TODO.txt

TODO:
allow limiting of min bound by confidence score or other criteria (and test)
joint_locs = self.joint_info[:self.dims]
        if min_conf is not None:
            mask = self.joint_confs < min_conf
            mask = np.expand_dims(mask, axis=2)
            mask = np.concatenate((mask,) * self.dims, axis=2)
            joint_locs[mask] = np.nanmax(joint_locs)
        return self.__get_bound(joint_locs, np.nanmin)

TODO elsewhere:
 - check and clip joint confidences

        # check all joint confs are between 0 and 1, and clip if not
        if np.max(self.joint_confs) >= 1.0 or np.min(self.joint_confs) <= 0.0:
            logger.debug(f"Joint confidence greater than 1.0 or less than 0.0: {self.joint_confs}")
            self.joint_confs = np.clip(self.joint_confs, 0, 1)

 - convert python lists to numpy arrays with specific nan values 

     def __numpy_float_converter(self, arr, dims=1, nan_vals=['']):
        ''' Takes in nested python lists, replaces specified values with NaNs,
        and returns a numpy array of dtype float32.

        Args:
            arr (list): an array represented as nested python lists
            dims (int): number of dimensions in the array
                Valid values are 1 and 2, with default=1
            nan_vals (list): values that should be replaced with NaN.
                Default is empty string only'''
        if type(arr) is np.array:
            return arr
        if dims == 1:
            arr = ['NaN' if e in nan_vals else e for e in arr]
        elif dims == 2:
            for i, entry in enumerate(arr):
                entry = ['NaN' if e in nan_vals else e for e in entry]
                arr[i] = entry
        elif dims == 3:
            for i, row in enumerate(arr):
                for j, col in enumerate(row):
                    col = ['NaN' if e in nan_vals else e for e in col]
                    arr[i][j] = col
        else:
            logger.warn("Only 1D and 2D arrays supported")
        return np.array(arr, dtype=np.float32)