From 7e2ca3286080990181a85064ce9f6b9914afc3c4 Mon Sep 17 00:00:00 2001
From: Dave Lawrence <davmlaw@gmail.com>
Date: Mon, 8 Jul 2024 13:14:28 +0930
Subject: [PATCH] issue #1090 VEP minor release changes (git hash) cause
 VEPVersionMismatchError

---
 .../0108_one_off_vep_version_major_only.py    | 26 +++++++++++++++++++
 annotation/vep_annotation.py                  | 12 +++++++--
 2 files changed, 36 insertions(+), 2 deletions(-)
 create mode 100644 annotation/migrations/0108_one_off_vep_version_major_only.py

diff --git a/annotation/migrations/0108_one_off_vep_version_major_only.py b/annotation/migrations/0108_one_off_vep_version_major_only.py
new file mode 100644
index 000000000..51d15ce79
--- /dev/null
+++ b/annotation/migrations/0108_one_off_vep_version_major_only.py
@@ -0,0 +1,26 @@
+# Generated by Django 4.2.10 on 2024-07-08 03:35
+
+from django.db import migrations
+
+
+def _major_version(version_str) -> str:
+    return version_str.split(".")[0]
+
+
+def _one_off_vep_version_major_only(apps, schema_editor):
+    VariantAnnotationVersion = apps.get_model('annotation', 'VariantAnnotationVersion')
+
+    for vav in VariantAnnotationVersion.objects.all():
+        vav.ensembl = _major_version(vav.ensembl)
+        vav.ensembl_io = _major_version(vav.ensembl_io)
+        vav.save()
+
+
+class Migration(migrations.Migration):
+    dependencies = [
+        ('annotation', '0107_one_off_change_hgvs_g_too_long_message'),
+    ]
+
+    operations = [
+        migrations.RunPython(_one_off_vep_version_major_only)
+    ]
diff --git a/annotation/vep_annotation.py b/annotation/vep_annotation.py
index dd53e43aa..cf01da529 100644
--- a/annotation/vep_annotation.py
+++ b/annotation/vep_annotation.py
@@ -236,14 +236,22 @@ def get_vep_version(genome_build: GenomeBuild, annotation_consortium):
 
 
 def vep_dict_to_variant_annotation_version_kwargs(vep_config, vep_version_dict: dict) -> dict:
-    def vep_int_version(vep_string_version):
+    def _vep_int_version(vep_string_version):
         m = re.match(r"v(\d+)", vep_string_version)
         return int(m.group(1))
 
+    # We strip off the hash ie '110.73b02d8' -> '110' so that we can re-use annotation
+    # if VEP do a minor bugfix change to their releases etc
+    def _major_version(version_str) -> str:
+        return version_str.split(".")[0]
+
+
     FIELD_CONVERSION = {
-        "vep": vep_int_version,
+        "vep": _vep_int_version,
         "cosmic": int,
         "dbsnp": int,
+        "ensembl": _major_version,
+        "ensembl_io": _major_version,
     }
 
     FIELD_LOOKUP = {