X-Git-Url: https://git.pterodactylus.net/?a=blobdiff_plain;f=src%2Fmain%2Fkotlin%2Fnet%2Fpterodactylus%2Fsone%2Fcore%2FDefaultElementLoader.kt;h=69872c52bed533c3f424b7731b20db7ad10c66ab;hb=ca05f37d6d77ebad800b252719b0ff03877fc968;hp=2849029394504e205cd8e9476fcf484c09908c27;hpb=a72774c302cdb2c35508380fb2ec445be2705efe;p=Sone.git diff --git a/src/main/kotlin/net/pterodactylus/sone/core/DefaultElementLoader.kt b/src/main/kotlin/net/pterodactylus/sone/core/DefaultElementLoader.kt index 2849029..69872c5 100644 --- a/src/main/kotlin/net/pterodactylus/sone/core/DefaultElementLoader.kt +++ b/src/main/kotlin/net/pterodactylus/sone/core/DefaultElementLoader.kt @@ -6,7 +6,6 @@ import com.google.common.cache.CacheBuilder import freenet.keys.FreenetURI import org.jsoup.Jsoup import org.jsoup.nodes.Document -import org.jsoup.nodes.TextNode import java.io.ByteArrayInputStream import java.net.URLDecoder import java.nio.charset.Charset @@ -100,11 +99,10 @@ private val Document.metaDescription: String? ?.second private val Document.firstNonHeadingParagraph: String? - get() = body().children() - .filter { it.children().all { it is TextNode } } + get() = body().select("div, p") + .filter { it.textNodes().isNotEmpty() } .map { it to it.text() } - .filterNot { it.second == "" } - .firstOrNull { !it.first.tagName().startsWith("h", ignoreCase = true) } + .firstOrNull { it.second != "" } ?.second private val Int.human get() = when (this) {