Function doc() doesn't output anything(on google colab)

Part 1 (2019)
1

Running the doc function (for example, doc(interp.plot_top_losses)) doesn’t show a panel like it does in the video(Lesson 1 around 1:10:41).
Running the function doesn’t throw an error either. Running ??doc does display the source for the function, so I don’t think there’s an import problem.

4 Likes
2

Hi,
Can you share the platform where you are running this, also details of the version might help.
I’m on the latest version, can run it without any issues (with the popup showing)

1 Like
3

I’m running on Google Colab.
For setup, I ran the installation script given here so I guess everything should be the latest version.

1 Like
4

Yeah Colab doesn’t support a lot of Jupyter stuff, so I suspect it’s not going to work there.

1 Like
5

I’m new to colab myself, but if you look on the “Overview of Colaboratory Features” notebook that opens when you first login to colab, it shows the following features under the “Tab-completion and exploring code” section:

Colab provides tab completion to explore attributes of Python objects, as well as to quickly view documentation strings. As an example, first run the following cell to import the numpy module.

import numpy as np

If you now insert your cursor after np.random. and press Tab , you will see the list of available completions within the np.random submodule.

np.random.

If you type an open parenthesis followed by the Tab key after any function or class in the module, you will see a pop-up of its documentation string:

np.random.rand(

To open the documentation in a persistent pane at the bottom of your screen, add a ? after the object or method name and execute the cell using Shift+Enter :

np.random?

3 Likes
6

Can you solve this problem? I also encountered the same problem.

7

I ran into this issue as well, and just realized that the output of doc (and ??) appears in a separate frame at the bottom, rather than inline. I’m not sure if the frame was hidden for me, or just two small to notice.

8

As of 2020-03-20 on Chrome, if you just mouse over the function, a small pop-up shows the input parameters.

As you start typing them auto completion is offered.

And if you press Ctrl while hovering, the function becomes a link that when clicked opens up the documentation of the function.

getting function parameters and documentation within Colab

Regarding this, Caggle (or any jupyter-like) cannot offer you info about something (autocomplete or hover info) if they don’t know what that something is. For that reason, while learning, I transform something like

 data = (ImageList.from_folder(path)
            .split_by_folder()              
            .label_from_folder()          
            .add_test_folder()             
            .transform(tfms, size=64)  
            .databunch())

Into (not literal translation, just as an example)

data = ImageItemList.from_df(path=path, df=df, cols='fpaths')
data = data.random_split_by_pct(valid_pct=0.2, seed=10)
data = data.label_from_df(cols='class_label')

Think that the first form has a sequence of operations
data= create_something.tranform_A.transform_B.pack_nicely()
and you can always break it down.

2 Likes