Merge pull request #75 from sudlab/ns-rse/66-parser

isoslam/processing.py

@@ -0,0 +1,153 @@
+"""Functions for Processing."""
+
+import argparse as arg
+import sys
+from pathlib import Path
+from typing import Any
+
+# from isoslam import __version__
+
+
+def create_parser() -> arg.ArgumentParser:
+    """
+    Create a parser for reading options.
+
+    Parser is created with multiple sub-parsers for eading options to run ``isoslam``.
+
+    Returns
+    -------
+    arg.ArgumentParser
+        Argument parser.
+    """
+    parser = arg.ArgumentParser(
+        description="Run various programs related to IsoSLAM. Add the name of the program you wish to run."
+    )
+    parser.add_argument(
+        "--version",
+        # version=f"Installed version of IsoSlam : {__version__}",
+        help="Report the installed version of IsoSLAM.",
+    )
+    parser.add_argument(
+        "-c",
+        "--config-file",
+        dest="config_file",
+        type=Path,
+        required=False,
+        help="Path to a YAML configuration file.",
+    )
+    parser.add_argument(
+        "-o",
+        "--output-dir",
+        dest="output_dir",
+        type=Path,
+        required=False,
+        help="Output directory to write results to.",
+    )
+    parser.add_argument(
+        "-l",
+        "--log-level",
+        dest="log_level",
+        type=str,
+        required=False,
+        help="Logging level to use, default is 'info' for verbose output use 'debug'.",
+    )
+
+    subparsers = parser.add_subparsers(title="program", description="Available programs listed below:", dest="program")
+
+    # Create sub-parsers for different stages
+    process_parser = subparsers.add_parser(
+        "process",
+        description="Process all files and run all summary plotting and statistics.",
+        help="Process all files and run all summary plotting and statistics.",
+    )
+    process_parser.add_argument(
+        "-b",
+        "--bam-file",
+        dest="bam_file",
+        type=Path,
+        required=False,
+        help="Path to '.bam' file that has undergone read assignment with 'featureCount'.",
+    )
+    process_parser.add_argument(
+        "-g", "--gtf-file", dest="gtf_file", type=Path, required=False, help="Path to '.gtf' transcript assembly file."
+    )
+    process_parser.add_argument(
+        "-d",
+        "--bed-file",
+        dest="bed_file",
+        type=Path,
+        required=False,
+        help="Path to '.bed' utron file. Must be bed6 format.",
+    )
+    process_parser.add_argument(
+        "-v", "--vcf-file", dest="vcf_file", type=Path, required=False, help="Path to '.vcf.gz' file."
+    )
+    process_parser.set_defaults(func=process)
+    # Additional parsers for future functionality
+    # summarize_counts_parser = subparsers.add_parser(
+    #     "summarize",
+    #     description="Summarize counts.",
+    #     help="Summarize counts.",
+    # )
+    # summarize_counts_parser.set_defaults(func=summarize_counts)
+    # plot_conversions_parser = subparsers.add_parser(
+    #     "plot_conversions",
+    #     description="Plot conversions.",
+    #     help="Plot conversions.",
+    # )
+    # plot_conversions_parser.set_defaults(func=plot_conversions)
+    return parser
+
+
+def process(args: arg.Namespace | None) -> None:
+    """
+    Process a set of files.
+
+    Parameters
+    ----------
+    args : arg.Namespace | None
+        Arguments function was invoked with.
+
+    Returns
+    -------
+    None
+        Function does not return anything.
+    """
+    return
+
+
+def entry_point(manually_provided_args: list[Any] | None = None, testing: bool = False) -> None | arg.Namespace:
+    """
+    Entry point for all IsoSLAM programs.
+
+    Main entry point for running 'isoslam' which allows the different processing, plotting and testing modules to be
+    run.
+
+    Parameters
+    ----------
+    manually_provided_args : None
+        Manually provided arguments.
+    testing : bool
+        Whether testing is being carried out.
+
+    Returns
+    -------
+    None
+        Function does not return anything.
+    """
+    parser = create_parser()
+    args = parser.parse_args() if manually_provided_args is None else parser.parse_args(manually_provided_args)
+
+    # If no module has been specified print help and exit
+    print(f"{args.program=}")
+    if not args.program:
+        parser.print_help()
+        sys.exit()
+
+    if testing:
+        return args
+
+    # Run the specified module(s)
+    args.func(args)
+
+    return None

pyproject.toml

@@ -10,7 +10,7 @@ name = "isoslam"
 description = "Tools for processing and summarising SLAMSeq experiments."
 readme = "README.md"
 license = {text = "MIT License"}
-requires-python = ">=3.9"
+requires-python = ">=3.10"
 dynamic = ["version"]
 authors = [
   {name = "Jack Riley", email = "[email protected]"},
@@ -19,7 +19,6 @@ authors = [
 ]
 classifiers = [
   "Programming Language :: Python :: 3",
-  "Programming Language :: Python :: 3.9",
   "Programming Language :: Python :: 3.10",
   "Programming Language :: Python :: 3.11",
   "Programming Language :: Python :: 3.12",
@@ -268,7 +267,7 @@ files = [
   "isoslam",
   "tests"
 ]
-python_version = "3.9"
+python_version = "3.10"
 strict = true
 enable_error_code = ["ignore-without-code", "redundant-expr", "truthy-bool"]
 warn_unreachable = true
@@ -278,4 +277,7 @@ exclude = [
 
 [[tool.mypy.overrides]]
 module = [ "numpy.*", ]
-ignore_missing_imports = true
+ignore_missing_imports = true
+
+[project.scripts]
+isoslam = "isoslam.processing:entry_point"

tests/test_processing.py

@@ -0,0 +1,102 @@
+"""Tests of the processing modules entry point, argument parsing and ability to correctly select programs."""
+
+from collections.abc import Callable
+from pathlib import Path
+
+import pytest
+
+from isoslam import processing
+
+
+@pytest.mark.parametrize(
+    ("option", "help_message"),
+    [
+        pytest.param(["-h"], "usage:", id="Single letter help"),
+        pytest.param(["--help"], "program", id="Multiple letter help"),
+        pytest.param(
+            ["process"],
+            "Process all files and run all summary plotting and statistics.",
+            id="process module",
+            marks=pytest.mark.xfail(reason="work in progress"),
+        ),
+    ],
+)
+def test_entry_point_help(option: str, help_message: str, capsys: pytest.CaptureFixture) -> None:
+    """Test the entry_point()'s help argument."""
+    try:
+        processing.entry_point(manually_provided_args=option)
+    except SystemExit:
+        pass
+    output = capsys.readouterr().out
+    assert help_message in output
+
+
+@pytest.mark.parametrize(
+    ("options", "expected_function", "expected_args"),
+    [
+        pytest.param(
+            ["--config", "dummy/config/dir/config.yaml", "process"],
+            processing.process,
+            {"config_file": Path("dummy/config/dir/config.yaml")},
+            id="dummy config and process",
+        ),
+        pytest.param(
+            ["--config", "dummy/config/dir/config.yaml", "process", "--bam-file", "data/bam/some.bam"],
+            processing.process,
+            {"config_file": Path("dummy/config/dir/config.yaml"), "bam_file": Path("data/bam/some.bam")},
+            id="dummy config, bam file and process",
+        ),
+        pytest.param(
+            [
+                "--config",
+                "dummy/config/dir/config.yaml",
+                "--output-dir",
+                "output",
+                "process",
+                "--bam-file",
+                "data/bam/some.bam",
+            ],
+            processing.process,
+            {
+                "config_file": Path("dummy/config/dir/config.yaml"),
+                "output_dir": Path("output"),
+                "bam_file": Path("data/bam/some.bam"),
+            },
+            id="dummy config and output, process with bam file",
+        ),
+        pytest.param(
+            [
+                "--config",
+                "dummy/config/dir/config.yaml",
+                "--output-dir",
+                "output",
+                "process",
+                "--bam-file",
+                "data/bam/some.bam",
+                "--gtf-file",
+                "data/gtf/some.gtf",
+                "--bed-file",
+                "data/bed/some.bed",
+                "--vcf-file",
+                "data/vcf/some.vcf.gz",
+            ],
+            processing.process,
+            {
+                "config_file": Path("dummy/config/dir/config.yaml"),
+                "output_dir": Path("output"),
+                "bam_file": Path("data/bam/some.bam"),
+                "gtf_file": Path("data/gtf/some.gtf"),
+                "bed_file": Path("data/bed/some.bed"),
+                "vcf_file": Path("data/vcf/some.vcf.gz"),
+            },
+            id="dummy config and output, process with all files",
+        ),
+    ],
+)
+def test_entry_point_sub_parsers(options: str, expected_function: Callable, expected_args: dict) -> None:
+    """Test the correct function is returned by each subparser."""
+    returned_args = processing.entry_point(manually_provided_args=options, testing=True)
+    assert returned_args.func == expected_function
+    returned_args_dict = vars(returned_args)
+    for argument, value in expected_args.items():
+        assert returned_args_dict[argument] == value