person_project_easy.py

import os
import json
import datasets

_BASE_URL = "https://huggingface.co./datasets/giulio98/person_project_easy/resolve/main/"

class SyntheticProjectConnectivityConfig(datasets.BuilderConfig):
    """BuilderConfig for SyntheticProjectConnectivity."""
    def __init__(self, chunk=None, **kwargs):
        """
        Args:
            chunk: (str) chunk identifier (e.g., "4", "8", etc.).
            **kwargs: Additional keyword arguments forwarded to super.
        """
        super().__init__(**kwargs)
        self.chunk = chunk


class SyntheticProjectConnectivity(datasets.GeneratorBasedBuilder):
    """
    Dataset builder for Synthetic Project Connectivity with phases phase1..phase8.
    """

    BUILDER_CONFIGS = [
        SyntheticProjectConnectivityConfig(
            name=f"{chunks_id}_chunks",
            chunk=str(chunks_id),
            version=datasets.Version("1.0.0"),
            description=f"Data for chunks {chunks_id}",
        )
        for chunks_id in [4, 8, 16, 32]
    ]

    def _info(self):
        return datasets.DatasetInfo(
            description="Dataset for synthetic project connectivity phases.",
            features=datasets.Features(
                {
                    "id": datasets.Value("string"),  # Unique ID field
                    "context": datasets.Value("string"),
                    "question": datasets.Value("string"),
                    "answers":  [datasets.Value("string")],  # List of answers
                }
            ),
            supervised_keys=None,
            homepage="https://example.com",
            citation="Your citation here",
        )

    def _split_generators(self, dl_manager: datasets.DownloadManager):
        """
        Define dataset splits using a base URL and dynamically appending phase names.
        """
        chunk_str = self.config.chunk
        if not chunk_str:
            raise ValueError("Chunk ID is not specified in the configuration.")
    
        # Construct the URL dynamically based on the phase
        data_url = f"{_BASE_URL}{chunk_str}_chunks.jsonl"
    
        # Download the file corresponding to the requested phase
        downloaded_file = dl_manager.download_and_extract(data_url)
    
        return [
            datasets.SplitGenerator(
                name=datasets.Split.TEST,
                gen_kwargs={"filepath": downloaded_file},
            )
        ]

    def _generate_examples(self, filepath):
        """
        Read lines from the JSONL file and yield examples.
        """
        with open(filepath, "r", encoding="utf-8") as f:
            for idx, line in enumerate(f):
                line = line.strip()
                if not line:
                    continue
                try:
                    data = json.loads(line)
                    yield idx, {
                        "id": data["id"],  # Include the unique ID field
                        "context": data["context"],
                        "question": data["question"],
                        "answers": data["answers"],  # Use answers list
                    }
                except json.JSONDecodeError as e:
                    raise ValueError(f"Error decoding JSON at line {idx} in {filepath}: {e}")