VimPlug/jedi-fork
1
0
Fork 0
forked from VimPlug/jedi
Code Pull Requests Activity

Fix _get_numpy_doc_string_cls: use cache

Browse Source 
I've noticed that Jedi tries to import numpydoc a lot when using
jedi-vim's goto method in jedi_vim.py itself (via printing in Neovim's
VimPathFinder.find_spec).

This patch uses the cache before trying the import again and again.
This commit is contained in:
Daniel Hahler
2018-05-05 19:18:14 +02:00
committed by Dave Halter
parent fc14aad8f2
commit 6748faa071
1 changed files with 3 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

5
jedi/evaluate/docstrings.py
View File

@@ -47,13 +47,14 @@ _numpy_doc_string_cache = None
def _get_numpy_doc_string_cls(): def _get_numpy_doc_string_cls():
global _numpy_doc_string_cache global _numpy_doc_string_cache
if isinstance(_numpy_doc_string_cache, ImportError):
raise _numpy_doc_string_cache
try: try:
from numpydoc.docscrape import NumpyDocString from numpydoc.docscrape import NumpyDocString
_numpy_doc_string_cache = NumpyDocString _numpy_doc_string_cache = NumpyDocString
except ImportError as e: except ImportError as e:
_numpy_doc_string_cache = e _numpy_doc_string_cache = e
if isinstance(_numpy_doc_string_cache, ImportError): raise
raise _numpy_doc_string_cache
return _numpy_doc_string_cache return _numpy_doc_string_cache