parse_res.py

import json
import numpy as np
import pandas as pd
import os
import subprocess

file_path = ""

if not file_path:
    file_path_answer = input("file to parse: ")
    file_path = file_path_answer


def parse_res(file_path: str) -> str:
    with open(file_path) as f:
        text = f.read()
    data = json.loads(text)
    print(data["input_path"])
    # I think this whole thing assumes we only have one grouping of data that the ocr outputs
    rec_texts = data["rec_texts"]
    print(rec_texts)

    arr1 = np.array(rec_texts)
    arr2 = arr1.reshape(arr1.size, 1)
    print(arr2)
    # df = pd.DataFrame(arr2).T
    # df = pd.DataFrame(arr1).T
    df = pd.DataFrame({"col1": arr1})

    name, ext = os.path.splitext(file_path)
    df.to_excel(excel_writer = f"{name}.xlsx", columns=['col1'])

    print("done.")
    return f"{name}.xlsx"

val = parse_res(file_path)
print (val)

open_answer = input("do you want to open the image and the spreadsheet? (Y/N): ")
if open_answer == "Y" or open_answer == "y":
    name, ext = os.path.splitext(file_path)
    base = name.rsplit("res", 1)[0]
    img_name = f"{base}ocr_res_img.png"
    subprocess.call(("open", f"{name}.xlsx"))
    subprocess.call(("open", img_name))