Week 8

Yayy, I passed the midterm evaluation …. Hello everyone, hope this blog finds you well.

We have reached into week 8 of the project and I have passed the first evaluation. Although I am satisfied with my work upto the midterm evaluation, I cannot become complacent. I have to keep improving and upgrading the use cases of my project.

Aim :

My aim for this week is to resolve bugs we had not tackled in the past weeks.

Work :

I made the following pull requests this week -

• PR#35 : Add Resize

  a = np.array([[1, 2], [3, 4]])
  a.resize([1, 4])
  assert_equal(a, np.array([[1, 2, 3, 4]]))
  

  As of now, our implementation does not preserve elements if expected size is not equal to the current size. You can have a look at this issue to know the reason - https://github.com/xtensor-stack/xtensor/issues/1445

• PR#36 : Add proper spaces for indentation
• PR#37 : Add dtypes as string attributes of numpy module

• PR#38 : Add the tolist function

  a = np.array([1.1, 2.2, 3.3])
  assert a.tolist() == [1.1, 2.2, 3.3]
  

• PR#39 : Added dtype argument for array creation

• PR#40 : Change default integer precision to int64

  This change has been made in the latest versions of numpy, where maximum precision for integer type is increased from 32 bits to 64 bits.

• PR#41 : Register all dtypes With this pull request we have started to support the following dtypes -

  #Boolean
  > pkpy::bool_
    
  #Integer
  > pkpy::int8
  > pkpy::int16
  > pkpy::int32
  > pkpy::int64
  > pkpy::int_
    
  #Floating
  > pkpy::float32
  > pkpy::float64
  > pkpy::float_
  

  Now the only dtypes left to support are, unsigned and complex datatypes.

• PR#42 : Correct output of np.shape

  With this pull request we correctly return a py::tuple instead of a std::vector.

Future Work :

In this upcoming week, we need to reduce disk space required for the project and complete on all the numpy dtypes.