Download model from github-raw

BakaBotTeam · Feb 25, 2024 · bef05ca · bef05ca
1 parent c898a2b
commit bef05ca
Show file tree

Hide file tree

Showing 3 changed files with 26 additions and 4 deletions.
diff --git a/src/main/kotlin/ltd/guimc/lgzbot/command/LGZBotCommand.kt b/src/main/kotlin/ltd/guimc/lgzbot/command/LGZBotCommand.kt
@@ -136,4 +136,11 @@ object LGZBotCommand : CompositeCommand(
         LL4JUtils.learn(type, string)
         sendMessage("Done.")
     }
+
+    @SubCommand("downloadModel")
+    @Description("从ADDetector仓库下载模型")
+    suspend fun CommandSender.llIIllIIllI() {
+        sendMessage("Downloading (current: ${LL4JUtils.version})")
+        LL4JUtils.downloadModel()
+    }
 }
diff --git a/src/main/kotlin/ltd/guimc/lgzbot/utils/LL4JUtils.kt b/src/main/kotlin/ltd/guimc/lgzbot/utils/LL4JUtils.kt
@@ -1,20 +1,23 @@
 package ltd.guimc.lgzbot.utils
 
-import huzpsb.ll4j.utils.data.DataSet
 import huzpsb.ll4j.model.Model
 import huzpsb.ll4j.nlp.token.Tokenizer
+import huzpsb.ll4j.utils.data.DataSet
 import java.io.BufferedReader
 import java.io.InputStreamReader
 
 object LL4JUtils {
     lateinit var model: Model
     lateinit var tokenizer: Tokenizer
+    var version = "FEB25"
 
     fun init() {
         val tokenizerFile = LL4JUtils.javaClass.getResourceAsStream("/ts.model")!!
         val modelFile = LL4JUtils.javaClass.getResourceAsStream("/anti-ad.model")!!
-        tokenizer = Tokenizer.load(tokenizerFile)
-        model = Model.read(BufferedReader(InputStreamReader(modelFile)))
+        tokenizer = Tokenizer.load(tokenizerFile.bufferedReader(Charsets.UTF_8))
+        modelFile.bufferedReader(Charsets.UTF_8).use {
+            model = Model.read(it)
+        }
     }
 
     fun predict(string: String): Boolean =
@@ -31,4 +34,13 @@ object LL4JUtils {
         dataSet.split.add(tokenizer.tokenize(type, string.replace("\n", "")))
         model.trainOn(dataSet)
     }
+
+    fun downloadModel() {
+        try {
+            model = Model.read(HttpUtils.getResponse("https://raw.githubusercontent.com/siuank/ADDetector4J/main/anti-ad.model"))
+            tokenizer = Tokenizer.load(HttpUtils.getResponse("https://raw.githubusercontent.com/siuank/ADDetector4J/main/t1.tokenized.txt").reader())
+            val time = GithubUtils.getLastCommit("siuank/ADDetector4J").commitTime
+            version = "${time.month.name.substring(0..3)}${time.dayOfMonth}"
+        } catch (_: Exception) {}
+    }
 }
diff --git a/src/test/java/Test.java b/src/test/java/Test.java
@@ -1,10 +1,13 @@
 import ltd.guimc.lgzbot.utils.LL4JUtils;
 
+import java.nio.charset.Charset;
 import java.util.Arrays;
 
 public class Test {
     public static void main(String[] args) {
         LL4JUtils.INSTANCE.init();
-        System.out.println(Arrays.toString(LL4JUtils.INSTANCE.predictAllResult("")));
+        System.out.println(Charset.defaultCharset());
+        System.out.println(Arrays.toString(LL4JUtils.INSTANCE.predictAllResult("shop.nyaproxy.xyz 1+ 21+NFA补货出mc美区激活码")));
+        System.out.println(LL4JUtils.INSTANCE.getModel().predict(LL4JUtils.INSTANCE.getTokenizer().tokenize(0, "shop.nyaproxy.xyz 1+ 21+NFA补货出mc美区激活码").values));
     }
 }