From d0d4025aaa6656226efb13987bd183ba93ac1f8f Mon Sep 17 00:00:00 2001
From: Dave Lawrence <davmlaw@gmail.com>
Date: Tue, 6 Feb 2024 16:56:47 +1030
Subject: [PATCH 1/2] #722 Babelfish VCF bugs

---
 src/hgvs/extras/babelfish.py        |  27 ++++++++++++++++-----------
 tests/data/cache-py3.hdp            | Bin 921577 -> 921714 bytes
 tests/test_hgvs_extras_babelfish.py |  12 +++++++++++-
 3 files changed, 27 insertions(+), 12 deletions(-)

diff --git a/src/hgvs/extras/babelfish.py b/src/hgvs/extras/babelfish.py
index b1a7b49d..77f0a95b 100644
--- a/src/hgvs/extras/babelfish.py
+++ b/src/hgvs/extras/babelfish.py
@@ -74,16 +74,24 @@ def hgvs_to_vcf(self, var_g):
         return chrom, start_i + 1, ref, alt, typ
 
     def vcf_to_g_hgvs(self, chrom, position, ref, alt):
+        # VCF spec https://samtools.github.io/hts-specs/VCFv4.1.pdf
+        # says for REF/ALT "Each base must be one of A,C,G,T,N (case insensitive)"
+        ref = ref.upper()
+        alt = alt.upper()
+
         ac = self.name_to_ac_map[chrom]
 
-        # Strip common prefix
-        if len(alt) > 1 and len(ref) > 1:
-            pfx = os.path.commonprefix([ref, alt])
-            lp = len(pfx)
-            if lp > 0:
-                ref = ref[lp:]
-                alt = alt[lp:]
-                position += lp
+        if ref != alt:
+            # Strip common prefix
+            if len(alt) > 1 and len(ref) > 1:
+                pfx = os.path.commonprefix([ref, alt])
+                lp = len(pfx)
+                if lp > 0:
+                    ref = ref[lp:]
+                    alt = alt[lp:]
+                    position += lp
+            elif alt == ".":
+                alt = ref
 
         if ref == "":  # Insert
             # Insert uses coordinates around the insert point.
@@ -93,9 +101,6 @@ def vcf_to_g_hgvs(self, chrom, position, ref, alt):
             start = position
             end = position + len(ref) - 1
 
-        if alt == ".":
-            alt = ref
-
         var_g = SequenceVariant(
             ac=ac,
             type="g",
diff --git a/tests/data/cache-py3.hdp b/tests/data/cache-py3.hdp
index e7ed304c4ce67c901eece222d8d9475b312ba12f..a7ad77a51c1b083d3f2b7595f4df91984d041335 100644
GIT binary patch
delta 94
zcmaDk-{R8(i-s1)7N!>F7M2#)7Pc1l7LFFq7OocV7M?A<55yU}wm%f-W%@Ba%#BTG
vx_vS$8{2)YUrb(KrU&rx$xW}m#wtDCijPl#?H)+<0Z3Gajmgn*N~s<IDSRK4

delta 50
zcmeygz~be6i-s1)7N!>F7M2#)7Pc1l7LFFq7OocV7M?A<55yT6w?7o;W%|L^!{|Px
GR1W~rY!M0o

diff --git a/tests/test_hgvs_extras_babelfish.py b/tests/test_hgvs_extras_babelfish.py
index 924dd3c2..c0d2d1c5 100644
--- a/tests/test_hgvs_extras_babelfish.py
+++ b/tests/test_hgvs_extras_babelfish.py
@@ -8,7 +8,17 @@
         "NC_000006.12:g.49949407=",
         [],
         ("6", 49949407, "A", ".", "identity"),
-        [("6", 49949407, "A", "A", "identity")],
+        [("6", 49949407, "A", "A", "identity"),
+         # Test case insensitivity
+         ("6", 49949407, "A", "a", "identity"),
+         ("6", 49949407, "a", "A", "identity"),]
+    ),
+    # Test multi-base identity
+    (
+        "NC_000006.12:g.49949407_49949408=",
+        [],
+        ("6", 49949407, "AA", ".", "identity"),
+        [("6", 49949407, "AA", "AA", "identity")],
     ),
     # snv
     (

From 30cd83ff270e1bcb22eb588bb437c8071ab6e4d3 Mon Sep 17 00:00:00 2001
From: Dave Lawrence <davmlaw@gmail.com>
Date: Thu, 21 Mar 2024 14:28:32 +1030
Subject: [PATCH 2/2] Update cache with tests

---
 tests/data/cache-py3.hdp | Bin 923986 -> 924046 bytes
 1 file changed, 0 insertions(+), 0 deletions(-)

diff --git a/tests/data/cache-py3.hdp b/tests/data/cache-py3.hdp
index 75fdcce4e51749a3921e4d14139368f03da73746..54d2f9a632400c42075d85d222126ba6b92ebc72 100644
GIT binary patch
delta 69
zcmcb#$)azwMMDc?3sVbo3rh=Y3tJ0&3r7oQ3s(zw3r`F07QUC_jB~fY66Z@}oUWhD
ZD#mtC>lc&PgXsZ$d@^iIj*e4G^#DaC7ODUM

delta 48
zcmeC%Y;kFmMMDc?3sVbo3rh=Y3tJ0&3r7oQ3s(zw3r`F07QUC_jEUP{iSwl~GL`B9
E0HA~qy8r+H