bea/HidekoBot
1
0
Fork 0
Code Issues 8 Pull Requests Packages Projects Releases Wiki Activity

Improve urban dictionary parsing

Browse Source
This commit is contained in:
Bea 2022-12-19 22:45:02 +01:00
parent 1421d52598
commit d5664eb646
1 changed files with 57 additions and 38 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

71
src/main/java/wtf/beatrice/hidekobot/commands/message/UrbanDictionaryCommand.java
View File

@ -76,33 +76,22 @@ public class UrbanDictionaryCommand implements MessageCommand
return;
}
List<String> htmlMeanings = new ArrayList<>();
List<String> htmlExamples = new ArrayList<>();
Elements definitions = doc.getElementsByClass("definition");
for(Element element : definitions)
{
Elements meanings = element.getElementsByClass("meaning");
for(Element meaning : meanings)
{
htmlMeanings.add(meaning.html());
break;// just one meaning per definition
}
Elements examples = element.getElementsByClass("example");
for(Element example : examples)
{
htmlExamples.add(example.html());
break; // just one example per definition
}
}
List<String> contributorsNames = new ArrayList<>();
List<String> submissionDates = new ArrayList<>();
List<String> plaintextMeanings = new ArrayList<>();
List<String> plaintextExamples = new ArrayList<>();
for(String htmlMeaning : htmlMeanings)
Elements definitions = doc.getElementsByClass("definition");
for(Element definition : definitions)
{
String text = htmlMeaning
Elements meaningSingleton = definition.getElementsByClass("meaning");
if(meaningSingleton.isEmpty())
{
plaintextMeanings.add(" ");
} else
{
Element meaning = meaningSingleton.get(0);
String text = meaning.html()
.replaceAll("<br\\s*?>", "\n") // keep newlines
.replaceAll("<.*?>", ""); // remove all other html tags
// discord only allows 1024 characters for embed fields
@ -110,9 +99,15 @@ public class UrbanDictionaryCommand implements MessageCommand
plaintextMeanings.add(text);
}
for(String htmlExample : htmlExamples)
Elements exampleSingleton = definition.getElementsByClass("example");
if(exampleSingleton.isEmpty())
{
String text = htmlExample
plaintextExamples.add(" ");
} else
{
Element example = exampleSingleton.get(0);
String text = example.html()
.replaceAll("<br\\s*?>", "\n") // keep newlines
.replaceAll("<.*?>", ""); // remove all other html tags
// discord only allows 1024 characters for embed fields
@ -120,16 +115,40 @@ public class UrbanDictionaryCommand implements MessageCommand
plaintextExamples.add(text);
}
Elements contributorSingleton = definition.getElementsByClass("contributor");
if(contributorSingleton.isEmpty())
{
contributorsNames.add("Unknown");
} else
{
Element contributor = contributorSingleton.get(0);
String htmlContributor = contributor.html();
String htmlContributorName = contributor.select("a").html();
String htmlSubmitDate = htmlContributor.substring(
htmlContributor.indexOf("</a>") + 4);
contributorsNames.add(htmlContributorName
.replaceAll("<.*?>", "")); // remove all html tags;
submissionDates.add(htmlSubmitDate
.replaceAll("<.*?>", "")); // remove all html tags;
}
}
// make it nice to look at, compared to the html value
term = term.replaceAll("\\+", " ");
term = WordUtils.capitalizeFully(term);
EmbedBuilder embedBuilder = new EmbedBuilder();
embedBuilder.setColor(Cache.getBotColor());
embedBuilder.setTitle("Urban Dictionary: " + term);
embedBuilder.setTitle(term + ", on Urban Dictionary", url);
embedBuilder.setAuthor(event.getAuthor().getAsTag(), null, event.getAuthor().getAvatarUrl());
embedBuilder.addField("Definition", plaintextMeanings.get(0), false);
embedBuilder.addField("Example", plaintextExamples.get(0), false);
embedBuilder.addField("Submission",
"*sent by " + contributorsNames.get(0) + " on " + submissionDates.get(0) + "*",
false);
event.getChannel().sendMessageEmbeds(embedBuilder.build()).queue();
}