rdfreader 1.0.5

Read the full contents of CTAB .rdf files in python. Captures RXN and MOL record using RDKit and reads additional data fields (including solvents/catalysts/agents).

RDF READER

User Guide

Installation

pip install rdfreader

Basic Usage

from rdfreader import RDFParser

rdf_file_name = "reactions.rdf"

with open(rdf_file_name, "r") as rdf_file:

    # create a RDFParser object, this is a generator that yields Reaction objects
    rdfreader = RDFParser(
        rdf_file,
        except_on_invalid_molecule=False,  # will return None instead of raising an exception if a molecule is invalid
        except_on_invalid_reaction=False,  # will return None instead of raising an exception if a reaction is invalid
    )

    for rxn in rdfreader:
        if rxn is None:
            continue # the parser failed to read the reaction, go to the next one

        # rxn is a Reaction object, it is several attributes, including:
        print(rxn.smiles) # reaction SMILES string
        print(rxn.properties) # a dictionary of properties extracted from the RXN record

        reactants = rxn.reactants # a list of Molecule objects
        products = rxn.products
        solvents = rxn.solvents
        catalysts = rxn.catalysts

        # Molecule objects have several attributes, including:
        print(reactants[0].smiles)
        print(reactants[0].properties) # a dictionary of properties extracted from the MOL record (often empty)
        reactants[0].rd_mol # an RDKit molecule object

Example Data

You can find example data in the test/resources directory. spresi-100.rdf contains 100 example records from SPRESI.

Important Note Regarding File Formats

If you are using files that have been saved with Windows-style carriage returns (^M^M, or \r\r), you may encounter issues when running this package.

To correct this issue, you can use the following sed command in a Linux-based terminal to convert double carriage returns to single ones in affected files:

sed -i 's/\r\r/\r/g' reactions.rdf

Developer Guide

The project is managed and packaged using poetry.

Installation

git clone https://github.com/ChemAILtd/rdfreader.git
poetry install  # create a virtual environment and install the project dependencies
pre-commit install  # install pre-commit hooks, these mostly manage codestyle

Contributions

Contributions are welcome via the fork and pull request model.

Before you commit changes, ensure these pass the hooks installed by pre-commit. This should be run automatically on each commit if you have run pre-commit install, but can be run manually from the terminal with pre-commit run.

Releases

Releases are managed by GitHub releases/workflow. The version number in the pyproject file should ideally be kept up to date to the current release but is ignored by the release workflow.

To release a new version:

• Update the pyproject.toml version number.
• Push the changes to GitHub and merge to main via a pull request.
• Use the github website to create a release. Tag the commit to be released with a version number, e.g. v1.2.3. The tag should be in v*.. and match the version number in the pyproject.toml file.
• When the release is published, a github workflow will run, build a wheel and publish it to PyPI.