Remove file loading for critical values

[prokolyvakis]

What?

• Add a Python class Consts that wraps up all the constants used by the diptest.py module.
  • The Consts class is implemented in a way that forces the variables to stay constant.
• Replace the np.loadtxt call and directly adding the aforementioned array in the Consts class.

Why?

Currently, the _CRIT_VALS numpy array is being read by calling the np.loadtxt method. Nevertheless, this adds a redundant call to the file system; in my laptop the reading op takes ~0.56 ms compared to the ~40ns when we directly add the array inside a Python module and import it.

How?

• Remove dip_crit.txt.
• Remove all the consts definitions from diptest.py module
• Add a consts.py module that wraps up all the constants defined in the diptest.py module.
  • Add the values of the _CRIT_VALS explicitly inside the module.
  • Restrict all the class variables from being changed.

Testing?

The implementation has been tested locally in a Darwin platform and all the unittests pass with success.

[RUrlus]

Hi @prokolyvakis, good idea! Thanks for the PR

[prokolyvakis]

Hi @RUrlus! Thanks for accepting the PR!