code/spec_rag/io.py

from __future__ import annotations

import json
from pathlib import Path
from typing import Iterable, List, Tuple

import numpy as np


def load_smiles(path: str | Path) -> List[str]:
    path = Path(path)
    if path.suffix.lower() in {".tsv", ".csv"}:
        import pandas as pd

        sep = "\t" if path.suffix.lower() == ".tsv" else ","
        df = pd.read_csv(path, sep=sep)
        if "smiles" not in df.columns:
            raise ValueError(f"{path} does not have a 'smiles' column")
        return df["smiles"].dropna().astype(str).tolist()
    if path.suffix.lower() == ".parquet":
        import pandas as pd

        df = pd.read_parquet(path)
        if "smiles" not in df.columns:
            raise ValueError(f"{path} does not have a 'smiles' column")
        return df["smiles"].dropna().astype(str).tolist()
    with path.open("r", encoding="utf-8") as f:
        return [line.strip() for line in f if line.strip()]


def save_smiles(path: str | Path, smiles: Iterable[str]) -> None:
    path = Path(path)
    with path.open("w", encoding="utf-8") as f:
        for smi in smiles:
            f.write(f"{smi}\n")


def load_embeddings(path: str | Path) -> np.ndarray:
    path = Path(path)
    if path.suffix == ".npz":
        data = np.load(path)
        if "embeddings" not in data:
            raise ValueError(f"Missing 'embeddings' in {path}")
        return data["embeddings"]
    return np.load(path)


def save_embeddings(path: str | Path, embeddings: np.ndarray) -> None:
    path = Path(path)
    np.save(path, embeddings.astype(np.float32))


def load_jsonl(path: str | Path) -> List[dict]:
    path = Path(path)
    rows: List[dict] = []
    with path.open("r", encoding="utf-8") as f:
        for line in f:
            line = line.strip()
            if not line:
                continue
            rows.append(json.loads(line))
    return rows


def save_jsonl(path: str | Path, rows: Iterable[dict]) -> None:
    path = Path(path)
    with path.open("w", encoding="utf-8") as f:
        for row in rows:
            f.write(json.dumps(row, ensure_ascii=True) + "\n")


def ensure_dir(path: str | Path) -> Path:
    path = Path(path)
    path.mkdir(parents=True, exist_ok=True)
    return path


def to_numpy(x) -> np.ndarray:
    if isinstance(x, np.ndarray):
        return x
    return np.asarray(x)


def chunked(items: List[str], batch_size: int) -> Iterable[Tuple[int, List[str]]]:
    for i in range(0, len(items), batch_size):
        yield i, items[i : i + batch_size]