preprocess.py

import numpy as np
import pandas as pd

df = pd.read_json("data_set_50.json")

meme_attribute = {}
meme_filename = []
meme_list = []

for col in df.columns:
    name = df.loc["filename", col]
    name = name.replace(".jpg", "")
    name = name.replace("High-Quality-", "")
    attribute_tmp = df.loc["file_attributes", col]
    if attribute_tmp == {}:
        continue
    attribute_str = attribute_tmp['image_label'].strip().strip('"')
    cleaned_attribute = attribute_str.replace(',', '')
    file_name = df.loc["filename", col]
    meme_attribute[name] = cleaned_attribute
    meme_filename.append(file_name)
    meme_list.append(name)