import os
from paddleocr import TextRecognition

ocr = TextRecognition(
    model_name="PP-OCRv5_server_rec"
    # model_dir="output/PP-OCRv5_server_rec_vin"
)

with open("train_data/val.txt", "r") as label_file:
    lines = label_file.readlines()
    total = len(lines)
    matches, mismatches = 0, 0
    for idx, line in enumerate(lines, start=1):
        file_name, label = line.split("\t")
        label = label.strip()
        path = os.path.join("train_data", os.path.join("images", file_name))
        result = ocr.predict(path)
        print(f"{idx}/{total} ", end="")
        if result[0]["rec_text"] and result[0]["rec_text"] == label:
            print(f"match {label}")
            matches += 1
        else:
            print(f"mismatch {label}")
            mismatches += 1

print(f"{matches} matches of {total}")
print(f"{mismatches} mismatches of {total}")