Method process

processor/mergers.py:30–69

(self, ctx: PipelineContext)

class AlignmentMerger(Merger):
    def process(self, ctx: PipelineContext):
        buffer: List[Image] = ctx.get_buffer()
        horizontal_alignment = ctx.getenum("horizontal_alignment", Alignment.CENTER, Alignment)
        vertical_alignment = ctx.getenum("vertical_alignment", Alignment.CENTER, Alignment)
        background: tuple = ctx.getcolor("background", (255, 255, 255, 0))  # 默认透明
        offsets = json.loads(ctx.get("offsets", "[]"))
        weights = json.loads(ctx.get("weights", "[]"))

        if not buffer:
            return
        # 计算画布大小（所有图片的最大宽高）
        max_width = max(img.width for img in buffer)
        max_height = max(img.height for img in buffer)
        # 创建画布
        canvas = Image.new("RGBA", (max_width, max_height), background)
        # buffer 重排序
        if weights:
            missing_count = len(buffer) - len(weights)
            weights = weights + [0] * missing_count
            combined = zip(weights, buffer)
            sorted_combined = sorted(combined, key=lambda x: x[0], reverse=False)
            buffer = [item for _, item in sorted_combined]

        # 遍历所有图片进行合并
        for i, img in enumerate(buffer):
            # 获取偏移量，索引超出时默认为 (0, 0)
            offset_x, offset_y = offsets[i] if i < len(offsets) else (0, 0)
            offset_x, offset_y = -offset_x, -offset_y
            # 处理水平对齐
            img_x = _calc_offset(img.width, max_width, horizontal_alignment)
            # 处理垂直对齐
            img_y = _calc_offset(img.height, max_height, vertical_alignment)
            # 处理偏移量
            img_x += offset_x
            img_y += offset_y
            # 确保图片是 RGBA 模式以正确处理透明度
            paste_img = img if img.mode == 'RGBA' else img.convert('RGBA')
            # 粘贴图片到画布（使用 alpha 通道作为蒙版）
            canvas.paste(paste_img, (img_x, img_y), paste_img)
        ctx.update_buffer([canvas]).save_buffer(self.name()).success()

    def name(self) -> str:
        return "alignment"