Initial commit

1 files changed, 115 insertions, 0 deletions

diff --git a/src/main/kotlin/ScraperMain.kt b/src/main/kotlin/ScraperMain.kt
new file mode 100644
index 0000000..674a87e
--- /dev/null
+++ b/src/main/kotlin/ScraperMain.kt
@@ -0,0 +1,115 @@
+import io.ktor.client.*
+import io.ktor.client.engine.cio.*
+import io.ktor.client.request.*
+import io.ktor.client.statement.*
+import kotlinx.serialization.*
+import kotlinx.serialization.descriptors.PolymorphicKind
+import kotlinx.serialization.descriptors.SerialDescriptor
+import kotlinx.serialization.descriptors.buildSerialDescriptor
+import kotlinx.serialization.encoding.Decoder
+import kotlinx.serialization.json.*
+import java.io.File
+
+@Serializable
+data class NpcListData(
+    val displayName: String,
+    val id: NpcId,
+    val tag: String? = null,
+)
+
+@Serializable
+data class ListData<T>(
+    val data: List<T>
+)
+
+
+val json = Json { ignoreUnknownKeys = true }
+val ListDataPattern = "new Listview\\((.*)\\);\n".toRegex()
+val NpcLocationDataPattern = "g_mapperData = (.*);\n".toRegex()
+val client = HttpClient(CIO)
+
+suspend fun getNpcList(zoneId: Int): List<NpcListData> {
+    return getNpcList("https://www.wowhead.com/npcs/react-a:1?filter=6;$zoneId;0") +
+            getNpcList("https://www.wowhead.com/npcs/react-a:0?filter=6;$zoneId;0") +
+            getNpcList("https://www.wowhead.com/npcs/react-a:-1?filter=6;$zoneId;0")
+}
+
+suspend fun getNpcList(url: String): List<NpcListData> {
+    val string =
+        client.get(url).bodyAsText()
+    val match = ListDataPattern.find(string)!!
+    val jsonData = match.groupValues[1]
+    val data = json.decodeFromString<ListData<NpcListData>>(jsonData).data
+    if (data.size == 1000) {
+        println("Warning: Encountered NPC limit for url $url")
+    }
+    return data
+}
+
+@Serializable
+@JvmInline
+value class ZoneId(val int: Int)
+
+@Serializable
+@JvmInline
+value class MapId(val int: Int)
+
+@Serializable
+@JvmInline
+value class NpcId(val int: Int)
+
+@Serializable
+data class MapCoordinates(
+    val count: Int,
+    val coords: List<List<Float>>,
+    val uiMapId: MapId? = null,
+    val uiMapName: String? = null,
+)
+
+@OptIn(InternalSerializationApi::class, ExperimentalSerializationApi::class)
+object NpcZoneMapSerializer : DeserializationStrategy<Map<ZoneId, List<MapCoordinates>>> {
+    override val descriptor: SerialDescriptor
+        get() = buildSerialDescriptor("NpcZoneMapSerializer", PolymorphicKind.SEALED)
+
+    override fun deserialize(decoder: Decoder): Map<ZoneId, List<MapCoordinates>> {
+        val d = decoder as JsonDecoder
+        val el = d.decodeJsonElement()
+        el as JsonObject
+
+        return el.mapNotNull { (t, u) ->
+            if (u is JsonArray)
+                ZoneId(t.toInt()) to d.json.decodeFromJsonElement<List<MapCoordinates>>(u)
+            else null
+        }.toMap()
+    }
+}
+
+suspend fun getNpcData(id: NpcId, name: String): Map<ZoneId, List<MapCoordinates>> {
+    val string = client.get("https://www.wowhead.com/npc=${id.int}/${name.replace(" ", "-").lowercase()}").bodyAsText()
+    val match = NpcLocationDataPattern.find(string)!!
+    val jsonData = match.groupValues[1]
+    try {
+        return json.decodeFromString(NpcZoneMapSerializer, jsonData)
+    } finally {
+        File("crash.html").writeText(string)
+    }
+}
+
+@Serializable
+data class CompleteNpcData(
+    val metadata: NpcListData,
+    val locations: Map<ZoneId, List<MapCoordinates>>,
+)
+
+@Serializable
+data class AllNpcData(
+    val npcData: Map<NpcId, CompleteNpcData>
+)
+
+suspend fun main() {
+    val zoneIdsToScrape = listOf(13646)
+    val npcsPerZone = zoneIdsToScrape.associateWith { getNpcList(it) }
+    val indexedNpcIds = npcsPerZone.flatMap { it.value }.associateBy { it.id }
+    val npcData = indexedNpcIds.values.associate { it.id to CompleteNpcData(it, getNpcData(it.id, it.displayName)) }
+    File("database.json").writeText(json.encodeToString(npcData))
+}
\ No newline at end of file