davisp / runfunc forked from simonw/optfunc
- Source
- Commits
- Network (11)
- Graphs
-
Branch:
master
| name | age | message | |
|---|---|---|---|
| |
.gitignore | Sun May 31 02:10:54 -0700 2009 | |
| |
LICENSE | Fri Jul 24 10:37:51 -0700 2009 | |
| |
MANIFEST.in | Fri Jul 24 10:50:48 -0700 2009 | |
| |
README.md | Fri Jul 24 10:50:48 -0700 2009 | |
| |
examples/ | Fri Jul 24 10:37:51 -0700 2009 | |
| |
ez_setup.py | Fri Jul 24 10:37:51 -0700 2009 | |
| |
runfunc.py | Fri Jul 24 10:37:51 -0700 2009 | |
| |
setup.py | Fri Jul 24 10:37:51 -0700 2009 | |
| |
test.py | Fri Jul 24 10:37:51 -0700 2009 |
runfunc
I liked optfunc but it hasn't been put on PyPI for actual use so I went ahead and rewrote it to use a different syntax to be able to include option help and input validators.
Basic Example
import sys
import runfunc as rf
class Help(rf.Help):
"""\
This is a blurb about this program. Its very entertaining to
watch and think about what we might do with it.
Look ma. Formatted text.
And stuff
"""
usage = "%prog [options] value"
value = rf.Check(int, "A value to multiply.")
factor = rf.Check(float, "A factor by which to multiply.", opt='f')
random = rf.Stream("r", "An input file to read from.", opt='r')
verbose = rf.Flag("Multiply verbosely.", opt='v')
def main(value, factor=2.0, random=sys.stdin, verbose=False):
newval = value * factor
if verbose:
print "%d * %s = %s" % (value, factor, newval)
else:
print newval
rf.run(main, Help())
Help Class
The basic outline for using runfunc is to create a subclass of runfunc.Help
and provide this class when running the function. This subclass is responsible
for specifying all of the possible arguments to be passed to the function. It
can contain extra arguments that are ignored so that the same subclass can be
reused for multiple command line scripts.
class Help(rf.Help):
"""\
This is a blurb about this program. Its very entertaining to
watch and think about what we might do with it.
Look ma. Formatted text.
And stuff
"""
usage = "%prog [options] value"
value = rf.Check(int, "A value to multiply.")
factor = rf.Check(float, "A factor by which to multiply.", opt='f')
random = rf.Stream("r", "An input file to read from.", opt='r')
verbose = rf.Flag("Multiply verbosely.", opt='v')
A couple points to note:
- The class's doc string will be used as the program's help description
- If a
usageattribute is present, it will be used for the program's help - Any other attributes that inherit from
runfunc.Argwill be available for use as a program argument.
Running a function
runfunc.run(callable, help_object, argv=None, check=True)
callableis callable object.help_objectis an instance of a class that inherits fromrunfunc.Help.argvis a list of arguments. If None,sys.argv[1:]is usedcheck=Truewill prevent the function from running when imported as a module.
Functions
When you specify a function, the arguments will be treated as required values on the command line. Any keyword arguments will be able to be specified as options the the program. For instance:
def foo(bar, val=None):
pass
Will have one required argument for bar and an option for val. Required
arguments are taken in order of specification on the command line. If you use
required arguments you should be sure to specify a helpful usage description
on your Help class.
All arguments to the function must be present as attributes on the
runfunc.Help instance passed to runfunc.run.
Validator Types
Check(func, desc, opt=None)
Use an arbitrary function to validate an input.
- func - A callable taking a single argument. Raises an exception on error.
- desc - Help message that describes the option
- opt - A single character option name.
Flag(desc, opt=None)
Passes true based on the presence of an option. Generally only useful for keyword arguments.
- desc - Help message that describes the option
- opt - A single character option name.
List(desc, opt=None, validator=None)
Append each value seen to a list. Validator is applied before appending each value.
- desc - Help message that describes the option
- opt - A single character option name.
- validator - A callable taking a single argument. Raises an exception on error.
Choice(choices, desc, opt=None, validator=None)
Limit input to a specified set of values. Validator is applied before testing for membership in the set.
- choices - A on object providing item access to test if an option is allowable.
- desc - Help message that describes the option
- opt - A single character option name.
- validator - A callable taking a single argument. Raises an exception on error.
Regexp(pattern, desc, opt=None, flags=0)
Require input to match a regular expression.
- pattern - A string acceptable for
re.compile(pattern, flags) - desc - Help message that describes the option
- opt - A single character option name.
- flags - Any modifiers for compiling the regular expression
Email(desc, opt=None)
Require input to look like a valid email address. Subclass of Regexp
- desc - Help message that describes the option
- opt - A single character option name.
IpAddr(desc, opt=None)
Require input to look like a valid IP Address. Subclass of Regexp.
- desc - Help message that describes the option
- opt - A single character option name.
Path(flags, desc, opt=None)
Require that input specifies a path. Tests are executed depending on the flags passed. FILE requires that the path looks like a file (no trailing path specifier). DIR requires the trailing path specifier. EXISTS will check if the path actually exists. PARENT requires that the path's parent exist. Any combination of flags can be passed though passing FILE | DIR would guarantee an error during validation.
- flags - An OR of FILE, DIR, EXISTS, PARENT
- desc - Help message that describes the option
- opt - A single character option name.
Stream(mode, desc, opt=None)
Open a path for use as a file object. A useful pattern is to use a default
value of sys.stdin, sys.stdout, or sys.stderr for common command line
semantics.
- mode - Passed to
open(path, mode)when opening the stream - desc - Help message that describes the option
- opt - A single character option name.
Custom Validators
Custom validators can be created by subclassing the Arg class and overriding the validate method.
An example for specifying a Range validator would look something like this:
import runfunc as rf
class Range(rf.Arg):
def __init__(self, func, left, right, desc, opt=None):
super(Range, self).__init__(desc, opt=opt)
self.func = func
self.left = left
self.right = right
def validate(self, option, optstr, value, parser):
value = self.func(value)
if value < self.left or value > self.right:
raise rf.OptionValueError("Value is outside range.")
setattr(parser.values, option.dest, value)
Notice that func is specifiable so that users can convert the raw value to an appropriate data type.
