fix(sefariasqlite): number and index Midrash Rabbah paragraphs

generator/sefariasqlite/src/jvmMain/kotlin/io/github/kdroidfilter/seforimlibrary/sefariasqlite/SefariaBookPayloadReader.kt

@@ -542,9 +542,13 @@ internal class SefariaBookPayloadReader(
 
             val sectionIndex = sectionNames.size - depth
             val isReferenceable = referenceableSections.getOrNull(sectionIndex) ?: true
+            val isNumberedParagraph = depth == 1 && hasNumberedParagraphs(bookEnTitle)
             val nextLinePrefix = if (
-                depth == 1 && isReferenceable && currentAddressType != "Integer" && nonEmptyCount > 1 &&
-                !hasSelfLabeledSegments(bookEnTitle)
+                isNumberedParagraph ||
+                (
+                    depth == 1 && isReferenceable && currentAddressType != "Integer" && nonEmptyCount > 1 &&
+                        !hasSelfLabeledSegments(bookEnTitle)
+                    )
             ) {
                 "($letter) "
             } else {
@@ -586,6 +590,7 @@ internal class SefariaBookPayloadReader(
                 append(", ")
             }
             val nextRefIndexOffset = childRefOffsets?.getOrNull(idx) ?: 0
+            val lineCountBefore = output.size
 
             recursiveSections(
                 sectionNames = sectionNames,
@@ -604,6 +609,12 @@ internal class SefariaBookPayloadReader(
                 referenceableSections = referenceableSections,
                 refIndexOffset = nextRefIndexOffset
             )
+
+            // The paragraph's TOC entry points at the paragraph line itself, so the
+            // text gets no extra heading line.
+            if (isNumberedParagraph && output.size > lineCountBefore) {
+                headings += Heading(title = "אות $letter", level = level, lineIndex = lineCountBefore)
+            }
         }
     }
 

generator/sefariasqlite/src/jvmMain/kotlin/io/github/kdroidfilter/seforimlibrary/sefariasqlite/SefariaImportText.kt

@@ -35,6 +35,24 @@ private val SELF_LABELED_SEGMENT_BOOKS = setOf("Mishnah Berurah")
 
 internal fun hasSelfLabeledSegments(bookEnTitle: String): Boolean = bookEnTitle in SELF_LABELED_SEGMENT_BOOKS
 
+// Books whose unnamed ("Integer") segments are the printed numbered paragraphs (אותיות)
+// readers navigate by: each segment gets its "(letter) " label and its own TOC entry
+// (kdroidFilter/Zayit#509).
+private val NUMBERED_PARAGRAPH_BOOKS = setOf(
+    "Bereshit Rabbah",
+    "Shemot Rabbah",
+    "Vayikra Rabbah",
+    "Bamidbar Rabbah",
+    "Devarim Rabbah",
+    "Ruth Rabbah",
+    "Esther Rabbah",
+    "Eikhah Rabbah",
+    "Kohelet Rabbah",
+    "Shir HaShirim Rabbah",
+)
+
+internal fun hasNumberedParagraphs(bookEnTitle: String): Boolean = bookEnTitle in NUMBERED_PARAGRAPH_BOOKS
+
 // Some simanim of the Mishnah Berurah (494-529) mark the seif katan as "{א}"
 // instead of the printed "(א)".
 private val BRACED_SEGMENT_MARKER_REGEX = Regex("""^\{([א-ת]{1,4})\}""")

generator/sefariasqlite/src/jvmTest/kotlin/io/github/kdroidfilter/seforimlibrary/sefariasqlite/SefariaNumberedParagraphsTest.kt

@@ -0,0 +1,96 @@
+package io.github.kdroidfilter.seforimlibrary.sefariasqlite
+
+import co.touchlab.kermit.Logger
+import kotlinx.coroutines.runBlocking
+import kotlinx.serialization.json.Json
+import java.nio.file.Files
+import kotlin.test.Test
+import kotlin.test.assertEquals
+
+/**
+ * Regression for kdroidFilter/Zayit#509: the Midrash Rabbah's paragraphs (אותיות)
+ * are unnamed "Integer" segments, so they had neither a number nor a TOC entry.
+ */
+class SefariaNumberedParagraphsTest {
+    @Test
+    fun midrashRabbahParagraphsGetLabelAndTocEntry() {
+        val payload = readBook("Bereshit Rabbah", "בראשית רבה")
+
+        assertEquals(
+            listOf(
+                "<h1>בראשית רבה</h1>",
+                "<h2>פרק א</h2>",
+                "(א) רבי הושעיה רבה פתח",
+                "(ב) רבי יהושע דסכנין",
+                "<h2>פרק ב</h2>",
+                "(א) והארץ היתה תהו",
+            ),
+            payload.lines,
+        )
+        assertEquals(
+            listOf(
+                Heading("בראשית רבה", 0, 0),
+                Heading("פרק א", 1, 1),
+                Heading("אות א", 2, 2),
+                Heading("אות ב", 2, 3),
+                Heading("פרק ב", 1, 4),
+                Heading("אות א", 2, 5),
+            ),
+            payload.headings,
+        )
+        // heRefs (the line id natural keys) are unchanged
+        assertEquals(
+            listOf("בראשית רבה א, א", "בראשית רבה א, ב", "בראשית רבה ב, א"),
+            payload.refEntries.map { it.heRef.replace(Regex(" +"), " ") },
+        )
+    }
+
+    @Test
+    fun otherBooksKeepUnlabeledIntegerSegments() {
+        val payload = readBook("FakeBook", "ספר")
+
+        assertEquals("רבי הושעיה רבה פתח", payload.lines[2])
+        assertEquals(listOf(0, 1, 1), payload.headings.map { it.level })
+    }
+
+    private fun readBook(title: String, heTitle: String) = runBlocking {
+        val tempDir = Files.createTempDirectory("seforim-numbered-paragraphs")
+        val schemaDir = Files.createDirectories(tempDir.resolve("schemas"))
+        val bookDir = Files.createDirectories(tempDir.resolve("json").resolve(title))
+        Files.writeString(schemaDir.resolve("${title.replace(' ', '_')}.json"), schemaJson(title, heTitle))
+        Files.writeString(bookDir.resolve("merged.json"), mergedJson(title, heTitle))
+
+        val reader = SefariaBookPayloadReader(
+            Json { ignoreUnknownKeys = true; coerceInputValues = true },
+            Logger.withTag("SefariaNumberedParagraphsTest"),
+        )
+        val schemaLookup = reader.buildSchemaLookup(schemaDir)
+        reader.readBooksInParallel(tempDir.resolve("json"), schemaDir, schemaLookup).single()
+    }
+
+    private fun schemaJson(title: String, heTitle: String) = """
+        {
+          "title": "$title",
+          "heTitle": "$heTitle",
+          "schema": {
+            "nodeType": "JaggedArrayNode",
+            "depth": 2,
+            "addressTypes": ["Perek", "Integer"],
+            "sectionNames": ["Chapter", "Paragraph"],
+            "title": "$title",
+            "heTitle": "$heTitle"
+          }
+        }
+    """.trimIndent()
+
+    private fun mergedJson(title: String, heTitle: String) = """
+        {
+          "title": "$title",
+          "heTitle": "$heTitle",
+          "text": [
+            ["רבי הושעיה רבה פתח", "רבי יהושע דסכנין"],
+            ["והארץ היתה תהו"]
+          ]
+        }
+    """.trimIndent()
+}