A little python utility to help you build command line prompts that can be styled using CSS.
Switched to six for compat Added travis test for python 3.4
Inputs can now specify duplicate keys and a list will be returned, for example:
from promptly import console
from promptly import Form
form = Form()
form.add.string(
'name',
'What is your name?',
default='Ollie'
)
form.add.string(
'name',
'What is your other name?',
default='Potato'
)
form.add.int(
'number',
'Pick a number'
)
console.run(form)
print(dict(form))
#
# {
# 'name': ['value1', 'value2'],
# 'number': 9
# }
#
Altered the console runners.console.ConsoleRunner.render() to fix inconsistent rendering in some teminals.
There was an issue on some terminals where passing the full prompt to
to input/raw_input (py3 and py2 respectively) would do some interesting
things with line wrapping and color codes. Part of the color code issue
was probably related to readline
and this reported documentation issue
re ansi escapes for readline (http://bugs.python.org/issue17337#msg183328)
http://bugs.python.org/issue17337
The fix applied however, skips sending the majority of the prompt through input/raw_input and instead writes it to stdout. Only the single line for the actual prompting is now sent to input/raw_input. This appears to address the issue.
Notifications can now be added into forms. The effect for the console runner will be to simply print the notification and then continue to the next question in the form. When converting the form to a dict, like branches, notifications will not be included in the data that is returned.
from promptly import console
from promptly import Form
form = Form()
form.add.notification('Welcome to a promptly form')
form.add.string('favorite_dog', 'Who is your favorite dog?', default='Lucy')
form.add.notification('Thanks for filling out the form!')
console.run(form)
print(dict(form))
In the case of the above you would get a prompt sequence like this:
···
Welcome to a promptly form
···
Who is your favorite dog?
> Lucy
···
Thanks for filling out the form!
···
{'favorite_dog': 'Lucy'}
Pyreadline does not supprt set_startup_hook(lambda: readline.insert_text(default))
as such the defaults on windows based machines would not show up.
The solution for this release was to alter the console renderer for string and integer input on those machines who have pyreadline installed.
In that case the prompt will appear as follows:
Who is your favorite dog? [Lucy]
>
Note that the default is tacked on to the end of the question.
Compare this to the style we get when we have a system that supports readline:
Who is your favorite dog?
> Lucy
Integers and String based prompts are rendered like this.
Added some love for our Windows friends. We now check weather or not readline is available and if not install pyreadline in that case.
This should be the last backwards incompatible update for a while. v0.5 saw a redesign of how forms are run. This was done in the hope that one day I have time to do a curses or urwid implementation. We will see. On the whole though it does make it more confirguable for individuals that do not like the default form rendering as Promptly now supports form runners.
What are form runners? Well put simply, in prior versions you would call:
# < v0.5
from promptly import From
form = Form()
form.add.string('favorite_food', 'What is your favorite food?')
form.run()
This worked well, but it bound the prompts to a single implementation of the
Form
object. v0.5
treats the Form
object as more of a collection and the
runners figure out how to deal with it. Lets take a look at the example from
above in in v0.5
:
# v0.5+
from promptly import From
form promptly import console
form = Form()
form.add.string('favorite_food', 'What is your favorite food?')
console.run(form)
Pretty much exactly the same, but we just hand the form off to the run to deal with, instead of the form.
Some additional changes, the promptly.inputs.*
have all been renamed
and simplified. Now they basically act as marker classes for input types.
They help the runner identify the kind of prompts to generate.
The logic, such as StringInput.build_prompt
, basically got moved into
promptly.renderers.console.StringPrompt
. If you were always using the
shortcut syntax for cerating your forms:
form.add.string(...)
form.add.bool(...)
form.add.int(...)
form.add.select(...)
form.add.multiselect(...)
Then you don't have to worry about anything, everything should still work fine for you. If you were using the more verbose style:
form.add(
'age',
IntegerInput('What is your age?',
default=1)
)
Things will break for you. It's probably better to always be using the shortcuts.
All of the input types now take "notifications" This is a convenient way to annotate your questions. Lets take a look at a prompt with notifications and the same prompt without notifications.
First, no notifications:
from promptly import From
form promptly import console
form = Form()
form.add.string('name', 'What is your name?', default='Lucy')
console.run(form, prefix='[promptly] ')
That will generate a prompt that looks like this:
[promptly] What is your name?
> Lucy
Now lets look at the same prompt with notifications:
from promptly import From
form promptly import console
form = Form()
form.add.string(
'name',
'What is your name?',
notifications=('This will help to identify you later', 'Identification is fun!')
default='Lucy')
console.run(form, prefix='[promptly] ')
That will generate a prompt that looks like this:
[promptly] What is your name?
This will help to identify you later
Identification is fun!
···
> Lucy
The notifications appear after the question, but before the user input.
The available CSS styles have also been updated to account for these. See the list below for the default styles available.
There is also convenience function for just dropping notifications to the console without running though a form. They will be styled according to the notification and prefix styles:
from promptly import console
console.notification('Hello World', prefix='[notice] ', stylesheet=None)
This will immediately write a message to sys.stdout.
Migration Guide
Migration Guide
my_form.add.choice
should be becomemy_form.add.select
ChoiceInput
should becomeSelectInput
- SelectInput (formerly ChoiceInput) and MultiSelectInput now take
an option_format callable. By default this callable is
promptly.utils.numeric_options
. This will take a list ['foo', 'bar'] and return a list: [(1, 'foo'), (2, 'bar')]. So if you only need numbers for your choices or multi-select input's you don't need to worry about, you get them for free. If you were passing your own in something like:zip(range(1,3), ['foo', 'bar'])
you no longer need to do that. In fact that will break things for you so you should replace it with just your list of choices
Forms can now branch. The branch input item takes a callable that will
be executed and is expected to return another Form
object. This Form
object will be merged into the currently running form at the location
where the branch was added. The callable signature is as follows:
my_branch_building_action(form, *args, **kwargs):
Example branch usage:
def handler(form, name):
branch = Form()
if form.age.value < 30:
branch.add.string('name', 'What is your name?')
else:
branch.add.string('name', 'What is your pet's name?', default=name)
return branch
form = Form()
form.add.int('age', 'What is your age?', default=age)
form.add.branch(handler, name='Lucy')
# The branch fields will be added here in terms of
# position in the form once the user reaches the branch
form.add.int('number', 'What is your favorite number?')
form.run()
A new input type has been added, MultiSelectInput
, a shortcut for creating
one is also available in the form of:
my_form.add.multiselect(key, label, choices, done_label='Done')
Note that done_label is optional.
MultiSelectInput lets the user choose multiple options from a SelectInput style display. It marks the currently selected items. If the user chooses the same option that has already been selected it will be deselected.
A final option is added to the list of choices provided to represent
the sentinel choice. The done_label
kwarg sets the value used here
By default it is set to Done. The user must select the sentinel choice
in order to continue on in the form.
from promptly import Form
form promptly import console
# Build our form
form = Form()
# add questions in the sequence you would like them to appear
form.add.string('name',
'What is your name?',
default='Aubrey')
form.add.int('age',
'What is your age?',
default=1)
# no options_format kwarg is provided for ChoiceInput
# so it will use the default numeric_options
form.add.select('color',
'What is your favorite color',
('red', 'green', 'blue'),
default=1)
form.add.bool('yaks', 'Do you like yaks?', default=True)
# Our form is created, lets prompt the user for the answers:
# promptly comes with a default set of styles or you can
# provide your own.
with open('/path/to/my/styles.css') as css:
console.run(form, prefix='[promptly] ', stylesheet=css.read())
# control has returned back to our script, lets see what the user said:
print(form.name.value)
print(form.age.value)
print(form.color.value) # this will be a (key, value) tuple
print(form.yaks.value)
if form.age.value < 12:
print(form.food.value)
# Or we can just convert the whole form into a dictionary:
d = dict(form)
print(d)
Promptly prompts are styles with a very limited subset of CSS. Only the following properties apply:
- color
- background-color
- font-weight
The avialable colors are limited to the color names provided by colorama:
Fore: BLACK, RED, GREEN, YELLOW, BLUE, MAGENTA, CYAN, WHITE, RESET.
Back: BLACK, RED, GREEN, YELLOW, BLUE, MAGENTA, CYAN, WHITE, RESET.
Style: DIM, NORMAL, BRIGHT, RESET_ALL
In other words:
.prefix {
color: white;
background-color: blue;
}
The font-weight property maps to colorama Style values in the following way:
font-weight: normal; -> Style.NORMAL
font-weight: bold; -> Style.BRIGHT
font-weight: lighter; -> Style.DIM
The CSS parser in promptly is VERY VERY primitive. It's just enough to parse what is below and that's all. It is by no means a full implementation of the CSS spec.
Below is the default stylesheet included with promptly. This stylesheet presents the exhaustive set of selectors that can be used to style your prompts. If it's not below, promptly doesn't support it.
Remember each selector can support:
color: <value>
background-color: <value>
font-weight: </value>
The default stylesheet below does not use every available option for obvious reasons. But you should feel free too if you so desire.
body
will set the default color and font-weight and background color.
The additional styles effectively cascade on top of body.
New selectors in v0.5
.action
represents the Cheveron before the user input is displayed.
.input
are the style for the user input.
.notification .footer
are the styles for the 3 dots that appear below
selection choices and after notifications.
body{
color:white;
font-weight:normal;
}
.action{
color:magenta;
font-weight:bold;
}
.input{
color:white;
font-weight:bold;
}
.prefix{
color:blue;
font-weight:bold;
}
.notification .label{
color:white;
font-weight:bold;
}
.notification .footer{
color:white;
font-weight:normal;
}
.string .label{
color:white;
}
.string .default-wrapper{
color:white;
font-weight:bold;
}
.string .default-value{
color:yellow;
}
.integer .label{
color:white;
}
.integer .default-wrapper{
color:white;
font-weight:bold;
}
.integer .default-value{
color:yellow;
}
.boolean .label{
color:white;
}
.boolean .default-wrapper{
color:white;
font-weight:bold;
}
.boolean .default-value{
color:yellow;
}
.boolean .other-value{
color:yellow;
}
.boolean .seperator{
color:white;
font-weight:bold;
}
.choices .label{
color:white;
}
.choices .default-wrapper{
color:white;
font-weight:bold;
}
.choices .default-value{
color:yellow;
}
.choices .option-key{
color:yellow;
}
.choices .seperator{
color:yellow;
font-weight:lighter;
}
.choices .option-value{
color:white;
font-weight:bold;
}
.choices .action{
color:magenta;
font-weight:bold;
}
.choices .selection{
color:white;
}