Delete DeepScholarBench.py

DeepScholarBench.py

@@ -1,337 +0,0 @@
-"""
-DeepScholarBench: Academic papers with extracted related works sections and recovered citations.
-
-This dataset contains academic papers from ArXiv with their related works sections and recovered citations,
-providing a rich resource for research generation and citation analysis tasks.
-"""
-
-import csv
-import datasets
-from typing import Dict, List, Any, Optional
-
-
-# Dataset URLs - these would typically point to hosted files
-_DESCRIPTION = """\
-A comprehensive dataset of academic papers with extracted related works sections and recovered citations, 
-designed for training and evaluating research generation systems.
-
-This dataset contains 63 academic papers from ArXiv with their related works sections and 1630 recovered citations, 
-providing a rich resource for research generation and citation analysis tasks.
-"""
-
-_CITATION = """\
-@misc{patel2025deepscholarbenchlivebenchmarkautomated,
-      title={DeepScholar-Bench: A Live Benchmark and Automated Evaluation for Generative Research Synthesis}, 
-      author={Liana Patel and Negar Arabzadeh and Harshit Gupta and Ankita Sundar and Ion Stoica and Matei Zaharia and Carlos Guestrin},
-      year={2025},
-      eprint={2412.19698},
-      archivePrefix={arXiv},
-      primaryClass={cs.CL},
-      url={https://arxiv.org/abs/2508.20033}, 
-}
-"""
-
-_HOMEPAGE = "https://github.com/guestrin-lab/deepscholar-bench"
-_LICENSE = "MIT"
-
-# URLs to the dataset files
-_URLS = {
-    "papers": "papers_with_related_works.csv",
-    "citations": "recovered_citations.csv",
-    "important_citations": "important_citations.csv",
-}
-
-
-class DeepScholarBenchConfig(datasets.BuilderConfig):
-    """BuilderConfig for DeepScholarBench dataset."""
-
-    def __init__(self, name: str, description: str, **kwargs):
-        """BuilderConfig for DeepScholarBench.
-        
-        Args:
-            name: Configuration name
-            description: Description of this configuration
-            **kwargs: Additional keyword arguments
-        """
-        super(DeepScholarBenchConfig, self).__init__(
-            name=name,
-            description=description,
-            version=datasets.Version("1.0.0"),
-            **kwargs
-        )
-
-
-class DeepScholarBench(datasets.GeneratorBasedBuilder):
-    """DeepScholarBench dataset."""
-
-    VERSION = datasets.Version("1.0.0")
-
-    BUILDER_CONFIGS = [
-        DeepScholarBenchConfig(
-            name="papers",
-            description="Academic papers with extracted related works sections (63 papers)",
-        ),
-        DeepScholarBenchConfig(
-            name="citations", 
-            description="Recovered citations with metadata (1630 citations)",
-        ),
-        DeepScholarBenchConfig(
-            name="important_citations",
-            description="Important citations with enhanced metadata (1050 citations)",
-        ),
-        DeepScholarBenchConfig(
-            name="full",
-            description="Complete dataset with papers, citations, and important citations",
-        ),
-    ]
-
-    DEFAULT_CONFIG_NAME = "papers"
-
-    def _info(self) -> datasets.DatasetInfo:
-        """Return the dataset info."""
-        
-        if self.config.name == "papers":
-            features = datasets.Features({
-                "arxiv_id": datasets.Value("string"),
-                "title": datasets.Value("string"),
-                "authors": datasets.Value("string"),
-                "abstract": datasets.Value("string"),
-                "categories": datasets.Value("string"),
-                "published_date": datasets.Value("string"),
-                "updated_date": datasets.Value("string"),
-                "abs_url": datasets.Value("string"),
-                "arxiv_link": datasets.Value("string"),
-                "publication_date": datasets.Value("string"),
-                "raw_latex_related_works": datasets.Value("string"),
-                "clean_latex_related_works": datasets.Value("string"),
-                "pdf_related_works": datasets.Value("string"),
-            })
-        elif self.config.name == "citations":
-            features = datasets.Features({
-                "parent_paper_title": datasets.Value("string"),
-                "parent_paper_arxiv_id": datasets.Value("string"),
-                "citation_shorthand": datasets.Value("string"),
-                "raw_citation_text": datasets.Value("string"),
-                "cited_paper_title": datasets.Value("string"),
-                "cited_paper_arxiv_link": datasets.Value("string"),
-                "cited_paper_abstract": datasets.Value("string"),
-                "has_metadata": datasets.Value("bool"),
-                "is_arxiv_paper": datasets.Value("bool"),
-                "bib_paper_authors": datasets.Value("string"),
-                "bib_paper_year": datasets.Value("float32"),
-                "bib_paper_month": datasets.Value("string"),
-                "bib_paper_url": datasets.Value("string"),
-                "bib_paper_doi": datasets.Value("string"),
-                "bib_paper_journal": datasets.Value("string"),
-                "original_title": datasets.Value("string"),
-                "search_res_title": datasets.Value("string"),
-                "search_res_url": datasets.Value("string"),
-                "search_res_content": datasets.Value("string"),
-            })
-        elif self.config.name == "important_citations":
-            features = datasets.Features({
-                "parent_paper_title": datasets.Value("string"),
-                "parent_paper_arxiv_id": datasets.Value("string"),
-                "citation_shorthand": datasets.Value("string"),
-                "raw_citation_text": datasets.Value("string"),
-                "cited_paper_title": datasets.Value("string"),
-                "cited_paper_arxiv_link": datasets.Value("string"),
-                "cited_paper_abstract": datasets.Value("string"),
-                "has_metadata": datasets.Value("bool"),
-                "is_arxiv_paper": datasets.Value("bool"),
-                "cited_paper_authors": datasets.Value("string"),
-                "bib_paper_year": datasets.Value("float32"),
-                "bib_paper_month": datasets.Value("string"),
-                "bib_paper_url": datasets.Value("string"),
-                "bib_paper_doi": datasets.Value("string"),
-                "bib_paper_journal": datasets.Value("string"),
-                "original_title": datasets.Value("string"),
-                "search_res_title": datasets.Value("string"),
-                "search_res_url": datasets.Value("string"),
-                "search_res_content": datasets.Value("string"),
-                "arxiv_id": datasets.Value("string"),
-                "arxiv_link": datasets.Value("string"),
-                "publication_date": datasets.Value("string"),
-                "title": datasets.Value("string"),
-                "abstract": datasets.Value("string"),
-                "raw_latex_related_works": datasets.Value("string"),
-                "related_work_section": datasets.Value("string"),
-                "pdf_related_works": datasets.Value("string"),
-                "cited_paper_content": datasets.Value("string"),
-            })
-        else:  # full config
-            features = datasets.Features({
-                "split": datasets.Value("string"),
-                "data": datasets.Value("string"),  # JSON encoded data
-            })
-
-        return datasets.DatasetInfo(
-            description=_DESCRIPTION,
-            features=features,
-            homepage=_HOMEPAGE,
-            license=_LICENSE,
-            citation=_CITATION,
-        )
-
-    def _split_generators(self, dl_manager: datasets.DownloadManager) -> List[datasets.SplitGenerator]:
-        """Return the dataset splits."""
-        
-        if self.config.name == "papers":
-            data_file = dl_manager.download_and_extract(_URLS["papers"])
-            return [
-                datasets.SplitGenerator(
-                    name=datasets.Split.TRAIN,
-                    gen_kwargs={
-                        "filepath": data_file,
-                        "split": "papers"
-                    },
-                ),
-            ]
-        elif self.config.name == "citations":
-            data_file = dl_manager.download_and_extract(_URLS["citations"])
-            return [
-                datasets.SplitGenerator(
-                    name=datasets.Split.TRAIN,
-                    gen_kwargs={
-                        "filepath": data_file,
-                        "split": "citations"
-                    },
-                ),
-            ]
-        elif self.config.name == "important_citations":
-            data_file = dl_manager.download_and_extract(_URLS["important_citations"])
-            return [
-                datasets.SplitGenerator(
-                    name=datasets.Split.TRAIN,
-                    gen_kwargs={
-                        "filepath": data_file,
-                        "split": "important_citations"
-                    },
-                ),
-            ]
-        else:  # full config
-            papers_file = dl_manager.download_and_extract(_URLS["papers"])
-            citations_file = dl_manager.download_and_extract(_URLS["citations"])
-            important_citations_file = dl_manager.download_and_extract(_URLS["important_citations"])
-            return [
-                datasets.SplitGenerator(
-                    name="papers",
-                    gen_kwargs={
-                        "filepath": papers_file,
-                        "split": "papers"
-                    },
-                ),
-                datasets.SplitGenerator(
-                    name="citations", 
-                    gen_kwargs={
-                        "filepath": citations_file,
-                        "split": "citations"
-                    },
-                ),
-                datasets.SplitGenerator(
-                    name="important_citations",
-                    gen_kwargs={
-                        "filepath": important_citations_file,
-                        "split": "important_citations"
-                    },
-                ),
-            ]
-
-    def _generate_examples(self, filepath: str, split: str):
-        """Generate examples from the dataset."""
-        
-        def _safe_bool_convert(value: str) -> bool:
-            """Safely convert string to boolean."""
-            if isinstance(value, str):
-                return value.lower() in ('true', 'yes', '1')
-            return bool(value)
-        
-        def _safe_float_convert(value: str) -> Optional[float]:
-            """Safely convert string to float."""
-            if not value or value.strip() == '' or value.lower() == 'nan':
-                return None
-            try:
-                return float(value)
-            except (ValueError, TypeError):
-                return None
-
-        if split == "papers":
-            with open(filepath, encoding="utf-8") as f:
-                reader = csv.DictReader(f)
-                for key, row in enumerate(reader):
-                    yield key, {
-                        "arxiv_id": row.get("arxiv_id", ""),
-                        "title": row.get("title", ""),
-                        "authors": row.get("authors", ""),
-                        "abstract": row.get("abstract", ""),
-                        "categories": row.get("categories", ""),
-                        "published_date": row.get("published_date", ""),
-                        "updated_date": row.get("updated_date", ""),
-                        "abs_url": row.get("abs_url", ""),
-                        "arxiv_link": row.get("arxiv_link", ""),
-                        "publication_date": row.get("publication_date", ""),
-                        "raw_latex_related_works": row.get("raw_latex_related_works", ""),
-                        "clean_latex_related_works": row.get("clean_latex_related_works", ""),
-                        "pdf_related_works": row.get("pdf_related_works", ""),
-                    }
-
-        elif split == "citations":
-            with open(filepath, encoding="utf-8") as f:
-                reader = csv.DictReader(f)
-                for key, row in enumerate(reader):
-                    yield key, {
-                        "parent_paper_title": row.get("parent_paper_title", ""),
-                        "parent_paper_arxiv_id": row.get("parent_paper_arxiv_id", ""),
-                        "citation_shorthand": row.get("citation_shorthand", ""),
-                        "raw_citation_text": row.get("raw_citation_text", ""),
-                        "cited_paper_title": row.get("cited_paper_title", ""),
-                        "cited_paper_arxiv_link": row.get("cited_paper_arxiv_link", ""),
-                        "cited_paper_abstract": row.get("cited_paper_abstract", ""),
-                        "has_metadata": _safe_bool_convert(row.get("has_metadata", "False")),
-                        "is_arxiv_paper": _safe_bool_convert(row.get("is_arxiv_paper", "False")),
-                        "bib_paper_authors": row.get("bib_paper_authors", ""),
-                        "bib_paper_year": _safe_float_convert(row.get("bib_paper_year", "")),
-                        "bib_paper_month": row.get("bib_paper_month", ""),
-                        "bib_paper_url": row.get("bib_paper_url", ""),
-                        "bib_paper_doi": row.get("bib_paper_doi", ""),
-                        "bib_paper_journal": row.get("bib_paper_journal", ""),
-                        "original_title": row.get("original_title", ""),
-                        "search_res_title": row.get("search_res_title", ""),
-                        "search_res_url": row.get("search_res_url", ""),
-                        "search_res_content": row.get("search_res_content", ""),
-                    }
-
-        elif split == "important_citations":
-            with open(filepath, encoding="utf-8") as f:
-                reader = csv.DictReader(f)
-                for key, row in enumerate(reader):
-                    yield key, {
-                        "parent_paper_title": row.get("parent_paper_title", ""),
-                        "parent_paper_arxiv_id": row.get("parent_paper_arxiv_id", ""),
-                        "citation_shorthand": row.get("citation_shorthand", ""),
-                        "raw_citation_text": row.get("raw_citation_text", ""),
-                        "cited_paper_title": row.get("cited_paper_title", ""),
-                        "cited_paper_arxiv_link": row.get("cited_paper_arxiv_link", ""),
-                        "cited_paper_abstract": row.get("cited_paper_abstract", ""),
-                        "has_metadata": _safe_bool_convert(row.get("has_metadata", "False")),
-                        "is_arxiv_paper": _safe_bool_convert(row.get("is_arxiv_paper", "False")),
-                        "cited_paper_authors": row.get("cited_paper_authors", ""),
-                        "bib_paper_year": _safe_float_convert(row.get("bib_paper_year", "")),
-                        "bib_paper_month": row.get("bib_paper_month", ""),
-                        "bib_paper_url": row.get("bib_paper_url", ""),
-                        "bib_paper_doi": row.get("bib_paper_doi", ""),
-                        "bib_paper_journal": row.get("bib_paper_journal", ""),
-                        "original_title": row.get("original_title", ""),
-                        "search_res_title": row.get("search_res_title", ""),
-                        "search_res_url": row.get("search_res_url", ""),
-                        "search_res_content": row.get("search_res_content", ""),
-                        "arxiv_id": row.get("arxiv_id", ""),
-                        "arxiv_link": row.get("arxiv_link", ""),
-                        "publication_date": row.get("publication_date", ""),
-                        "title": row.get("title", ""),
-                        "abstract": row.get("abstract", ""),
-                        "raw_latex_related_works": row.get("raw_latex_related_works", ""),
-                        "related_work_section": row.get("related_work_section", ""),
-                        "pdf_related_works": row.get("pdf_related_works", ""),
-                        "cited_paper_content": row.get("cited_paper_content", ""),
-                    }