Use ktxml for parsing feed XML (#212)

* Add ktxml dependency * Add common RSS and Atom content parsers * Add common feed parser For the time being the feed parser interface is converted to open class, until the platform specific feed parsers are removed. After that we can change this to normal class * Stop providing platform specific feed parsers * Remove platform specific feed parsers * Remove open modifier for `FeedParser` * Convert content parsers to Kotlin objects * Rename `xmlContent` param name to `feedContent`
msasikanth · Jan 8, 2024 · 9a25c14 · 9a25c14
1 parent f388291
commit 9a25c14
Show file tree

Hide file tree

Showing 17 changed files with 110 additions and 716 deletions.
diff --git a/core/network/build.gradle.kts b/core/network/build.gradle.kts
@@ -43,6 +43,7 @@ kotlin {
       // TODO: Extract logging abstraction into separate module
       implementation(libs.napier)
       implementation(libs.sentry)
+      implementation(libs.ktxml)
     }
     commonTest.dependencies { implementation(libs.kotlin.test) }
 

diff --git a/...twork/src/androidMain/kotlin/dev/sasikanth/rss/reader/core/network/di/NetworkComponent.kt b/...twork/src/androidMain/kotlin/dev/sasikanth/rss/reader/core/network/di/NetworkComponent.kt
@@ -16,18 +16,13 @@
 
 package dev.sasikanth.rss.reader.core.network.di
 
-import dev.sasikanth.rss.reader.core.network.parser.AndroidFeedParser
-import dev.sasikanth.rss.reader.core.network.parser.FeedParser
 import dev.sasikanth.rss.reader.di.scopes.AppScope
 import io.ktor.client.HttpClient
 import io.ktor.client.engine.okhttp.OkHttp
 import me.tatarka.inject.annotations.Provides
 
 actual interface NetworkComponent {
 
-  val AndroidFeedParser.bind: FeedParser
-    @Provides @AppScope get() = this
-
   @Provides
   @AppScope
   fun providesHttpClient(): HttpClient {

diff --git a/.../src/androidMain/kotlin/dev/sasikanth/rss/reader/core/network/parser/AndroidFeedParser.kt b/.../src/androidMain/kotlin/dev/sasikanth/rss/reader/core/network/parser/AndroidFeedParser.kt
diff --git a/...etwork/src/commonMain/kotlin/dev/sasikanth/rss/reader/core/network/fetcher/FeedFetcher.kt b/...etwork/src/commonMain/kotlin/dev/sasikanth/rss/reader/core/network/fetcher/FeedFetcher.kt
@@ -93,7 +93,7 @@ class FeedFetcher(private val httpClient: HttpClient, private val feedParser: Fe
         throw UnsupportedOperationException()
       }
     } else {
-      val feedPayload = feedParser.parse(xmlContent = responseContent, feedUrl = url)
+      val feedPayload = feedParser.parse(feedContent = responseContent, feedUrl = url)
       FeedFetchResult.Success(feedPayload)
     }
   }

diff --git a/.../core/network/parser/AndroidAtomParser.kt → .../core/network/parser/AtomContentParser.kt b/.../core/network/parser/AndroidAtomParser.kt → .../core/network/parser/AtomContentParser.kt
@@ -16,7 +16,6 @@
 
 package dev.sasikanth.rss.reader.core.network.parser
 
-import android.net.Uri
 import com.mohamedrejeb.ksoup.html.parser.KsoupHtmlOptions
 import com.mohamedrejeb.ksoup.html.parser.KsoupHtmlParser
 import dev.sasikanth.rss.reader.core.model.remote.FeedPayload
@@ -32,23 +31,24 @@ import dev.sasikanth.rss.reader.core.network.parser.FeedParser.Companion.TAG_PUB
 import dev.sasikanth.rss.reader.core.network.parser.FeedParser.Companion.TAG_SUBTITLE
 import dev.sasikanth.rss.reader.core.network.parser.FeedParser.Companion.TAG_TITLE
 import dev.sasikanth.rss.reader.core.network.parser.FeedParser.Companion.TAG_UPDATED
+import io.ktor.http.Url
 import kotlinx.datetime.Clock
-import org.xmlpull.v1.XmlPullParser
+import org.kobjects.ktxml.api.EventType
+import org.kobjects.ktxml.api.XmlPullParser
 
-internal class AndroidAtomParser(private val parser: XmlPullParser, private val feedUrl: String) :
-  Parser() {
+internal object AtomContentParser : ContentParser() {
 
-  override fun parse(): FeedPayload {
-    parser.require(XmlPullParser.START_TAG, namespace, TAG_ATOM_FEED)
+  override fun parse(feedUrl: String, parser: XmlPullParser): FeedPayload {
+    parser.require(EventType.START_TAG, parser.namespace, TAG_ATOM_FEED)
 
     val posts = mutableListOf<PostPayload?>()
 
     var title: String? = null
     var description: String? = null
     var link: String? = null
 
-    while (parser.next() != XmlPullParser.END_TAG) {
-      if (parser.eventType != XmlPullParser.START_TAG) continue
+    while (parser.next() != EventType.END_TAG) {
+      if (parser.eventType != EventType.START_TAG) continue
       when (val name = parser.name) {
         TAG_TITLE -> {
           title = readTagText(name, parser)
@@ -70,30 +70,36 @@ internal class AndroidAtomParser(private val parser: XmlPullParser, private val
       }
     }
 
-    val domain = Uri.parse(link).host!!
-    val iconUrl = FeedParser.feedIcon(domain)
+    val domain = Url(link!!)
+    val host =
+      if (domain.host != "localhost") {
+        domain.host
+      } else {
+        throw NullPointerException("Unable to get host domain")
+      }
+    val iconUrl = FeedParser.feedIcon(host)
 
     return FeedPayload(
       name = FeedParser.cleanText(title ?: link, decodeUrlEncoding = true)!!,
       description = FeedParser.cleanText(description, decodeUrlEncoding = true).orEmpty(),
       icon = iconUrl,
-      homepageLink = link!!,
+      homepageLink = link,
       link = feedUrl,
       posts = posts.filterNotNull()
     )
   }
 
   private fun readAtomEntry(parser: XmlPullParser, hostLink: String): PostPayload? {
-    parser.require(XmlPullParser.START_TAG, null, "entry")
+    parser.require(EventType.START_TAG, null, "entry")
 
     var title: String? = null
     var link: String? = null
     var content: String? = null
     var date: String? = null
     var image: String? = null
 
-    while (parser.next() != XmlPullParser.END_TAG) {
-      if (parser.eventType != XmlPullParser.START_TAG) continue
+    while (parser.next() != EventType.END_TAG) {
+      if (parser.eventType != EventType.START_TAG) continue
 
       when (val tagName = parser.name) {
         TAG_TITLE -> {
@@ -148,13 +154,13 @@ internal class AndroidAtomParser(private val parser: XmlPullParser, private val
 
   private fun readAtomLink(tagName: String, parser: XmlPullParser): String? {
     var link: String? = null
-    parser.require(XmlPullParser.START_TAG, namespace, tagName)
-    val relType = parser.getAttributeValue(namespace, ATTR_REL)
+    parser.require(EventType.START_TAG, parser.namespace, tagName)
+    val relType = parser.getAttributeValue(parser.namespace, ATTR_REL)
     if (relType == ATTR_VALUE_ALTERNATE || relType.isNullOrBlank()) {
-      link = parser.getAttributeValue(namespace, ATTR_HREF)
+      link = parser.getAttributeValue(parser.namespace, ATTR_HREF)
     }
     parser.nextTag()
-    parser.require(XmlPullParser.END_TAG, namespace, tagName)
+    parser.require(EventType.END_TAG, parser.namespace, tagName)
     return link
   }
 }
diff --git a/.../rss/reader/core/network/parser/Parser.kt → ...ader/core/network/parser/ContentParser.kt b/.../rss/reader/core/network/parser/Parser.kt → ...ader/core/network/parser/ContentParser.kt
@@ -1,5 +1,5 @@
 /*
- * Copyright 2023 Sasikanth Miriyampalli
+ * Copyright 2024 Sasikanth Miriyampalli
  *
  * Licensed under the Apache License, Version 2.0 (the "License");
  * you may not use this file except in compliance with the License.
@@ -17,43 +17,45 @@
 package dev.sasikanth.rss.reader.core.network.parser
 
 import dev.sasikanth.rss.reader.core.model.remote.FeedPayload
-import org.xmlpull.v1.XmlPullParser
+import org.kobjects.ktxml.api.EventType
+import org.kobjects.ktxml.api.XmlPullParser
 
-abstract class Parser {
+abstract class ContentParser {
 
-  val namespace: String? = null
-
-  abstract fun parse(): FeedPayload
+  abstract fun parse(feedUrl: String, parser: XmlPullParser): FeedPayload
 
   fun readAttrText(attrName: String, parser: XmlPullParser): String? {
-    val url = parser.getAttributeValue(namespace, attrName)
+    val url = parser.getAttributeValue(parser.namespace, attrName)
     skip(parser)
     return url
   }
 
   fun readTagText(tagName: String, parser: XmlPullParser): String {
-    parser.require(XmlPullParser.START_TAG, namespace, tagName)
+    parser.require(EventType.START_TAG, parser.namespace, tagName)
     val title = readText(parser)
-    parser.require(XmlPullParser.END_TAG, namespace, tagName)
+    parser.require(EventType.END_TAG, parser.namespace, tagName)
     return title
   }
 
   private fun readText(parser: XmlPullParser): String {
     var result = ""
-    if (parser.next() == XmlPullParser.TEXT) {
+    if (parser.next() == EventType.TEXT) {
       result = parser.text
       parser.nextTag()
     }
     return result
   }
 
   fun skip(parser: XmlPullParser) {
-    parser.require(XmlPullParser.START_TAG, namespace, null)
+    parser.require(EventType.START_TAG, parser.namespace, null)
     var depth = 1
     while (depth != 0) {
       when (parser.next()) {
-        XmlPullParser.END_TAG -> depth--
-        XmlPullParser.START_TAG -> depth++
+        EventType.END_TAG -> depth--
+        EventType.START_TAG -> depth++
+        else -> {
+          // no-op
+        }
       }
     }
   }

diff --git a/.../network/src/commonMain/kotlin/dev/sasikanth/rss/reader/core/network/parser/FeedParser.kt b/.../network/src/commonMain/kotlin/dev/sasikanth/rss/reader/core/network/parser/FeedParser.kt
@@ -16,12 +16,49 @@
 package dev.sasikanth.rss.reader.core.network.parser
 
 import dev.sasikanth.rss.reader.core.model.remote.FeedPayload
+import dev.sasikanth.rss.reader.di.scopes.AppScope
+import dev.sasikanth.rss.reader.exceptions.XmlParsingError
+import dev.sasikanth.rss.reader.util.DispatchersProvider
 import dev.sasikanth.rss.reader.util.decodeUrlEncodedString
+import io.github.aakira.napier.LogLevel
+import io.github.aakira.napier.log
 import io.ktor.http.URLBuilder
 import io.ktor.http.URLProtocol
 import io.ktor.http.set
-
-interface FeedParser {
+import kotlinx.coroutines.withContext
+import me.tatarka.inject.annotations.Inject
+import org.kobjects.ktxml.api.XmlPullParserException
+import org.kobjects.ktxml.mini.MiniXmlPullParser
+
+@Inject
+@AppScope
+class FeedParser(private val dispatchersProvider: DispatchersProvider) {
+
+  suspend fun parse(feedContent: String, feedUrl: String): FeedPayload {
+    return try {
+      withContext(dispatchersProvider.io) {
+        // Currently MiniXmlPullParser fails to parse XML if it contains
+        // the <?xml ?> tag in the first line. So we are removing it until
+        // the issue gets resolved.
+        // https://github.com/kobjects/ktxml/issues/5
+        val xmlDeclarationPattern = Regex("<\\?xml .*\\?>")
+        val parser =
+          MiniXmlPullParser(source = xmlDeclarationPattern.replaceFirst(feedContent, "").iterator())
+
+        parser.nextTag()
+
+        return@withContext when (parser.name) {
+          RSS_TAG -> RssContentParser.parse(feedUrl, parser)
+          ATOM_TAG -> AtomContentParser.parse(feedUrl, parser)
+          HTML_TAG -> throw HtmlContentException()
+          else -> throw UnsupportedOperationException("Unknown feed type: ${parser.name}")
+        }
+      }
+    } catch (e: XmlPullParserException) {
+      log(LogLevel.ERROR, throwable = e) { "Failed to parse the XML" }
+      throw XmlParsingError(e.stackTraceToString())
+    }
+  }
 
   companion object {
     const val RSS_TAG = "rss"
@@ -103,8 +140,6 @@ interface FeedParser {
       return pattern.containsMatchIn(url)
     }
   }
-
-  suspend fun parse(xmlContent: String, feedUrl: String): FeedPayload
 }
 
 internal class HtmlContentException : Exception()