TYP: check_untyped_defs core.indexes.base

[simonjayhawkins]

pandas\core\indexes\base.py:975: error: Argument 1 to "format_object_attrs" has incompatible type "Index"; expected "Sequence[Any]" [arg-type]
pandas\core\indexes\base.py:1612: error: "Index" has no attribute "levels"; maybe "nlevels"? [attr-defined]
pandas\core\indexes\base.py:1613: error: "Index" has no attribute "codes" [attr-defined]
pandas\core\indexes\base.py:3763: error: "Index" has no attribute "levels"; maybe "nlevels"? [attr-defined]
pandas\core\indexes\base.py:3779: error: "Index" has no attribute "codes" [attr-defined]
pandas\core\indexes\base.py:3787: error: "Index" has no attribute "codes" [attr-defined]
pandas\core\indexes\base.py:3790: error: "Index" has no attribute "codes" [attr-defined]
pandas\core\indexes\base.py:3793: error: "Index" has no attribute "levels"; maybe "nlevels"? [attr-defined]
pandas\core\indexes\base.py:3837: error: "Index" has no attribute "codes" [attr-defined]
pandas\core\indexes\base.py:3840: error: "Index" has no attribute "codes" [attr-defined]
pandas\core\indexes\base.py:4807: error: Too many arguments for "get_indexer_non_unique" of "Index" [call-arg]
pandas\core\indexes\base.py:5403: error: Cannot assign to a method [assignment]
pandas\core\indexes\base.py:5404: error: Cannot assign to a method [assignment]
pandas\core\indexes\base.py:5405: error: Unsupported left operand type for < ("Type[Index]") [operator]
pandas\core\indexes\base.py:5406: error: Unsupported left operand type for > ("Type[Index]") [operator]
pandas\core\indexes\base.py:5407: error: Unsupported left operand type for <= ("Type[Index]") [operator]
pandas\core\indexes\base.py:5408: error: Unsupported left operand type for >= ("Type[Index]") [operator]
pandas\core\indexes\base.py:5415: error: Unsupported left operand type for + ("Type[Index]") [operator]
pandas\core\indexes\base.py:5416: error: "Type[Index]" has no attribute "radd" [attr-defined]
pandas\core\indexes\base.py:5417: error: Unsupported left operand type for - ("Type[Index]") [operator]
pandas\core\indexes\base.py:5418: error: "Type[Index]" has no attribute "rsub" [attr-defined]
pandas\core\indexes\base.py:5419: error: "Type[Index]" has no attribute "rpow" [attr-defined]
pandas\core\indexes\base.py:5420: error: Unsupported left operand type for ** ("Type[Index]") [operator]
pandas\core\indexes\base.py:5422: error: Unsupported left operand type for / ("Type[Index]") [operator]
pandas\core\indexes\base.py:5423: error: "Type[Index]" has no attribute "rtruediv" [attr-defined]
pandas\core\indexes\base.py:5425: error: Unsupported left operand type for % ("Type[Index]") [operator]
pandas\core\indexes\base.py:5426: error: "Type[Index]" has no attribute "rmod" [attr-defined]
pandas\core\indexes\base.py:5427: error: Unsupported left operand type for // ("Type[Index]") [operator]
pandas\core\indexes\base.py:5428: error: "Type[Index]" has no attribute "rfloordiv" [attr-defined]
pandas\core\indexes\base.py:5429: error: Unsupported left operand type for divmod ("Type[Index]") [operator]
pandas\core\indexes\base.py:5430: error: "Type[Index]" has no attribute "rdivmod" [attr-defined]
pandas\core\indexes\base.py:5431: error: Unsupported left operand type for * ("Type[Index]") [operator]
pandas\core\indexes\base.py:5432: error: "Type[Index]" has no attribute "rmul" [attr-defined]
pandas\core\indexes\base.py:5449: error: Unsupported operand type for unary - ("Type[Index]") [operator]
pandas\core\indexes\base.py:5450: error: Unsupported operand type for unary + ("Type[Index]") [operator]
pandas\core\indexes\base.py:5451: error: "Type[Index]" has no attribute "abs" [attr-defined]
pandas\core\indexes\base.py:5452: error: "Type[Index]" has no attribute "inv" [attr-defined]
pandas\core\indexes\base.py:5564: error: "Type[Index]" has no attribute "all" [attr-defined]
pandas\core\indexes\base.py:5567: error: "Type[Index]" has no attribute "any" [attr-defined]
pandas\core\indexes\base.py:5576: error: "Type[Index]" has no attribute "all" [attr-defined]
pandas\core\indexes\base.py:5577: error: "Type[Index]" has no attribute "any" [attr-defined]

[jbrockmendel]

AnyArrayLike doesnt satisfy Sequence?

[simonjayhawkins]

not at the moment, xref #28770

[simonjayhawkins]

It looks like the issue here actually is that the docstring states iterable, but len is used in the function code and is the only requirement.

>>> from pandas.io.formats.printing import format_object_attrs
>>>
>>> format_object_attrs(pd.array([1, 2, 3] * 100, dtype="Int64"))
[('dtype', "'Int64'"), ('length', 300)]
>>>
>>> format_object_attrs(pd.Index([1, 2, 3] * 100))
[('dtype', "'int64'"), ('length', 300)]
>>>
>>> format_object_attrs("foo" * 100)
[('length', 300)]
>>>
>>> class MyClass:
...     def __len__(self):
...         return 300
...
>>> format_object_attrs(MyClass())
[('length', 300)]
>>>
>>> format_object_attrs(42)
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "C:\Users\simon\Anaconda3\envs\pandas-1.1.2\lib\site-packages\pandas\io\formats\printing.py", line 536, in format_object_attrs
    if len(obj) > max_seq_items:
TypeError: object of type 'int' has no len()
>>>

I'll type as Sized for now to be as permissive as possible to faciliate re-usability, but AFAICT, format_object_attrs is only ever called with an Index.

(so alternatively could just use Index as type for obj, and maybe reduce the hasattr usages or replace with isinstance(obj, MultiIndex))

[jreback]

can you rebase

[jbrockmendel]

restarted the (clearly-unrelated) failing build. AFAICT on windows LinePlot._is_ts_plot is messing up, but no insight as to why

[jbrockmendel]

LGTM cc @jreback

[jreback]

thanks @simonjayhawkins really like this cleaning & checking of untyped defs, keep em coming!