Created
November 17, 2013 07:57
-
-
Save kumarsoumya/7510623 to your computer and use it in GitHub Desktop.
Read the comments section at top.
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
/* | |
* @author: Soumya Kumar | |
* | |
* The following program scrapes the webpage "https://lslab.ics.purdue.edu/icsWeb/AvailableStations" | |
* and parses it get the current lab availability information. | |
* | |
* Dependency: Jsoup Parser | |
* | |
* */ | |
/************************************************************************************/ | |
/* Sample Output - Taken at 2:54 AM on 11/17/2013 | |
* BCC204=-1.0 | |
* MRDH146S=-1.0 | |
* BRES201=-1.0 | |
* RHPH272=-1.0 | |
* RHPH316=-1.0 | |
* HAMP3144=-1.0 | |
* HAMP2215=-1.0 | |
* PHYS026=-1.0 | |
* PHYS022=-1.0 | |
* PHYS117=-1.0 | |
* PHYS116=-1.0 | |
* PHYS014=-1.0 | |
* PHYS290=-1.0 | |
* MATHB010=73.52941176470588 | |
* MATH311=-1.0 | |
* SC283=-1.0 | |
* SCG046=-1.0 | |
* SC189=-1.0 | |
* SC183=-1.0 | |
* SC246=-1.0 | |
* SC277=-1.0 | |
* SC179=-1.0 | |
* SC231=-1.0 | |
* SC289=-1.0 | |
* WTHR214=-1.0 | |
* WTHR212=-1.0 | |
* WTHR114=-1.0 | |
* WTHR307=-1.0 | |
* WTHR305=-1.0 | |
* HEAV227=-1.0 | |
* GRIS121=-1.0 | |
* GRIS120=-1.0 | |
* BRNGB286=-1.0 | |
* BRNGB282=-1.0 | |
* BRNGB280=-1.0 | |
* BRNGB275=-1.0 | |
* BRNGB274=-1.0 | |
* BRNGB291=-1.0 | |
* MTHW116=-1.0 | |
* MTHW301=-1.0 | |
* STEW135=-1.0 | |
* STEW102=-1.0 | |
* HIKSG950=71.1864406779661 | |
* HIKSG980D=71.42857142857143 | |
* HIKSG980C=100.0 | |
* LILYG428=-1.0 | |
* LILY2400=-1.0 | |
* LYNN2133=-1.0 | |
* MCUTC216=-1.0 | |
*/ | |
/*******************************************************************************************************/ | |
import org.jsoup.nodes.Document; | |
import org.jsoup.Jsoup; | |
import org.jsoup.nodes.Element; | |
import org.jsoup.select.Elements; | |
public class HTMLLabScrapper { | |
/* | |
* -1.0 means closed | |
* -2.0 means class in session | |
* else return percentage available computers | |
*/ | |
/* | |
* @return percentage of lab availibity or 1.0 or 2.0 | |
* | |
*/ | |
public static double getLabInfo(String labUrl) { | |
Document doc=null; | |
try { | |
doc = Jsoup.connect(labUrl).get(); | |
} catch (Exception e) { | |
System.out.println("Error connecting using Jsoup"); | |
} | |
int count = 0; | |
String temp[]; | |
double avail=0; //number of available computers | |
double inUse=0; //number of computer in use | |
Elements links = doc.select("span"); | |
for (Element link : links) { | |
count++; | |
if (count==2) { | |
temp=link.text().split(" "); | |
if (temp[3].equals("OPEN.")) | |
; | |
else if(temp[3].equals("CLOSED.")) | |
return -1.0; | |
else | |
return -2.0; | |
} | |
if (count==3) { | |
temp=link.text().split(" "); | |
inUse=Double.parseDouble(temp[2]); | |
} | |
if (count==4) { | |
temp=link.text().split(" "); | |
avail=Double.parseDouble(temp[2]); | |
} | |
} | |
return (avail-inUse)/avail*100; | |
} | |
public static void main(String[] args) { | |
String[] labUrls = | |
{ | |
"https://lslab.ics.purdue.edu/icsWeb/LabInfo?building=BCC&room=204", | |
"https://lslab.ics.purdue.edu/icsWeb/LabInfo?building=MRDH&room=146S", | |
"https://lslab.ics.purdue.edu/icsWeb/LabInfo?building=BRES&room=201", | |
"https://lslab.ics.purdue.edu/icsWeb/LabInfo?building=RHPH&room=272", | |
"https://lslab.ics.purdue.edu/icsWeb/LabInfo?building=RHPH&room=316", | |
"https://lslab.ics.purdue.edu/icsWeb/LabInfo?building=HAMP&room=3144", | |
"https://lslab.ics.purdue.edu/icsWeb/LabInfo?building=HAMP&room=2215", | |
"https://lslab.ics.purdue.edu/icsWeb/LabInfo?building=PHYS&room=026", | |
"https://lslab.ics.purdue.edu/icsWeb/LabInfo?building=PHYS&room=022", | |
"https://lslab.ics.purdue.edu/icsWeb/LabInfo?building=PHYS&room=117", | |
"https://lslab.ics.purdue.edu/icsWeb/LabInfo?building=PHYS&room=116", | |
"https://lslab.ics.purdue.edu/icsWeb/LabInfo?building=PHYS&room=014", | |
"https://lslab.ics.purdue.edu/icsWeb/LabInfo?building=PHYS&room=290", | |
"https://lslab.ics.purdue.edu/icsWeb/LabInfo?building=MATH&room=B010", | |
"https://lslab.ics.purdue.edu/icsWeb/LabInfo?building=MATH&room=311", | |
"https://lslab.ics.purdue.edu/icsWeb/LabInfo?building=SC&room=283", | |
"https://lslab.ics.purdue.edu/icsWeb/LabInfo?building=SC&room=G046", | |
"https://lslab.ics.purdue.edu/icsWeb/LabInfo?building=SC&room=189", | |
"https://lslab.ics.purdue.edu/icsWeb/LabInfo?building=SC&room=183", | |
"https://lslab.ics.purdue.edu/icsWeb/LabInfo?building=SC&room=246", | |
"https://lslab.ics.purdue.edu/icsWeb/LabInfo?building=SC&room=277", | |
"https://lslab.ics.purdue.edu/icsWeb/LabInfo?building=SC&room=179", | |
"https://lslab.ics.purdue.edu/icsWeb/LabInfo?building=SC&room=231", | |
"https://lslab.ics.purdue.edu/icsWeb/LabInfo?building=SC&room=289", | |
"https://lslab.ics.purdue.edu/icsWeb/LabInfo?building=WTHR&room=214", | |
"https://lslab.ics.purdue.edu/icsWeb/LabInfo?building=WTHR&room=212", | |
"https://lslab.ics.purdue.edu/icsWeb/LabInfo?building=WTHR&room=114", | |
"https://lslab.ics.purdue.edu/icsWeb/LabInfo?building=WTHR&room=307", | |
"https://lslab.ics.purdue.edu/icsWeb/LabInfo?building=WTHR&room=305", | |
"https://lslab.ics.purdue.edu/icsWeb/LabInfo?building=HEAV&room=227", | |
"https://lslab.ics.purdue.edu/icsWeb/LabInfo?building=GRIS&room=121", | |
"https://lslab.ics.purdue.edu/icsWeb/LabInfo?building=GRIS&room=120", | |
"https://lslab.ics.purdue.edu/icsWeb/LabInfo?building=BRNG&room=B286", | |
"https://lslab.ics.purdue.edu/icsWeb/LabInfo?building=BRNG&room=B282", | |
"https://lslab.ics.purdue.edu/icsWeb/LabInfo?building=BRNG&room=B280", | |
"https://lslab.ics.purdue.edu/icsWeb/LabInfo?building=BRNG&room=B275", | |
"https://lslab.ics.purdue.edu/icsWeb/LabInfo?building=BRNG&room=B274", | |
"https://lslab.ics.purdue.edu/icsWeb/LabInfo?building=BRNG&room=B291", | |
"https://lslab.ics.purdue.edu/icsWeb/LabInfo?building=MTHW&room=116", | |
"https://lslab.ics.purdue.edu/icsWeb/LabInfo?building=MTHW&room=301", | |
"https://lslab.ics.purdue.edu/icsWeb/LabInfo?building=STEW&room=135", | |
"https://lslab.ics.purdue.edu/icsWeb/LabInfo?building=STEW&room=102", | |
"https://lslab.ics.purdue.edu/icsWeb/LabInfo?building=HIKS&room=G950", | |
"https://lslab.ics.purdue.edu/icsWeb/LabInfo?building=HIKS&room=G980D", | |
"https://lslab.ics.purdue.edu/icsWeb/LabInfo?building=HIKS&room=G980C", | |
"https://lslab.ics.purdue.edu/icsWeb/LabInfo?building=LILY&room=G428", | |
"https://lslab.ics.purdue.edu/icsWeb/LabInfo?building=LILY&room=2400", | |
"https://lslab.ics.purdue.edu/icsWeb/LabInfo?building=LYNN&room=2133", | |
"https://lslab.ics.purdue.edu/icsWeb/LabInfo?building=MCUT&room=C216" | |
}; | |
String labName; | |
String[] temp; | |
for (int i=0; i<labUrls.length; i++) { | |
labName = labUrls[i].substring(53, labUrls[i].length()); | |
temp = labName.split("&room="); | |
System.out.print(temp[0]); | |
System.out.print(temp[1]); | |
System.out.println("="+getLabInfo(labUrls[i])); | |
} | |
} | |
} | |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment