This script converts pdf "factuur" from https://www.ns.nl/mijnns#/betaaloverzicht to JSON array with date, transporter, discount, from, to and price. Hacky, but it works. Station parsing needs to be done better.

ns_factuur_extract.py

import pytesseract  # type: ignore
from PIL import Image  # type: ignore
from time import sleep
import re
import fitz  # type: ignore
import json
import glob

for p in glob.glob("*.pdf"):
    pdf_file = fitz.open(p)

    zoom_size = 3

    mat = fitz.Matrix(zoom_size, zoom_size)

    page = pdf_file.load_page(1)
    pix = page.get_pixmap(matrix=mat)
    pix.save("ss.png")
    pdf_file.close()

    # Open an image
    image = Image.open("ss.png")  # Update with your image path

    # Get image dimensions
    width, height = image.size

    images = []

    step_size = 120 * zoom_size

    entries = []

    last_height = 0
    for i in range(1910 * zoom_size, height * 10, step_size):
        n = image.crop((0, last_height / 10, width, i / 10))
        # n.save("ss.png")
        images.append(n)
        last_height = i
        # sleep(0.5)

    print("Converted to images", len(images))

    for image in images:
        text = pytesseract.image_to_string(image, lang="eng").strip()

        # If text does not start with date (DD-MM-YYYY), ignore it
        if not re.match(r"\d{2}-\d{2}-\d{4}", text):
            continue
        if "Fiets" in text:
            continue
        if len(text) < 20:
            continue
        print("Matched", text)

        # Get date
        date = re.findall(r"\d{2}-\d{2}-\d{4}", text)[0]
        print("Date", date)

        transporter = text.split(" ")[1]
        print("Transporter", transporter)

        # Next is the discount, there are two variants: "20% korting in de spits" and "40% korting buiten de spits"
        discount = text.split(" ")[2]
        print("Discount", discount)

        AVAILABLE_STATIONS = [
            "Oss West",
            "'s-Hertogenbosch",
            "Tilburg",
            "Eindhoven Strijp-S",
            "Oss",
            "Eindhoven Centraal",
        ]

        # Find stations in text using regex
        stations = re.findall(r"(?=(" + "|".join(AVAILABLE_STATIONS) + r"))", text)
        print("Stations", stations)

        # After the euro sign, there is the price
        price = float(text.split("€")[1].strip().replace(",", "."))
        print("Price", price)

        obj = {
            "date": date,
            "transporter": transporter,
            "discount": discount,
            "from": stations[0],
            "to": stations[1],
            "price": price,
        }
        print(obj)
        entries.append(obj)

    current_entries = open("entries.json", "r").read()
    current_entries = json.loads(current_entries)
    # check if array is empty
    if len(current_entries) == 0:
        current_entries = []  # type: ignore

    entries = current_entries + entries  # type: ignore

    with open("entries.json", "w") as f:
        f.write(json.dumps(entries))