# pandas: For data manipulation and transformation.
# pyarrow: For reading and writing Parquet files.
import pyarrow.parquet as pq
import os
import json

table = pq.read_table('data/test-00000-of-00001.parquet')
df = table.to_pandas() # DataFrame

print("INICIO", df)

with open("images/images.json") as json_file:
  data = json.load(json_file)
  image_rows = data["imagesToUpload"]

for img in image_rows:
  image_path = f"images/{img["name"]}"
  image_text = img["text"]
  with open(image_path, "rb") as image:
    image_bytes = image.read()
    new_row = {'image': {'bytes': image_bytes, 'path': image_path}, 'text': image_text}
    df = df._append(new_row, ignore_index=True)
  os.remove(image_path)

print("FIN", df)
df.to_parquet('data/test-00000-of-00001.parquet')