xdz / process.py
marksaroufim's picture
yolo
cd08c07
import os
import pandas as pd
import pyarrow as pa
import pyarrow.parquet as pq
import argparse
import re
import base64
def encode_file(file_path):
"""Encode text files or base64 encode image files."""
if file_path.endswith('.jpg'):
with open(file_path, "rb") as image_file:
return base64.b64encode(image_file.read()).decode('utf-8')
else:
try:
with open(file_path, 'r', encoding='utf-8') as file:
return file.read()
except UnicodeDecodeError as e:
print(f"Error decoding file {file_path}: {e}")
return None
def extract_images(markdown_content):
"""Extract PHOTO_IDs from markdown files and return as a list."""
return re.findall(r'\{\{PHOTO_ID:(\d+)\|WIDTH:\d+\}\}', markdown_content)
def collect_data(directory):
data = {}
image_files = {re.search(r'(\d+)', filename).group(1): filename
for filename in os.listdir(directory) if filename.endswith('.jpg')}
markdown_files = [f for f in os.listdir(directory) if f.endswith('.md') or f.endswith('.sol.md')]
for mfile in markdown_files:
# Adjust the pattern if problem IDs include characters before "sol"
problem_id = re.sub(r'sol$', '', mfile.split('.')[0]) # Strip "sol" from end
if problem_id not in data:
data[problem_id] = {
'Problem ID': problem_id,
'Problem': None,
'in': None,
'Solution': None,
'cpp': None,
'out': None,
'Images': []
}
# Now associate other files with these problem IDs
for filename in os.listdir(directory):
problem_id = re.sub(r'sol$', '', filename.split('.')[0])
if problem_id in data:
file_type = filename.split('.')[-1]
file_path = os.path.join(directory, filename)
content = encode_file(file_path) if not filename.endswith('.jpg') else None
if file_type in ['in', 'out', 'cpp']:
data[problem_id][file_type] = content
if file_type == "md":
if "sol" in filename:
data[problem_id]['Solution'] = content
else:
data[problem_id]['Problem'] = content
image_ids = extract_images(content)
data[problem_id]['Images'] += [image_files[id] for id in image_ids if id in image_files]
data[problem_id]['Images'] = list(set(data[problem_id]['Images'])) # Remove duplicates
return list(data.values())
def create_parquet_file(data, output_file):
df = pd.DataFrame(data)
table = pa.Table.from_pandas(df)
pq.write_table(table, output_file)
def main():
parser = argparse.ArgumentParser(description='Convert dataset to Parquet format.')
parser.add_argument('directory', type=str, help='Directory containing the dataset files.')
parser.add_argument('-o', '--output', type=str, default='output_dataset.parquet', help='Output Parquet file name.')
args = parser.parse_args()
data = collect_data(args.directory)
create_parquet_file(data, args.output)
if __name__ == "__main__":
main()