convert.py

import json
import numpy as np
import torch


def convert_to_braille_unicode(str_input: str, path: str = "./src/utils/number_map.json") -> str:
    with open(path, "r") as fl:
        data = json.load(fl)

    if str_input in data.keys():
        str_output = data[str_input]
    return str_output


def parse_xywh_and_class(boxes: torch.Tensor) -> list:
    """
    boxes input tensor
        boxes (torch.Tensor) or (numpy.ndarray): A tensor or numpy array containing the detection boxes,
            with shape (num_boxes, 6).
        orig_shape (torch.Tensor) or (numpy.ndarray): Original image size, in the format (height, width).
    Properties:
        xyxy (torch.Tensor) or (numpy.ndarray): The boxes in xyxy format.
        conf (torch.Tensor) or (numpy.ndarray): The confidence values of the boxes.
        cls (torch.Tensor) or (numpy.ndarray): The class values of the boxes.
        xywh (torch.Tensor) or (numpy.ndarray): The boxes in xywh format.
        xyxyn (torch.Tensor) or (numpy.ndarray): The boxes in xyxy format normalized by original image size.
        xywhn (torch.Tensor) or (numpy.ndarray): The boxes in xywh format normalized by original image size.
    """

    # copy values from troublesome "boxes" object to numpy array
    new_boxes = np.zeros(boxes.shape)
    new_boxes[:, :4] = boxes.xywh.cpu().numpy()  # first 4 channels are xywh
    new_boxes[:, 4] = boxes.conf.cpu().numpy()  # 5th channel is confidence
    new_boxes[:, 5] = boxes.cls.cpu().numpy()  # 6th channel is class which is last channel

    # sort according to y coordinate
    new_boxes = new_boxes[new_boxes[:, 1].argsort()]

    # find threshold index to break the line
    y_threshold = np.mean(new_boxes[:, 3]) // 2
    boxes_diff = np.diff(new_boxes[:, 1])
    threshold_index = np.where(boxes_diff > y_threshold)[0]

    # cluster according to threshold_index
    boxes_clustered = np.split(new_boxes, threshold_index + 1)
    boxes_return = []
    for cluster in boxes_clustered:
        # sort according to x coordinate
        cluster = cluster[cluster[:, 0].argsort()]
        boxes_return.append(cluster)

    return boxes_return


def arrange_braille_to_2x3(box_classes: list) -> list:
    """
    将检测到的盲文字符类别数组转为 2x3 点阵格式。
    :param box_classes: 检测到的盲文字符类别列表 (长度必须是6的倍数)
    :return: 2x3 盲文点阵列表
    """
    # 检查输入长度是否为6的倍数
    if len(box_classes) % 6 != 0:
        raise ValueError("输入的盲文字符数组长度必须是6的倍数")

    braille_2x3_list = []

    # 每次取6个字符并将它们排成2x3格式
    for i in range(0, len(box_classes), 6):
        # reshape为3x2矩阵然后转置为2x3矩阵
        braille_char = np.array(box_classes[i:i + 6]).reshape(3, 2).T
        braille_2x3_list.append(braille_char)

    return braille_2x3_list