Allow readonly arrays in numba jit signature

[jvdd]

This PR is a continuation of the work contributed in PR #13

Contributions of this PR:

• allow readonly arrays in the numba jit signature
  -> this was an issue for sample_entropy and detrended_fluctuation
• minor refactoring of the tests - moving duplicate code to a utils.py file
• adding 2 additional tests that test whether non-writable (i.e., readonly) arrays can be used as input for a range of datatypes

Other contribution:

• set nopython to True for the _higuchi_fd function - this was the only @jit decorated function for which this flag was not set.

---

Illustration of the issue that this PR tackles:

In [1]: import antropy as ant; import numpy as np

In [2]: ant.__version__
Out[2]: '0.1.5'

In [3]: x = np.random.randn(3_000).astype("float32")
   ...: x.flags.writeable = False
   ...: ant.detrended_fluctuation(x)
Out[3]: 0.5552594723035992

In [4]: x = np.random.randn(3_000).astype("float64")
   ...: x.flags.writeable = False
   ...: ant.detrended_fluctuation(x)
---------------------------------------------------------------------------
TypeError
    ...
TypeError: No matching definition for argument type(s) readonly array(float64, 1d, C)

In [5]: x = np.random.rand(3_000).astype("float32")
   ...: x.flags.writeable = False
   ...: ant.sample_entropy(x)
Out[5]: 2.17591547251706

In [6]: x = np.random.rand(3_000).astype("float64")
   ...: x.flags.writeable = False
   ...: ant.sample_entropy(x)
---------------------------------------------------------------------------
TypeError
    ...
TypeError: No matching definition for argument type(s) readonly array(float64, 1d, C), int64, float64

[jvdd]

Should be ready for review! @raphaelvallat :)

[raphaelvallat]

Beautiful work 🚀 Merging now! Thanks again :)