---
dataset_info:
  features:
  - name: idx
    dtype: int64
  - name: document_id
    dtype: int64
  - name: document_name
    dtype: string
  - name: chunk_type
    dtype: string
  - name: chunk_id
    dtype: int64
  - name: text
    dtype: string
  - name: metadata
    struct:
    - name: doc_id
      dtype: string
    - name: document_name
      dtype: string
    - name: source_description_file
      dtype: string
    - name: relation_pdf
      dtype: string
    - name: norm_code
      dtype: string
    - name: chunk_id
      dtype: string
    - name: section_paths
      list: string
    - name: title
      dtype: string
    - name: main_chapter
      dtype: string
    - name: token_count
      dtype: int64
    - name: block_start
      dtype: string
    - name: block_end
      dtype: string
    - name: page_start
      dtype: string
    - name: page_end
      dtype: string
    - name: pdf_page_start
      dtype: int64
    - name: pdf_page_end
      dtype: int64
    - name: prev_chunk_id
      dtype: string
    - name: next_chunk_id
      dtype: string
    - name: is_forced_split
      dtype: bool
    - name: split_reason
      dtype: string
    - name: is_forced_oversize
      dtype: bool
    - name: oversize_tokens
      dtype: 'null'
    - name: has_table
      dtype: bool
    - name: has_figure
      dtype: bool
    - name: has_unsplittable_table
      dtype: bool
    - name: has_unsplittable_figure
      dtype: bool
    - name: split_part
      dtype: int64
    - name: split_parts
      dtype: int64
    - name: is_front_matter
      dtype: bool
    - name: chunk_title
      dtype: string
    - name: content_description
      dtype: string
    - name: main_topics
      list: string
    - name: mentioned_docs
      list: string
    - name: equipment_or_components
      list: string
    - name: content_mix
      list: string
    - name: qa_audit_decision
      dtype: 'null'
    - name: qa_potential
      dtype: 'null'
    - name: qa_issues
      list: 'null'
  splits:
  - name: train
    num_bytes: 4869340
    num_examples: 321
  download_size: 2135990
  dataset_size: 4869340
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---