Umi-OCR项目（1）

最近接触到了一个项目，我在想能不能做出点东西出来。
目标：识别一张带表格的图片，要求非表格内容和表格内容都要识别得很好，并且可视化输出为word文档。
下面是第一步的测试代码，测试是否能够调用ocr能力。

import requests
import jsonurl = "http://127.0.0.1:1224/api/ocr"
data = {"base64": "iVBORw0KGgoAAAANSUhEUgAAAC4AAAAXCAIAAAD7ruoFAAAACXBIWXMAABnWAAAZ1gEY0crtAAAAEXRFWHRTb2Z0d2FyZQBTbmlwYXN0ZV0Xzt0AAAHjSURBVEiJ7ZYrcsMwEEBXnR7FLuj0BPIJHJOi0DAZ2qSsMCxEgjYrDQqJdALrBJ2ASndRgeNI8ledutOCLrLl1e7T/mRkjIG/IXe/DWBldRTNEoQSpgNURe5puiiaJehrMuJSXSTgbaby0A1WzLrCCQCmyn0FwoN0V06QONWAt1nUxfnjHYA8p65GjhDKxcjedVH6JOejBPwYh21eE0Wzfe0tqIsEkGXcVcpoMH4CRZ+P0lsQp/pWJ4ripf1XFDFe8GHSHlYcSo9Es31t60RdFlN1RUmrma5oTzTVB8ZUaeeYEC9GmL6kNkDw9BANAQYo3xTNdqUkvHq+rYhDKW0Bj3RSEIpmyWyBaZaMTCrCK+tJ5Jsa07fs3E7esE66HzralRLgJKp0/BD6fJRSxvmDsb6joqkcFXGqMVVFFEHDL2gTxwCAaTabnkFUWhDCHTd9iYrGcAL1ZnqIp5Vpiqh7bCfua7FA4qN0INMcN1+cgCzj+UFxtbmvwdZvGIrI41JiqhZBWhhF8WxorkYPpQwJiWYJeA3rXE4hzcwJ+B96F9zCFHC0FcVegghvFul7oeEE8PvHeJqC0w0AUbbFIT8JnEwGbPKcS2OxU3HMTqD0r4wgEIuiKJ7i4MS16+og8/+bPZRPLa+6Ld2DSzcAAAAASUVORK5CYII=",# 可选参数示例"options": {"data.format": "text", # list}
}
headers = {"Content-Type": "application/json"}
data_str = json.dumps(data)
response = requests.post(url, data=data_str, headers=headers)
response.raise_for_status()
res_dict = json.loads(response.text)
print(res_dict)