From 06aea81fec7383c43bccf57b2fd8fa05618521d9 Mon Sep 17 00:00:00 2001
From: John Marshall <john.marshall@populationgenomics.org.au>
Date: Mon, 11 Dec 2023 14:03:20 +1300
Subject: [PATCH] Ensure populated CramPaths have both path and index_path
 (#542)

The resulting self.index_path will only be set if the
CramPath() constructor also specifies an index_path.

Use exists() to benefit from whole-directory existence caching.
---
 .bumpversion.cfg              |  2 +-
 .github/workflows/docker.yaml |  2 +-
 cpg_workflows/inputs.py       | 14 +++++++++-----
 setup.py                      |  2 +-
 4 files changed, 12 insertions(+), 8 deletions(-)

diff --git a/.bumpversion.cfg b/.bumpversion.cfg
index 45d9918e1..d908532e1 100644
--- a/.bumpversion.cfg
+++ b/.bumpversion.cfg
@@ -1,5 +1,5 @@
 [bumpversion]
-current_version = 1.18.12
+current_version = 1.18.13
 commit = True
 tag = False
 
diff --git a/.github/workflows/docker.yaml b/.github/workflows/docker.yaml
index 72d96ebca..2e0c4a6b4 100644
--- a/.github/workflows/docker.yaml
+++ b/.github/workflows/docker.yaml
@@ -15,7 +15,7 @@ permissions:
   contents: read
   
 env:
-  VERSION: 1.18.12
+  VERSION: 1.18.13
 
 jobs:
   docker:
diff --git a/cpg_workflows/inputs.py b/cpg_workflows/inputs.py
index 084a9252c..fa1e04ec9 100644
--- a/cpg_workflows/inputs.py
+++ b/cpg_workflows/inputs.py
@@ -10,6 +10,7 @@
 
 from .metamist import AnalysisType, Assay, MetamistError, get_metamist, parse_reads
 from .targets import Cohort, PedigreeInfo, SequencingGroup, Sex
+from .utils import exists
 
 _cohort: Cohort | None = None
 
@@ -140,25 +141,28 @@ def _populate_analysis(cohort: Cohort) -> None:
         for sequencing_group in dataset.get_sequencing_groups():
             if (analysis := gvcf_by_sgid.get(sequencing_group.id)) and analysis.output:
                 # assert file exists
-                assert analysis.output.exists(), (
+                assert exists(analysis.output), (
                     'gvcf file does not exist',
                     analysis.output,
                 )
                 sequencing_group.gvcf = GvcfPath(path=analysis.output)
-            elif sequencing_group.make_gvcf_path().exists():
+            elif exists(sequencing_group.make_gvcf_path()):
                 logging.warning(
                     f'We found a gvcf file in the expected location {sequencing_group.make_gvcf_path()},'
                     'but it is not logged in metamist. Skipping. You may want to update the metadata and try again. '
                 )
             if (analysis := cram_by_sgid.get(sequencing_group.id)) and analysis.output:
                 # assert file exists
-                assert analysis.output.exists(), (
+                assert exists(analysis.output), (
                     'cram file does not exist',
                     analysis.output,
                 )
-                sequencing_group.cram = CramPath(path=analysis.output)
+                crai_path = analysis.output.with_suffix('.cram.crai')
+                if not exists(crai_path):
+                    crai_path = None
+                sequencing_group.cram = CramPath(analysis.output, crai_path)
 
-            elif sequencing_group.make_cram_path().exists():
+            elif exists(sequencing_group.make_cram_path()):
                 logging.warning(
                     f'We found a cram file in the expected location {sequencing_group.make_cram_path()},'
                     'but it is not logged in metamist. Skipping. You may want to update the metadata and try again. '
diff --git a/setup.py b/setup.py
index 40e662c7a..acedbc1ee 100644
--- a/setup.py
+++ b/setup.py
@@ -5,7 +5,7 @@
 setup(
     name='cpg-workflows',
     # This tag is automatically updated by bumpversion
-    version='1.18.12',
+    version='1.18.13',
     description='CPG workflows for Hail Batch',
     long_description=open('README.md').read(),
     long_description_content_type='text/markdown',