Snakemake version 8+ support

.github/workflows/cibuild.yml

@@ -8,7 +8,7 @@ jobs:
     strategy:
       max-parallel: 4
       matrix:
-        python-version: ["3.9", "3.10", "3.11", "3.12"]
+        python-version: ["3.11", "3.12", "3.13"]
     steps:
     - uses: actions/checkout@v1
     - name: Set up Python ${{ matrix.python-version }}

CHANGELOG.md

@@ -10,6 +10,7 @@ This project adheres to [Semantic Versioning](http://semver.org/).
 - Colons are now used instead of commas to separate SNP alleles in microhap alleles (#192).
 - Implemented Python 3.12 support by integrating happer package and increasing minimum version of MicroHapDB dependency (#193).
 - Updated working directory organization to provide additional structure (#194).
+- Snakemake is now invoked using its new Python API, requiring a minimum version of 8.0 and Python 3.11+ (#195).
 
 ### Fixed
 - Bug with handling marker vs. locus identifiers when running `mhpl8r seq` (#190).

microhapulator/cli/pipe.py

@@ -18,7 +18,9 @@
 from os import cpu_count, symlink
 from pathlib import Path
 from shutil import copy
-from snakemake import snakemake
+from snakemake.api import SnakemakeApi as smk_api
+from snakemake.exceptions import WorkflowError
+from snakemake.settings import types as smk_types
 import sys
 
 
@@ -29,10 +31,10 @@ def main(args):
     config = dict(
         samples=samples,
         readfiles=workingfiles,
-        mhrefr=Path(args.markerrefr).resolve(),
-        mhdefn=Path(args.markerdefn).resolve(),
-        hg38path=Path(args.hg38).resolve(),
-        hg38index=Path(args.hg38idx).resolve(),
+        mhrefr=str(Path(args.markerrefr).resolve()),
+        mhdefn=str(Path(args.markerdefn).resolve()),
+        hg38path=str(Path(args.hg38).resolve()),
+        hg38index=str(Path(args.hg38idx).resolve()),
         thresh_static=args.static,
         thresh_dynamic=args.dynamic,
         thresh_file=args.config,
@@ -43,19 +45,28 @@ def main(args):
         thresh_gap_alert=args.gap_alert,
         hspace=args.hspace,
     )
-    snakefile = files("microhapulator") / "workflows" / "analysis.smk"
-    success = snakemake(
-        snakefile,
-        cores=args.threads,
-        printshellcmds=True,
-        dryrun=args.dryrun,
-        config=config,
-        workdir=args.workdir,
-    )
-    if not success:
+    try:
+        run_snakemake(config, args.workdir, cores=args.threads, dryrun=args.dryrun)
+        return 0
+    except WorkflowError:
         return 1
 
 
+def run_snakemake(config, workdir, cores=1, dryrun=False):
+    outcfg = smk_types.OutputSettings(printshellcmds=True)
+    with smk_api(outcfg) as smk:
+        workflow = smk.workflow(
+            config_settings=smk_types.ConfigSettings(config=config),
+            storage_settings=smk_types.StorageSettings(),
+            resource_settings=smk_types.ResourceSettings(cores=cores),
+            snakefile=files("microhapulator") / "workflows" / "analysis.smk",
+            workdir=Path(workdir),
+        )
+        dag = workflow.dag(smk_types.DAGSettings(targets=["report"]))
+        mode = "dryrun" if dryrun else "local"
+        dag.execute_workflow(executor=mode)
+
+
 def validate_panel_config(markerseqs, markerdefn):
     print("[MicroHapulator] validating panel configuration", file=sys.stderr)
     index = MicrohapIndex.from_files(markerdefn, markerseqs)

microhapulator/workflows/analysis.smk

@@ -211,7 +211,9 @@ rule apply_filters:
     params:
         static=f"--static {config['thresh_static']}",
         dynamic=f"--dynamic {config['thresh_dynamic']}",
-        threshfile="" if config["thresh_file"] == "" else f"--config {config['thresh_file']}",
+        threshfile=(
+            "" if config["thresh_file"] in ("", None) else f"--config {config['thresh_file']}"
+        ),
     shell:
         "mhpl8r filter {input} --out {output} {params.static} {params.dynamic} {params.threshfile}"
 
@@ -270,7 +272,7 @@ rule plot_haplotype_calls:
         result = TypingResult(fromfile=input.result)
         mhapi.plot_haplotype_calls(
             result,
-            "analysis/{}/03typing/callplots".format(wildcards.sample),
+            f"analysis/{wildcards.sample}/03typing/callplots",
             sample=wildcards.sample,
         )
 

microhapulator/workflows/preproc-paired.smk

@@ -40,8 +40,7 @@ rule fastqc:
         outfiles = sorted(Path(params.outdir).glob("*.html"))
         for end, outfile in enumerate(outfiles, 1):
             outfile = Path(outfile)
-            # Snakemake f-strings break with Python 3.12: https://github.com/snakemake/snakemake/issues/2648
-            linkfile = "{}/R{}-fastqc.html".format(params.outdir, end)
+            linkfile = f"{params.outdir}/R{end}-fastqc.html"
             symlink(outfile.name, linkfile)
 
 

setup.py

@@ -55,7 +55,7 @@
         "pulp==2.3.1",
         "scipy>=1.7",
         "seaborn>=0.13.2",
-        "snakemake>=7.15.2,<8.0",
+        "snakemake>=8",
         "termgraph>=0.5",
         "tqdm>=4.0",
     ],