You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Improvement Description
This came from a discussion I had with @ebolyen on Slack recently and he can fill in technical blanks I'm omitting here.
To re-word the --p-exclude-ids description to clearly describe expected input and behavior.
Current Behavior
Using feature-table filter-samples below but same applies to all 3 filtering actions. Currently, it reads:
--p-exclude-ids / --p-no-exclude-ids
If true, the samples selected by `metadata` or `where`
parameters will be excluded from the filtered table
instead of being retained. [default: False]
It is unclear what inputs are expected i.e. true, True, F etc.
Without an input, it suggests that it would behave as --p-no-exclude-ids defaulting to False...but it doesn't in all cases, see below.
There are 6 potential options for the user here which is too many...
not setting the parameter at all, defaults to False -> doesn't exclude the ids
Setting --p-exclude-ids \ with no arg should default to False based on its description, however, it works just like --p-exclude-ids True instead.
Setting --p-exclude-ids with a True arg -> excludes ids
Setting --p-exclude-ids with a False arg -> doesn't exclude ids
Setting --p-no-exclude-ids with a True arg -> doesn't exclude ids
Setting --p-no-exclude-ids with a False arg == --p-exclude-ids False 🙃
This gets confusing from a user perspective, especially with the double negatives.
Proposed Behavior
Ideally, the user would only have to make a True/False choice, and there wouldn't be a --p-no-exclude-ids at all. Ex.
--p-exclude-ids
If `True`, the samples selected by `metadata` or `where`
parameters will be excluded from the filtered table,
`False` will retain only those samples. [required] [default: True]
Or make it so that --p-exclude-ids and --p-no-exclude-ids take no argument, but one is always required.
Comments
In discussion with Evan, it doesn't sound like there'd be an easy solution like I described above due to the way the parser is set up. Instead, we could just remove the --p-no-exclude-ids option all together from the help files alone and it could just be an easter egg that is there without being shown.
The text was updated successfully, but these errors were encountered:
Improvement Description
This came from a discussion I had with @ebolyen on Slack recently and he can fill in technical blanks I'm omitting here.
To re-word the
--p-exclude-ids
description to clearly describe expected input and behavior.Current Behavior
Using
feature-table filter-samples
below but same applies to all 3 filtering actions. Currently, it reads:It is unclear what inputs are expected i.e.
true
,True
,F
etc.Without an input, it suggests that it would behave as
--p-no-exclude-ids
defaulting toFalse
...but it doesn't in all cases, see below.There are 6 potential options for the user here which is too many...
--p-exclude-ids \
with no arg should default to False based on its description, however, it works just like--p-exclude-ids True
instead.--p-exclude-ids
with a True arg -> excludes ids--p-exclude-ids
with a False arg -> doesn't exclude ids--p-no-exclude-ids
with a True arg -> doesn't exclude ids--p-no-exclude-ids
with a False arg ==--p-exclude-ids False
🙃This gets confusing from a user perspective, especially with the double negatives.
Proposed Behavior
Ideally, the user would only have to make a True/False choice, and there wouldn't be a
--p-no-exclude-ids
at all. Ex.Or make it so that
--p-exclude-ids
and--p-no-exclude-ids
take no argument, but one is always required.Comments
In discussion with Evan, it doesn't sound like there'd be an easy solution like I described above due to the way the parser is set up. Instead, we could just remove the
--p-no-exclude-ids
option all together from the help files alone and it could just be an easter egg that is there without being shown.The text was updated successfully, but these errors were encountered: