1 add docker commands

config.yaml

@@ -3,4 +3,5 @@ tool_version: 1.0.0
 variant_analysis: True
 gene_analysis: True
 disease_analysis: False
-tool_specific_configuration_options:
+tool_specific_configuration_options:
+  environment: apptainer # either apptainer or docker

pyproject.toml

@@ -9,6 +9,7 @@ packages = [{include = "pheval_ai_marrvel", from = "src"}]
 [tool.poetry.dependencies]
 python = "^3.10"
 pheval = "^0.3.2"
+docker = "^7.1.0"
 
 [tool.poetry.scripts]
 pheval-ai = "pheval_ai_marrvel.cli:main"

src/pheval_ai_marrvel/constants.py

@@ -0,0 +1,4 @@
+VCF_FILE = "/input/vcf.gz"
+HPO_TXT = "/input/hpo.txt"
+DATA_DEPENDENCIES = "/run/data_dependencies"
+OUTPUT_DIR = "/out"

src/pheval_ai_marrvel/run/create_docker_commands.py

@@ -0,0 +1,152 @@
+from dataclasses import dataclass
+from pathlib import Path
+from typing import List
+
+import docker
+from docker import DockerClient
+from pheval.utils.file_utils import all_files
+from pheval.utils.phenopacket_utils import PhenopacketUtil, phenopacket_reader
+
+from pheval_ai_marrvel.constants import DATA_DEPENDENCIES, HPO_TXT, OUTPUT_DIR, VCF_FILE
+
+
+@dataclass
+class AIMARRVELVolumes:
+    """
+    Volumes to mount to run AI MARRVEL
+
+    Attributes:
+        vcf_path (str): Path to VCF file
+        data_dependencies (str): Path to data dependencies
+        hpo_txt (str): Path to hpo txt file
+        output_dir (str): Path to output directory
+    """
+
+    vcf_path: str
+    data_dependencies: str
+    hpo_txt: str
+    output_dir: str
+
+
+@dataclass
+class SampleData:
+    """
+    Sample data.
+    Attributes:
+        sample_id (str): Sample ID
+        genome_assembly (str): Genome assembly
+        vcf_name (Path): VCF file name
+    """
+
+    sample_id: str
+    genome_assembly: str
+    vcf_name: Path
+
+
+def get_sample_data(phenopacket_path: Path, vcf_dir: Path) -> SampleData:
+    """
+    Get sample data.
+
+    Args:
+        phenopacket_path (Path): Path to phenopacket file
+        vcf_dir (Path): Path to VCF directory
+
+    Returns:
+        SampleData:The sample data
+    """
+    phenopacket_util = PhenopacketUtil(phenopacket_reader(phenopacket_path))
+    vcf_data = phenopacket_util.vcf_file_data(phenopacket_path, vcf_dir)
+    return SampleData(
+        sample_id=phenopacket_util.sample_id(),
+        genome_assembly=vcf_data.file_attributes["genomeAssembly"],
+        vcf_name=vcf_data.uri,
+    )
+
+
+def create_volumes(
+    vcf_path: Path, data_dependencies: Path, hpo_txt: Path, output_dir: Path
+) -> AIMARRVELVolumes:
+    """
+    Create volumes to mount AI MARRVEL.
+    Args:
+        vcf_path (Path): Path to VCF file
+        data_dependencies (Path): Path to data dependencies
+        hpo_txt (Path): Path to hpo txt file
+        output_dir (Path): Path to output directory
+    Returns:
+        AIMARRVELVolumes:The volumes to mount AI MARRVEL
+    """
+    return AIMARRVELVolumes(
+        vcf_path=f"{str(vcf_path)}:{VCF_FILE}",
+        data_dependencies=f"{str(data_dependencies)}:{DATA_DEPENDENCIES}",
+        hpo_txt=f"{str(hpo_txt)}:{HPO_TXT}",
+        output_dir=f"{str(output_dir)}:{OUTPUT_DIR}",
+    )
+
+
+def create_docker_command(sample_data: SampleData) -> List[str]:
+    """
+    Create docker command to run AI MARRVEL.
+    Args:
+        sample_data (SampleData): The sample data
+    Returns:
+        List[str]: The docker command to run AI MARRVEL
+    """
+    return ["/run/proc.sh", sample_data.sample_id, sample_data.genome_assembly, "30G"]
+
+
+def run_docker_sample(
+    phenopacket_path: Path,
+    vcf_dir: Path,
+    data_dependencies: Path,
+    hpo_txt: Path,
+    output_dir: Path,
+    client: DockerClient,
+) -> None:
+    """
+    Run docker command for a sample.
+    Args:
+        phenopacket_path (Path): Path to phenopacket file
+        vcf_dir (Path): Path to VCF directory
+        data_dependencies (str): Path to data dependencies
+        hpo_txt (str): Path to hpo txt file
+        output_dir (str): Path to output directory
+        client (DockerClient): Docker client
+    """
+    sample_data = get_sample_data(phenopacket_path, vcf_dir)
+    docker_mounts = create_volumes(sample_data.vcf_name, data_dependencies, hpo_txt, output_dir)
+    vol = [
+        docker_mounts.vcf_path,
+        docker_mounts.hpo_txt,
+        docker_mounts.data_dependencies,
+        docker_mounts.output_dir,
+    ]
+    docker_command = create_docker_command(sample_data)
+    container = client.containers.run(
+        "chaozhongliu/aim-lite",
+        " ".join(docker_command),
+        volumes=[x for x in vol if x is not None],
+        detach=True,
+    )
+    for line in container.logs(stream=True):
+        print(line.strip())
+
+
+def run_docker(testdata_dir: Path, input_dir: Path, output_dir: Path) -> None:
+    """
+    Run AI MARRVEL with docker on a corpus.
+    Args:
+        testdata_dir (Path): Path to test data directory
+        input_dir (Path): Path to input directory
+        output_dir (Path): Path to output directory
+    """
+    client = docker.from_env()
+    for phenopacket_path in all_files(testdata_dir.joinpath("phenopackets")):
+        run_docker_sample(
+            phenopacket_path,
+            testdata_dir.joinpath("vcf"),
+            input_dir,
+            output_dir,
+            testdata_dir.joinpath(f"hpo_ids/{phenopacket_path.stem}.txt"),
+            client,
+        )

src/pheval_ai_marrvel/run/run.py

@@ -2,10 +2,15 @@
 from pathlib import Path
 
 from pheval_ai_marrvel.run.create_apptainer_commands import create_apptainer_commands
+from pheval_ai_marrvel.run.create_docker_commands import run_docker
 
 
 def run_commands(
-    tool_input_commands_dir: Path, testdata_dir: Path, input_dir: Path, output_dir: Path
+    tool_input_commands_dir: Path,
+    testdata_dir: Path,
+    input_dir: Path,
+    output_dir: Path,
+    environment: str,
 ) -> None:
     """
     Run the apptainer commands.
@@ -16,9 +21,12 @@ def run_commands(
         input_dir (Path): Path to the input directory.
         output_dir (Path): Path to the output directory.
     """
-    create_apptainer_commands(tool_input_commands_dir, testdata_dir, input_dir, output_dir)
-    batch_file = tool_input_commands_dir.joinpath(f"{testdata_dir.name}_commands.txt")
-    subprocess.run(
-        ["bash", str(batch_file)],
-        shell=False,
-    )
+    if environment == "apptainer":
+        create_apptainer_commands(tool_input_commands_dir, testdata_dir, input_dir, output_dir)
+        batch_file = tool_input_commands_dir.joinpath(f"{testdata_dir.name}_commands.txt")
+        subprocess.run(
+            ["bash", str(batch_file)],
+            shell=False,
+        )
+    elif environment == "docker":
+        run_docker(testdata_dir, input_dir, output_dir)

src/pheval_ai_marrvel/runner.py

@@ -6,6 +6,7 @@
 from pheval_ai_marrvel.post_process.post_process import post_process_results
 from pheval_ai_marrvel.prepare.prepare import prepare_inputs
 from pheval_ai_marrvel.run.run import run_commands
+from pheval_ai_marrvel.tool_specific_configuration_options import AIMARRVELConfigurations
 
 
 @dataclass
@@ -29,11 +30,15 @@ def run(self):
         Run AI-MARRVEL to produce the raw output.
         """
         print("running with AI-MARRVEL")
+        config = AIMARRVELConfigurations.parse_config(
+            self.input_dir_config.tool_specific_configuration_options
+        )
         run_commands(
             tool_input_commands_dir=self.tool_input_commands_dir,
             testdata_dir=self.testdata_dir,
             input_dir=self.input_dir,
             output_dir=self.output_dir,
+            environment=config.environment,
         )
 
     def post_process(self):

src/pheval_ai_marrvel/tool_specific_configuration_options.py

@@ -0,0 +1,12 @@
+from pydantic import BaseModel, Field
+
+
+class AIMARRVELConfigurations(BaseModel):
+    """
+    Class for defining the AI MARRVEL configurations in tool_specific_configurations field,
+    within the input_dir config.yaml
+    Args:
+        environment (str): Environment to run AI MARRVEL, i.e., docker/apptainer
+    """
+
+    environment: str = Field(...)