From: David ‘Bombe’ Roden <bombe@pterodactylus.net>
Date: Fri, 2 Sep 2022 15:18:21 +0000 (+0200)
Subject: 🚸 Improve text extraction from freesites
X-Git-Url: https://git.pterodactylus.net/?p=Sone.git;a=commitdiff_plain;h=17a659821355e6396f464e50a9b4048c0ea01ff7

🚸 Improve text extraction from freesites
---

diff --git a/src/main/kotlin/net/pterodactylus/sone/core/DefaultElementLoader.kt b/src/main/kotlin/net/pterodactylus/sone/core/DefaultElementLoader.kt
index 2849029..88cb1f5 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
@@ -101,7 +100,6 @@ private val Document.metaDescription: String?
 
 private val Document.firstNonHeadingParagraph: String?
 	get() = body().children()
-		.filter { it.children().all { it is TextNode } }
 		.map { it to it.text() }
 		.filterNot { it.second == "" }
 		.firstOrNull { !it.first.tagName().startsWith("h", ignoreCase = true) }
diff --git a/src/test/kotlin/net/pterodactylus/sone/core/DefaultElementLoaderTest.kt b/src/test/kotlin/net/pterodactylus/sone/core/DefaultElementLoaderTest.kt
index 1b03d7a..8a3d40f 100644
--- a/src/test/kotlin/net/pterodactylus/sone/core/DefaultElementLoaderTest.kt
+++ b/src/test/kotlin/net/pterodactylus/sone/core/DefaultElementLoaderTest.kt
@@ -126,14 +126,14 @@ class DefaultElementLoaderTest {
 	}
 
 	@Test
-	fun `element loader can not extract description if html is more complicated`() {
+	fun `element loader can extract description if html is more complicated`() {
 		runWithCallback(textKey) { elementLoader, _, callback, _ ->
 			callback.loaded(FreenetURI(textKey), "text/html; charset=UTF-8", read("element-loader3.html"))
 			val linkedElement = elementLoader.loadElement(textKey)
 			assertThat(linkedElement, isLinkedElement(equalTo(textKey), allOf(
 				hasEntry("type", "html"), hasEntry("size", 204), hasEntry("sizeHuman", "204 B"),
 				hasEntry("title", "Some Nice Page Title"),
-				hasEntry("description", null)
+				hasEntry("description", "This is the first paragraph of the very nice freesite.")
 			)))
 		}
 	}