use operator on a class method fails

[jenda1]

I tried to use aiostream.core.operator on a class method:

from aiostream.core import operator

class test(object):
    @operator
    async def testit(self):
        pass

but it fails with error:

$ python3.7 test.py
Traceback (most recent call last):
  File "test.py", line 3, in <module>
    class test(object):
  File "test.py", line 4, in test
    @operator
  File "/home/jlana/Dropbox/vyuka2/venv/lib/python3.7/site-packages/aiostream/core.py", line 333, in operator
    return decorator if func is None else decorator(func)
  File "/home/jlana/Dropbox/vyuka2/venv/lib/python3.7/site-packages/aiostream/core.py", line 279, in decorator
    init.__signature__ = signature.replace(parameters=new_parameters)
  File "/usr/lib/python3.7/inspect.py", line 2843, in replace
    return_annotation=return_annotation)
  File "/usr/lib/python3.7/inspect.py", line 2788, in __init__
    raise ValueError(msg)
ValueError: duplicate parameter name: 'self'

As workaround it is possible to use different name and call the operator with the class instance:

class test(object):
    n = 123

    @operator
    async def testit(myself):
        yield myself.n

async def main():
    t = test()
    await stream.print(t.testit(t))

[vxgmichel]

Hi @jenda1, thank you for creating this issue.

The problem here is that the decorated generator becomes a Stream subclass, in order to provide support for pipe-lining (|), slicing ([]) and concatenation (+). That means testit is not a method and this is why you have to pass the t instance explicitly in your workaround.

Instead, you could consider this alternative:

@operator
async def stream_from_test_instance(test):
    yield test.n


class test:
    n = 123

    def testit(self):
        return stream_from_test_instance(self)

Or, since streams are reusable:

class test:

    def __init__(self, n):
        self.n = n
        self.stream = stream_from_test_instance(self)

    def testit(self):
        return self.stream

The error message you got isn't really explaining what's happening though. Let's try to detect self or cls arguments and raise an exception with a proper explanation.

Hope that helps, thanks again.

[jenda1]

ok, thanks for explanation. - Jenda po 27. 8. 2018 v 12:07 odesílatel Vincent Michel <notifications@github.com> napsal:

…

Hi @jenda1 <https://github.com/jenda1>, thank you for creating this issue. The problem here is that the decorated generator becomes a Stream subclass, in order to provide support for pipe-lining (|), slicing ([]) and concatenation (+). That means testit is *not* a method and this is why you have to pass the t instance explicitly in your workaround. Instead, you could consider this alternative: @operatorasync def stream_from_test_instance(test): yield test.n class test: n = 123 def testit(self): return stream_from_test_instance(self) Or, since streams are reusable: class test: def __init__(self, n): self.n = n self.stream = stream_from_test_instance(self) def testit(self): return self.stream The error message you got isn't really explaining what's happening though. I think I'll try to detect self or cls arguments and raise an exception with a proper explanation. Hope that helps, thanks again. — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#23 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ARJE-rfymZ_eAE27ukYs50GsEBw9eUO8ks5uU8TGgaJpZM4WLsYA> .

[vxgmichel]

Let's try to detect self or cls arguments and raise an exception with a proper explanation.

Done ! (see PR #24)