5SWQoInRhcmdldCIpLmlubmVySFRNTCA9ICJUaGlzIHdvbid0IGJlIGhlcmUgaWYgamF2YXNjcmlwdCBpcyBkaXNhYmxlZCIgOwogICAgICB9ICAgIAogICAgPC9zY3JpcHQ+CiAgPC9ib2R5Pgo8L2h0bWw+Cg==","size":349},"headersSize":971,"statusText":"OK","url":"https://rud.is/splash-js-test.html?2"},"_splash_processing_state":"finished","startedDateTime":"2018-08-18T22:27:57.552544Z"}]}}
This is the documentation for the
splishlibrary.Package Information
The splish package contains a single class splish.Splash with methods for interacting with a ScrapingHub Splash instance.
If you haven’t hit the above link yet and are unfamilar with Splash, the TLDR is that it’s an alternative to Selenium in that it’s a full browser and executes javascript. The full rendering engine is based on Qt Webkit and Splash instances have a REST API that provides a ton of flexibility when needed and ease of use for more casual scraping tasks.
You can get it up and running locally with Docker via:
If you've built the source and run
sbt packInstall, you can start playing withsplashon the command line via~/local/bin/splash-main. Here's the help:The first thing we need to do is make a connection to the server
We can test that connection and get some other information as well:
The library makes use of [
uJson](http://www.lihaoyi.com/upickle/#uJson) for more complex return types and a few methods return arequests[Response](https://github.com/lihaoyi/requests-scala/blob/master/requests/src/requests/Model.scala#L235-L276) object due to the result of a call to more dynamic endpoints being un-knowable at call time (Splash allows you to useLuato perform complex page interaction and you can return images, plaintext, HTML or JSON via the Lua interface).The classic use case for Splash is to feed it a URL and get HTML back after it’s had time to process any javascript. The URL in the following example relies on javascript to add content to the page:
Here’s what that looks like just using the
requestslibrary:Most of the other Splash API endpoints have corresponding methods in the library (the image-oriented ones are on the TODO list). We can get the same page in both Splash JSON:
and HAR formats: