jedi-gen is a cache generator for deoplete-jedi.
deoplete-jedi will take long time to load information for some large packages, e.g. numpy
, tensorflow
.
At sometime, the procession can be fail, cause jedi has a timeout.
So, We can manually generate the information file for deoplete-jedi.
First, we need to install jedi-gen at our python environment.
If you use virtualenv
, you must be at the virtual environment.
git clone https://github.com/zchee/deoplete-jedi
cd deoplete-jedi
python setup.py install
Now, we can generate the information file for any models we needed.
jedi-gen tensorflow > /path/to/tensorflow.json
Or, giving output path.
jedi-gen tensorflow -o /path/to/tensorflow.json
For macos, deoplete-jedi cache directory is ~/.cache/deoplete/jedi
.
Enjoying the completion server, after you move the cache file to the directory.
If you only need to generate a sub-model cache information file, you can do as following:
jedi-gen tensorflow.nn. -o /path/to/tensorflow.nn.json
Warning: there is tensorflow.nn.
, not tensorflow.nn