BUG: Series dot product __rmatmul__ doesn't allow matrix vector multiplication

[minggli]

• closes BUG: np.array @ pd.Series swaps arguments #21530
• tests added / passed
• passes git diff upstream/master -u -- "*.py" | flake8 --diff
• whatsnew entry

the new __rmatmul__ implementation in Series seems to be missing matrix vector multiplication case as raised in the issue in question. Only inner product of two vectors was supported in __rmatmul__ method.

This PR uses the same implementation in DataFrame to add support in Series and test case.

matmul operator (@, @=) was added in python 3.5 in https://www.python.org/dev/peps/pep-0465/

class A(object):
    def __matmul__(self, other):
        print('__matmul__ is called in A.')

class B(object):
    def __rmatmul__(self, other):
        print('__rmatmul__ is called in B.')

A() @ B()
B() @ A()
del A.__matmul__
A() @ B()

>>> A() @ B()
__matmul__ is called in A.
>>> B() @ A()
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
TypeError: unsupported operand type(s) for @: 'B' and 'A'
>>> del A.__matmul__
>>> A() @ B()
__rmatmul__ is called in B.

[gfyoung]

@minggli : Looks like a good start! Can we also add a whatsnew too?

[codecov]

Codecov Report

Merging #21578 into master will not change coverage.
The diff coverage is 100%.

@@           Coverage Diff           @@
##           master   #21578   +/-   ##
=======================================
  Coverage    91.9%    91.9%           
=======================================
  Files         153      153           
  Lines       49549    49549           
=======================================
  Hits        45539    45539           
  Misses       4010     4010

Flag	Coverage Δ
#multiple	90.3% <100%> (ø)	⬆️
#single	41.78% <0%> (ø)	⬆️

Impacted Files	Coverage Δ
pandas/core/series.py	94.19% <100%> (ø)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 028c9c0...d5a0d28. Read the comment docs.

[minggli]

raised an issue #21581 regarding error message when __rmatmul__ is called in dataframe and series. as dot method is implemented treating dataframe or series as left. when __rmatmul__ is called the error message doesn't work well.

this is an existing error.

need advice. fix it in the same PR and how to do that or should we handle it separately? @jreback

[jreback]

what about the np.array @ Series ? is that always calling matmul (and not rmatmul)?

[jreback]

for both 1d and 2d. the OP example was 2 d, can you add that as well.

[minggli]

np.array @ Series should call __rmatmul__, so either list or np.array should be the same, i.e. calling __rmatmul__
np.array @ DataFrame however only calls np.array.__matmul__

actioned.

[jreback]

for both 1d and 2d. the OP example was 2 d, can you add that as well.

[jreback]

thanks @minggli very nice

[minggli]

Thanks @jreback and @gfyoung