create_simple_dataset.py

#!/usr/bin/env python3
import json
import os
import shutil
from tqdm import tqdm

# First, let's clean up our previous attempts
if os.path.exists('default'):
    shutil.rmtree('default')
if os.path.exists('default-fixed'):
    shutil.rmtree('default-fixed')

# Create a simple dataset with just 10 examples to test
os.makedirs('simple_dataset', exist_ok=True)

# Directory containing the PDF files
data_dir = "data"

# Get all PDF files (just the first 10 for testing)
pdf_files = sorted(os.listdir(data_dir))[:10]

# Create a simple metadata.jsonl file
with open('simple_dataset/metadata.jsonl', 'w') as f:
    for pdf_file in tqdm(pdf_files, desc="Creating metadata"):
        file_path = os.path.join(data_dir, pdf_file)
        if os.path.isfile(file_path) and pdf_file.endswith('.pdf'):
            file_size = os.path.getsize(file_path)
            
            # Create a simple metadata entry
            metadata = {
                "file_name": pdf_file,
                "file_path": file_path,
                "file_size": file_size,
                "content_type": "application/pdf"
            }
            
            f.write(json.dumps(metadata) + '\n')

# Create a simple README.md file for the dataset
with open('simple_dataset/README.md', 'w') as f:
    f.write("""# JFK Document Collection Sample

This is a sample of the JFK Document Release 2025 collection. The full collection contains over 2,000 declassified documents related to the assassination of President John F. Kennedy.

## Files

The documents are PDF files named according to the National Archives naming convention (e.g., `104-10003-10041.pdf`).

## Usage

You can access the full collection at: https://huggingface.co/datasets/varunh/jfk-files-2025

For a more user-friendly interface, visit: https://huggingface.co/datasets/varunh/jfk-files-2025/blob/main/index.html
""")

print("Created simple dataset with 10 examples")