Issues installing mecab-python3 using pip

Question

Today I've been attempting-- and failing-- to install this guy (MeCab library for Python 3.5+) for the sake of building a simple personalized Japanese readability analysis tool (as a learner of the language and data nerd).

Of course, the first thing I tried was the simple pip install mecab-python3, to see this wall of text (process and errors). The error message is (probably) mojibaked which makes it impossible to know what actually happened.

I repeatedly googled important snippets of it that were readable in search of an answer to the problem, but no results resembled exactly what my problem appears to be.

I then tried downloading the source from pip and manually building it. It was doing fine until the final python setup.py build command, which resulted in:

running build_py
creating build
creating build\lib.win32-3.8
creating build\lib.win32-3.8\MeCab
copying src\MeCab\__init__.py -> build\lib.win32-3.8\MeCab
running build_ext
error: [WinError 2] 指定されたファイルが見つかりません。

If you can't read Japanese, basically what that says is "the specified file could not be found." Not sure which file this refers to, and what to do about it.

I'm using Windows 10.

My apologies if the answer should be obvious. I know what it's like to be good at something and watch other people be inept at it, so I'm sorry in advance for the potential frustration.

Thank you!

Answer 1

I'm the maintainer of mecab-python3 and fugashi. Both of them have Windows wheels now - they were released for mecab-python3 a few days after you posted this - so installing them should work now, unless you're using mecab-python3 and Anaconda (there's some DLL issue I haven't tracked down).

If you still have trouble with the software please file an issue on Github.

Answer 2

The creators of the mecab extension did not build binary wheels for windows. However may You can try this instead of that one: Fugashi's Cython MeCab wrapper for easy Japanese tokenization.