submodule not showing

[hyukim17]

I have a directory structure

folder_a
-- folder_b/
---- file_a.py
---- file_b.py

The page for folder_b shows only file_a as a sub module. Any ideas why this may be the case? Everything imports correctly, as I checked in a terminal.

[BurntSushi]

First, please do consult the docs. The first section describes how pdoc determines the public interface of a module: http://pdoc.burntsushi.net/pdoc

Second, the directory structure is largely irrelevant.

Third, your directory structure has no __init__.py files. I would expect there to at least be a __init__.py in your folder_b.

Fourth, as per the docs, the items that pdoc shows in the documentation are largely determined by inspecting the module itself. In order to debug problems, I need to be able to see your source so that I can reproduce your problem.

[hyukim17]

Sure. It actually has an init.py file, but I forgot to include it. It's baffling because there's really no difference between file_a and file_b structurally. I'll debug it for more info.

[BurntSushi]

Presumably pdoc is documenting folder_b as a module. So it looks in __init__.py. What is in __init__.py? If you only import file_a and not file_b, then pdoc will only show file_a.

[hyukim17]

Found the problem.
It was because when pdoc imports modules, something in the environment causes colorama to not work.

In particular,

import sys

from colorama import init

init(strip=not sys.stdout.isatty())

fails.

Going forward, it would be nice to have a log that would write down errors that are raised.

[BurntSushi]

Yes, I agree a log would be nice. By default, it's hidden because it's amazing how much garbage modules spit out when they're imported: https://github.com/BurntSushi/pdoc/blob/master/pdoc/__init__.py#L368 (And yes, some even block on stdin... sigh)

[hyukim17]

As it stands, it seems that monkey-patching the source to capture stdout and stderr on a case-by-case basis is the best way to debug import errors.