app.py

import logging
import pathlib

import gradio as gr
import numpy as np
import pandas as pd
from gt4sd.properties.molecules import MOLECULE_PROPERTY_PREDICTOR_FACTORY

from utils import draw_grid_predict

logger = logging.getLogger(__name__)
logger.addHandler(logging.NullHandler())

MOLFORMER_VERSIONS = {
    "molformer_classification": ["bace", "bbbp", "hiv"],
    "molformer_regression": [
        "alpha",
        "cv",
        "g298",
        "gap",
        "h298",
        "homo",
        "lipo",
        "lumo",
        "mu",
        "r2",
        "u0",
    ],
    "molformer_multitask_classification": ["clintox", "sider", "tox21"],
}

REMOVE = ["docking", "docking_tdc", "molecule_one", "askcos", "plogp"]
REMOVE.extend(["similarity_seed", "activity_against_target", "organtox"])
REMOVE.extend(MOLFORMER_VERSIONS.keys())

MODEL_PROP_DESCRIPTION = {
    "Tox21": "NR-AR, NR-AR-LBD, NR-AhR, NR-Aromatase, NR-ER, NR-ER-LBD, NR-PPAR-gamma, SR-ARE, SR-ATAD5, SR-HSE, SR-MMP, SR-p53",
    "Sider": "Hepatobiliary disorders,Metabolism and nutrition disorders,Product issues,Eye disorders,Investigations,Musculoskeletal disorders,Gastrointestinal disorders,Social circumstances,Immune system disorders,Reproductive system and breast disorders,Bening & malignant,General disorders,Endocrine disorders,Surgical & medical procedures,Vascular disorders,Blood & lymphatic disorders,Skin & subcutaneous disorders,Congenital & genetic disorders,Infections,Respiratory & thoracic disorders,Psychiatric disorders,Renal & urinary disorders,Pregnancy conditions,Ear disorders,Cardiac disorders,Nervous system disorders,Injury & procedural complications",
    "Clintox": "FDA approval, Clinical trial failure",
}


def main(property: str, smiles: str, smiles_file: str):
    if "Molformer" in property:
        version = property.split(" ")[-1].split("(")[-1].split(")")[0]
        property = property.split(" ")[0]
    algo, config = MOLECULE_PROPERTY_PREDICTOR_FACTORY[property.lower()]
    kwargs = (
        {"algorithm_version": "v0"} if property in MODEL_PROP_DESCRIPTION.keys() else {}
    )
    if property.lower() in MOLFORMER_VERSIONS.keys():
        kwargs["algorithm_version"] = version

    model = algo(config(**kwargs))
    if smiles != "" and smiles_file is not None:
        raise ValueError("Pass either smiles or smiles_file, not both.")
    elif smiles != "":
        smiles = [smiles]
    elif smiles_file is not None:
        smiles = pd.read_csv(smiles_file.name, header=None, sep="\t")[0].tolist()
    props = np.array(list(map(model, smiles))).round(2)

    # Expand to 2D array if needed
    if len(props.shape) == 1:
        props = np.expand_dims(np.array(props), -1)

    if property in MODEL_PROP_DESCRIPTION.keys():
        property_names = MODEL_PROP_DESCRIPTION[property].split(",")
    else:
        property_names = [property]

    return draw_grid_predict(
        smiles, props, property_names=property_names, domain="Molecules"
    )


if __name__ == "__main__":
    # Preparation (retrieve all available algorithms)
    properties = list(MOLECULE_PROPERTY_PREDICTOR_FACTORY.keys())[::-1]
    for prop in REMOVE:
        prop_to_idx = dict(zip(properties, range(len(properties))))
        properties.pop(prop_to_idx[prop])
    properties = list(map(lambda x: x.capitalize(), properties))

    # MolFormer options
    for key in MOLFORMER_VERSIONS.keys():
        properties.extend(
            [f"{key.capitalize()} ({version})" for version in MOLFORMER_VERSIONS[key]]
        )

    # Load metadata
    metadata_root = pathlib.Path(__file__).parent.joinpath("model_cards")

    examples = [
        ["Qed", "", metadata_root.joinpath("examples.smi")],
        [
            "Esol",
            "CN1CCN(CCCOc2ccc(N3C(=O)C(=Cc4ccc(Oc5ccc([N+](=O)[O-])cc5)cc4)SC3=S)cc2)CC1",
            None,
        ],
    ]

    with open(metadata_root.joinpath("article.md"), "r") as f:
        article = f.read()
    with open(metadata_root.joinpath("description.md"), "r") as f:
        description = f.read()

    demo = gr.Interface(
        fn=main,
        title="Molecular properties",
        inputs=[
            gr.Dropdown(properties, label="Property", value="Scscore"),
            gr.Textbox(
                label="Single SMILES",
                placeholder="CC(C#C)N(C)C(=O)NC1=CC=C(Cl)C=C1",
                lines=1,
            ),
            gr.File(
                file_types=[".smi"],
                label="Multiple SMILES (tab-separated, `.smi` file)",
            ),
        ],
        outputs=gr.HTML(label="Output"),
        article=article,
        description=description,
        examples=examples,
    )
    demo.launch(debug=True, show_error=True)