Type hints in mmif.Mmif

[marcverhagen]

In mmif.Mmif there is a method with the following signature

def get_documents_by_type(self, doc_type: Union[str, DocType]) -> List[Document]:

This always passed pytype tests fine until I used the above method in the summarizer, at which time I got the following error:

[13:49:55] (346-summarizer)> pytype mmif/utils/summarizer/
Computing dependencies
Analyzing 6 sources with 16 local dependencies
ninja: Entering directory `.pytype'
[5/6] check mmif.utils.summarizer.summary
FAILED: [code=1] /Users/marc/Desktop/projects/clams/code/clamsproject/mmif-python/.pytype/pyi/mmif/utils/summarizer/summary.pyi
/usr/local/opt/python@3.11/bin/python3.11 -m pytype.main --imports_info /Users/marc/Desktop/projects/clams/code/clamsproject/mmif-python/.pytype/imports/mmif.utils.summarizer.summary.imports --module-name mmif.utils.summarizer.summary --platform darwin -V 3.11 -o /Users/marc/Desktop/projects/clams/code/clamsproject/mmif-python/.pytype/pyi/mmif/utils/summarizer/summary.pyi --analyze-annotated --nofail --quick /Users/marc/Desktop/projects/clams/code/clamsproject/mmif-python/mmif/utils/summarizer/summary.py
/Users/marc/Desktop/projects/clams/code/clamsproject/mmif-python/mmif/utils/summarizer/summary.py:151:16: error: in video_documents: Function Mmif.get_documents_by_type was called with the wrong arguments [wrong-arg-types]
         Expected: (self, doc_type: Union[mmif.vocabulary.document_types.DocumentTypes, str])
  Actually passed: (self, doc_type: mmif.vocabulary.base_types.DocumentTypesBase)

        return self.mmif.get_documents_by_type(DocumentTypes.VideoDocument)
               ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

It makes sense to me that using DocumentTypes is frowned upon by pytype and replacing the signature with the following got rid of the error:

def get_documents_by_type(self, doc_type: Union[str, DocumentTypes.Document]) -> List[Document]:

But then the CLI script crashes on some typing issue:

12:06:59] (346-summarizer)> mmif -h
Traceback (most recent call last):
  File "/usr/local/bin/mmif", line 5, in <module>
    from mmif.__init__ import cli
  File "/Users/marc/Desktop/projects/clams/code/clamsproject/mmif-python/mmif/__init__.py", line 12, in <module>
    from mmif.serialize import *
  File "/Users/marc/Desktop/projects/clams/code/clamsproject/mmif-python/mmif/serialize/__init__.py", line 3, in <module>
    from .mmif import *
  File "/Users/marc/Desktop/projects/clams/code/clamsproject/mmif-python/mmif/serialize/mmif.py", line 146, in <module>
    class Mmif(MmifObject):
  File "/Users/marc/Desktop/projects/clams/code/clamsproject/mmif-python/mmif/serialize/mmif.py", line 494, in Mmif
    def get_documents_by_type(self, doc_type: Union[str, DocumentTypes.Document]) -> List[Document]:
                                              ~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/usr/local/Cellar/python@3.11/3.11.6/Frameworks/Python.framework/Versions/3.11/lib/python3.11/typing.py", line 355, in inner
    return cached(*args, **kwds)
           ^^^^^^^^^^^^^^^^^^^^^
  File "/usr/local/Cellar/python@3.11/3.11.6/Frameworks/Python.framework/Versions/3.11/lib/python3.11/typing.py", line 481, in __getitem__
    return self._getitem(self, parameters)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/usr/local/Cellar/python@3.11/3.11.6/Frameworks/Python.framework/Versions/3.11/lib/python3.11/typing.py", line 694, in Union
    parameters = tuple(_type_check(p, msg) for p in parameters)
                 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/usr/local/Cellar/python@3.11/3.11.6/Frameworks/Python.framework/Versions/3.11/lib/python3.11/typing.py", line 694, in <genexpr>
    parameters = tuple(_type_check(p, msg) for p in parameters)
                       ^^^^^^^^^^^^^^^^^^^
  File "/usr/local/Cellar/python@3.11/3.11.6/Frameworks/Python.framework/Versions/3.11/lib/python3.11/typing.py", line 190, in _type_check
    if arg in (Any, LiteralString, NoReturn, Never, Self, TypeAlias):
       ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/Users/marc/Desktop/projects/clams/code/clamsproject/mmif-python/mmif/vocabulary/base_types.py", line 193, in __eq__
    return self._eq_internal(other, self.fuzzy_eq and other.fuzzy_eq)
                                                      ^^^^^^^^^^^^^^
AttributeError: type object 'Any' has no attribute 'fuzzy_eq'

The only thing I could come up with to make all this go away is to make the type hint useless.

def get_documents_by_type(self, doc_type: Any) -> List[Document]:

For now that is what I am doing in the summarizer branch, but it does not feel right. Note that this may also become an issue with the get_documents_locations and get_document_locations methods.

[marcverhagen]

The issue seems resolved more satisfactory by using DocumentsTypeBase for the type hint.

Will do some testing and then close this issue.