使用trp模块无法提取Amazon Textract中Polygon的实际坐标值？

我有一份Amazon Textract响应，想要从中提取Polygon的实际坐标值。我的代码如下：

doc = Document(response)
polygons= []

# Iterate over elements in the document
for page in doc.pages:
    for field in page.form.fields:
        polygons.append(field.geometry.polygon)

执行后得到的输出如下：

[[<trp.Polygon at 0x22671e9fb80>,
  <trp.Polygon at 0x22671e9fd90>,
  <trp.Polygon at 0x22671e9fe50>,
  <trp.Polygon at 0x22671e9feb0>],
 [<trp.Polygon at 0x22671e9bd30>,
  <trp.Polygon at 0x22671e9bfa0>,
  <trp.Polygon at 0x22671e9bac0>,
  <trp.Polygon at 0x22671e9bb20>],
 [<trp.Polygon at 0x22672e017f0>,
  <trp.Polygon at 0x22672e018b0>,
  <trp.Polygon at 0x22672e01700>,
  <trp.Polygon at 0x22672e01910>],
...
  <trp.Polygon at 0x22672e02070>],
 [<trp.Polygon at 0x22672dfa1c0>,
  <trp.Polygon at 0x22672dfa8e0>,
  <trp.Polygon at 0x22672dfa9a0>,
  <trp.Polygon at 0x22672dfa7f0>]]

我尝试使用field.geometry.polygon.text但没有效果，请问如何提取Polygon的实际坐标值？

你看到的<trp.Polygon>是trp库封装的多边形对象，它的坐标点存储在points属性里，每个点对象包含x和y两个数值属性。修改代码即可提取坐标：

doc = Document(response)
polygon_coords = []

# 遍历文档元素提取坐标
for page in doc.pages:
    for field in page.form.fields:
        # 获取当前字段的多边形对象
        polygon = field.geometry.polygon
        # 提取每个点的x、y坐标，组成列表
        coords = [(point.x, point.y) for point in polygon.points]
        polygon_coords.append(coords)

执行后，polygon_coords里会存储每个多边形的坐标点列表，格式类似：

[[(x1, y1), (x2, y2), (x3, y3), (x4, y4)], [(x1, y1), ...], ...]

如果需要单独提取横坐标或纵坐标，直接访问对应属性即可，比如point.x获取横坐标，point.y获取纵坐标。

内容的提问来源于stack exchange，提问作者InterestingPenguin80