tab complete results in duplicate entries

Here’s a screenshot of completing for a = np.ndarray(3)

They don’t seem to be duplicated in the dir method:

Counter(a.__dir__()).most_common(4)
=> [('__repr__', 1), ('__hash__', 1), ('__str__', 1), ('__lt__', 1)]

Sys info

$ python -c "import IPython; print(IPython.sys_info())"
{'commit_hash': 'ca5443062',
 'commit_source': 'installation',
 'default_encoding': 'UTF-8',
 'ipython_path': '/Users/wbeard/miniconda3/envs/test/lib/python3.6/site-packages/IPython',
 'ipython_version': '6.2.1',
 'os_name': 'posix',
 'platform': 'Darwin-16.7.0-x86_64-i386-64bit',
 'sys_executable': '/Users/wbeard/miniconda3/envs/test/bin/python',
 'sys_platform': 'darwin',
 'sys_version': '3.6.3 | packaged by conda-forge | (default, Dec  9 2017, '
                '16:20:51) \n'
                '[GCC 4.2.1 Compatible Apple LLVM 6.1.0 (clang-602.0.53)]'}

Conda install file

$ cat environment.yml | egrep -v "(^ *#.*|^$)"
name: test
dependencies:
- python=3
- flake8
- pandas
- jupyter
- notebook
- pip:
  - line-profiler
  - jupyter_contrib_nbextensions
  - python-dotenv>=0.5.1

[paulochf]

I’ve just got onto this issue.

Problem comes from ipykernel==4.7.0. I did pip install ipykernel==4.6.1 and it got fixed.

[takluyver]

I think this should be fixed in master by #10969.