Keep full comp-het where only half is new

.bumpversion.cfg

@@ -1,5 +1,5 @@
 [bumpversion]
-current_version = 3.0.1
+current_version = 3.0.2
 commit = True
 tag = False
 

.github/workflows/clinvar_runner.yaml

@@ -32,5 +32,5 @@ jobs:
           curl --fail --silent --show-error -X POST \
               -H "Authorization: Bearer $TOKEN" \
               -H "Content-Type:application/json" \
-              -d '{"output": "generate_clinvar_${{ steps.date.outputs.date }}", "dataset": "talos", "accessLevel": "full", "repo": "automated-interpretation-pipeline", "commit": "${{ github.sha }}", "cwd": "reanalysis", "script": ["./clinvar_runner.py"], "description": "Generate Latest Clinvar Summaries", "image": "australia-southeast1-docker.pkg.dev/cpg-common/images/cpg_aip:3.0.1", "config": {"workflow": {"sequencing_type": "genome"}, "cohorts": {"talos": {"clinvar_filter": ["victorian clinical genetics services, murdoch childrens research institute"]}}}, "wait": false}' \
+              -d '{"output": "generate_clinvar_${{ steps.date.outputs.date }}", "dataset": "talos", "accessLevel": "full", "repo": "automated-interpretation-pipeline", "commit": "${{ github.sha }}", "cwd": "reanalysis", "script": ["./clinvar_runner.py"], "description": "Generate Latest Clinvar Summaries", "image": "australia-southeast1-docker.pkg.dev/cpg-common/images/cpg_aip:3.0.2", "config": {"workflow": {"sequencing_type": "genome"}, "cohorts": {"talos": {"clinvar_filter": ["victorian clinical genetics services, murdoch childrens research institute"]}}}, "wait": false}' \
               https://server-a2pko7ameq-ts.a.run.app

.github/workflows/docker.yaml

@@ -14,7 +14,7 @@ permissions:
   contents: read
 
 env:
-  VERSION: 3.0.1
+  VERSION: 3.0.2
 
 jobs:
   docker:

reanalysis/html_builder.py

@@ -4,6 +4,7 @@
 
 from argparse import ArgumentParser
 from dataclasses import dataclass
+from itertools import chain
 from pathlib import Path
 from typing import Any
 
@@ -469,6 +470,9 @@ def check_date_filter(
     Check if there's a date filter in the config
     if there is, load the results JSON and filter out variants
 
+    Extra consideration - if one part of a comp-het variant pair is new,
+    retain both sides in the report
+
     Args:
         results (str): path to the results file
         filter_date (str | None): path to the results file
@@ -483,11 +487,23 @@ def check_date_filter(
 
     # Filter out variants based on date
     for content in results_dict.results.values():
-        # keep only this run's new variants
-        content.variants = [
+        # keep only this run's new variants, or partners thereof
+        vars_to_keep = [
             variant for variant in content.variants if variant.first_seen == filter_date
         ]
 
+        pairs_to_keep = set(
+            chain.from_iterable(var.support_vars for var in vars_to_keep)
+        )
+        content.variants = [
+            variant
+            for variant in content.variants
+            if (
+                variant.first_seen == filter_date
+                or variant.var_data.coordinates.string_format in pairs_to_keep
+            )
+        ]
+
     # pop off all the samples with no variants
     for sample_id in list(results_dict.results.keys()):
         if not results_dict.results[sample_id].variants:

reanalysis/reanalysis_global.toml

@@ -74,7 +74,7 @@ default_memory = 'highmem'
 
 [images]
 gatk = 'australia-southeast1-docker.pkg.dev/cpg-common/images/gatk:4.2.6.1'
-aip = 'australia-southeast1-docker.pkg.dev/cpg-common/images/cpg_aip:3.0.1'
+aip = 'australia-southeast1-docker.pkg.dev/cpg-common/images/cpg_aip:3.0.2'
 vep_110 = "australia-southeast1-docker.pkg.dev/cpg-common/images/vep_110:release_110.1"
 cpg_workflows = "australia-southeast1-docker.pkg.dev/cpg-common/images/cpg_workflows:latest"
 

reanalysis/version.py

@@ -3,4 +3,4 @@
 """
 
 # Do not edit this file manually
-__version__ = '3.0.1'
+__version__ = '3.0.2'

setup.py

@@ -31,7 +31,7 @@ def read_reqs(filename: str) -> list[str]:
     name='automated-interpretation-pipeline',
     description='CPG Variant Prioritisation',
     long_description=readme,
-    version='3.0.1',
+    version='3.0.2',
     author='Matthew Welland, CPG',
     author_email=(
         '[email protected], '

test/test_html_builder.py

@@ -60,6 +60,59 @@ def test_check_date_filter(tmp_path):
     )
 
 
+def test_check_date_filter_pair(tmp_path):
+    """
+    test for the case when a comp-het pair is both recovered by one meeting the
+    specified date and the other not
+    """
+    # needs to be a valid ResultData
+    result_dict = ResultData(
+        **{
+            'metadata': {'categories': {'1': '1'}, 'run_datetime': '2021-01-01'},
+            'results': {
+                'sample1': {
+                    'metadata': {'ext_id': 'sample1', 'family_id': 'sample1'},
+                    'variants': [
+                        {
+                            'sample': 'sample1',
+                            'first_seen': '2021-01-01',
+                            'var_data': VAR_1,
+                            'support_vars': {VAR_2.coordinates.string_format},
+                        },
+                        {
+                            'sample': 'sample1',
+                            'first_seen': '2023-03-03',
+                            'var_data': VAR_2,
+                            'support_vars': {VAR_1.coordinates.string_format},
+                        },
+                    ],
+                },
+                'sample2': {
+                    'metadata': {'ext_id': 'sample1', 'family_id': 'sample1'},
+                    'variants': [
+                        {
+                            'sample': 'sample1',
+                            'first_seen': '2022-02-02',
+                            'var_data': VAR_2,
+                        }
+                    ],
+                },
+            },
+        }
+    )
+    result_path = str(tmp_path / 'results.json')
+    with open(result_path, 'w', encoding='utf-8') as handle:
+        handle.write(ResultData.model_validate(result_dict).model_dump_json(indent=4))
+
+    filtered_results = check_date_filter(result_path, '2021-01-01')
+    assert 'sample1' in filtered_results.results
+    assert 'sample2' not in filtered_results.results
+    assert len(filtered_results.results['sample1'].variants) == 2
+    assert sorted(
+        var.first_seen for var in filtered_results.results['sample1'].variants
+    ) == ['2021-01-01', '2023-03-03']
+
+
 def test_check_date_filter_none_pass(tmp_path):
     """