datamol.data

The data module aims to provide a fast and convenient access to various molecular datasets.

---

cdk2(as_df=True, mol_column='mol')

Return the RDKit CDK2 dataset from RDConfig.RDDocsDir, 'Book/data/cdk2.sdf'.

Parameters:

Name	Type	Description	Default
as_df	bool	Whether to return a list mol or a pandas DataFrame.	True
mol_column	Optional[str]	Name of the mol column. Only relevant if as_df is True.	'mol'

Source code in datamol/data/__init__.py

def cdk2(as_df: bool = True, mol_column: Optional[str] = "mol"):
    """Return the RDKit CDK2 dataset from `RDConfig.RDDocsDir, 'Book/data/cdk2.sdf'`.

    Args:
        as_df: Whether to return a list mol or a pandas DataFrame.
        mol_column: Name of the mol column. Only relevant if `as_df` is True.
    """

    with open_datamol_data_file("cdk2.sdf", open_binary=True) as f:
        data = read_sdf(f, as_df=as_df, mol_column=mol_column)
    return data

chembl_drugs(as_df=True)

A list of ~2.5k molecules from ChEMBL (all approved drugs) in SMILES format. Includes metadata indicating year of first approval, molecule chembl id, molecule type and pref_name.

List was generated with 'Get_ChEMBL_Approved_Drugs.ipynb' on 2023-10-18. The notebook works with the chembl_webresource_client api to collect chembl IDs and metadata, then focuses on small molecules with valid SMILES and first approval date.

Source code in datamol/data/__init__.py

def chembl_drugs(as_df: bool = True) -> Union[List[Mol], pd.DataFrame]:
    """A list of ~2.5k molecules from ChEMBL (all approved drugs) in SMILES format.
    Includes metadata indicating year of first approval, molecule chembl id, molecule type and pref_name.

    List was generated with ['Get_ChEMBL_Approved_Drugs.ipynb'](https://github.com/datamol-io/datamol/notebooks/Get_ChEMBL_Approved_Drugs.ipynb) on 2023-10-18.
    The notebook works with the chembl_webresource_client api to collect chembl IDs and metadata, then focuses on small molecules with valid SMILES and first approval date.
    """
    with open_datamol_data_file("chembl_approved_drugs.parquet", open_binary=True) as f:
        data = pd.read_parquet(f)

    if not as_df:
        data = from_df(data)

    return data

chembl_samples(as_df=True)

A list of ~2k molecules from ChEMBL.

Originally, proposed by Patrick Walters at https://github.com/PatWalters/practical_cheminformatics_posts/tree/b4dae239a8b942dab3a41e637ac55d4491aee96f/molskill.

Source code in datamol/data/__init__.py

def chembl_samples(as_df: bool = True) -> Union[List[Mol], pd.DataFrame]:
    """A list of ~2k molecules from ChEMBL.

    Originally, proposed by Patrick Walters at <https://github.com/PatWalters/practical_cheminformatics_posts/tree/b4dae239a8b942dab3a41e637ac55d4491aee96f/molskill>.
    """

    with open_datamol_data_file("chembl_samples.csv") as f:
        data = pd.read_csv(f)

    if not as_df:
        data = from_df(data)

    return data

freesolv(as_df=True)

Return the FreeSolv dataset as a dataframe.

The dataset contains 642 molecules and the following columns: ['iupac', 'smiles', 'expt', 'calc'].

Warning

This dataset is only meant to be used as a toy dataset for pedagogic and testing purposes. It is not a dataset for benchmarking, analysis or model training.

Source code in datamol/data/__init__.py

def freesolv(as_df: bool = True) -> Union[List[Mol], pd.DataFrame]:
    """Return the FreeSolv dataset as a dataframe.

    The dataset contains 642 molecules and the following columns:
    `['iupac', 'smiles', 'expt', 'calc']`.

    Warning:
        This dataset is only meant to be used as a toy dataset for pedagogic and
        testing purposes. **It is not** a dataset for benchmarking, analysis or
        model training.
    """

    with open_datamol_data_file("freesolv.csv") as f:
        data = pd.read_csv(f)

    if not as_df:
        data = from_df(data)

    return data

solubility(as_df=True, mol_column='mol')

Return the RDKit solubility dataset from RDConfig.RDDocsDir, 'Book/data/solubility.{train|test}.sdf'.

The dataframe or the list of molecules with contain a split column, either train or test.

Parameters:

Name	Type	Description	Default
as_df	bool	Whether to return a list mol or a pandas DataFrame.	True
mol_column	Optional[str]	Name of the mol column. Only relevant if as_df is True.	'mol'

Source code in datamol/data/__init__.py

def solubility(as_df: bool = True, mol_column: Optional[str] = "mol"):
    """Return the RDKit solubility dataset from `RDConfig.RDDocsDir, 'Book/data/solubility.{train|test}.sdf'`.

    The dataframe or the list of molecules with contain a `split` column, either `train` or `test`.

    Args:
        as_df: Whether to return a list mol or a pandas DataFrame.
        mol_column: Name of the mol column. Only relevant if `as_df` is True.
    """

    with open_datamol_data_file("solubility.train.sdf", open_binary=True) as f:
        train = read_sdf(f, as_df=True, mol_column="mol", smiles_column=None)

    with open_datamol_data_file("solubility.test.sdf", open_binary=True) as f:
        test = read_sdf(f, as_df=True, mol_column="mol", smiles_column=None)

    train = cast(pd.DataFrame, train)
    test = cast(pd.DataFrame, test)

    train["split"] = "train"
    test["split"] = "test"

    # NOTE(hadim): LMAO RDkit consistency xD
    test = test.rename(columns={"SMILES": "smiles"})

    data = pd.concat([train, test], ignore_index=True)

    if as_df:
        if mol_column is None:
            data = data.drop(columns=["mol"])

        render_mol_df(data)
        return data

    return from_df(data, mol_column=mol_column)