dataset_json_generate.py

import pandas as pd
import json

df = pd.read_excel('road accident dataset.xlsx',sheet_name="Merged")

# print(df)
# print(type(df))

# for index, row in df.iterrows():
#     # Access row data using row['column_name'] or row[column_index]
#     print(row)
#     print(type(row))

# for column_name, column_data in df.items():
#     print(column_name, column_data)

# print(df['article_content'])

# json_dataset = [
# ]

json_dataset = {
    "data" : []
}

for index, row in df.iterrows():
    # Iterate through each column in the row
    text_input = ""
    json_output = {}
    for column_name, cell_value in row.items():
        if column_name == "article_content":
            text_input=cell_value
        else:
            json_output[column_name] = cell_value
        print(f'Row: {index}, Column: {column_name}, Value: {cell_value}')
    
    json_dataset["data"].append({
        "text_input" : text_input,
        "output" : str(json_output)
    })

    # obj = {
    #     "inputs" : [],
    #     "outputs" : []
    # }

    # obj["inputs"].append(text_input)
    # obj["outputs"].append(str(json_output))
    # json_dataset.append(obj)

print(json_dataset)

with open("dataset.json", "w") as json_file:
    json.dump(json_dataset, json_file, indent=4)