Android: allow passing a Jsoup Document to html converter (#91)

* Android: allow passing a Jsoup `Document` to html converter

Also, centralise, extract and make public the HTML to Document parser.

* Fix tests

Author: jmartinesp

platforms/android/library-compose/src/main/java/io/element/android/wysiwyg/compose/StyledHtmlConverter.kt

@@ -12,6 +12,7 @@ import android.content.Context
 import io.element.android.wysiwyg.compose.internal.toStyleConfig
 import io.element.android.wysiwyg.display.MentionDisplayHandler
 import io.element.android.wysiwyg.utils.HtmlConverter
+import org.jsoup.nodes.Document
 import timber.log.Timber
 
 /**
@@ -41,9 +42,11 @@ class StyledHtmlConverter(
         return htmlConverter?.fromHtmlToSpans(html) ?: errorNotConfigured()
     }
 
+    override fun fromDocumentToSpans(dom: Document): CharSequence {
+        return htmlConverter?.fromDocumentToSpans(dom) ?: errorNotConfigured()
+    }
+
     private fun errorNotConfigured(): Nothing {
         error("ComposableHtmlConverter must be configured with a RichTextEditorStyle before use")
     }
-
-
 }

platforms/android/library/src/main/java/io/element/android/wysiwyg/internal/utils/AndroidHtmlConverter.kt

@@ -9,13 +9,20 @@
 package io.element.android.wysiwyg.internal.utils
 
 import io.element.android.wysiwyg.utils.HtmlConverter
+import io.element.android.wysiwyg.utils.HtmlToDomParser
 import io.element.android.wysiwyg.utils.HtmlToSpansParser
+import org.jsoup.nodes.Document
 
 internal class AndroidHtmlConverter(
-    private val provideHtmlToSpansParser: (html: String) -> HtmlToSpansParser
+    private val provideHtmlToSpansParser: (dom: Document) -> HtmlToSpansParser
 ) : HtmlConverter {
 
-    override fun fromHtmlToSpans(html: String): CharSequence =
-        provideHtmlToSpansParser(html).convert()
+    override fun fromHtmlToSpans(html: String): CharSequence {
+        val dom = HtmlToDomParser.document(html)
+        return fromDocumentToSpans(dom)
+    }
 
+    override fun fromDocumentToSpans(dom: Document): CharSequence {
+        return provideHtmlToSpansParser(dom).convert()
+    }
 }

platforms/android/library/src/main/java/io/element/android/wysiwyg/utils/HtmlConverter.kt

@@ -8,15 +8,15 @@
 
 package io.element.android.wysiwyg.utils
 
-import android.app.Application
 import android.content.Context
 import io.element.android.wysiwyg.display.MentionDisplayHandler
 import io.element.android.wysiwyg.internal.utils.AndroidHtmlConverter
 import io.element.android.wysiwyg.view.StyleConfig
+import org.jsoup.nodes.Document
 
 interface HtmlConverter {
-
     fun fromHtmlToSpans(html: String): CharSequence
+    fun fromDocumentToSpans(dom: Document): CharSequence
 
     object Factory {
         fun create(
@@ -27,10 +27,10 @@ interface HtmlConverter {
             isMention: ((text: String, url: String) -> Boolean)? = null,
         ): HtmlConverter {
             val resourcesProvider = AndroidResourcesHelper(context)
-            return AndroidHtmlConverter(provideHtmlToSpansParser = { html ->
+            return AndroidHtmlConverter(provideHtmlToSpansParser = { dom ->
                 HtmlToSpansParser(
                     resourcesHelper = resourcesProvider,
-                    html = html,
+                    dom = dom,
                     styleConfig = styleConfig,
                     mentionDisplayHandler = mentionDisplayHandler,
                     isEditor = isEditor,
@@ -39,6 +39,4 @@ interface HtmlConverter {
             })
         }
     }
-
-
 }

platforms/android/library/src/main/java/io/element/android/wysiwyg/utils/HtmlToDomParser.kt

@@ -0,0 +1,22 @@
+package io.element.android.wysiwyg.utils
+
+import org.jsoup.Jsoup
+import org.jsoup.nodes.Document
+import org.jsoup.nodes.Document.OutputSettings
+import org.jsoup.safety.Safelist
+
+object HtmlToDomParser {
+    fun document(html: String): Document {
+        val outputSettings = OutputSettings().prettyPrint(false).indentAmount(0)
+        val cleanHtml = Jsoup.clean(html, "", safeList, outputSettings)
+        return Jsoup.parse(cleanHtml)
+    }
+
+    private val safeList = Safelist()
+        .addTags(
+            "a", "b", "strong", "i", "em", "u", "del", "code", "ul", "ol", "li", "pre",
+            "blockquote", "p", "br"
+        )
+        .addAttributes("a", "href", "data-mention-type", "contenteditable")
+        .addAttributes("ol", "start")
+}

platforms/android/library/src/main/java/io/element/android/wysiwyg/utils/HtmlToSpansParser.kt

@@ -32,13 +32,11 @@ import io.element.android.wysiwyg.view.spans.PillSpan
 import io.element.android.wysiwyg.view.spans.PlainAtRoomMentionDisplaySpan
 import io.element.android.wysiwyg.view.spans.QuoteSpan
 import io.element.android.wysiwyg.view.spans.UnorderedListSpan
-import org.jsoup.Jsoup
 import org.jsoup.internal.StringUtil
-import org.jsoup.nodes.Document.OutputSettings
+import org.jsoup.nodes.Document
 import org.jsoup.nodes.Element
 import org.jsoup.nodes.Node
 import org.jsoup.nodes.TextNode
-import org.jsoup.safety.Safelist
 import timber.log.Timber
 import kotlin.math.roundToInt
 
@@ -51,27 +49,16 @@ import kotlin.math.roundToInt
  */
 internal class HtmlToSpansParser(
     private val resourcesHelper: ResourcesHelper,
-    private val html: String,
+    private val dom: Document,
     private val styleConfig: StyleConfig,
     private val mentionDisplayHandler: MentionDisplayHandler?,
     private val isEditor: Boolean,
     private val isMention: ((text: String, url: String) -> Boolean)? = null,
 ) {
-    private val safeList = Safelist()
-        .addTags(
-            "a", "b", "strong", "i", "em", "u", "del", "code", "ul", "ol", "li", "pre",
-            "blockquote", "p", "br"
-        )
-        .addAttributes("a", "href", "data-mention-type", "contenteditable")
-        .addAttributes("ol", "start")
-
     /**
      * Convert the HTML string into a [Spanned] text.
      */
     fun convert(): Spanned {
-        val outputSettings = OutputSettings().prettyPrint(false).indentAmount(0)
-        val cleanHtml = Jsoup.clean(html, "", safeList, outputSettings)
-        val dom = Jsoup.parse(cleanHtml)
         val text = buildSpannedString {
             val body = dom.body()
             parseChildren(body)

platforms/android/library/src/test/kotlin/io/element/android/wysiwyg/utils/BasicHtmlConverter.kt

@@ -8,10 +8,14 @@
 
 package io.element.android.wysiwyg.utils
 
+import org.jsoup.nodes.Document
+
 /**
  * HTML converter that is not depend on Android, for unit tests.
  */
 class BasicHtmlConverter: HtmlConverter {
-
     override fun fromHtmlToSpans(html: String): CharSequence = html.replace("<[^>]*>".toRegex(), "")
+    override fun fromDocumentToSpans(dom: Document): CharSequence {
+        return fromHtmlToSpans(dom.html())
+    }
 }

platforms/android/library/src/test/kotlin/io/element/android/wysiwyg/utils/HtmlToSpansParserTest.kt

@@ -304,7 +304,7 @@ class HtmlToSpansParserTest {
         val styleConfig = createFakeStyleConfig()
         return HtmlToSpansParser(
             resourcesHelper = AndroidResourcesHelper(context = app),
-            html = html,
+            dom = HtmlToDomParser.document(html),
             styleConfig = styleConfig,
             mentionDisplayHandler = mentionDisplayHandler,
             isEditor = isEditor,