Expectmine pipeline

It is recommended to visit the documentation

Project setup

To get started, we first create a virtual environment and install the necessary dependencies. We can then use the makefile commands to navigate in the project.

Linux or macOS

python -m venv env
source env/bin/activate
make deps
make env-vars

Windows

python -m venv env
.\env\Scripts\activate
make deps
make env-vars

Commands

As this step relies on Make, you either have to install [Make for Windows](https://gnuwin32.sourceforge.net/packages/make.htm)
or do the steps manually.

Every necessary step in the repo can be executed through a command in the Makefile. The commands can be grouped in the following groups:

Project scripts

• make deps: Installs all necessary dependencies in the virtual environment.
• make clean: Removes the environment and pytest_cache.
• make lint: Lints the project using flake8 and pyright.
• make format: Formats code and imports using isort and black.
• make test: Runs tests and calculates test coverage using pytest.

Code generation

• make step: Starts the workflow of creating a new step.
• make readme: Updates the README.md using the sphinx documentation.
• make env-vars: Generates an empty .env file with all necessary environment variables.
• make docs-html: Generates the html documentation using .md and .rst files.

Documentation

• make docs-start: Starts hot reloading of the sphinx documentation to easily work on it.
• make open-docs: Opens the local documentation in the browser.

Repo structure

The repository is divided into the following parts:

• docs/ Contains the sphinx documentation
  • build/ Contains the html output of the generated documentation
  • source/ Used to build the documentation
• scripts/ Contains all scripts used for code generation
• expectmine/ Contains the entire pipeline
  • io/ All Io logic, concerned with getting user input to the pipeline and steps.
  • logger/ Contains all loggers
  • pipeline/ Contains the pipeline code
  • steps/ Contains all steps used by the pipeline
  • storage/ Contains all storage logic
  • utils/ Util functions used by all modules
• tests/ Used for testing
• Makefile Used for running commands
• pyproject.toml Configures pyright isort and flake8
• README.md Auto generated readme
• requirements.txt Requirements used by the project

Your first pipeline

Initializing a simple pipeline

Creating a pipeline to execute in python directly is super easy, just import the get_quickstart_config method and initialize a pipeline directly from it.

from pathlib import Path

from expectmine.pipeline.pipeline import Pipeline
from expectmine.pipeline.utils import get_quickstart_config

pipeline = Pipeline(*get_quickstart_config(output_path=Path("pipeline_output")))

Adding input files

To add input files to the pipeline, just give it an array of paths to the individual files. The files you provide will serve as input to the first step.

pipeline.set_input([Path('file1.mzml')])

Adding steps

To add a step, just provide the type of step you want to add and an io object, which is used to configure the step.

pipeline.add_step(
    MZmine3,
    {
        "mzmine3_path": Path(
            "/Applications/MZmine.app/Contents/MacOS/MZmine", absolute=True
        ),
        "batchfile": Path("batchfile.xml"),
    }
)

Run the pipeline

After calling run the pipeline will then run and output files for each step in the directory you provided as output_path.

pipeline.run()

Summary

from pathlib import Path

from expectmine.pipeline.pipeline import Pipeline
from expectmine.pipeline.utils import get_quickstart_config
from expectmine.steps.steps import SiriusFingerprint
from expectmine.steps.steps.mzmine3.mzmine3 import MZmine3

pipeline = Pipeline(*get_quickstart_config(output_path=Path("pipeline_output")))

pipeline.set_input([Path("file1.mzml"), Path("file2.mzml")])

pipeline.add_step(
    MZmine3,
    {
        "mzmine3_path": Path(
            "/Applications/MZmine.app/Contents/MacOS/MZmine", absolute=True
        ),
        "batchfile": Path("batchfile.xml"),
    }
)

pipeline.add_step(
    SiriusFingerprint,
    {
        "sirius_path": Path(
            "/Applications/sirius.app/Contents/MacOS/sirius", absolute=True
        ),
        "set_max_mz": False,
        "instrument": "orbitrap",
    }
)

pipeline.run()