Rosetta Code:Village Pump/tasks descriptions only

From Rosetta Code
tasks descriptions only
This is a particular discussion thread among many which consider Rosetta Code.

Summary

How to get a list of programming tasks *without* solutions

Discussion

I'm going on a long plane trip. I'd like to take a printout of all the tasks but without the solutions. I have in mind to solve some of the solutions on paper whilst airborne, taking the list of solutions and the programming language manual. Axtens (talk) 07:04, 27 March 2015 (UTC)

Not sure it is what you were thinking, but I want this for a specific Language. How can that be done?
--RLRandallx 21:55, 20 May 2016 (UTC)
Sounds like a good idea for a task "Extract task description from an RC task" (i.e. down to the first {{header|...}} :-)
--Paddy3118 (talk) 10:18, 27 March 2015 (UTC)
I'd approach this by downloading all the tasks (approximately: all the rosetta wiki links from http://rosettacode.org/wiki/Category:Programming_Tasks which do not have a ':' in the url's path), and then clip them off starting at the line which contains "mw-headline"). That should get you close enough?
In other words, something like this:
<lang bash>wget -k http://rosettacode.org/wiki/Category:Programming_Tasks

lynx -force_html -dump Category%3AProgramming_Tasks | awk '/http:..rosettacode.org.wiki/{print $2}' | grep -v ':.*:' | xargs wget --wait=1 -kp find rosettacode.org/wiki -type f | xargs perl -i -0777 -pe 's/\n[^\n]*mw-headline.*//s'</lang>

Change perl's -i option to -i.bak if you feel like you might want to save the originals.
You'll probably want a .html extension on those files, though, so also:
<lang bash>find rosettacode.org/wiki -type f | while read f; do mv $f $f.html; done</lang>
This will leave the table of contents in place, if but you could remove that if you like, using a similar approach. (Make a copy of your work before experimenting, so you do not overburden the site. When the site is overburdened, nobody can pull down content from it - including you.)
I hope this helps. --Rdm (talk) 14:58, 27 March 2015 (UTC)
I feel this task should be a builtin task for the website. Could someone please add it to the Main Menu? The essential code is above.
--RLRandallx 21:07, 20 May 2016 (UTC)
I was just fiddling inside of my webbrowser and came up with the following javascript-ish pseudocode. A solution using JScript or VBScript inside (gasp) Internet Explorer would seem to be the go. (Doing it on Linux might be a bit trickier.)
<lang javascript>

pages = document.getElementById("mw-pages") anchors = pages.getElementsByTagName("a") document.location = anchors[0].href // each one

start = document.getElementsByClassName("infobox")[0] cursor = start desc = "" while (cursor.tagName !== "TABLE") {

 there = cursor; 
 desc = desc + there.innerText + "\n"; 
 cursor = cursor.nextElementSibling 

} // desc contains the task description</lang>

The while loop should handle the different amounts of text between the infobox markup and the beginning of the contents table. Axtens (talk) 09:53, 16 April 2015 (UTC)
Here's a SlimerJS solution. Not perfect but seems to work reasonably well (it's funny watching all the little windows open and close.) A little annoyingly, SlimerJS doesn't implement innerText, so there are the odd occasions where there's markup in the file. Also, occasionally there's a TABLE that's not the contents before the contents, thus truncating the description.
<lang javascript>var fs = require('fs');

function innerCall(nam, ref) { var ipage = require("webpage").create(); ipage.open(ref, function () { var description = ipage.evaluate( function () { var start = document.getElementsByClassName("infobox")[0]; var cursor = start.nextElementSibling; var desc = ""; while (cursor.tagName !== "TABLE") { there = cursor; desc = desc + there.innerHTML + "\n"; cursor = cursor.nextElementSibling; } return desc; }); var fileName = nam + ".txt"; fileName = fileName.replace(/\//g,"_"); fs.write(fileName,description); //console.log(nam + " " + ref + "\n" + description); ipage.close();

       });
       

}


var page = require("webpage").create(); page.open("http://rosettacode.org/wiki/Category:Programming_Tasks", function () {

   var anchors = page.evaluate(function () {
       return document.getElementById("mw-pages").getElementsByTagName("a");
   });
   for (var i = 0; i < anchors.length; i++ ) {
       //console.log(anchors[i].innerHTML + ' ' + anchors[i].href);
       innerCall(anchors[i].innerHTML, anchors[i].href);
       slimer.wait(100);
   }
   page.close();
   //phantom.exit();

});</lang>

I would have used PhantomJs but had trouble building it and lost patience rapidly. Axtens (talk) 15:21, 16 April 2015 (UTC)
Ah silly me, I should have done a 'sudo apt install phantomjs'. Trying that now. Meanwhile, I'm happy to make this a challenge. Worthwhile? Axtens (talk) 15:41, 16 April 2015 (UTC)
One could try to make this itself a programming task 🤔 Rainb (talk) 12:45, 30 January 2019 (UTC)
Indeed one could; and maybe call it something like Rosetta_Code/Tasks_without_examples, except that already exists. --Thundergnat (talk) 16:42, 30 January 2019 (UTC)