Merge dev changes

.github/workflows/ci.yml

@@ -8,7 +8,7 @@ jobs:
     strategy:
       fail-fast: true
       matrix:
-        python-version: ["3.9", "3.13"]
+        python-version: ["3.11", "3.14"]
         os: [ubuntu-22.04] #, macos-latest, windows-latest]
     runs-on: ${{ matrix.os }}
     steps:
@@ -34,7 +34,7 @@ jobs:
       - uses: actions/checkout@v5
       - uses: actions/setup-python@v5
         with:
-          python-version: "3.13"
+          python-version: "3.14"
       - name: Install dependencies
         run: |
           python -m pip install --upgrade pip
@@ -47,7 +47,7 @@ jobs:
       - name: Run mypy
         run: poetry run mypy sentieon_cli --exclude scripts
       - name: Run flake8
-        run: poetry run flake8 . --extend-ignore E231,E221, --exclude .github/scripts/license_message.py,sentieon_cli/scripts/gvcf_combine.py,sentieon_cli/scripts/vcf_mod.py,sentieon_cli/scripts/hybrid_anno.py,sentieon_cli/scripts/hybrid_select.py,sentieon_cli/archive.py,sentieon_cli/scripts/trimalt.py,playground,tests # false+ from python 3.12
+        run: poetry run flake8 .
       - name: Run the automated tests
         run: poetry run pytest -v
       - name: Run doct tests

.github/workflows/main.yml

@@ -14,8 +14,8 @@ jobs:
       fail-fast: true
       max-parallel: 1
       matrix:
-        python-version: ["3.9", "3.13"]
-        sentieon-version: ["202503.01"]
+        python-version: ["3.11", "3.14"]
+        sentieon-version: ["202503.02"]
         os: [ubuntu-22.04] #, macos-latest, windows-latest]
     runs-on: ${{ matrix.os }}
     steps:
@@ -97,8 +97,8 @@ jobs:
       fail-fast: true
       max-parallel: 1
       matrix:
-        python-version: ["3.9", "3.13"]
-        sentieon-version: ["202503.01"]
+        python-version: ["3.11", "3.14"]
+        sentieon-version: ["202503.02"]
         os: [ubuntu-22.04] #, macos-latest, windows-latest]
     runs-on: ${{ matrix.os }}
     steps:
@@ -219,8 +219,8 @@ jobs:
       fail-fast: true
       max-parallel: 1
       matrix:
-        python-version: ["3.9", "3.13"]
-        sentieon-version: ["202503.01"]
+        python-version: ["3.11", "3.14"]
+        sentieon-version: ["202503.02"]
         os: [ubuntu-22.04] #, macos-latest, windows-latest]
     runs-on: ${{ matrix.os }}
     steps:

.github/workflows/release.yml

@@ -13,7 +13,7 @@ jobs:
       - uses: actions/checkout@v3
       - uses: actions/setup-python@v4
         with:
-          python-version: "3.13"
+          python-version: "3.14"
       - name: Install dependencies
         run: |
           python -m pip install --upgrade pip
@@ -29,4 +29,4 @@ jobs:
           poetry build -f sdist
           gh release upload ${{github.event.release.tag_name}} dist/*.tar.gz
         env:
-          GITHUB_TOKEN: ${{ github.TOKEN }}
+          GITHUB_TOKEN: ${{ github.TOKEN }}

README.md

@@ -8,8 +8,8 @@ A command-line interface for the Sentieon software
 
 Download the latest tar.gz file from the GitHub release page, https://github.com/sentieon/sentieon-cli/releases/ and install the package with pip:
 ```sh
-curl -LO https://github.com/Sentieon/sentieon-cli/releases/download/v1.5.2/sentieon_cli-1.5.2.tar.gz
-pip install sentieon_cli-1.5.2.tar.gz
+curl -LO https://github.com/Sentieon/sentieon-cli/releases/download/v1.6.0/sentieon_cli-1.6.0.tar.gz
+pip install sentieon_cli-1.6.0.tar.gz
 ```
 
 ## Installation with Poetry

pyproject.toml

@@ -1,14 +1,14 @@
 
 [project]
 name = "sentieon_cli"
-version = "1.5.2"
+version = "1.6.0"
 description = "entry point for sentieon command-line tools"
 authors = [
     {name = "Don Freed", email = "don.freed@sentieon.com"},
     {name = "Brent", email = "bpederse@gmail.com"},
 ]
 readme = "README.md"
-requires-python = ">= 3.8"
+requires-python = ">= 3.11"
 
 classifiers = [
     "Development Status :: 5 - Production/Stable",
@@ -18,9 +18,10 @@ classifiers = [
 include = ["LICENSE"]
 
 dependencies = [
-    "colorlog (>=6.7.0,<7.0.0)",
-    "importlib-resources (>=5.0,<6.0)",
-    "packaging (>=23.2,<26.0)",
+    "packaging (>=23.2)",
+    "mappy (>=2.30,<3.0)",
+    "scipy (>=1.17.1,<2.0.0)",
+    "levenshtein (>=0.27.3,<0.28.0)",
 ]
 
 [project.scripts]
@@ -32,17 +33,41 @@ ignore_missing_imports = true
 [tool.black]
 line-length = 79
 
+[tool.flake8]
+max-line-length = 79
+extend-ignore = ["E231", "E221"]
+exclude = [
+    ".github/scripts/license_message.py",
+    "sentieon_cli/scripts/combine_cnv.py",
+    "sentieon_cli/scripts/gvcf_combine.py",
+    "sentieon_cli/scripts/vcf_mod.py",
+    "sentieon_cli/scripts/hybrid_anno.py",
+    "sentieon_cli/scripts/hybrid_select.py",
+    "sentieon_cli/scripts/indel2cnv.py",
+    "sentieon_cli/scripts/trimalt.py",
+    "sentieon_cli/archive.py",
+    "playground",
+    "tests",
+]
+
 [tool.poetry.group.dev.dependencies]
 pytest = "^7.0.0"
+pytest-asyncio = ">=0.21.0"
 black = "*"
 flake8 = "*"
+Flake8-pyproject = "*"
 mypy = "*"
 
 [tool.poetry.group.test.dependencies]
 pytest = "^7.0.0"
+pytest-asyncio = ">=0.21.0"
 black = "*"
 mypy = "*"
 
+[tool.pytest.ini_options]
+testpaths = ["tests"]
+asyncio_mode = "strict"
+
 [build-system]
 requires = ["poetry-core"]
 build-backend = "poetry.core.masonry.api"

sentieon_cli/base_pangenome.py

@@ -8,7 +8,7 @@
 import pathlib
 from typing import List, Optional
 
-from importlib_resources import files
+from importlib.resources import files
 
 from . import command_strings as cmds
 from .job import Job

sentieon_cli/command_strings.py

@@ -112,7 +112,8 @@ def cmd_pyexec_vcf_mod_haploid_patch(
     """
     merge dnascope and dnascope-hp variants
     """
-    assert tech.upper() in ("HIFI", "ONT")
+    if tech.upper() not in ("HIFI", "ONT"):
+        raise ValueError(f"Unsupported tech '{tech}', must be 'HIFI' or 'ONT'")
 
     cmd = f"sentieon pyexec {kwargs['vcf_mod_py']} -t {cores} "
     cmd += "haploid_patch "

sentieon_cli/dag.py

@@ -29,10 +29,13 @@ def add_job(self, job: Job, dependencies: Optional[Set[Job]] = None):
         """Add a job to the DAG"""
         if dependencies:
             for dependency in dependencies:
-                assert (
-                    dependency in self.waiting_jobs
-                    or dependency in self.ready_jobs
-                )
+                if (
+                    dependency not in self.waiting_jobs
+                    and dependency not in self.ready_jobs
+                ):
+                    raise ValueError(
+                        f"Dependency '{dependency}' is not in the DAG"
+                    )
 
         if isinstance(dependencies, set) and len(dependencies) > 0:
             self.waiting_jobs[job] = dependencies.copy()

sentieon_cli/dnascope.py

@@ -14,7 +14,7 @@
 
 import packaging.version
 
-from importlib_resources import files
+from importlib.resources import files
 
 from . import command_strings as cmds
 from .dag import DAG
@@ -49,7 +49,6 @@
     total_memory,
 )
 
-
 ALN_MIN_VERSIONS = {
     "sentieon driver": packaging.version.Version("202308"),
     "samtools": packaging.version.Version("1.16"),
@@ -256,7 +255,9 @@ def __init__(self) -> None:
         self.no_split_alignment = False
 
     def validate(self) -> None:
-        assert self.output_vcf
+        if not self.output_vcf:
+            self.logger.error("output_vcf is required")
+            sys.exit(2)
 
         # uniquify pipeline attributes
         self.sr_r1_fastq = self.r1_fastq
@@ -458,8 +459,12 @@ def build_dag(self) -> DAG:
 
     def sr_align_inputs(self) -> Tuple[List[pathlib.Path], Set[Job], Job]:
         """Align input BAM/CRAM/uBAM/uCRAM files with bwa"""
-        assert self.reference
-        assert self.model_bundle
+        if not self.reference:
+            self.logger.error("reference is required")
+            sys.exit(2)
+        if not self.model_bundle:
+            self.logger.error("model_bundle is required")
+            sys.exit(2)
 
         if not self.skip_version_check:
             for cmd, min_version in ALN_MIN_VERSIONS.items():
@@ -526,8 +531,12 @@ def sr_align_fastq(
         self,
     ) -> Tuple[List[pathlib.Path], Set[Job], Optional[Job]]:
         """Align fastq files to the reference genome using bwa"""
-        assert self.reference
-        assert self.model_bundle
+        if not self.reference:
+            self.logger.error("reference is required")
+            sys.exit(2)
+        if not self.model_bundle:
+            self.logger.error("model_bundle is required")
+            sys.exit(2)
 
         res: List[pathlib.Path] = []
         jobs: Set[Job] = set()
@@ -625,7 +634,9 @@ def dedup_and_metrics(
         Optional[Job],
     ]:
         """Perform dedup and metrics collection"""
-        assert self.output_vcf
+        if not self.output_vcf:
+            self.logger.error("output_vcf is required")
+            sys.exit(2)
         suffix = "bam" if self.bam_format else "cram"
 
         # Create the metrics directory
@@ -790,8 +801,12 @@ def sr_call_variants(
         deduped: List[pathlib.Path],
     ) -> Tuple[Job, Job, Job, Optional[Job], Optional[Job], Optional[Job]]:
         """Call SNVs, indels, and SVs using DNAscope"""
-        assert self.model_bundle
-        assert self.output_vcf
+        if not self.model_bundle:
+            self.logger.error("model_bundle is required")
+            sys.exit(2)
+        if not self.output_vcf:
+            self.logger.error("output_vcf is required")
+            sys.exit(2)
 
         if not self.skip_version_check:
             for cmd, min_version in VARIANTS_MIN_VERSIONS.items():

sentieon_cli/dnascope_hybrid.py

@@ -11,7 +11,7 @@
 
 import packaging.version
 
-from importlib_resources import files
+from importlib.resources import files
 
 from .logging import get_logger
 from .archive import ar_load
@@ -47,7 +47,6 @@
 )
 from .transfer import build_transfer_jobs
 
-
 logger = get_logger(__name__)
 
 
@@ -520,9 +519,15 @@ def build_dag(self) -> DAG:
         self.logger.info("Building the DAG")
         dag = DAG()
 
-        assert self.output_vcf
-        assert self.reference
-        assert self.model_bundle
+        if not self.output_vcf:
+            self.logger.error("output_vcf is required")
+            sys.exit(2)
+        if not self.reference:
+            self.logger.error("reference is required")
+            sys.exit(2)
+        if not self.model_bundle:
+            self.logger.error("model_bundle is required")
+            sys.exit(2)
 
         rg_info = RgInfo(
             self.lr_aln_readgroups,
@@ -712,9 +717,15 @@ def call_variants(
         """
         Call SNVs and indels using the DNAscope hybrid pipeline
         """
-        assert self.output_vcf
-        assert self.reference
-        assert self.model_bundle
+        if not self.output_vcf:
+            self.logger.error("output_vcf is required")
+            sys.exit(2)
+        if not self.reference:
+            self.logger.error("reference is required")
+            sys.exit(2)
+        if not self.model_bundle:
+            self.logger.error("model_bundle is required")
+            sys.exit(2)
 
         ref_fai = pathlib.Path(str(self.reference) + ".fai")