py-tps 0.1.7

安装

安装的主要方式是从pip:

pip install py_tps

源代码安装通过诗歌进行管理。https://pypi.org/project/poetry/ 来源可以从https://gitlab.com/ryan-feather/py_tps

要安装，首先安装诗歌。

conda用户的注释

要将conda环境与poetry一起使用，必须首先将poetry配置为使用conda环境和 不创建新的virtualenv环境。示例：

poetry config settings.virtualenvs.path $CONDA_ENV_PATH
poetry config settings.virtualenvs.create 0

然后运行：

poetry install
poetry build
cd dist
pip install *.whl

用法

from py_tps import TPSFile, TPSImage, TPSCurve, TPSPoints
import numpy as np

# construct and write
# numpy is used for numeric data
points = TPSPoints(np.asarray([[33,44],
                               [44,45]]))
curve = TPSCurve(points)
image = TPSImage('T1.JPG', curves=[curve], id_number=0, comment="This is a test", scale=0.0045)
tps_file = TPSFile([image]) # can have many images, but here we do just one
tps_file.write_to_file('TestFile.TPS')

创建包含以下内容的文件：

LM=0
CURVES=1
POINTS=2
33 44
44 45
IMAGE=T1.JPG
ID=0
COMMENT=This is a test
SCALE=0.0045

现在阅读：

tps_file_in =TPSFile.read_file('TestFile.TPS')
# now can access images list, and so on.  Attributes have the same names as they do in the objects above.
print(tps_file.images[0].image) # T1.JPG
print(tps_file.images[0].curves[0].tps_points.points) # [[33,44],[44,45]]

地标可以设置，而不是曲线。如果我们这样做：

image = TPSImage('T1.JPG', landmarks=points, id_number=0, comment="This is a test", scale=0.0045)

它创建包含以下内容的文件：

LM=2
33 44
44 45
IMAGE=T1.JPG
ID=0
COMMENT=This is a test
SCALE=0.0045