practice on the version control and reproducing the experiments
Rules:
Keep the random seed unchanged (15213)
For the data, use Python scripts instead of the cells from the jupyter-notebook to generate data.
Never delete those Python scripts even though the data was deleted.
The input from the Python scripts is the raw DICOM files from /raid/Data.
Try not to use dependent functions from other files. If I have to use it, use it from the master branch.
Be cautious to run operations or modify the functions in the master branch.
Do push the code and backup the data generation scripts for each experiment.
For each function, write the dependency functions on the comments, so that if I have to modify some functions from the master branch, the second thing I must do is to run the unit test and check if the dependent functions are working properly.
Last updated