""" A validator for ASCII MOCs, as per MOC 1.1, (see http://ivoa.net/documents/MOC/). To you use it, just send an ASCII MOC to stdin; more or less readable messages will come out. To run the embedded tests, call it with a "test" command line argument. This script requires pyparsing installed (debian package: python-pyparsing or python3-pyparsing; or do pip install pyparsing). Written 2019-06-07 by Markus Demleitner This code is in the public domain. """ from pyparsing import (nums, ParserElement, Regex, Optional, NotAny, ZeroOrMore, OneOrMore, Group, ParseException, ParseSyntaxException, ParseFatalException) def get_grammar(): """returns a pyparsing grammar for ASCII MOCs. This grammar's form is slightly different from what's printed in the spec because of implementation considerations; it should be easy in all cases to see the equivalence; the only really ugly part is the negative lookahead to tell the order from a pixel. It certainly would have been nicer to have two different separators for pix and ordpix. """ ParserElement.setDefaultWhitespaceChars('') int = Regex("[0-9]+").setName("integer") separator = Regex(r"[ ,\n\r]").setName("separator") pix = (Group(Optional( int + Optional('-' + int)) + NotAny('/')) .setName("pix") .setResultsName('pix', listAllMatches=True)) pixs = (pix + ZeroOrMore(OneOrMore(separator) + pix)).setName("pixs") ordpix = (Group(int.setResultsName('order', listAllMatches=True) + '/' + pixs) .setName("ordpix") .setResultsName('ordpix', listAllMatches=True)) ordpix.setParseAction(validate_ordpix) moc = ordpix + ZeroOrMore( OneOrMore(separator) + ordpix).setName("moc") return moc def validate_ordpix(s, loc, toks): """ Raise a ParseFatalException if the parsed ordpix is not semantically valid. So far this just checks that pixel numbers are legal for the order. """ try: d = toks.asDict() ordpix = d['ordpix'][0] order = int(ordpix['order'][0]) for pix in ordpix['pix']: validate_pix(order, pix) except (Exception) as ex: raise ParseFatalException(s, loc, str(ex)) def validate_pix(order, pix): """ Raise an Exception if the given parsed pix is legal at the specified order. pix is either a list of one integer string, or a list of [int_string, '-', 'int_string]. """ max_pix = num_pix(order) - 1 if len(pix) == 1: # pix is a single pixel val = int(pix[0]) if not (val <= max_pix): raise Exception("Illegal pixel {}; order {} pixels must be <= to {}" .format(val, order, max_pix)) elif len(pix) == 3: # pix is a range. minval = int(pix[0]) maxval = int(pix[2]) if not (minval < maxval <= max_pix): raise Exception( "Illegal pixel range {}-{}; min must be < max," " and order {} pixels must be <= to {}" .format(minval, maxval, order, max_pix)) def num_pix(order): """ Returns the number of HEALPix pixels at the given order. The legal range of pixel numbers at order n is [0, num_pix(n)-1] """ return 12 * 4**order def validate(moc, grammar=get_grammar()): """returns either "OK" or a string with "Error" depending on whether moc matches the grammar for version 1.1 ASCII MOCs. """ try: res = grammar.parseString(moc, parseAll=True) except (ParseException, ParseSyntaxException, ParseFatalException) as ex: return "Error: {}".format(ex) return "OK" ########## End of validator. Below here: validator validation GOOD_MOCS = [ '1/', '0/1', '0/1-11', '0/1,2 4\n5,\n6,,,, \n\r11', '0/1 2/ 23 12,9 5/1-13,, ,29', '1/1,2,4 2/12-14,21,23,25 8/', # pixel range legal boundary cases '0/1 2/ 23 12,9 5/1-13,, ,12287', '0/1 2/ 23 12,9 5/1-13,, ,22,25/13510798882111487', '0/1 2/ 23 12,9 5/1-13,, ,22,29/3458764513820540927', ] BAD_MOCS = [ # These come with expected error messages (',', 'Error: Expected integer (at char 0), (line:1, col:1)'), ('1,', 'Error: Expected "/" (at char 1), (line:1, col:2)'), (',1', 'Error: Expected integer (at char 0), (line:1, col:1)'), ('aber/1', 'Error: Expected integer (at char 0), (line:1, col:1)'), ('1/aber', 'Error: Expected end of text (at char 2), (line:1, col:3)'), ('1/12-13-14', 'Error: Expected end of text (at char 7), (line:1, col:8)'), ('/12', 'Error: Expected integer (at char 0), (line:1, col:1)'), ('1/12 /29', 'Error: Expected end of text (at char 4), (line:1, col:5)'), ('1 /12', 'Error: Expected "/" (at char 1), (line:1, col:2)'), ('1/12 2 /12', 'Error: Expected end of text (at char 6), (line:1, col:7)'), # pixel range errors ('0/1,2 4\n5,\n6,,,, \n\r12', 'Error: Illegal pixel 12; order 0 pixels must be <= to 11' ' (at char 0), (line:1, col:1)'), ('0/15 2/ 23 12,9 5/1-13,, ,29', 'Error: Illegal pixel 15; order 0 pixels must be <= to 11' ' (at char 0), (line:1, col:1)'), ('0/1 2/ 230 12,9 5/1-13,, ,29', 'Error: Illegal pixel 230; order 2 pixels must be <= to 191' ' (at char 4), (line:1, col:5)'), ('0/1 2/ 23 12,9 5/18-13,, ,29', 'Error: Illegal pixel range 18-13; min must be < max, and order 5' ' pixels must be <= to 12287 (at char 15), (line:1, col:16)'), ('0/1 2/ 23 12,9 5/1-13,, ,12288', 'Error: Illegal pixel 12288; order 5 pixels must be <= to 12287' ' (at char 15), (line:1, col:16)'), ('0/1 2/ 23 12,9 5/1-13,, ,22,25/13510798882111488', 'Error: Illegal pixel 13510798882111488; order 25 pixels must be' ' <= to 13510798882111487 (at char 28), (line:1, col:29)'), ('0/1 2/ 23 12,9 5/1-13,, ,22,29/3458764513820540928', 'Error: Illegal pixel 3458764513820540928; order 29 pixels must be' ' <= to 3458764513820540927 (at char 28), (line:1, col:29)'), ('1/1,2,4 2/12-14,3000,23,25 8/', 'Error: Illegal pixel 3000; order 2 pixels must be <= to 191' ' (at char 8), (line:1, col:9)'), ] def _test(): for literal in GOOD_MOCS: res = validate(literal) if res != "OK": print("Failure for MOC {}: Should be ok, but rejected as {}".format( literal, res)) for literal, message in BAD_MOCS: res = validate(literal) if res != message: print("Failure for MOC {}: Message should have been '{}' but" " was '{}'.".format(literal, message, res)) if __name__ == "__main__": import sys if len(sys.argv) > 1 and sys.argv[1] == "test": _test() elif len(sys.argv) > 1: with open(sys.argv[1], 'r') as f: mocstring = f.read() res = validate(mocstring) print(res) else: res = validate(sys.stdin.read()) print(res)