Christophe Guillon authored 9 months ago and Maxence Naud committed 8 months ago

For constructors, now manage correctly:
- np array 0-rank arrays scalar,
- python int|float scalars,
- np scalars np.dtype() (different from 0-rank array scalars),
- contruction with dims=..., now a kw only argument,
- np arrays of rank >= 1 as before.

For setters, now manage correctly conversion without
loss of precision and do not raise in case of overflow as we
would expect with C++ interface.
Added support in ctors/get/set for some missing types, now
support [u]int(8|16|32|64).

Christophe Guillon authored 9 months ago and Maxence Naud committed 9 months ago

Add python enum values for all defined dtypes as per the
table of dtype names (lower case of the dtype names).
Add __str__() binding such that str(tensor.dtype()) return
bare enumeration names, suitable for comparison with numpy
dtype names.

Christophe Guillon authored 9 months ago and Maxence Naud committed 9 months ago

Define a Tensor with undefined dims by explicitly setting mSize to 0.
Provide undefined() method which must be used instead of empty()
to test whether a Tensor as associated dimensions and can be used
for forwarding dimensions downward.
This disambiguates from defined scalar Tensor which actually have
empty dimensions, but size of 1.
Note that as soon as a Tensor with undefined dimensions is resized,
it's dimensions are permanently defined (though may still change).

This change solves issues in operators forwardDims() when
input values are scalar, i.e. !undefined() and empty().

- Fix datatype call
- Change 'DataType' to 'dtype' for Python interface

- [Add] Attributes: __str__(), __repr__() and dict() methods
- [Add] GraphView: __repr__(), __len__() methods
- [Add] Node: __repr__() method
- [Add] Operator: __repr__() method and attr property
- improve Tensor __str__ method if it is empty
- change DataType to lower case for python

- add Tensor operators +,-,*,/
- add __repr__ funciton to Tensor
- add operator constructor to each python binding
- fix identity_op Base class declaration in binding

- add Tensor operators +,-,*,/
- add __repr__ funciton to Tensor
- add operator constructor to each python binding
- fix identity_op Base class declaration in binding

Fix numpy -> tensor with latest tensorImpl changes (copyFromHost instead of setRawPtr) + Use of std::int32_t instead of in and std::int64_t instead of long to fit Data.hpp convention.