/usr/lib/python3/dist-packages/CTDopts/CTDopts.py is in python3-ctdopts 1.2-3.
This file is owned by root:root, with mode 0o644.
The actual contents of the file can be viewed below.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 | import argparse
from collections import OrderedDict, Mapping
from itertools import chain
from xml.etree.ElementTree import Element, SubElement, tostring, parse
from xml.dom.minidom import parseString
import warnings
# dummy classes for input-file and output-file CTD types.
class _ASingleton(type):
"""
A metaclass for singletons
"""
_instances = {}
def __call__(cls, *args, **kwargs):
if cls not in cls._instances:
cls._instances[cls] = super(_ASingleton, cls).__call__(*args, **kwargs)
return cls._instances[cls]
class _Null(object, metaclass=_ASingleton):
"""
A null singleton for non-initialized fields to distinguish between initialized=None and non-initialized members
"""
class _InFile(str):
"""Dummy class for input-file CTD type. I think most users would want to just get the file path
string but if it's required to open these files for reading or writing, one could do it in these
classes in a later release. Otherwise, it's equivalent to str with the information that we're
dealing with a file argument.
"""
pass
class _OutFile(str):
"""Same thing, a dummy class for output-file CTD type."""
pass
# module globals for some common operations (python types to CTD-types back and forth)
TYPE_TO_CTDTYPE = {int: 'int', float: 'float', str: 'string', bool: 'boolean',
_InFile: 'input-file', _OutFile: 'output-file'}
CTDTYPE_TO_TYPE = {'int': int, 'float': float, 'double': float, 'string': str, 'boolean': bool, 'bool': bool,
'input-file': _InFile, 'output-file': _OutFile, int: int, float: float, str: str,
bool: bool, _InFile: _InFile, _OutFile: _OutFile}
PARAM_DEFAULTS = {'advanced': False, 'required': False, 'restrictions': None, 'description': None,
'supported_formats': None, 'tags': None, 'position': None} # unused. TODO.
# a boolean type caster to circumvent bool('false')==True when we cast CTD 'value' attributes to their correct type
CAST_BOOLEAN = lambda x: bool(x) if not isinstance(x, str) else (x in ('true', 'True', '1'))
# instead of using None or _Null, we define non-present 'position' attribute values as -1
NO_POSITION = -1
# Module-level functions for querying and manipulating argument dictionaries.
def get_nested_key(arg_dict, key_list):
"""Looks up a nested key in an arbitrarily nested dictionary. `key_list` should be an iterable:
get_nested_key(args, ['group', 'subgroup', 'param']) returns args['group']['subgroup']['param']
"""
key_list = [key_list] if isinstance(key_list, str) else key_list # just to be safe.
res = arg_dict
for key in key_list:
res = res[key]
else:
return res
def set_nested_key(arg_dict, key_list, value):
"""Inserts a value into an arbitrarily nested dictionary, creating nested sub-dictionaries on
the way if needed:
set_nested_key(args, ['group', 'subgroup', 'param'], value) sets args['group']['subgroup']['param'] = value
"""
key_list = [key_list] if isinstance(key_list, str) else key_list # just to be safe.
res = arg_dict
for key in key_list[:-1]:
if key not in res:
res[key] = {} # OrderedDict()
res = res[key]
else:
res[key_list[-1]] = value
def flatten_dict(arg_dict, as_string=False):
"""Creates a flattened dictionary out of a nested dictionary. New keys will be tuples, with the
nesting information. Ie. arg_dict['group']['subgroup']['param1'] will be
result[('group', 'subgroup', 'param1')] in the flattened dictionary.
`as_string` joins the nesting levels into a single string with a semicolon, so the same entry
would be under result['group:subgroup:param1']
"""
result = {}
def flattener(subgroup, level):
# recursive closure that accesses and modifies result dict and registers nested elements
# as it encounters them
for key, value in subgroup.items():
if isinstance(value, Mapping): # collections.Mapping instead of dict for generality
flattener(value, level + [key])
else:
result[tuple(level + [key])] = value
flattener(arg_dict, [])
if as_string:
return {':'.join(keylist): value for keylist, value in result.items()}
else:
return result
def override_args(*arg_dicts):
"""Takes any number of (nested or flat) argument dictionaries and combines them, giving preference
to the last one if more than one have the same entry. Typically would be used like:
combined_args = override_args(args_from_ctd, args_from_commandline)
"""
overridden_args = dict(chain(*(iter(flatten_dict(d).items()) for d in arg_dicts)))
result = {}
for keylist, value in overridden_args.items():
set_nested_key(result, keylist, value)
return result
def _translate_ctd_to_param(attribs):
"""Translates a CTD <ITEM> or <ITEMLIST> XML-node's attributes to keyword arguments that Parameter's
constructor expects. One should be able to call Parameter(*result) with the output of this function.
For list parameters, adding is_list=True and getting <LISTITEM> values is needed after translation,
as they are not stored as XML attributes.
"""
# right now value is a required field, but it shouldn't be for required parameters.
if 'value' in attribs: # TODO 1_6_3, this line will be deleted.
attribs['default'] = attribs.pop('value') # rename 'value' to 'default' (Parameter constructor takes 'default')
if 'supported_formats' in attribs: # supported_formats in CTD xml is called file_formats in CTDopts
attribs['file_formats'] = attribs.pop('supported_formats') # rename that attribute too
if 'restrictions' in attribs: # find out whether restrictions are choices ('this,that') or numeric range ('3:10')
if ',' in attribs['restrictions']:
attribs['choices'] = attribs['restrictions'].split(',')
elif ':' in attribs['restrictions']:
n_min, n_max = attribs['restrictions'].split(':')
n_min = None if n_min == '' else n_min
n_max = None if n_max == '' else n_max
attribs['num_range'] = (n_min, n_max)
else:
# there is nothing we can split with... so we will assume that this is a restriction of one possible
# value... anyway, the user should be warned about it
warnings.warn("Restriction [%s] of a single value found for parameter [%s]. \n"
"Restrictions should be comma separated value lists or colon separated values to "
"indicate numeric ranges (e.g., 'true,false', '0:14', '1:', ':2.8')\n"
"Will use a restriction with one possible value of choice." %
(attribs['restrictions'], attribs['name']))
attribs['choices'] = [attribs['restrictions']]
# TODO: advanced. Should it be stored as a tag, or should we extend Parameter class to have that attribute?
# what we can do is keep it as a tag in the model, and change Parameter._xml_node() so that if it finds
# 'advanced' among its tag-list, make it output it as a separate attribute.
return attribs
class ArgumentError(Exception):
"""Base exception class for argument related problems.
"""
def __init__(self, parameter):
self.parameter = parameter
self.param_name = ':'.join(self.parameter.get_lineage(name_only=True))
class ArgumentMissingError(ArgumentError):
"""Exception for missing required arguments.
"""
def __init__(self, parameter):
super(ArgumentMissingError, self).__init__(parameter)
def __str__(self):
return 'Required argument %s missing' % self.param_name
class ArgumentTypeError(ArgumentError):
"""Exception for arguments that can't be casted to the type defined in the model.
"""
def __init__(self, parameter, value):
super(ArgumentTypeError, self).__init__(parameter)
self.value = value
def __str__(self):
return "Argument %s is of wrong type. Expected: %s, got %s" % (
self.param_name, TYPE_TO_CTDTYPE[self.parameter.type], self.value)
class ArgumentRestrictionError(ArgumentError):
"""Exception for arguments violating numeric, file format or controlled vocabulary restrictions.
"""
def __init__(self, parameter, value):
super(ArgumentRestrictionError, self).__init__(parameter)
self.value = value
def __str__(self):
return 'Argument restrictions for %s failed. Restriction: %s. Value: %s' % (
self.param_name, self.parameter.restrictions.ctd_restriction_string(), self.value)
class ModelError(Exception):
"""Exception for errors related to CTDModel building
"""
def __init__(self):
super(ModelError, self).__init__()
class ModelParsingError(ModelError):
"""Exception for errors related to CTD parsing
"""
def __init__(self, message):
super(ModelParsingError, self).__init__()
self.message = message
def __str__(self):
return "An error occurred while parsing the CTD file: %s" % self.message
def __repr__(self):
return str(self)
class UnsupportedTypeError(ModelError):
"""Exception for attempting to use unsupported types in the model
"""
def __init__(self, wrong_type):
super(UnsupportedTypeError, self).__init__()
self.wrong_type = wrong_type
def __str__(self):
return 'Unsupported type encountered during model construction: %s' % self.wrong_type
class DefaultError(ModelError):
def __init__(self, parameter):
super(DefaultError, self).__init__()
self.parameter = parameter
def __str__(self):
pass
class _Restriction(object):
"""Superclass for restriction classes (numeric, file format, controlled vocabulary).
"""
def __init__(self):
pass
# if Python had virtual methods, this one would have a _single_check() virtual method, as all
# subclasses have to implement for check() to go through. check() expects them to be present,
# and validates normal and list parameters accordingly.
def check(self, value):
"""Checks whether `value` satisfies the restriction conitions. For list parameters it checks
every element individually.
"""
if isinstance(value, list): # check every element of list (in case of list parameters)
return all((self._single_check(v) for v in value))
else:
return self._single_check(value)
class _NumericRange(_Restriction):
"""Class for numeric range restrictions. Stores valid numeric ranges, checks values against
them and outputs CTD restrictions attribute strings.
"""
def __init__(self, n_type, n_min=None, n_max=None):
super(_NumericRange, self).__init__()
self.n_type = n_type
self.n_min = self.n_type(n_min) if n_min is not None else None
self.n_max = self.n_type(n_max) if n_max is not None else None
def ctd_restriction_string(self):
n_min = str(self.n_min) if self.n_min is not None else ''
n_max = str(self.n_max) if self.n_max is not None else ''
return '%s:%s' % (n_min, n_max)
def _single_check(self, value):
if self.n_min is not None and value < self.n_min:
return False
elif self.n_max is not None and value > self.n_max:
return False
else:
return True
def __repr__(self):
return 'numeric range: %s to %s' % (self.n_min, self.n_max)
class _FileFormat(_Restriction):
"""Class for file format restrictions. Stores valid file formats, checks filenames against them
and outputs CTD supported_formats attribute strings.
"""
def __init__(self, formats):
super(_FileFormat, self).__init__()
if isinstance(formats, str): # to handle ['txt', 'csv', 'tsv'] and '*.txt,*.csv,*.tsv'
formats = [x.replace('*.', '').strip() for x in formats.split(',')]
self.formats = formats
def ctd_restriction_string(self):
return ','.join(('*.' + f for f in self.formats))
def _single_check(self, value):
for f in self.formats:
if value.endswith('.' + f):
return True
return False
def __repr__(self):
return 'file formats: %s' % (', '.join(self.formats))
class _Choices(_Restriction):
"""Class for controlled vocabulary restrictions. Stores controlled vocabulary elements, checks
values against them and outputs CTD restrictions attribute strings.
"""
def __init__(self, choices):
super(_Choices, self).__init__()
if isinstance(choices, str): # If it actually has to run, a user is screwing around...
choices = choices.replace(', ', ',').split(',')
self.choices = choices
def _single_check(self, value):
return value in self.choices
def ctd_restriction_string(self):
return ','.join(self.choices)
def __repr__(self):
return 'choices: %s' % (', '.join(map(str, self.choices)))
class Parameter(object):
def __init__(self, name, parent, **kwargs):
"""Required positional arguments: `name` string and `parent` ParameterGroup object
Optional keyword arguments:
`type`: Python type object, or a string of a valid CTD types.
For all valid values, see: CTDopts.CTDTYPE_TO_TYPE.keys()
`default`: default value. Will be casted to the above type (default None)
`is_list`: bool, indicating whether this is a list parameter (default False)
`required`: bool, indicating whether this is a required parameter (default False)
`description`: string containing parameter description (default None)
`tags`: list of strings or comma separated string (default [])
`num_range`: (min, max) tuple. None in either position makes it unlimited
`choices`: list of allowed values (controlled vocabulary)
`file_formats`: list of allowed file extensions
`short_name`: string for short name annotation
`position`: index (1-based) of the position on which the parameter appears on the command-line
"""
self.name = name
self.parent = parent
self.short_name = kwargs.get('short_name', _Null)
try:
self.type = CTDTYPE_TO_TYPE[kwargs.get('type', str)]
except:
raise UnsupportedTypeError(kwargs.get('type'))
self.tags = kwargs.get('tags', [])
if isinstance(self.tags, str): # so that tags can be passed as ['tag1', 'tag2'] or 'tag1,tag2'
self.tags = list(filter(bool, self.tags.split(','))) # so an empty string doesn't produce ['']
self.required = CAST_BOOLEAN(kwargs.get('required', False))
self.is_list = CAST_BOOLEAN(kwargs.get('is_list', False))
self.description = kwargs.get('description', None)
self.advanced = CAST_BOOLEAN(kwargs.get('advanced', False))
self.position = int(kwargs.get('position', str(NO_POSITION)))
default = kwargs.get('default', _Null)
self._validate_numerical_defaults(default)
# TODO 1_6_3: right now the CTD schema requires the 'value' attribute to be present for every parameter.
# So every time we build a model from a CTD file, we find at least a default='' or default=[]
# for every parameter. This should change soon, but for the time being, we have to get around this
# and disregard such default attributes. The below two lines will be deleted after fixing 1_6_3.
if default == '' or (self.is_list and default == []):
default = _Null
# enforce that default is the correct type if exists. Elementwise for lists
if default is _Null:
self.default = _Null
elif default is None:
self.default = None
else:
if self.is_list:
self.default = list(map(self.type, default))
else:
self.default = self.type(default)
# same for choices. I'm starting to think it's really unpythonic and we should trust input. TODO
if self.type == bool:
assert self.is_list is False, "Boolean flag can't be a list type"
self.required = False # override whatever we found. Boolean flags can't be required...
self.default = CAST_BOOLEAN(default)
# Default value should exist IFF argument is not required.
# TODO: if we can have optional list arguments they don't have to have a default? (empty list)
# TODO: CTD Params 1.6.3 have a required value attrib. That's very wrong for parameters that are required.
# ... until that's ironed out, we have to comment this part out.
#
# ACTUALLY now that I think of it, letting required fields have value attribs set too
# can be useful for users who want to abuse CTD and build models from argument-storing CTDs.
# I know some users will do this (who are not native CTD users just want to convert their stuff
# with minimal effort) so we might as well let them.
#
# if self.required:
# assert self.default is None, ('Required field `%s` has default value' % self.name)
# else:
# assert self.default is not None, ('Optional field `%s` has no default value' % self.name)
self.restrictions = None
if 'num_range' in kwargs:
try:
self.restrictions = _NumericRange(self.type, *kwargs['num_range'])
except ValueError:
num_range = kwargs['num_range']
raise ModelParsingError("Provided range [%s, %s] is not of type %s" %
(num_range[0], num_range[1], self.type))
elif 'choices' in kwargs:
self.restrictions = _Choices(list(map(self.type, kwargs['choices'])))
elif 'file_formats' in kwargs:
self.restrictions = _FileFormat(kwargs['file_formats'])
# perform some basic validation on the provided default values...
# an empty string IS NOT a float/int!
def _validate_numerical_defaults(self, default):
if default is not None and default is not _Null:
if self.type is int or self.type is float:
defaults_to_validate = []
errors_so_far = []
if self.is_list:
# for lists, validate each provided element
defaults_to_validate.extend(default)
else:
defaults_to_validate.append(default)
for default_to_validate in defaults_to_validate:
try:
if self.type is int:
int(default_to_validate)
else:
float(default_to_validate)
except ValueError:
errors_so_far.append(default_to_validate)
if len(errors_so_far) > 0:
raise ModelParsingError("Invalid default value(s) provided for parameter %(name)s of type %(type)s:"
" '%(default)s'"
% {"name": self.name,
"type": self.type,
"default": ', '.join(map(str, errors_so_far))})
def get_lineage(self, name_only=False, short_name=False):
"""Returns a list of zero or more ParameterGroup objects plus this Parameter object at the end,
ie. the nesting lineage of the Parameter object. With `name_only` setting on, it only returns
the names of said objects. For top level parameters, it's a list with a single element.
"""
lineage = []
i = self
while i.parent is not None:
# Exclude ParameterGroup here, since they do not have a short_name attribute (lzimmermann)
lineage.append(i.short_name if short_name and not isinstance(i, ParameterGroup) else i.name if name_only else i)
i = i.parent
lineage.reverse()
return lineage
def __repr__(self):
info = []
info.append('PARAMETER %s%s' % (self.name, ' (required)' if self.required else ''))
info.append(' type: %s%s%s' % ('list of ' if self.is_list else '', TYPE_TO_CTDTYPE[self.type],
's' if self.is_list else ''))
if self.default:
info.append(' default: %s' % self.default)
if self.tags:
info.append(' tags: %s' % ', '.join(self.tags))
if self.restrictions:
info.append(' restrictions on %s' % self.restrictions)
if self.description:
info.append(' description: %s' % self.description)
return '\n'.join(info)
def _xml_node(self, arg_dict=None):
if arg_dict is not None: # if we call this function with an argument dict, get value from there
try:
value = get_nested_key(arg_dict, self.get_lineage(name_only=True))
except KeyError:
value = self.default
else: # otherwise take the parameter default
value = self.default
# XML attributes to be created (depending on whether they are needed or not):
# name, value, type, description, tags, restrictions, supported_formats
attribs = OrderedDict() # LXML keeps the order, ElemenTree doesn't. We use ElementTree though.
attribs['name'] = self.name
if not self.is_list: # we'll deal with list parameters later, now only normal:
# TODO: once Param_1_6_3.xsd gets fixed, we won't have to set an empty value='' attrib.
# but right now value is a required attribute.
attribs['value'] = '' if value is _Null else str(value)
if self.type is bool: # for booleans str(True) returns 'True' but the XS standard is lowercase
attribs['value'] = 'true' if value else 'false'
attribs['type'] = TYPE_TO_CTDTYPE[self.type]
if self.description:
attribs['description'] = self.description
if self.tags:
attribs['tags'] = ','.join(self.tags)
# Choices and NumericRange restrictions go in the 'restrictions' attrib, FileFormat has
# its own attribute 'supported_formats' for whatever historic reason.
if isinstance(self.restrictions, _Choices) or isinstance(self.restrictions, _NumericRange):
attribs['restrictions'] = self.restrictions.ctd_restriction_string()
elif isinstance(self.restrictions, _FileFormat):
attribs['supported_formats'] = self.restrictions.ctd_restriction_string()
if self.is_list: # and now list parameters
top = Element('ITEMLIST', attribs)
# (lzimmermann) I guess _Null has to be exluded here, too
if value is not None and value is not _Null:
for d in value:
SubElement(top, 'LISTITEM', {'value': str(d)})
return top
else:
return Element('ITEM', attribs)
def _cli_node(self, parent_name, prefix='--'):
lineage = self.get_lineage(name_only=True)
top_node = Element('clielement', {"optionIdentifier": prefix+':'.join(lineage)})
SubElement(top_node, 'mapping', {"referenceName": parent_name+"."+self.name})
return top_node
def is_positional(self):
return self.position != NO_POSITION
class ParameterGroup(object):
def __init__(self, name, parent, description=None):
self.name = name
self.parent = parent
self.description = description
self.parameters = OrderedDict()
def add(self, name, **kwargs):
"""Registers a parameter in a ParameterGroup. Required: `name` string.
Optional keyword arguments:
`type`: Python type object, or a string of a valid CTD types.
For all valid values, see: CTDopts.CTDTYPE_TO_TYPE.keys()
`default`: default value. Will be casted to the above type (default None)
`is_list`: bool, indicating whether this is a list parameter (default False)
`required`: bool, indicating whether this is a required parameter (default False)
`description`: string containing parameter description (default None)
`tags`: list of strings or comma separated string (default [])
`num_range`: (min, max) tuple. None in either position makes it unlimited
`choices`: list of allowed values (controlled vocabulary)
`short_name`: string for short name annotation
"""
# TODO assertion if name already exists? It just overrides now, but I'm not sure if allowing this behavior is OK
self.parameters[name] = Parameter(name, self, **kwargs)
return self.parameters[name]
def add_group(self, name, description=None):
"""Registers a child parameter group under a ParameterGroup. Required: `name` string. Optional: `description`
"""
# TODO assertion if name already exists? It just overrides now, but I'm not sure if allowing this behavior is OK
self.parameters[name] = ParameterGroup(name, self, description)
return self.parameters[name]
def _get_children(self):
children = []
for child in self.parameters.values():
if isinstance(child, Parameter):
children.append(child)
elif isinstance(child, ParameterGroup):
children.extend(child._get_children())
return children
def _xml_node(self, arg_dict=None):
xml_attribs = {'name': self.name}
if self.description:
xml_attribs['description'] = self.description
top = Element('NODE', xml_attribs)
# TODO: if a Parameter comes after an ParameterGroup, the CTD won't validate. BTW, that should be changed.
# Of course this should never happen if the argument tree is built properly but it would be
# nice to take care of it if a user happens to randomly define his arguments and groups.
# So first we could sort self.parameters (Items first, Groups after them).
for arg in self.parameters.values():
top.append(arg._xml_node(arg_dict))
return top
def _cli_node(self, parent_name="", prefix='--'):
"""
Generates a list of clielements of that group
:param arg_dict: dafualt values for elements
:return: list of clielements
"""
for arg in self.parameters.values():
yield arg._cli_node(parent_name=parent_name+"."+self.name, prefix=prefix)
def __repr__(self):
info = []
info.append('PARAMETER GROUP %s (' % self.name)
for subparam in self.parameters.values():
info.append(subparam.__repr__())
info.append(')')
return '\n'.join(info)
class Mapping(object):
def __init__(self, reference_name=None):
self.reference_name = reference_name
class CLIElement(object):
def __init__(self, option_identifier=None, mappings=[]):
self.option_identifier = option_identifier
self.mappings = mappings
class CLI(object):
def __init__(self, cli_elements=[]):
self.cli_elements = cli_elements
class CTDModel(object):
def __init__(self, name=None, version=None, from_file=None, **kwargs):
"""The parameter model of a tool.
`name`: name of the tool
`version`: version of the tool
`from_file`: create the model from a CTD file at provided path
Other (self-explanatory) keyword arguments:
`docurl`, `description`, `manual`, `executableName`, `executablePath`, `category`
"""
if from_file is not None:
self._load_from_file(from_file)
else:
self.name = name
self.version = version
# TODO: check whether optional attributes in kwargs are all allowed or just ignore the rest?
self.opt_attribs = kwargs # description, manual, docurl, category (+executable stuff).
self.parameters = ParameterGroup('1', None, 'Parameters of %s' % self.name) # openMS legacy, top group named "1"
self.cli = []
def _load_from_file(self, filename):
"""Builds a CTDModel from a CTD XML file.
"""
root = parse(filename).getroot()
assert root.tag == 'tool', "Invalid CTD file, root is not <tool>" # TODO: own exception
self.opt_attribs = {}
self.cli = []
for tool_required_attrib in ['name', 'version']:
assert tool_required_attrib in root.attrib, "CTD tool is missing a %s attribute" % tool_required_attrib
setattr(self, tool_required_attrib, root.attrib[tool_required_attrib])
for tool_opt_attrib in ['docurl', 'category']:
if tool_opt_attrib in root.attrib:
self.opt_attribs[tool_opt_attrib] = root.attrib[tool_opt_attrib]
for tool_element in root:
if tool_element.tag in ['manual', 'description', 'executableName', 'executablePath']:
# ignoring: cli, logs, relocators. cli and relocators might be useful later.
self.opt_attribs[tool_element.tag] = tool_element.text
if tool_element.tag == 'cli':
self._build_cli(tool_element.findall('clielement'))
if tool_element.tag == 'PARAMETERS':
# tool_element.attrib['version'] == '1.6.2' # check whether the schema matches the one CTDOpts uses?
params_container_node = tool_element.find('NODE')
# we have to check the case in which the parent node contains
# item/itemlist elements AND node element children
params_container_node_contains_items = params_container_node.find('ITEM') is not None or params_container_node.find('ITEMLIST')
# assert params_container_node.attrib['name'] == self.name
# check params_container_node's first ITEM child's tool version information again? (OpenMS legacy?)
params = params_container_node.find('NODE') # OpenMS legacy again, NODE with name="1" on top
# check for the case when we have PARAMETERS/NODE/ITEM
if params is None or params_container_node_contains_items:
self.parameters = self._build_param_model(params_container_node, base=None)
else:
# OpenMS legacy again, PARAMETERS/NODE/NODE/ITEM
self.parameters = self._build_param_model(params, base=None)
def _build_cli(self, xml_cli_elements):
for xml_cli_element in xml_cli_elements:
mappings = []
for xml_mapping in xml_cli_element.findall('mapping'):
mappings.append(Mapping(xml_mapping.attrib['referenceName'] if 'referenceName' in xml_mapping.attrib else None))
self.cli.append(CLIElement(xml_cli_element.attrib['optionIdentifier'] if 'optionIdentifier' in xml_cli_element.attrib else None, mappings))
def _build_param_model(self, element, base):
if element.tag == 'NODE':
validate_contains_keys(element.attrib, ['name'], 'NODE')
if base is None: # top level group (<NODE name="1">) has to be created on its own
current_group = ParameterGroup(element.attrib['name'], base, element.attrib.get('description', ''))
else: # other groups can be registered as a subgroup, as they'll always have parent base nodes
current_group = base.add_group(element.attrib['name'], element.attrib.get('description', ''))
for child in element:
self._build_param_model(child, current_group)
return current_group
elif element.tag == 'ITEM':
setup = _translate_ctd_to_param(dict(element.attrib))
validate_contains_keys(setup, ['name'], 'ITEM')
base.add(**setup) # register parameter in model
elif element.tag == 'ITEMLIST':
setup = _translate_ctd_to_param(dict(element.attrib))
setup['default'] = [listitem.attrib['value'] for listitem in element]
setup['is_list'] = True
validate_contains_keys(setup, ['name'], 'ITEMLIST')
base.add(**setup) # register list parameter in model
def add(self, name, **kwargs):
"""Registers a top level parameter to the model. Required: `name` string.
Optional keyword arguments:
`type`: Python type object, or a string of a valid CTD types.
For all valid values, see: CTDopts.CTDTYPE_TO_TYPE.keys()
`default`: default value. Will be casted to the above type (default None)
`is_list`: bool, indicating whether this is a list parameter (default False)
`required`: bool, indicating whether this is a required parameter (default False)
`description`: string containing parameter description (default None)
`tags`: list of strings or comma separated string (default [])
`num_range`: (min, max) tuple. None in either position makes it unlimited
`choices`: list of allowed values (controlled vocabulary)
`file_formats`: list of allowed file extensions
`short_name`: string for short name annotation
"""
return self.parameters.add(name, **kwargs)
def add_group(self, name, description=None):
"""Registers a top level parameter group to the model. Required: `name` string. Optional: `description`
"""
return self.parameters.add_group(name, description)
def list_parameters(self):
"""Returns a list of all Parameter objects registered in the model.
"""
# root node will list all its children (recursively, if they are nested in ParameterGroups)
return self.parameters._get_children()
def get_defaults(self):
"""Returns a nested dictionary with all parameters of the model having default values.
"""
params_w_default = (p for p in self.list_parameters() if p.default is not _Null)
defaults = {}
for param in params_w_default:
set_nested_key(defaults, param.get_lineage(name_only=True), param.default)
return defaults
def validate_args(self, args_dict, enforce_required=0, enforce_type=0, enforce_restrictions=0):
"""Validates an argument dictionary against the model, and returns a type-casted argument
dictionary with defaults for missing arguments. Valid values for `enforce_required`,
`enforce_type` and `enforce_restrictions` are 0, 1 and 2, where the different levels are:
* 0: doesn't enforce anything,
* 1: raises a warning
* 2: raises an exception
"""
# iterate over model parameters, look them up in the argument dictionary, convert to correct type,
# use default if argument is not present and raise exception if required argument is missing.
validated_args = {} # OrderedDict()
all_params = self.list_parameters()
for param in all_params:
lineage = param.get_lineage(name_only=True)
try:
arg = get_nested_key(args_dict, lineage)
# boolean values are the only ones that don't get casted correctly with, say, bool('false')
typecast = param.type if param.type is not bool else CAST_BOOLEAN
try:
validated_value = list(map(typecast, arg)) if param.is_list else typecast(arg)
except ValueError: # type casting failed
validated_value = arg # just keep it as a string (or list of strings)
if enforce_type: # but raise a warning or exception depending on enforcement level
if enforce_type == 1:
warnings.warn('Argument %s is of wrong type. Expected %s, got: %s' %
(':'.join(lineage), TYPE_TO_CTDTYPE[param.type], arg))
else:
raise ArgumentTypeError(param, arg)
if enforce_restrictions and param.restrictions and not param.restrictions.check(validated_value):
if enforce_restrictions == 1:
warnings.warn('Argument restrictions for %s violated. Restriction: %s. Value: %s' %
(':'.join(lineage), param.restrictions.ctd_restriction_string(), validated_value))
else:
raise ArgumentRestrictionError(param, validated_value)
set_nested_key(validated_args, lineage, validated_value)
except KeyError: # argument was not found, checking whether required and using defaults if not
if param.required:
if not enforce_required:
continue # this argument will be missing from the dict as required fields have no default value
elif enforce_required == 1:
warnings.warn('Required argument %s missing' % ':'.join(lineage), UserWarning)
else:
raise ArgumentMissingError(param)
else:
set_nested_key(validated_args, lineage, param.default)
return validated_args
def parse_cl_args(self, cl_args=None, prefix='--', short_prefix="-", get_remaining=False):
"""Parses command line arguments `cl_args` (either a string or a list like sys.argv[1:])
assuming that parameter names are prefixed by `prefix` (default '--').
Returns a nested dictionary with found arguments. Note that parameters have to be registered
in the model to be parsed and returned.
Remaining (unmatchable) command line arguments can be accessed if the method is called with
`get_remaining`. In this case, the method returns a tuple, whose first element is the
argument dictionary, the second a list of unmatchable command line options.
"""
cl_parser = argparse.ArgumentParser()
for param in self.list_parameters():
lineage = param.get_lineage(name_only=True)
short_lineage = param.get_lineage(name_only=True, short_name=True)
cl_arg_kws = {} # argument processing info passed to argparse in keyword arguments, we build them here
if param.type is bool: # boolean flags are not followed by a value, only their presence is required
cl_arg_kws['action'] = 'store_true'
else:
# we take every argument as string and cast them only later in validate_args() if
# explicitly asked for. This is because we don't want to deal with type exceptions
# at this stage, and prefer the multi-leveled strictness settings in validate_args()
cl_arg_kws['type'] = str
if param.is_list:
# or '+' rather? Should we allow empty lists here? If default is a proper list with elements
# that we want to clear, this would be the only way to do it so I'm inclined to use '*'
cl_arg_kws['nargs'] = '*'
if param.default is not _Null():
cl_arg_kws['default'] = param.default
if param.required:
cl_arg_kws['required'] = True
# hardcoded 'group:subgroup:param1'
if all(a is not _Null for a in short_lineage):
cl_parser.add_argument(short_prefix+':'.join(short_lineage), prefix + ':'.join(lineage), **cl_arg_kws)
else:
cl_parser.add_argument(prefix + ':'.join(lineage), **cl_arg_kws)
cl_arg_list = cl_args.split() if isinstance(cl_args, str) else cl_args
#if no arguments are given print help
if not cl_arg_list:
cl_arg_list.append("-h")
parsed_args, rest = cl_parser.parse_known_args(cl_arg_list)
res_args = {} # OrderedDict()
for param_name, value in vars(parsed_args).items():
# None values are created by argparse if it didn't find the argument or default=None, we skip params
# that dont have a default value
if value is not None or value == self.parameters.parameters[param_name].default:
set_nested_key(res_args, param_name.split(':'), value)
return res_args if not get_remaining else (res_args, rest)
def generate_ctd_tree(self, arg_dict=None, log=None, cli=False, prefix='--'):
"""Generates an XML ElementTree from the model and returns the top <tool> Element object,
that can be output to a file (CTDModel.write_ctd() does everything needed if the user
doesn't need access to the actual element-tree).
Calling this function without any arguments generates the tool-describing CTD with default
values. For parameter-storing and logging optional arguments can be passed:
`arg_dict`: nested dictionary with values to be used instead of defaults.
`log`: dictionary with the following optional keys:
'time_start' and 'time_finish': proper XML date strings (eg. datetime.datetime.now(pytz.utc).isoformat())
'status': exit status
'output': standard output or whatever output the user intends to log
'warning': warning logs
'error': standard error or whatever error log the user wants to store
`cli`: boolean whether or not cli elements should be generated (needed for GenericKNIMENode for example)
"""
tool_attribs = OrderedDict()
tool_attribs['version'] = self.version
tool_attribs['name'] = self.name
tool_attribs['xmlns:xsi'] = "http://www.w3.org/2001/XMLSchema-instance"
tool_attribs['xsi:schemaLocation'] = "https://github.com/genericworkflownodes/CTDopts/raw/master/schemas/CTD_0_3.xsd"
opt_attribs = ['docurl', 'category']
for oo in opt_attribs:
if oo in self.opt_attribs:
tool_attribs[oo] = self.opt_attribs[oo]
tool = Element('tool', tool_attribs) # CTD root
opt_elements = ['manual', 'description', 'executableName', 'executablePath']
for oo in opt_elements:
if oo in self.opt_attribs:
SubElement(tool, oo).text = self.opt_attribs[oo]
if log is not None:
# log is supposed to be a dictionary, with the following keys (none of them being required):
# time_start, time_finish, status, output, warning, error
# generate
log_node = SubElement(tool, 'log')
if 'time_start' in log: # expect proper XML date string like datetime.datetime.now(pytz.utc).isoformat()
log_node.attrib['executionTimeStart'] = log['time_start']
if 'time_finish' in log:
log_node.attrib['executionTimeStop'] = log['time_finish']
if 'status' in log:
log_node.attrib['executionStatus'] = log['status']
if 'output' in log:
SubElement(log_node, 'executionMessage').text = log['output']
if 'warning' in log:
SubElement(log_node, 'executionWarning').text = log['warning']
if 'error' in log:
SubElement(log_node, 'executionError').text = log['error']
# XML.ETREE SYNTAX
params = SubElement(tool, 'PARAMETERS', {
'version': '1.6.2',
'xmlns:xsi': "http://www.w3.org/2001/XMLSchema-instance",
'xsi:noNamespaceSchemaLocation': "https://github.com/genericworkflownodes/CTDopts/raw/master/schemas/Param_1_6_2.xsd"
})
# This seems to be some OpenMS hack (defining name, description, version for the second time)
# but I'll stick to it for consistency
top_node = SubElement(params, 'NODE', name=self.name, description=self.opt_attribs.get('description', ''))
SubElement(top_node, 'ITEM',
name='version',
value=self.version,
type='string',
description='Version of the tool that generated this parameters file.',
tags='advanced')
# all the above was boilerplate, now comes the actual parameter tree generation
args_top_node = self.parameters._xml_node(arg_dict)
top_node.append(args_top_node)
if cli:
cli_node = SubElement(tool, "cli")
for e in self.parameters._cli_node(parent_name=self.name, prefix=prefix):
cli_node.append(e)
# # LXML w/ pretty print syntax
# return tostring(tool, pretty_print=True, xml_declaration=True, encoding="UTF-8")
# xml.etree syntax (no pretty print available, so we use xml.dom.minidom stuff)
return tool
def write_ctd(self, out_file, arg_dict=None, log=None, cli=False):
"""Generates a CTD XML from the model and writes it to `out_file`, which is either a string
to a file path or a stream with a write() method.
Calling this function without any arguments besides `out_file` generates the tool-describing
CTD with default values. For parameter-storing and logging optional arguments can be passed:
`arg_dict`: nested dictionary with values to be used instead of defaults.
`log`: dictionary with the following optional keys:
'time_start' and 'time_finish': proper XML date strings (eg. datetime.datetime.now(pytz.utc).isoformat())
'status': exit status
'output': standard output or whatever output the user intends to log
'warning': warning logs
'error': standard error or whatever error log the user wants to store
`cli`: boolean whether or not cli elements should be generated (needed for GenericKNIMENode for example)
"""
xml_content = parseString(tostring(self.generate_ctd_tree(arg_dict, log, cli), encoding="UTF-8")).toprettyxml()
if isinstance(out_file, str): # if out_file is a string, we create and write the file
with open(out_file, 'w') as f:
f.write(xml_content)
else: # otherwise we assume it's a writable stream and write into that.
out_file.write(xml_content)
def args_from_file(filename):
"""Takes a CTD file and returns a nested dictionary with all argument values found. It's not
linked to a model, so there's no type casting or validation done on the arguments. This is useful
for users who just want to access arguments in CTD files without having to deal with building a CTD model.
If type casting or validation is required, two things can be done to hack one's way around it:
Build a model from the same file and call get_defaults() on it. This takes advantage from the
fact that when building a model from a CTD, the value attributes are used as defaults. Although
one shouldn't build a model from an argument storing CTD (as opposed to tool describing CTDs)
there's no technical obstacle to do so.
"""
def get_args(element, base=None):
# recursive argument lookup if encountering <NODE>s
if element.tag == 'NODE':
current_group = {} # OrderedDict()
for child in element:
get_args(child, current_group)
if base is not None:
base[element.attrib['name']] = current_group
else:
# top level <NODE name='1'> is the only one called with base=None.
# As the argument parsing is recursive, whenever the top node finishes, we are done
# with the parsing and have to return the results.
return current_group
elif element.tag == 'ITEM':
if 'value' in element.attrib:
base[element.attrib['name']] = element.attrib['value']
elif element.tag == 'ITEMLIST':
if element.getchildren():
base[element.attrib['name']] = [listitem.attrib['value'] for listitem in element]
root = parse(filename).getroot()
param_root = root if root.tag == 'PARAMETERS' else root.find('PARAMETERS')
parameters = param_root.find('NODE').find('NODE')
return get_args(parameters, base=None)
def parse_cl_directives(cl_args, write_tool_ctd='write_tool_ctd', write_param_ctd='write_param_ctd',
input_ctd='input_ctd', prefix='--'):
'''Parses command line CTD processing directives. `write_tool_ctd`, `write_param_ctd` and `input_ctd`
string are customizable, and will be parsed for in command line. `prefix` should be one or two dashes,
default is '--'.
Returns a dictionary with keys
'write_tool_ctd': if flag set, either True or the filename provided in command line. Otherwise None.
'write_param_ctd': if flag set, either True or the filename provided in command line. Otherwise None.
'input_ctd': filename if found, otherwise None
'''
parser = argparse.ArgumentParser()
parser.add_argument(prefix + write_tool_ctd, nargs='*')
parser.add_argument(prefix + write_param_ctd, nargs='*')
parser.add_argument(prefix + input_ctd, type=str)
cl_arg_list = cl_args.split() if isinstance(cl_args, str) else cl_args # string or list of args
directives, rest = parser.parse_known_args(cl_arg_list)
directives = vars(directives)
transform = lambda x: None if x is None else True if x == [] else x[0]
parsed_directives = {}
parsed_directives['write_tool_ctd'] = transform(directives[write_tool_ctd])
parsed_directives['write_param_ctd'] = transform(directives[write_param_ctd])
parsed_directives['input_ctd'] = directives[input_ctd]
return parsed_directives
# TODO: ElementTree does not provide line information... maybe refactor using lxml or other parser that does support it?
def validate_contains_keys(dictionary, keys, element_tag):
for key in keys:
assert key in dictionary, "Missing required attribute '%s' in %s element. Present attributes: %s" % \
(key, element_tag,
', '.join(['{0}="{1}"'.format(k, v) for k, v in dictionary.items()]))
|