Important changes to repositories hosted on mbed.com
Mbed hosted mercurial repositories are deprecated and are due to be permanently deleted in July 2026.
To keep a copy of this software download the repository Zip archive or clone locally using Mercurial.
It is also possible to export all your personal repositories from the account settings page.
Fork of mbed-sdk-tools by
Diff: utils.py
- Revision:
- 22:9e85236d8716
- Parent:
- 21:4fdf0dd04f6f
- Child:
- 24:25bff2709c20
diff -r 4fdf0dd04f6f -r 9e85236d8716 utils.py --- a/utils.py Fri Jul 15 22:58:15 2016 +0100 +++ b/utils.py Sat Jul 16 00:34:03 2016 +0100 @@ -17,6 +17,8 @@ import sys import inspect import os +import argparse +import math from os import listdir, remove, makedirs from shutil import copyfile from os.path import isdir, join, exists, split, relpath, splitext @@ -227,5 +229,81 @@ # Read a JSON file and return its Python representation, transforming all the strings from Unicode # to ASCII. The order of keys in the JSON file is preserved. def json_file_to_dict(fname): - with open(fname, "rt") as f: - return dict_to_ascii(json.load(f, object_pairs_hook=OrderedDict)) + try: + with open(fname, "rt") as f: + return dict_to_ascii(json.load(f, object_pairs_hook=OrderedDict)) + except (ValueError, IOError): + sys.stderr.write("Error parsing '%s':\n" % fname) + raise + +# Wowza, double closure +def argparse_type(casedness, prefer_hyphen=False) : + def middle(list, type_name): + # validate that an argument passed in (as string) is a member of the list of possible + # arguments. Offer a suggestion if the case of the string, or the hyphens/underscores + # do not match the expected style of the argument. + def parse_type(string): + if prefer_hyphen: newstring = casedness(string).replace("_","-") + else: newstring = casedness(string).replace("-","_") + if string in list: + return string + elif string not in list and newstring in list: + raise argparse.ArgumentTypeError("{0} is not a supported {1}. Did you mean {2}?".format(string, type_name, newstring)) + else: + raise argparse.ArgumentTypeError("{0} is not a supported {1}. Supported {1}s are:\n{2}".format(string, type_name, columnate(list))) + return parse_type + return middle + +# short cuts for the argparse_type versions +argparse_uppercase_type = argparse_type(str.upper, False) +argparse_lowercase_type = argparse_type(str.lower, False) +argparse_uppercase_hyphen_type = argparse_type(str.upper, True) +argparse_lowercase_hyphen_type = argparse_type(str.lower, True) + +def argparse_force_type(case): + def middle(list, type_name): + # validate that an argument passed in (as string) is a member of the list of possible + # arguments after converting it's case. Offer a suggestion if the hyphens/underscores + # do not match the expected style of the argument. + def parse_type(string): + for option in list: + if case(string) == case(option): + return option + raise argparse.ArgumentTypeError("{0} is not a supported {1}. Supported {1}s are:\n{2}".format(string, type_name, columnate(list))) + return parse_type + return middle + +# these two types convert the case of their arguments _before_ validation +argparse_force_uppercase_type = argparse_force_type(str.upper) +argparse_force_lowercase_type = argparse_force_type(str.lower) + +# An argument parser combinator that takes in an argument parser and creates a new parser that +# accepts a comma separated list of the same thing. +def argparse_many(fn): + def wrap(string): + return [fn(s) for s in string.split(",")] + return wrap + +# An argument parser that verifies that a string passed in corresponds to a file +def argparse_filestring_type(string) : + if exists(string) : + return string + else : + raise argparse.ArgumentTypeError("{0}"" does not exist in the filesystem.".format(string)) + +# render a list of strings as a in a bunch of columns +def columnate(strings, seperator=", ", chars=80): + col_width = max(len(s) for s in strings) + total_width = col_width + len(seperator) + columns = math.floor(chars / total_width) + output = "" + for i, s in zip(range(len(strings)), strings): + append = s + if i != len(strings) - 1: + append += seperator + if i % columns == columns - 1: + append += "\n" + else: + append = append.ljust(total_width) + output += append + return output