File: NodeSet.py

package info (click to toggle)
clustershell 1.9.3-1
  • links: PTS, VCS
  • area: main
  • in suites: forky, sid, trixie
  • size: 2,228 kB
  • sloc: python: 20,978; makefile: 149
file content (1591 lines) | stat: -rw-r--r-- 61,797 bytes parent folder | download
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
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
1176
1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
1199
1200
1201
1202
1203
1204
1205
1206
1207
1208
1209
1210
1211
1212
1213
1214
1215
1216
1217
1218
1219
1220
1221
1222
1223
1224
1225
1226
1227
1228
1229
1230
1231
1232
1233
1234
1235
1236
1237
1238
1239
1240
1241
1242
1243
1244
1245
1246
1247
1248
1249
1250
1251
1252
1253
1254
1255
1256
1257
1258
1259
1260
1261
1262
1263
1264
1265
1266
1267
1268
1269
1270
1271
1272
1273
1274
1275
1276
1277
1278
1279
1280
1281
1282
1283
1284
1285
1286
1287
1288
1289
1290
1291
1292
1293
1294
1295
1296
1297
1298
1299
1300
1301
1302
1303
1304
1305
1306
1307
1308
1309
1310
1311
1312
1313
1314
1315
1316
1317
1318
1319
1320
1321
1322
1323
1324
1325
1326
1327
1328
1329
1330
1331
1332
1333
1334
1335
1336
1337
1338
1339
1340
1341
1342
1343
1344
1345
1346
1347
1348
1349
1350
1351
1352
1353
1354
1355
1356
1357
1358
1359
1360
1361
1362
1363
1364
1365
1366
1367
1368
1369
1370
1371
1372
1373
1374
1375
1376
1377
1378
1379
1380
1381
1382
1383
1384
1385
1386
1387
1388
1389
1390
1391
1392
1393
1394
1395
1396
1397
1398
1399
1400
1401
1402
1403
1404
1405
1406
1407
1408
1409
1410
1411
1412
1413
1414
1415
1416
1417
1418
1419
1420
1421
1422
1423
1424
1425
1426
1427
1428
1429
1430
1431
1432
1433
1434
1435
1436
1437
1438
1439
1440
1441
1442
1443
1444
1445
1446
1447
1448
1449
1450
1451
1452
1453
1454
1455
1456
1457
1458
1459
1460
1461
1462
1463
1464
1465
1466
1467
1468
1469
1470
1471
1472
1473
1474
1475
1476
1477
1478
1479
1480
1481
1482
1483
1484
1485
1486
1487
1488
1489
1490
1491
1492
1493
1494
1495
1496
1497
1498
1499
1500
1501
1502
1503
1504
1505
1506
1507
1508
1509
1510
1511
1512
1513
1514
1515
1516
1517
1518
1519
1520
1521
1522
1523
1524
1525
1526
1527
1528
1529
1530
1531
1532
1533
1534
1535
1536
1537
1538
1539
1540
1541
1542
1543
1544
1545
1546
1547
1548
1549
1550
1551
1552
1553
1554
1555
1556
1557
1558
1559
1560
1561
1562
1563
1564
1565
1566
1567
1568
1569
1570
1571
1572
1573
1574
1575
1576
1577
1578
1579
1580
1581
1582
1583
1584
1585
1586
1587
1588
1589
1590
1591
#
# Copyright (C) 2007-2016 CEA/DAM
# Copyright (C) 2007-2017 Aurelien Degremont <aurelien.degremont@cea.fr>
# Copyright (C) 2015-2017 Stephane Thiell <sthiell@stanford.edu>
#
# This file is part of ClusterShell.
#
# ClusterShell is free software; you can redistribute it and/or
# modify it under the terms of the GNU Lesser General Public
# License as published by the Free Software Foundation; either
# version 2.1 of the License, or (at your option) any later version.
#
# ClusterShell is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
# Lesser General Public License for more details.
#
# You should have received a copy of the GNU Lesser General Public
# License along with ClusterShell; if not, write to the Free Software
# Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA

"""
Cluster node set module.

A module to efficiently deal with node sets and node groups.
Instances of NodeSet provide similar operations than the builtin set() type,
see http://www.python.org/doc/lib/set-objects.html

Usage example
=============
  >>> # Import NodeSet class
  ... from ClusterShell.NodeSet import NodeSet
  >>>
  >>> # Create a new nodeset from string
  ... nodeset = NodeSet("cluster[1-30]")
  >>> # Add cluster32 to nodeset
  ... nodeset.update("cluster32")
  >>> # Remove from nodeset
  ... nodeset.difference_update("cluster[2-5,8-31]")
  >>> # Print nodeset as a pdsh-like pattern
  ... print nodeset
  cluster[1,6-7,32]
  >>> # Iterate over node names in nodeset
  ... for node in nodeset:
  ...     print node
  cluster1
  cluster6
  cluster7
  cluster32
"""

import fnmatch
import re
import string
import sys

from ClusterShell.Defaults import config_paths, DEFAULTS
import ClusterShell.NodeUtils as NodeUtils

# Import all RangeSet module public objects
from ClusterShell.RangeSet import RangeSet, RangeSetND, AUTOSTEP_DISABLED
from ClusterShell.RangeSet import RangeSetParseError


# Python 3 compatibility
try:
    basestring
except NameError:
    basestring = str

# Define default GroupResolver object used by NodeSet
DEF_GROUPS_CONFIGS = config_paths('groups.conf')
ILLEGAL_GROUP_CHARS = set("@,!&^*")
_DEF_RESOLVER_STD_GROUP = NodeUtils.GroupResolverConfig(DEF_GROUPS_CONFIGS,
                                                        ILLEGAL_GROUP_CHARS)
# Standard group resolver
RESOLVER_STD_GROUP = _DEF_RESOLVER_STD_GROUP
# Special constants for NodeSet's resolver parameter
#   RESOLVER_NOGROUP => avoid any group resolution at all
#   RESOLVER_NOINIT  => reserved use for optimized copy()
RESOLVER_NOGROUP = -1
RESOLVER_NOINIT = -2
# 1.5 compat (deprecated)
STD_GROUP_RESOLVER = RESOLVER_STD_GROUP
NOGROUP_RESOLVER = RESOLVER_NOGROUP


class NodeSetException(Exception):
    """Base NodeSet exception class."""

class NodeSetError(NodeSetException):
    """Raised when an error is encountered."""

class NodeSetParseError(NodeSetError):
    """Raised when NodeSet parsing cannot be done properly."""
    def __init__(self, part, msg):
        if part:
            msg = "%s: \"%s\"" % (msg, part)
        NodeSetError.__init__(self, msg)
        # faulty part; this allows you to target the error
        self.part = part

class NodeSetParseRangeError(NodeSetParseError):
    """Raised when bad range is encountered during NodeSet parsing."""
    def __init__(self, rset_exc):
        NodeSetParseError.__init__(self, str(rset_exc), "bad range")

class NodeSetExternalError(NodeSetError):
    """Raised when an external error is encountered."""


class NodeSetBase(object):
    """
    Base class for NodeSet.

    This class allows node set base object creation from specified string
    pattern and rangeset object.  If optional copy_rangeset boolean flag is
    set to True (default), provided rangeset object is copied (if needed),
    otherwise it may be referenced (should be seen as an ownership transfer
    upon creation).

    This class implements core node set arithmetic (no string parsing here).

    Example:
       >>> nsb = NodeSetBase('node%s-ipmi', RangeSet('1-5,7'), False)
       >>> str(nsb)
       'node[1-5,7]-ipmi'
       >>> nsb = NodeSetBase('node%s-ib%s', RangeSetND([['1-5,7', '1-2']]), False)
       >>> str(nsb)
       'node[1-5,7]-ib[1-2]'
    """
    def __init__(self, pattern=None, rangeset=None, copy_rangeset=True,
                 autostep=None, fold_axis=None):
        """New NodeSetBase object initializer"""
        self._autostep = autostep
        self._length = 0
        self._patterns = {}
        self.fold_axis = fold_axis  #: iterable over nD 0-indexed axis
        if self.fold_axis is None and DEFAULTS.fold_axis:
            self.fold_axis = DEFAULTS.fold_axis  # non-empty tuple
        if pattern:
            self._add(pattern, rangeset, copy_rangeset)
        elif rangeset:
            raise ValueError("missing pattern")

    def get_autostep(self):
        """Get autostep value (property)"""
        return self._autostep

    def set_autostep(self, val):
        """Set autostep value (property)"""
        if val is None:
            self._autostep = None
        else:
            # Work around the pickling issue of sys.maxint (+inf) in py2.4
            self._autostep = min(int(val), AUTOSTEP_DISABLED)

        # Update our RangeSet/RangeSetND objects
        for pat, rset in self._patterns.items():
            if rset:
                rset.autostep = self._autostep

    autostep = property(get_autostep, set_autostep)

    def _iter(self):
        """Iterator on internal item tuples
            (pattern, indexes, autostep)."""
        for pat, rset in sorted(self._patterns.items()):
            if rset:
                autostep = rset.autostep
                if rset.dim() == 1:
                    assert isinstance(rset, RangeSet)
                    for idx in rset:
                        yield pat, (idx,), autostep
                else:
                    for rvec in rset:
                        yield pat, rvec, autostep
            else:
                yield pat, None, None

    def _iterbase(self):
        """Iterator on single, one-item NodeSetBase objects."""
        for pat, ivec, autostep in self._iter():
            rset = None     # 'no node index' by default
            if ivec is not None:
                assert len(ivec) > 0
                if len(ivec) == 1:
                    rset = RangeSet.fromone(ivec[0], autostep)
                else:
                    rset = RangeSetND([ivec], autostep)
            yield NodeSetBase(pat, rset)

    def __iter__(self):
        """Iterator on single nodes as string."""
        # Does not call self._iterbase() + str() for better performance.
        for pat, ivec, _ in self._iter():
            if ivec is not None:
                yield pat % ivec
            else:
                yield pat % ()

    # define striter() alias for convenience (to match RangeSet.striter())
    striter = __iter__

    # define nsiter() as an object-based iterator that could be used for
    # __iter__() in the future...

    def nsiter(self):
        """Object-based NodeSet iterator on single nodes."""
        for pat, ivec, autostep in self._iter():
            nodeset = self.__class__()
            if ivec is not None:
                if len(ivec) == 1:
                    nodeset._add_new(pat, RangeSet.fromone(ivec[0]))
                else:
                    nodeset._add_new(pat, RangeSetND([ivec], autostep))
            else:
                nodeset._add_new(pat, None)
            yield nodeset

    def contiguous(self):
        """Object-based NodeSet iterator on contiguous node sets.

        Contiguous node set contains nodes with same pattern name and a
        contiguous range of indexes, like foobar[1-100]."""
        for pat, rangeset in sorted(self._patterns.items()):
            if rangeset:
                for cont_rset in rangeset.contiguous():
                    nodeset = self.__class__()
                    nodeset._add_new(pat, cont_rset)
                    yield nodeset
            else:
                nodeset = self.__class__()
                nodeset._add_new(pat, None)
                yield nodeset

    def __len__(self):
        """Get the number of nodes in NodeSet."""
        cnt = 0
        for rangeset in self._patterns.values():
            if rangeset:
                cnt += len(rangeset)
            else:
                cnt += 1
        return cnt

    def _iter_nd_pat(self, pat, rset):
        """
        Take a pattern and a RangeSetND object and iterate over nD computed
        nodeset strings while following fold_axis constraints.
        """
        try:
            dimcnt = rset.dim()
            if self.fold_axis is None:
                # fold along all axis (default)
                fold_axis = range(dimcnt)
            else:
                # set of user-provided fold axis (support negative numbers)
                fold_axis = [int(x) % dimcnt for x in self.fold_axis
                             if -dimcnt <= int(x) < dimcnt]
        except (TypeError, ValueError) as exc:
            raise NodeSetParseError("fold_axis=%s" % self.fold_axis, exc)

        for rgvec in rset.vectors():
            rgnargs = []    # list of str rangeset args
            for axis, rangeset in enumerate(rgvec):
                # build an iterator over rangeset strings to add
                if len(rangeset) > 1:
                    if axis not in fold_axis: # expand
                        rgstrit = rangeset.striter()
                    else:
                        rgstrit = ["[%s]" % rangeset]
                else:
                    rgstrit = [str(rangeset)]

                # aggregate/expand along previous computed axis...
                t_rgnargs = []
                for rgstr in rgstrit: # 1-time when not expanding
                    if not rgnargs:
                        t_rgnargs.append([rgstr])
                    else:
                        for rga in rgnargs:
                            t_rgnargs.append(rga + [rgstr])
                rgnargs = t_rgnargs

            # get nodeset patterns formatted with range strings
            for rgargs in rgnargs:
                yield pat % tuple(rgargs)

    def __str__(self):
        """Get ranges-based pattern of node list."""
        results = []
        try:
            for pat, rset in sorted(self._patterns.items()):
                if not rset:
                    results.append(pat % ())
                elif rset.dim() == 1:
                    # check if allowed to fold even for 1D pattern
                    if self.fold_axis is None or \
                            list(x for x in self.fold_axis if -1 <= int(x) < 1):
                        rgs = str(rset)
                        cnt = len(rset)
                        if cnt > 1:
                            rgs = "[%s]" % rgs
                        results.append(pat % rgs)
                    else:
                        results.extend((pat % rgs for rgs in rset.striter()))
                elif rset.dim() > 1:
                    results.extend(self._iter_nd_pat(pat, rset))
        except TypeError:
            raise NodeSetParseError(pat, "Internal error: node pattern and "
                                         "ranges mismatch")
        return ",".join(results)

    def copy(self):
        """Return a shallow copy."""
        cpy = self.__class__()
        cpy.fold_axis = self.fold_axis
        cpy._autostep = self._autostep
        cpy._length = self._length
        dic = {}
        for pat, rangeset in self._patterns.items():
            if rangeset is None:
                dic[pat] = None
            else:
                dic[pat] = rangeset.copy()
        cpy._patterns = dic
        return cpy

    def __contains__(self, other):
        """Is node contained in NodeSet ?"""
        return self.issuperset(other)

    def _binary_sanity_check(self, other):
        # check that the other argument to a binary operation is also
        # a NodeSet, raising a TypeError otherwise.
        if not isinstance(other, NodeSetBase):
            raise TypeError("Binary operation only permitted between "
                            "NodeSetBase")

    def issubset(self, other):
        """Report whether another nodeset contains this nodeset."""
        self._binary_sanity_check(other)
        return other.issuperset(self)

    def issuperset(self, other):
        """Report whether this nodeset contains another nodeset."""
        self._binary_sanity_check(other)
        status = True
        for pat, erangeset in other._patterns.items():
            rangeset = self._patterns.get(pat)
            if rangeset:
                status = rangeset.issuperset(erangeset)
            else:
                # might be an unnumbered node (key in dict but no value)
                status = pat in self._patterns
            if not status:
                break
        return status

    def __eq__(self, other):
        """NodeSet equality comparison."""
        # See comment for for RangeSet.__eq__()
        if not isinstance(other, NodeSetBase):
            return NotImplemented
        return len(self) == len(other) and self.issuperset(other)

    # inequality comparisons using the is-subset relation
    __le__ = issubset
    __ge__ = issuperset

    def __lt__(self, other):
        """x.__lt__(y) <==> x<y"""
        self._binary_sanity_check(other)
        return len(self) < len(other) and self.issubset(other)

    def __gt__(self, other):
        """x.__gt__(y) <==> x>y"""
        self._binary_sanity_check(other)
        return len(self) > len(other) and self.issuperset(other)

    def _extractslice(self, index):
        """Private utility function: extract slice parameters from slice object
        `index` for an list-like object of size `length`."""
        length = len(self)
        if index.start is None:
            sl_start = 0
        elif index.start < 0:
            sl_start = max(0, length + index.start)
        else:
            sl_start = index.start
        if index.stop is None:
            sl_stop = sys.maxsize
        elif index.stop < 0:
            sl_stop = max(0, length + index.stop)
        else:
            sl_stop = index.stop
        if index.step is None:
            sl_step = 1
        elif index.step < 0:
            # We support negative step slicing with no start/stop, ie. r[::-n].
            if index.start is not None or index.stop is not None:
                raise IndexError("illegal start and stop when negative step "
                                 "is used")
            # As RangeSet elements are ordered internally, adjust sl_start
            # to fake backward stepping in case of negative slice step.
            stepmod = (length + -index.step - 1) % -index.step
            if stepmod > 0:
                sl_start += stepmod
            sl_step = -index.step
        else:
            sl_step = index.step
        if not isinstance(sl_start, int) or not isinstance(sl_stop, int) \
            or not isinstance(sl_step, int):
            raise TypeError("slice indices must be integers")
        return sl_start, sl_stop, sl_step

    def __getitem__(self, index):
        """Return the node at specified index or a subnodeset when a slice is
        specified."""
        if isinstance(index, slice):
            inst = NodeSetBase()
            sl_start, sl_stop, sl_step = self._extractslice(index)
            sl_next = sl_start
            if sl_stop <= sl_next:
                return inst
            length = 0
            for pat, rangeset in sorted(self._patterns.items()):
                if rangeset:
                    cnt = len(rangeset)
                    offset = sl_next - length
                    if offset < cnt:
                        num = min(sl_stop - sl_next, cnt - offset)
                        inst._add(pat, rangeset[offset:offset + num:sl_step])
                    else:
                        #skip until sl_next is reached
                        length += cnt
                        continue
                else:
                    cnt = num = 1
                    if sl_next > length:
                        length += cnt
                        continue
                    inst._add(pat, None)
                # adjust sl_next...
                sl_next += num
                if (sl_next - sl_start) % sl_step:
                    sl_next = sl_start + \
                        ((sl_next - sl_start)/sl_step + 1) * sl_step
                if sl_next >= sl_stop:
                    break
                length += cnt
            return inst
        elif isinstance(index, int):
            if index < 0:
                length = len(self)
                if index >= -length:
                    index = length + index # - -index
                else:
                    raise IndexError("%d out of range" % index)
            length = 0
            for pat, rangeset in sorted(self._patterns.items()):
                if rangeset:
                    cnt = len(rangeset)
                    if index < length + cnt:
                        # return a subrangeset of size 1 to manage padding
                        if rangeset.dim() == 1:
                            return pat % rangeset[index-length:index-length+1]
                        else:
                            sub = rangeset[index-length:index-length+1]
                            for rgvec in sub.vectors():
                                return pat % (tuple(rgvec))
                else:
                    cnt = 1
                    if index == length:
                        return pat
                length += cnt
            raise IndexError("%d out of range" % index)
        else:
            raise TypeError("NodeSet indices must be integers")

    def _add_new(self, pat, rangeset):
        """Add nodes from a (pat, rangeset) tuple.
        Predicate: pattern does not exist in current set. RangeSet object is
        referenced (not copied)."""
        assert pat not in self._patterns
        self._patterns[pat] = rangeset

    def _add(self, pat, rangeset, copy_rangeset=True):
        """Add nodes from a (pat, rangeset) tuple.
        `pat' may be an existing pattern and `rangeset' may be None.
        RangeSet or RangeSetND objects are copied if re-used internally
        when provided and if copy_rangeset flag is set.
        """
        if pat in self._patterns:
            # existing pattern: get RangeSet or RangeSetND entry...
            pat_e = self._patterns[pat]
            # sanity checks
            if (pat_e is None) is not (rangeset is None):
                raise NodeSetError("Invalid operation")
            # entry may exist but set to None (single node)
            if pat_e:
                pat_e.update(rangeset)
        else:
            # new pattern...
            if rangeset and copy_rangeset:
                # default is to inherit rangeset autostep value
                rangeset = rangeset.copy()
                # but if set, self._autostep does override it
                if self._autostep is not None:
                    # works with rangeset 1D or nD
                    rangeset.autostep = self._autostep
            self._add_new(pat, rangeset)

    def union(self, other):
        """
        s.union(t) returns a new set with elements from both s and t.
        """
        self_copy = self.copy()
        self_copy.update(other)
        return self_copy

    def __or__(self, other):
        """
        Implements the | operator. So s | t returns a new nodeset with
        elements from both s and t.
        """
        if not isinstance(other, NodeSetBase):
            return NotImplemented
        return self.union(other)

    def add(self, other):
        """
        Add node to NodeSet.
        """
        self.update(other)

    def update(self, other):
        """
        s.update(t) updates nodeset s with elements added from t.
        """
        for pat, rangeset in other._patterns.items():
            self._add(pat, rangeset)

    def updaten(self, others):
        """
        s.updaten(list) updates nodeset s with elements added from given list.
        """
        for other in others:
            self.update(other)

    def clear(self):
        """
        Remove all nodes from this nodeset.
        """
        self._patterns.clear()

    def __ior__(self, other):
        """
        Implements the ``|=`` operator. So ``s |= t`` returns nodeset s
        with elements added from t. (Python version 2.5+ required)
        """
        self._binary_sanity_check(other)
        self.update(other)
        return self

    def intersection(self, other):
        """
        s.intersection(t) returns a new set with elements common to s
        and t.
        """
        self_copy = self.copy()
        self_copy.intersection_update(other)
        return self_copy

    def __and__(self, other):
        """
        Implements the & operator. So ``s & t`` returns a new nodeset with
        elements common to s and t.
        """
        if not isinstance(other, NodeSet):
            return NotImplemented
        return self.intersection(other)

    def intersection_update(self, other):
        """
        ``s.intersection_update(t)`` updates nodeset s keeping only
        elements also found in t.
        """
        if other is self:
            return

        tmp_ns = NodeSetBase()

        for pat, irangeset in other._patterns.items():
            rangeset = self._patterns.get(pat)
            if rangeset:
                irset = rangeset.intersection(irangeset)
                # ignore pattern if empty rangeset
                if len(irset) > 0:
                    tmp_ns._add(pat, irset, copy_rangeset=False)
            elif not irangeset and pat in self._patterns:
                # intersect two nodes with no rangeset
                tmp_ns._add(pat, None)

        # Substitute
        self._patterns = tmp_ns._patterns

    def __iand__(self, other):
        """
        Implements the &= operator. So ``s &= t`` returns nodeset s keeping
        only elements also found in t. (Python version 2.5+ required)
        """
        self._binary_sanity_check(other)
        self.intersection_update(other)
        return self

    def difference(self, other):
        """
        ``s.difference(t)`` returns a new NodeSet with elements in s but not
        in t.
        """
        self_copy = self.copy()
        self_copy.difference_update(other)
        return self_copy

    def __sub__(self, other):
        """
        Implement the - operator. So ``s - t`` returns a new nodeset with
        elements in s but not in t.
        """
        if not isinstance(other, NodeSetBase):
            return NotImplemented
        return self.difference(other)

    def difference_update(self, other, strict=False):
        """
        ``s.difference_update(t)`` removes from s all the elements found in t.

        :raises KeyError: an element cannot be removed (only if strict is
            True)
        """
        # the purge of each empty pattern is done afterward to allow self = ns
        purge_patterns = []

        # iterate first over exclude nodeset rangesets which is usually smaller
        for pat, erangeset in other._patterns.items():
            # if pattern is found, deal with it
            rangeset = self._patterns.get(pat)
            if rangeset:
                # sub rangeset, raise KeyError if not found
                rangeset.difference_update(erangeset, strict)

                # check if no range left and add pattern to purge list
                if len(rangeset) == 0:
                    purge_patterns.append(pat)
            else:
                # unnumbered node exclusion
                if pat in self._patterns:
                    purge_patterns.append(pat)
                elif strict:
                    raise KeyError(pat)

        for pat in purge_patterns:
            del self._patterns[pat]

    def __isub__(self, other):
        """
        Implement the -= operator. So ``s -= t`` returns nodeset s after
        removing elements found in t. (Python version 2.5+ required)
        """
        self._binary_sanity_check(other)
        self.difference_update(other)
        return self

    def remove(self, elem):
        """
        Remove element elem from the nodeset. Raise KeyError if elem
        is not contained in the nodeset.

        :raises KeyError: elem is not contained in the nodeset
        """
        self.difference_update(elem, True)

    def symmetric_difference(self, other):
        """
        ``s.symmetric_difference(t)`` returns the symmetric difference of
        two nodesets as a new NodeSet.

        (ie. all nodes that are in exactly one of the nodesets.)
        """
        self_copy = self.copy()
        self_copy.symmetric_difference_update(other)
        return self_copy

    def __xor__(self, other):
        """
        Implement the ^ operator. So ``s ^ t`` returns a new NodeSet with
        nodes that are in exactly one of the nodesets.
        """
        if not isinstance(other, NodeSet):
            return NotImplemented
        return self.symmetric_difference(other)

    def symmetric_difference_update(self, other):
        """
        ``s.symmetric_difference_update(t)`` updates nodeset s keeping all
        nodes that are in exactly one of the nodesets.
        """
        purge_patterns = []

        # iterate over our rangesets
        for pat, rangeset in self._patterns.items():
            brangeset = other._patterns.get(pat)
            if brangeset:
                rangeset.symmetric_difference_update(brangeset)
            else:
                if pat in other._patterns:
                    purge_patterns.append(pat)

        # iterate over other's rangesets
        for pat, brangeset in other._patterns.items():
            rangeset = self._patterns.get(pat)
            if not rangeset and not pat in self._patterns:
                self._add(pat, brangeset)

        # check for patterns cleanup
        for pat, rangeset in self._patterns.items():
            if rangeset is not None and len(rangeset) == 0:
                purge_patterns.append(pat)

        # cleanup
        for pat in purge_patterns:
            del self._patterns[pat]

    def __ixor__(self, other):
        """
        Implement the ^= operator. So ``s ^= t`` returns nodeset s after
        keeping all nodes that are in exactly one of the nodesets.
        (Python version 2.5+ required)
        """
        self._binary_sanity_check(other)
        self.symmetric_difference_update(other)
        return self


def _strip_escape(nsstr):
    """
    Helper to prepare a nodeset string for parsing: trim boundary
    whitespaces and escape special characters.
    """
    return nsstr.strip().replace('%', '%%')

def _rsets4nsb(rsets, autostep):
    """
    Helper to convert a list of RangeSet objects into the proper object
    for NodeSetBase: RangeSet, RangeSetND or None (no node index).
    """
    if len(rsets) > 1:
        return RangeSetND([rsets], None, autostep, copy_rangeset=False)
    elif len(rsets) == 1:
        return rsets[0]


class ParsingEngine(object):
    """
    Class that is able to transform a source into a NodeSetBase.
    """

    OP_CODES = {',': 'update',
                '!': 'difference_update',
                '&': 'intersection_update',
                '^': 'symmetric_difference_update'}

    OP_CODES_PAT = '[%s]' % re.escape(''.join(OP_CODES.keys()))

    BRACKET_OPEN = '['
    BRACKET_CLOSE = ']'

    def __init__(self, group_resolver, node_wildcard_enable=True):
        """
        Initialize Parsing Engine.
        """
        self.group_resolver = group_resolver
        self.base_node_re = re.compile(r"(\D*)(\d*)")
        self.node_wc = node_wildcard_enable  # node wildcard support

    def parse(self, nsobj, autostep):
        """
        Parse provided object if possible and return a NodeSetBase object.
        """
        # passing None is supported
        if nsobj is None:
            return NodeSetBase()

        # is nsobj a NodeSetBase instance?
        if isinstance(nsobj, NodeSetBase):
            return nsobj

        # or is nsobj a string?
        if isinstance(nsobj, basestring):
            try:
                return self.parse_string(str(nsobj), autostep)
            except (NodeUtils.GroupSourceQueryFailed, RuntimeError) as exc:
                raise NodeSetParseError(nsobj, str(exc))

        raise TypeError("Unsupported NodeSet input %s" % type(nsobj))

    def parse_string(self, nsstr, autostep, namespace=None):
        """Parse provided string in optional namespace.

        This method parses string, resolves all node groups, and
        computes set operations.

        Return a NodeSetBase object.
        """
        alln_cache = None  # used to compute 'all nodes' only once
        nodeset = NodeSetBase()
        nsstr = _strip_escape(nsstr)

        for opc, pat, rgnd in self._scan_string(nsstr, autostep):
            # Parser main debugging:
            #print "OPC %s PAT %s RANGESETS %s" % (opc, pat, rgnd)
            if self.group_resolver and pat[0] == '@':
                ns_group = NodeSetBase()
                for nodegroup in NodeSetBase(pat, rgnd):
                    # parse/expand nodes group: get group string and namespace
                    ns_str_ext, ns_nsp_ext = self.parse_group_string(nodegroup,
                                                                     namespace)
                    if ns_str_ext: # may still contain groups
                        # recursively parse and aggregate result
                        ns_group.update(self.parse_string(ns_str_ext,
                                                          autostep,
                                                          ns_nsp_ext))
                # perform operation
                getattr(nodeset, opc)(ns_group)

            elif self.group_resolver and self.node_wc and ('*' in pat or
                                                           '?' in pat):
                # We support ranges with wildcard mask by testing all nodes
                # against each expanded mask (wcmasks).
                wcmasks = (str(wcn) for wcn in NodeSetBase(pat, rgnd, False))

                # Our reference set is 'all nodes', we need to build it from
                # NodeSetBase to iterate over each individual node.
                if alln_cache is None:
                    self.node_wc = False  # avoid infinite recursion
                    try:
                        nsb = NodeSetBase()
                        for res in self.all_nodes(namespace):
                            nsb.update(self.parse_string(res, autostep,
                                                         namespace))
                        alln_cache = set(str(node) for node in nsb)
                    finally:
                        self.node_wc = True

                alln = alln_cache.copy()

                # A wildcarded nodeset can be seen as a single nodeset, so we
                # compute the union of nodes matching the wildcard mask(s) and
                # use the resulting NodeSetBase object as argument of the next
                # operation (opc).
                wcns = NodeSetBase()
                for wcmask in wcmasks:
                    # Expand nodes matching any of the wildcard mask
                    for node in fnmatch.filter(alln, wcmask):
                        alln.remove(node)  # remove matching node for next iter
                        wcp, wcr = self._scan_string_single(node, autostep)
                        wcrgnd = _rsets4nsb(wcr, autostep)
                        wcns.update(NodeSetBase(wcp, wcrgnd, False))
                getattr(nodeset, opc)(wcns)

            else:
                getattr(nodeset, opc)(NodeSetBase(pat, rgnd, False))

        return nodeset

    def parse_string_single(self, nsstr, autostep):
        """Parse provided string and return a NodeSetBase object."""
        pat, rangesets = self._scan_string_single(_strip_escape(nsstr),
                                                  autostep)
        if len(rangesets) > 1:
            rgobj = RangeSetND([rangesets], None, autostep, copy_rangeset=False)
        elif len(rangesets) == 1:
            rgobj = rangesets[0]
        else: # non-indexed nodename
            rgobj = None
        return NodeSetBase(pat, rgobj, False)

    def parse_group(self, group, namespace=None, autostep=None):
        """Parse provided single group name (without @ prefix)."""
        assert self.group_resolver is not None
        nodestr = self.group_resolver.group_nodes(group, namespace)
        return self.parse(",".join(nodestr), autostep)

    def parse_group_string(self, nodegroup, namespace=None):
        """Parse provided raw nodegroup string in optional namespace.

        Warning: 1 pass only, may still return groups.

        Return a tuple (grp_resolved_string, namespace).
        """
        assert nodegroup[0] == '@'
        assert self.group_resolver is not None
        grpstr = group = nodegroup[1:]
        if grpstr.find(':') >= 0:
            # specified namespace does always override
            namespace, group = grpstr.split(':', 1)
        if group == '*': # @* or @source:* magic
            reslist = self.all_nodes(namespace)
        elif group.startswith('@'): # @@source group name list
            reslist = self.grouplist(grpstr[1:])
        else:
            reslist = self.group_resolver.group_nodes(group, namespace)
        return ','.join(reslist), namespace

    def grouplist(self, namespace=None):
        """
        Return a sorted list of groups from current resolver (in optional
        group source / namespace).
        """
        grpset = NodeSetBase()
        for grpstr in self.group_resolver.grouplist(namespace):
            # We scan each group string to expand any range seen...
            grpstr = _strip_escape(grpstr)
            for opc, pat, rgnd in self._scan_string(grpstr, None):
                getattr(grpset, opc)(NodeSetBase(pat, rgnd, False))
        return list(grpset)

    def all_nodes(self, namespace=None):
        """Get all nodes from group resolver as a list of strings."""
        # namespace is the optional group source
        assert self.group_resolver is not None
        alln = []
        try:
            # Ask resolver to provide all nodes.
            alln = self.group_resolver.all_nodes(namespace)
        except NodeUtils.GroupSourceNoUpcall:
            try:
                # As the resolver is not able to provide all nodes directly,
                # failback to list + map(s) method:
                for grp in self.grouplist(namespace):
                    alln += self.group_resolver.group_nodes(grp, namespace)
            except NodeUtils.GroupSourceNoUpcall:
                # We are not able to find "all" nodes, definitely.
                msg = "Not enough working methods (all or map + list) to " \
                      "get all nodes"
                raise NodeSetExternalError(msg)
        except NodeUtils.GroupSourceQueryFailed as exc:
            raise NodeSetExternalError("Failed to get all nodes: %s" % exc)
        return alln

    def _next_op(self, pat):
        """Opcode parsing subroutine."""
        mobj = re.search(ParsingEngine.OP_CODES_PAT, pat)
        if mobj:
            return mobj.span()[0], mobj.group()
        else:
            return -1, None

    def _scan_string_single(self, nsstr, autostep):
        """Single node scan, returns (pat, list of rangesets)"""
        # single node parsing
        pfx_nd = [mobj.groups() for mobj in self.base_node_re.finditer(nsstr)]
        pfx_nd = pfx_nd[:-1]
        if not pfx_nd:
            raise NodeSetParseError(nsstr, "parse error")

        pat = ""
        rangesets = []
        for pfx, idx in pfx_nd:
            if idx:
                # optimization: process single index padding directly
                pad = 0
                if int(idx) != 0:
                    idxs = idx.lstrip("0")
                    if len(idx) - len(idxs) > 0:
                        pad = len(idx)
                    idxint = int(idxs)
                else:
                    if len(idx) > 1:
                        pad = len(idx)
                    idxint = 0
                if idxint > 1e100:
                    raise NodeSetParseRangeError( \
                        RangeSetParseError(idx, "invalid rangeset index"))
                # optimization: use numerical RangeSet constructor
                pat += "%s%%s" % pfx
                rangesets.append(RangeSet.fromone(idxint, pad, autostep))
            else:
                # undefined pad means no node index
                pat += pfx
        return pat, rangesets

    def _scan_string(self, nsstr, autostep):
        """Parsing engine's string scanner method (iterator)."""
        next_op_code = ','  # if no operator, default one is to update nodeset
        while nsstr:
            # Ignore whitespace(s) for convenience
            nsstr = nsstr.lstrip()

            rsets = []
            op_code = next_op_code

            op_idx, next_op_code = self._next_op(nsstr)
            bracket_idx = nsstr.find(self.BRACKET_OPEN)

            # Check if the operator is after the bracket, or if there
            # is no operator at all but some brackets.
            if bracket_idx >= 0 and (op_idx > bracket_idx or op_idx < 0):
                # In this case, we have a pattern of potentially several
                # nodes.
                # Fill prefix, range and suffix from pattern
                # eg. "forbin[3,4-10]-ilo" -> "forbin", "3,4-10", "-ilo"
                newpat = ""
                sfx = nsstr
                while bracket_idx >= 0 and (op_idx > bracket_idx or op_idx < 0):
                    pfx, sfx = sfx.split(self.BRACKET_OPEN, 1)
                    try:
                        rng, sfx = sfx.split(self.BRACKET_CLOSE, 1)
                    except ValueError:
                        raise NodeSetParseError(nsstr, "missing bracket")

                    # illegal closing bracket checks
                    if pfx.find(self.BRACKET_CLOSE) > -1:
                        raise NodeSetParseError(pfx, "illegal closing bracket")

                    if len(sfx) > 0:
                        bra_end = sfx.find(self.BRACKET_CLOSE)
                        bra_start = sfx.find(self.BRACKET_OPEN)
                        if bra_start == -1:
                            bra_start = bra_end + 1
                        if bra_end >= 0 and bra_end < bra_start:
                            msg = "illegal closing bracket"
                            raise NodeSetParseError(sfx, msg)

                    pfxlen, sfxlen = len(pfx), len(sfx)

                    if sfxlen > 0:
                        try:
                            # amending trailing digits generates /steps
                            sfx, rng = self._amend_trailing_digits(sfx, rng)
                        except RangeSetParseError as ex:
                            raise NodeSetParseRangeError(ex)

                    if pfxlen > 0:
                        try:
                            # this method supports /steps
                            pfx, rng = self._amend_leading_digits(pfx, rng)
                        except RangeSetParseError as ex:
                            raise NodeSetParseRangeError(ex)
                        if pfx:
                            # scan any nonempty pfx as a single node (no bracket)
                            pfx, pfxrvec = self._scan_string_single(pfx, autostep)
                            rsets += pfxrvec

                    # readahead for sanity check
                    bracket_idx = sfx.find(self.BRACKET_OPEN,
                                           bracket_idx - pfxlen)
                    op_idx, next_op_code = self._next_op(sfx)

                    if len(sfx) > 0 and sfx[0] == '[':
                        msg = "illegal reopening bracket"
                        raise NodeSetParseError(sfx, msg)

                    newpat += "%s%%s" % pfx
                    try:
                        rsets.append(RangeSet(rng, autostep))
                    except RangeSetParseError as ex:
                        raise NodeSetParseRangeError(ex)

                # Check if we have a next op-separated node or pattern
                op_idx, next_op_code = self._next_op(sfx)
                if op_idx < 0:
                    nsstr = None
                else:
                    sfx, nsstr = sfx.split(next_op_code, 1)
                    # Detected character operator so right operand is mandatory
                    if not nsstr:
                        msg = "missing nodeset operand with '%s' " \
                              "operator" % next_op_code
                        raise NodeSetParseError(None, msg)

                # Ignore whitespace(s)
                sfx = sfx.rstrip()
                if sfx:
                    sfx, sfxrvec = self._scan_string_single(sfx, autostep)
                    newpat += sfx
                    rsets += sfxrvec
            else:
                # In this case, either there is no comma and no bracket,
                # or the bracket is after the comma, then just return
                # the node.
                if op_idx < 0:
                    node = nsstr
                    nsstr = None # break next time
                else:
                    node, nsstr = nsstr.split(next_op_code, 1)
                    # Detected character operator so both operands are mandatory
                    if not node or not nsstr:
                        msg = "missing nodeset operand with '%s' " \
                              "operator" % next_op_code
                        raise NodeSetParseError(node or nsstr, msg)

                # Check for illegal closing bracket
                if node.find(self.BRACKET_CLOSE) > -1:
                    raise NodeSetParseError(node, "illegal closing bracket")

                # Ignore whitespace(s)
                node = node.rstrip()
                newpat, rsets = self._scan_string_single(node, autostep)

            op = ParsingEngine.OP_CODES[op_code]
            yield op, newpat, _rsets4nsb(rsets, autostep)

    def _amend_leading_digits(self, outer, inner):
        """Helper to get rid of leading bracket digits.

        Take a bracket outer prefix string and an inner range set string and
        return amended strings.
        """
        outerstrip = outer.rstrip(string.digits)
        outerlen, outerstriplen = len(outer), len(outerstrip)
        if outerstriplen < outerlen:
            # get outer bracket leading digits
            outerdigits = outer[outerstriplen:]
            inner = ','.join(
                '-'.join(outerdigits + bound for bound in elem.split('-'))
                for elem in (str(subrng)
                             for subrng in RangeSet(inner).contiguous()))
        return outerstrip, inner

    def _amend_trailing_digits(self, outer, inner):
        """Helper to get rid of trailing bracket digits.

        Take a bracket outer suffix string and an inner range set string and
        return amended strings.
        """
        outerstrip = outer.lstrip(string.digits)
        outerlen, outerstriplen = len(outer), len(outerstrip)
        if outerstriplen < outerlen:
            # step syntax is not compatible with trailing digits
            if '/' in inner:
                msg = "illegal trailing digits after range with steps"
                raise NodeSetParseError(outer, msg)
            # get outer bracket trailing digits
            outerdigits = outer[0:outerlen-outerstriplen]
            outlen = len(outerdigits)
            def shiftstep(orig, power):
                """Add needed step after shifting range indexes"""
                if '-' in orig:
                    return orig + '/1' + '0' * power
                return orig # do not use /step for single index
            inner = ','.join(shiftstep(s, outlen) for s in
                             ('-'.join(bound + outerdigits
                                       for bound in elem.split('-'))
                              for elem in inner.split(',')))
        return outerstrip, inner


class NodeSet(NodeSetBase):
    """
    Iterable class of nodes with node ranges support.

    NodeSet creation examples:

       >>> nodeset = NodeSet()               # empty NodeSet
       >>> nodeset = NodeSet("cluster3")     # contains only cluster3
       >>> nodeset = NodeSet("cluster[5,10-42]")
       >>> nodeset = NodeSet("cluster[0-10/2]")
       >>> nodeset = NodeSet("cluster[0-10/2],othername[7-9,120-300]")

    NodeSet provides methods like update(), intersection_update() or
    difference_update() methods, which conform to the Python Set API.
    However, unlike RangeSet or standard Set, NodeSet is somewhat not
    so strict for convenience, and understands NodeSet instance or
    NodeSet string as argument. Also, there is no strict definition of
    one element, for example, it IS allowed to do:

        >>> nodeset = NodeSet("blue[1-50]")
        >>> nodeset.remove("blue[36-40]")
        >>> print nodeset
        blue[1-35,41-50]

    Additionally, the NodeSet class recognizes the "extended string
    pattern" which adds support for union (special character ","),
    difference ("!"), intersection ("&") and symmetric difference ("^")
    operations. String patterns are read from left to right, by
    proceeding any character operators accordingly.

    Extended string pattern usage examples:

        >>> nodeset = NodeSet("node[0-10],node[14-16]") # union
        >>> nodeset = NodeSet("node[0-10]!node[8-10]")  # difference
        >>> nodeset = NodeSet("node[0-10]&node[5-13]")  # intersection
        >>> nodeset = NodeSet("node[0-10]^node[5-13]")  # xor
    """

    _VERSION = 2

    def __init__(self, nodes=None, autostep=None, resolver=None,
                 fold_axis=None):
        """Initialize a NodeSet object.

        The `nodes` argument may be a valid nodeset string or a NodeSet
        object. If no nodes are specified, an empty NodeSet is created.

        The optional `autostep` argument is passed to underlying
        :class:`.RangeSet.RangeSet` objects and aims to enable and make use of
        the range/step syntax (eg. ``node[1-9/2]``) when converting NodeSet to
        string (using folding). To enable this feature, autostep must be set
        there to the min number of indexes that are found at equal distance of
        each other inside a range before NodeSet starts to use this syntax. For
        example, `autostep=3` (or less) will pack ``n[2,4,6]`` into
        ``n[2-6/2]``. Default autostep value is None which means "inherit
        whenever possible", ie. do not enable it unless set in NodeSet objects
        passed as `nodes` here or during arithmetic operations.
        You may however use the special ``AUTOSTEP_DISABLED`` constant to force
        turning off autostep feature.

        The optional `resolver` argument may be used to override the group
        resolving behavior for this NodeSet object. It can either be set to a
        :class:`.NodeUtils.GroupResolver` object, to the ``RESOLVER_NOGROUP``
        constant to disable any group resolution, or to None (default) to use
        standard NodeSet group resolver (see :func:`.set_std_group_resolver()`
        at the module level to change it if needed).

        nD nodeset only: the optional `fold_axis` parameter, if specified, set
        the public instance member `fold_axis` to an iterable over nD 0-indexed
        axis integers. This parameter may be used to disengage some nD folding.
        That may be useful as all cluster tools don't support folded-nD nodeset
        syntax. Pass ``[0]``, for example, to only fold along first axis (that
        is, to fold first dimension using ``[a-b]`` rangeset syntax whenever
        possible). Using `fold_axis` ensures that rangeset won't be folded on
        unspecified axis, but please note however, that using `fold_axis` may
        lead to suboptimal folding, this is because NodeSet algorithms are
        optimized for folding along all axis (default behavior).
        """
        NodeSetBase.__init__(self, autostep=autostep, fold_axis=fold_axis)

        # Set group resolver.
        if resolver in (RESOLVER_NOGROUP, RESOLVER_NOINIT):
            self._resolver = None
        else:
            self._resolver = resolver or RESOLVER_STD_GROUP

        # Initialize default parser.
        if resolver == RESOLVER_NOINIT:
            self._parser = None
        else:
            self._parser = ParsingEngine(self._resolver)
            self.update(nodes)

    @classmethod
    def _fromlist1(cls, nodelist, autostep=None, resolver=None):
        """Class method that returns a new NodeSet with single nodes from
        provided list (optimized constructor)."""
        inst = NodeSet(autostep=autostep, resolver=resolver)
        for single in nodelist:
            inst.update(inst._parser.parse_string_single(single, autostep))
        return inst

    @classmethod
    def fromlist(cls, nodelist, autostep=None, resolver=None):
        """Class method that returns a new NodeSet with nodes from provided
        list."""
        inst = NodeSet(autostep=autostep, resolver=resolver)
        inst.updaten(nodelist)
        return inst

    @classmethod
    def fromall(cls, groupsource=None, autostep=None, resolver=None):
        """Class method that returns a new NodeSet with all nodes from optional
        groupsource."""
        inst = NodeSet(autostep=autostep, resolver=resolver)
        try:
            if not inst._resolver:
                raise NodeSetExternalError("Group resolver is not defined")
            else:
                # fill this nodeset with all nodes found by resolver
                inst.updaten(inst._parser.all_nodes(groupsource))
        except NodeUtils.GroupResolverError as exc:
            errmsg = "Group source error (%s: %s)" % (exc.__class__.__name__,
                                                      exc)
            raise NodeSetExternalError(errmsg)
        return inst

    def __getstate__(self):
        """Called when pickling: remove references to group resolver."""
        odict = self.__dict__.copy()
        odict['_version'] = NodeSet._VERSION
        del odict['_resolver']
        del odict['_parser']
        return odict

    def __setstate__(self, dic):
        """Called when unpickling: restore parser using non group
        resolver."""
        self.__dict__.update(dic)
        self._resolver = None
        self._parser = ParsingEngine(None)
        if getattr(self, '_version', 1) <= 1:
            self.fold_axis = None
            # if setting state from first version, a conversion is needed to
            # support native RangeSetND
            old_patterns = self._patterns
            self._patterns = {}
            for pat, rangeset in sorted(old_patterns.items()):
                if rangeset:
                    assert isinstance(rangeset, RangeSet)
                    rgs = str(rangeset)
                    if len(rangeset) > 1:
                        rgs = "[%s]" % rgs
                    self.update(pat % rgs)
                else:
                    self.update(pat)

    def copy(self):
        """Return a shallow copy of a NodeSet."""
        cpy = self.__class__(resolver=RESOLVER_NOINIT)
        dic = {}
        for pat, rangeset in self._patterns.items():
            if rangeset is None:
                dic[pat] = None
            else:
                dic[pat] = rangeset.copy()
        cpy._patterns = dic
        cpy.fold_axis = self.fold_axis
        cpy._autostep = self._autostep
        cpy._resolver = self._resolver
        cpy._parser = self._parser
        return cpy

    __copy__ = copy # For the copy module

    def _find_groups(self, node, namespace, allgroups):
        """Find groups of node by namespace."""
        if allgroups:
            # find node groups using in-memory allgroups
            for grp, nodeset in allgroups.items():
                if node in nodeset:
                    yield grp
        else:
            # find node groups using resolver
            try:
                for group in self._resolver.node_groups(node, namespace):
                    yield group
            except NodeUtils.GroupSourceQueryFailed as exc:
                msg = "Group source query failed: %s" % exc
                raise NodeSetExternalError(msg)

    def _groups2(self, groupsource=None, autostep=None):
        """Find node groups this nodeset belongs to. [private]"""
        if not self._resolver:
            raise NodeSetExternalError("No node group resolver")
        try:
            # Get all groups in specified group source.
            allgrplist = self._parser.grouplist(groupsource)
        except NodeUtils.GroupSourceError:
            # If list query failed, we still might be able to regroup
            # using reverse.
            allgrplist = None
        groups_info = {}
        allgroups = {}
        # Check for external reverse presence, and also use the
        # following heuristic: external reverse is used only when number
        # of groups is greater than the NodeSet size.
        if self._resolver.has_node_groups(groupsource) and \
            (not allgrplist or len(allgrplist) >= len(self)):
            # use external reverse
            pass
        else:
            if not allgrplist: # list query failed and no way to reverse!
                return groups_info # empty
            try:
                # use internal reverse: populate allgroups
                for grp in allgrplist:
                    nodelist = self._resolver.group_nodes(grp, groupsource)
                    allgroups[grp] = NodeSet(",".join(nodelist),
                                             resolver=self._resolver)
            except NodeUtils.GroupSourceQueryFailed as exc:
                # External result inconsistency
                raise NodeSetExternalError("Unable to map a group " \
                        "previously listed\n\tFailed command: %s" % exc)

        # For each NodeSetBase in self, find its groups.
        for node in self._iterbase():
            for grp in self._find_groups(node, groupsource, allgroups):
                if grp not in groups_info:
                    nodes = self._parser.parse_group(grp, groupsource, autostep)
                    groups_info[grp] = (1, nodes)
                else:
                    i, nodes = groups_info[grp]
                    groups_info[grp] = (i + 1, nodes)
        return groups_info

    def groups(self, groupsource=None, noprefix=False):
        """Find node groups this nodeset belongs to.

        Return a dictionary of the form:
            group_name => (group_nodeset, contained_nodeset)

        Group names are always prefixed with "@". If groupsource is provided,
        they are prefixed with "@groupsource:", unless noprefix is True.
        """
        groups = self._groups2(groupsource, self._autostep)
        result = {}
        for grp, (_, nsb) in groups.items():
            if groupsource and not noprefix:
                key = "@%s:%s" % (groupsource, grp)
            else:
                key = "@" + grp
            result[key] = (NodeSet(nsb, resolver=self._resolver),
                           self.intersection(nsb))
        return result

    def regroup(self, groupsource=None, autostep=None, overlap=False,
                noprefix=False):
        """Regroup nodeset using node groups.

        Try to find fully matching node groups (within specified groupsource)
        and return a string that represents this node set (containing these
        potential node groups). When no matching node groups are found, this
        method returns the same result as str()."""
        groups = self._groups2(groupsource, autostep)
        if not groups:
            return str(self)

        # Keep only groups that are full.
        fulls = []
        for k, (i, nodes) in groups.items():
            assert i <= len(nodes)
            if i == len(nodes):
                fulls.append((i, k))

        rest = NodeSet(self, resolver=RESOLVER_NOGROUP)
        regrouped = NodeSet(resolver=RESOLVER_NOGROUP)

        # Build regrouped NodeSet by selecting largest groups first.
        for _, grp in sorted(fulls, key=lambda x: (-x[0], x[1])):
            if not overlap and groups[grp][1] not in rest:
                continue
            if groupsource and not noprefix:
                regrouped.update("@%s:%s" % (groupsource, grp))
            else:
                regrouped.update("@" + grp)
            rest.difference_update(groups[grp][1])
            if not rest:
                return str(regrouped)

        if regrouped:
            return "%s,%s" % (regrouped, rest)

        return str(rest)

    def issubset(self, other):
        """
        Report whether another nodeset contains this nodeset.
        """
        nodeset = self._parser.parse(other, self._autostep)
        return NodeSetBase.issuperset(nodeset, self)

    def issuperset(self, other):
        """
        Report whether this nodeset contains another nodeset.
        """
        nodeset = self._parser.parse(other, self._autostep)
        return NodeSetBase.issuperset(self, nodeset)

    def __getitem__(self, index):
        """
        Return the node at specified index or a subnodeset when a slice
        is specified.
        """
        base = NodeSetBase.__getitem__(self, index)
        if not isinstance(base, NodeSetBase):
            return base
        # return a real NodeSet
        inst = NodeSet(autostep=self._autostep, resolver=self._resolver)
        inst._patterns = base._patterns
        return inst

    def split(self, nbr):
        """
        Split the nodeset into nbr sub-nodesets (at most). Each
        sub-nodeset will have the same number of elements more or
        less 1. Current nodeset remains unmodified.

        >>> for nodeset in NodeSet("foo[1-5]").split(3):
        ...     print nodeset
        foo[1-2]
        foo[3-4]
        foo5
        """
        assert(nbr > 0)

        # We put the same number of element in each sub-nodeset.
        slice_size = len(self) // nbr
        left = len(self) % nbr

        begin = 0
        for i in range(0, min(nbr, len(self))):
            length = slice_size + int(i < left)
            yield self[begin:begin + length]
            begin += length

    def update(self, other):
        """
        s.update(t) returns nodeset s with elements added from t.
        """
        nodeset = self._parser.parse(other, self._autostep)
        NodeSetBase.update(self, nodeset)

    def intersection_update(self, other):
        """
        s.intersection_update(t) returns nodeset s keeping only
        elements also found in t.
        """
        nodeset = self._parser.parse(other, self._autostep)
        NodeSetBase.intersection_update(self, nodeset)

    def difference_update(self, other, strict=False):
        """
        s.difference_update(t) removes from s all the elements
        found in t. If strict is True, raise KeyError if an
        element in t cannot be removed from s.
        """
        nodeset = self._parser.parse(other, self._autostep)
        NodeSetBase.difference_update(self, nodeset, strict)

    def symmetric_difference_update(self, other):
        """
        s.symmetric_difference_update(t) returns nodeset s keeping all
        nodes that are in exactly one of the nodesets.
        """
        nodeset = self._parser.parse(other, self._autostep)
        NodeSetBase.symmetric_difference_update(self, nodeset)


def expand(pat):
    """
    Commodity function that expands a nodeset pattern into a list of nodes.
    """
    return list(NodeSet(pat))

def fold(pat):
    """
    Commodity function that clean dups and fold provided pattern with ranges
    and "/step" support.
    """
    return str(NodeSet(pat))

def grouplist(namespace=None, resolver=None):
    """
    Commodity function that retrieves the list of raw groups for a specified
    group namespace (or use default namespace).
    Group names are not prefixed with "@".
    """
    return ParsingEngine(resolver or RESOLVER_STD_GROUP).grouplist(namespace)

def std_group_resolver():
    """
    Get the current resolver used for standard "@" group resolution.
    """
    return RESOLVER_STD_GROUP

def set_std_group_resolver(new_resolver):
    """
    Override the resolver used for standard "@" group resolution. The
    new resolver should be either an instance of
    NodeUtils.GroupResolver or None. In the latter case, the group
    resolver is restored to the default one.
    """
    global RESOLVER_STD_GROUP
    RESOLVER_STD_GROUP = new_resolver or _DEF_RESOLVER_STD_GROUP

def set_std_group_resolver_config(groupsconf, illegal_chars=None):
    """
    Helper to create and set std group resolver from a config file path.

    By default, the GroupResolverConfig object is created using
    illegal_chars=NodeSet.ILLEGAL_GROUP_CHARS.

    This method does nothing if groupsconf is not defined.
    """
    if groupsconf:
        if illegal_chars is None:
            illegal_chars = ILLEGAL_GROUP_CHARS
        group_resolver = NodeUtils.GroupResolverConfig(groupsconf,
                                                       illegal_chars)
        set_std_group_resolver(group_resolver)