fix(api): cards - list media

Fixed bug which was causing AnkiDroid API to not list media on cards.

Due to card rendering being moved to the backend, and Anki
using different Regex handling code, we needed to change the
implementation of `files_in_str`:

[sound:] tags were stripped, for example

https://github.com/ankitects/anki/blob/64ca90934bc26ddf7125913abc9dd9de8cb30c2b/pylib/anki/media.py#L136-L150

Fixes 17062

Author: ujjol1234

AnkiDroid/src/androidTest/java/com/ichi2/anki/tests/ContentProviderTest.kt

@@ -21,6 +21,7 @@ package com.ichi2.anki.tests
 import android.content.ContentResolver
 import android.content.ContentUris
 import android.content.ContentValues
+import android.database.Cursor
 import android.database.CursorWindow
 import android.net.Uri
 import anki.notetypes.StockNotetype
@@ -49,12 +50,15 @@ import com.ichi2.libanki.getStockNotetype
 import com.ichi2.libanki.sched.Scheduler
 import com.ichi2.testutils.common.assertThrows
 import com.ichi2.utils.emptyStringArray
+import kotlinx.serialization.json.Json
 import net.ankiweb.rsdroid.exceptions.BackendNotFoundException
 import org.hamcrest.MatcherAssert.assertThat
+import org.hamcrest.Matchers.allOf
 import org.hamcrest.Matchers.containsString
 import org.hamcrest.Matchers.equalTo
 import org.hamcrest.Matchers.greaterThan
 import org.hamcrest.Matchers.greaterThanOrEqualTo
+import org.hamcrest.Matchers.hasItem
 import org.json.JSONObject.NULL
 import org.junit.After
 import org.junit.Assert.assertEquals
@@ -67,6 +71,7 @@ import org.junit.Rule
 import org.junit.Test
 import timber.log.Timber
 import kotlin.test.assertNotNull
+import kotlin.test.assertTrue
 import kotlin.test.junit.JUnitAsserter.assertNotNull
 
 /**
@@ -1345,6 +1350,48 @@ class ContentProviderTest : InstrumentedTest() {
             } ?: fail("query returned null")
     }
 
+    @Test
+    fun testMediaFilesAddedCorrectlyInReviewInfo() {
+        val imageFileName = "img.jpg"
+        val audioFileName = "test.mp3"
+        addNoteUsingBasicNoteType("""Hello <img src="$imageFileName"> [sound:$audioFileName]""")
+            .firstCard(col)
+            .update {
+                queue = QueueType.New
+                due = col.sched.today
+            }
+
+        queryReviewInfo { cursor ->
+            val media =
+                cursor
+                    .getString(cursor.getColumnIndex(FlashCardsContract.ReviewInfo.MEDIA_FILES))
+                    .let { Json.decodeFromString<List<String>>(it) }
+
+            assertThat(
+                "media files returned",
+                media,
+                allOf(
+                    hasItem(imageFileName),
+                    hasItem(audioFileName),
+                ),
+            )
+        }
+    }
+
+    private fun queryReviewInfo(block: (Cursor) -> Unit) {
+        contentResolver
+            .query(
+                FlashCardsContract.ReviewInfo.CONTENT_URI,
+                null,
+                null,
+                null,
+                null,
+            )?.use { cursor ->
+                assertTrue("has rows") { cursor.moveToFirst() }
+                block(cursor)
+            }
+    }
+
     private fun reopenCol(): com.ichi2.libanki.Collection {
         Timber.i("closeCollection: %s", "ContentProviderTest: reopenCol")
         CollectionManager.closeCollectionBlocking()

AnkiDroid/src/androidTest/java/com/ichi2/anki/tests/InstrumentedTest.kt

@@ -149,6 +149,13 @@ abstract class InstrumentedTest {
         }
     }
 
+    @DuplicatedCode("This should be refactored into a shared library later")
+    fun Card.update(block: Card.() -> Unit): Card {
+        block(this)
+        col.updateCard(this)
+        return this
+    }
+
     @DuplicatedCode("This is copied from RobolectricTest. This will be refactored into a shared library later")
     protected fun Card.moveToReviewQueue() {
         this.queue = QueueType.Rev

AnkiDroid/src/main/java/com/ichi2/anki/provider/CardContentProvider.kt

@@ -1137,7 +1137,7 @@ class CardContentProvider : ContentProvider() {
                 FlashCardsContract.ReviewInfo.NEXT_REVIEW_TIMES -> rb.add(nextReviewTimesJson.toString())
                 FlashCardsContract.ReviewInfo.MEDIA_FILES ->
                     rb.add(
-                        JSONArray(col.media.filesInStr(currentCard.question(col) + currentCard.answer(col))),
+                        JSONArray(col.media.filesInStr(currentCard.renderOutput(col))),
                     )
                 else -> throw UnsupportedOperationException("Queue \"$column\" is unknown")
             }

AnkiDroid/src/main/java/com/ichi2/libanki/Media.kt

@@ -20,7 +20,10 @@ package com.ichi2.libanki
 import androidx.annotation.WorkerThread
 import anki.media.CheckMediaResponse
 import com.google.protobuf.kotlin.toByteString
+import com.ichi2.libanki.TemplateManager.TemplateRenderContext.TemplateRenderOutput
 import com.ichi2.libanki.exception.EmptyMediaException
+import com.ichi2.libanki.utils.LibAnkiAlias
+import com.ichi2.libanki.utils.NotInLibAnki
 import timber.log.Timber
 import java.io.File
 
@@ -63,18 +66,36 @@ open class Media(
     /**
      * Extract media filenames from an HTML string.
      *
-     * @param string The string to scan for media filenames ([sound:...] or <img...>).
-     * @return A list containing all the sound and image filenames found in the input string.
+     * @param currentCard The card to scan for media filenames ([sound:...] or <img...>).
+     * @return A distinct, unordered list containing all the sound and image filenames found in the card.
      */
-    fun filesInStr(string: String): List<String> =
-        col.backend
-            .extractAvTags(string, true)
-            .avTagsList
-            .filter {
-                it.hasSoundOrVideo()
-            }.map {
-                it.soundOrVideo
+    @LibAnkiAlias("files_in_str")
+    fun filesInStr(
+        renderOutput: TemplateRenderOutput,
+        includeRemote: Boolean = false,
+    ): List<String> {
+        val files = mutableListOf<String>()
+        val processedText = LaTeX.mungeQA(renderOutput.questionText + renderOutput.answerText, col, true)
+
+        for (pattern in htmlMediaRegexps) {
+            val matches = pattern.findAll(processedText)
+            for (match in matches) {
+                val fname = pattern.extractFilename(match) ?: continue
+                val isLocal = !Regex("(?i)https?|ftp://").containsMatchIn(fname)
+                if (isLocal || includeRemote) {
+                    files.add(fname)
+                }
             }
+        }
+
+        // not in libAnki: the rendered output no longer contains [sound:] tags
+        files.addAll(
+            (renderOutput.questionAvTags + renderOutput.answerAvTags)
+                .filterIsInstance<SoundOrVideoTag>()
+                .map { it.filename },
+        )
+        return files.distinct()
+    }
 
     fun findUnusedMediaFiles(): List<File> = check().unusedList.map { File(dir, it) }
 
@@ -134,4 +155,32 @@ open class Media(
     private fun restoreTrash() {
         col.backend.restoreTrash()
     }
+
+    companion object {
+        /**
+         * Given a media regex, return the index of the filename in the result
+         */
+        @NotInLibAnki
+        private fun Regex.extractFilename(match: MatchResult): String? {
+            val index =
+                when (htmlMediaRegexps.indexOf(this)) {
+                    0, 2 -> 3
+                    1, 3 -> 2
+                    else -> throw IllegalStateException(pattern)
+                }
+            return match.groups[index]?.value
+        }
+
+        val htmlMediaRegexps =
+            listOf(
+                // src element quoted case (img/audio)
+                Regex("(?i)(<(?:img|audio)\\b[^>]* src=(['\"])([^>]+?)\\2[^>]*>)"), // Group 3 = fname
+                // unquoted src (img/audio)
+                Regex("(?i)(<(?:img|audio)\\b[^>]* src=(?!['\"])([^ >]+)[^>]*?>)"), // Group 2 = fname
+                // quoted data attribute (object)
+                Regex("(?i)(<object\\b[^>]* data=(['\"])([^>]+?)\\2[^>]*>)"), // Group 3 = fname
+                // unquoted data attribute (object)
+                Regex("(?i)(<object\\b[^>]* data=(?!['\"])([^ >]+)[^>]*?>)"), // Group 2 = fname
+            )
+    }
 }