From 87c78db6ce48a548de39d72ce263450aa133ce93 Mon Sep 17 00:00:00 2001
From: Terry <tmc46@cornell.edu>
Date: Tue, 10 Dec 2024 16:40:32 -0500
Subject: [PATCH 1/2] Added method HaplotypeGraph.hapidToSeqLength()

---
 .../maizegenetics/phgv2/api/HaplotypeGraph.kt | 15 +++++++++++
 .../cli/CreateFastaFromHvcfRangeFastaTest.kt  | 27 +++++++++++++++++++
 2 files changed, 42 insertions(+)

diff --git a/src/main/kotlin/net/maizegenetics/phgv2/api/HaplotypeGraph.kt b/src/main/kotlin/net/maizegenetics/phgv2/api/HaplotypeGraph.kt
index f803fb62..db3cabea 100644
--- a/src/main/kotlin/net/maizegenetics/phgv2/api/HaplotypeGraph.kt
+++ b/src/main/kotlin/net/maizegenetics/phgv2/api/HaplotypeGraph.kt
@@ -562,4 +562,19 @@ class HaplotypeGraph(hvcfFiles: List<String>) {
         return altHeaderMap[hapid]?.refChecksum ?: ""
     }
 
+    /**
+     * Returns a map of haplotype id to sequence length.
+     * This is calculated from the AltHeaderMetaData regions.
+     */
+    fun hapidToSeqLength(): Map<String, Int> {
+
+        return altHeaderMap.map { (hapid, altHeader) ->
+            val seqLength = altHeader.regions.sumOf { (start, end) ->
+                end.position - start.position + 1
+            }
+            Pair(hapid, seqLength)
+        }.toMap()
+
+    }
+
 }
diff --git a/src/test/kotlin/net/maizegenetics/phgv2/cli/CreateFastaFromHvcfRangeFastaTest.kt b/src/test/kotlin/net/maizegenetics/phgv2/cli/CreateFastaFromHvcfRangeFastaTest.kt
index 9ac05a3e..e31f1b32 100644
--- a/src/test/kotlin/net/maizegenetics/phgv2/cli/CreateFastaFromHvcfRangeFastaTest.kt
+++ b/src/test/kotlin/net/maizegenetics/phgv2/cli/CreateFastaFromHvcfRangeFastaTest.kt
@@ -1,8 +1,10 @@
 package net.maizegenetics.phgv2.cli
 
 import com.github.ajalt.clikt.testing.test
+import net.maizegenetics.phgv2.api.HaplotypeGraph
 import net.maizegenetics.phgv2.brapi.createSmallSeqTiledb
 import net.maizegenetics.phgv2.utils.getChecksum
+import net.maizegenetics.phgv2.utils.seqFromAGC
 import org.apache.logging.log4j.LogManager
 import org.junit.jupiter.api.BeforeAll
 import org.junit.jupiter.api.Test
@@ -27,6 +29,8 @@ class CreateFastaFromHvcfRangeFastaTest {
 
         private val dbPath = "${exportHvcfDir}/tiledb_ref_range_fasta"
 
+        val HVCF_PATTERN = Regex("""(\.hvcf|\.h\.vcf|\.hvcf\.gz|\.h\.vcf\.gz)$""")
+
         @BeforeAll
         @JvmStatic
         fun setup() {
@@ -77,4 +81,27 @@ class CreateFastaFromHvcfRangeFastaTest {
 
     }
 
+    @Test
+    fun testHaplotypeGraphHapidToSeqLength() {
+
+        val inputFiles =
+            File(multiInputDir)
+                .walk()
+                .filter {  HVCF_PATTERN.containsMatchIn(it.name) }
+                .map { it.absolutePath }
+                .toList()
+
+        require(inputFiles.isNotEmpty()) { "At least one HVCF file should be specified." }
+
+        val graph = HaplotypeGraph(inputFiles)
+
+        graph.altHeaders()
+            .map { it.key }
+            .forEach { hapid ->
+                val temp = seqFromAGC(dbPath, graph, hapid, range)
+                seq = temp.first
+            }
+
+    }
+
 }
\ No newline at end of file

From 14d235b8c808db0f64b763df29e657ad0119d014 Mon Sep 17 00:00:00 2001
From: Terry <tmc46@cornell.edu>
Date: Wed, 11 Dec 2024 10:24:25 -0500
Subject: [PATCH 2/2] Added unit test for HaplotypeGraph.hapidToSeqLength()

---
 .../cli/CreateFastaFromHvcfRangeFastaTest.kt  | 28 +++++++++++++++----
 1 file changed, 22 insertions(+), 6 deletions(-)

diff --git a/src/test/kotlin/net/maizegenetics/phgv2/cli/CreateFastaFromHvcfRangeFastaTest.kt b/src/test/kotlin/net/maizegenetics/phgv2/cli/CreateFastaFromHvcfRangeFastaTest.kt
index e31f1b32..3252ed14 100644
--- a/src/test/kotlin/net/maizegenetics/phgv2/cli/CreateFastaFromHvcfRangeFastaTest.kt
+++ b/src/test/kotlin/net/maizegenetics/phgv2/cli/CreateFastaFromHvcfRangeFastaTest.kt
@@ -3,6 +3,7 @@ package net.maizegenetics.phgv2.cli
 import com.github.ajalt.clikt.testing.test
 import net.maizegenetics.phgv2.api.HaplotypeGraph
 import net.maizegenetics.phgv2.brapi.createSmallSeqTiledb
+import net.maizegenetics.phgv2.utils.Position
 import net.maizegenetics.phgv2.utils.getChecksum
 import net.maizegenetics.phgv2.utils.seqFromAGC
 import org.apache.logging.log4j.LogManager
@@ -87,7 +88,7 @@ class CreateFastaFromHvcfRangeFastaTest {
         val inputFiles =
             File(multiInputDir)
                 .walk()
-                .filter {  HVCF_PATTERN.containsMatchIn(it.name) }
+                .filter { HVCF_PATTERN.containsMatchIn(it.name) }
                 .map { it.absolutePath }
                 .toList()
 
@@ -95,11 +96,26 @@ class CreateFastaFromHvcfRangeFastaTest {
 
         val graph = HaplotypeGraph(inputFiles)
 
-        graph.altHeaders()
-            .map { it.key }
-            .forEach { hapid ->
-                val temp = seqFromAGC(dbPath, graph, hapid, range)
-                seq = temp.first
+        val hapidToSeqLength = graph.hapidToSeqLength()
+
+        graph.ranges()
+            .forEach { range ->
+                graph.hapIdToSampleGametes(range).keys
+                    .forEach { hapid ->
+                        seqFromAGC(
+                            dbPath,
+                            graph,
+                            hapid,
+                            Pair(Position(range.contig, range.start), Position(range.contig, range.end))
+                        )
+                            .let { (seq, _) ->
+                                assertEquals(
+                                    hapidToSeqLength.getValue(hapid),
+                                    seq.length,
+                                    "hapid: $hapid seq length: ${seq.length} != ${hapidToSeqLength.getValue(hapid)}"
+                                )
+                            }
+                    }
             }
 
     }