---
license: mit
dataset_info:
  features:
  - name: instance_id
    dtype: string
  - name: doc_type
    dtype: string
  - name: source
    dtype: string
  - name: url
    dtype: string
  - name: edu_pred_input
    dtype: string
  - name: ground_truth
    dtype: string
  splits:
  - name: test
    num_bytes: 32221618
    num_examples: 248
  download_size: 6102151
  dataset_size: 32221618
configs:
- config_name: default
  data_files:
  - split: test
    path: test/data.parquet
---

# Dataset Card for StructBench

## Dataset Summary

StructBench is a benchmark for evaluating fine-grained document structure analysis. It provides a high-quality test set of 248 documents in diverse formats, including 203 Web pages and 47 PDFs.

To ensure reliable ground truth, all documents were:

- Parsed and sentence-segmented

- Manually annotated by human experts for discourse structure

In addition to the structured annotations, raw Web pages and PDF files are included.

## Dataset Structure

- **test/**: evaluation-only split
  - **data.parquet**: data samples
  - **raw_pdf_files/**: original PDF files
  - **raw_web_htmls/**: original WEB htmls

## Tasks

- Discouse Analysis
- Document Structure Parsing
- Document Understanding

## Usage

Clone the dataset:

```bash
git clone https://huggingface.co/datasets/deeplang-ai/StructBench
```

Or load with Hugging Face datasets library:

```python
from datasets import load_dataset
dataset = load_dataset("deeplang-ai/StructBench", split="test")
```

## Citation

```bibtex
@misc{zhou2025contextedusfaithfulstructured,
      title={From Context to EDUs: Faithful and Structured Context Compression via Elementary Discourse Unit Decomposition}, 
      author={Yiqing Zhou and Yu Lei and Shuzheng Si and Qingyan Sun and Wei Wang and Yifei Wu and Hao Wen and Gang Chen and Fanchao Qi and Maosong Sun},
      year={2025},
      eprint={2512.14244},
      archivePrefix={arXiv},
      primaryClass={cs.CL},
      url={https://arxiv.org/abs/2512.14244}, 
}
```