Spyder Console autocomplete issues vs VSCode

[mg3146]

Hi - this could simply be a setting, but I find that the console's autocomplete to be fairly limited vs when I run the same code in VS Code Jupyter notebooks when working with my own library/code base.

Spyder is installed via conda into the same env that I run my code base from.

For example, when working with nested classes, the Spyder console autocomplete seems to stop after the first level. Ie. in VSCode it'll autocomplete something like a=A.B.C. In spyder I'd have to do x = A.B followed by a=x.C

Any suggestions? Thx

[dalthviz]

Hi @mg3146 thank you for the feedback! I think over the Spyder IPython Console preferences/Advanced settings there are a couple of options that maybe could improve the completions you get from the consoles. You can find those options over Preferences > IPython Console > Advaced Settings. One is called Jedi completion and the other is Greedy completion. Note that both options are disabled by default due to some possible unwanted side effects like slowing down the console if you are working with big sized variables (Jedi completion) and security considerations (things get evaluated as you ask for a completion when using Greedy completion):

Let us know if the info above helps!

[mg3146]

Thanks @dalthviz . I've tried Jedi completion, but I'm not really finding much of a difference. For example, here is a blank console with just pandas.

and if I then set it to a variable and hit tab.

and my settings just for good order:

[dalthviz]

I think then your use case needs for the Greedy completion to be enabled. Checking locally, when doing the df.head().<Tab> completion while having Greddy completion enabled, I see the following:

That's what you were expecting, right? Let us know!

[mg3146]

Greedy did the trick, wayyyy better now. Thanks!

[dalthviz]

Awesome! I think I will close this then but if you have further feedback let us know!