crag_sampler/sampler.py

import bz2
from typing import Iterator, Dict, Any, List, Optional
import pandas as pd
import os
import hashlib
import json
from sklearn.model_selection import StratifiedKFold
import numpy as np
from multiprocessing import Pool, cpu_count
from functools import partial
import subprocess

from .utils import (
    read_jsonl_fields_fast,
    process_answer_types,
    create_stratified_subsamples,
    subsample_jsonl_file,
)


class CragSampler:
    """Main class for handling CRAG dataset sampling operations."""

    def __init__(
        self,
        input_file: str,
        required_fields: Optional[List[str]] = None,
        use_cache: bool = True,
    ):
        """Initialize CragSampler.

        Args:
            input_file: Path to input JSONL file (can be bz2 compressed)
            required_fields: List of field names to extract. If None, uses default fields
            use_cache: Whether to use/create cache file
        """
        self.input_file = input_file
        self.required_fields = required_fields or [
            "domain",
            "answer",
            "question_type",
            "static_or_dynamic",
        ]
        self.use_cache = use_cache
        self.df = self._load_data()

    def _load_data(self) -> pd.DataFrame:
        """Load and process data from JSONL file."""
        df = read_jsonl_fields_fast(
            self.input_file, self.required_fields, self.use_cache
        )
        return process_answer_types(df)

    def create_subsamples(
        self,
        n_subsamples: int = 5,
        stratify_columns: Optional[List[str]] = None,
        output_path: Optional[str] = None,
        force_compute: bool = False,
    ) -> Dict:
        """Create stratified subsamples of the dataset.

        Args:
            n_subsamples: Number of subsamples to create
            stratify_columns: Columns to use for stratification. If None, uses defaults
            output_path: Path to save/load the JSON output
            force_compute: If True, always compute subsamples even if file exists

        Returns:
            Dictionary containing the subsamples information
        """
        if stratify_columns is None:
            stratify_columns = [
                "domain",
                "answer_type",
                "question_type",
                "static_or_dynamic",
            ]

        if output_path is None:
            output_path = os.path.join(
                os.path.dirname(self.input_file),
                f"{os.path.splitext(os.path.basename(self.input_file))[0]}_subsamples.json",
            )

        return create_stratified_subsamples(
            self.df,
            n_subsamples=n_subsamples,
            stratify_columns=stratify_columns,
            output_path=output_path,
            force_compute=force_compute,
        )

    def write_subsamples(
        self,
        subsamples_file: str,
        output_dir: Optional[str] = None,
        compress: bool = True,
        n_processes: Optional[int] = None,
        overwrite: bool = False,
    ) -> None:
        """Write subsamples to separate files.

        Args:
            subsamples_file: Path to JSON file containing subsample indices
            output_dir: Directory to save subsample files
            compress: Whether to compress output files with bz2
            n_processes: Number of processes to use
            overwrite: If False, skip existing output files
        """
        subsample_jsonl_file(
            self.input_file,
            subsamples_file,
            output_dir=output_dir,
            compress=compress,
            n_processes=n_processes,
            overwrite=overwrite,
        )