scheme.py 22.9 KB
Newer Older
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
# coding: utf-8
# /*##########################################################################
#
# Copyright (c) 2017 European Synchrotron Radiation Facility
#
# Permission is hereby granted, free of charge, to any person obtaining a copy
# of this software and associated documentation files (the "Software"), to deal
# in the Software without restriction, including without limitation the rights
# to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
# copies of the Software, and to permit persons to whom the Software is
# furnished to do so, subject to the following conditions:
#
# The above copyright notice and this permission notice shall be included in
# all copies or substantial portions of the Software.
#
# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
# IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
# FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
# AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
# LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
# OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
# THE SOFTWARE.
#
# ###########################################################################*/

__authors__ = ["H.Payno"]
__license__ = "MIT"
__date__ = "17/12/2018"


from xml.etree.ElementTree import TreeBuilder, Element, ElementTree
from collections import defaultdict
from itertools import count, chain
import json
import pprint
import base64
import pickle
import logging
39
import typing
40
from .node import Node
41
from .link import Link
42
43
44
45
46
47
48
49
50
from ast import literal_eval

_logger = logging.getLogger(__name__)


class Scheme(object):
    """
    class to define a workflow scheme from nodes and links

51
52
53
54
55
56
57
    :param typing.Iterable nodes: set of Node contained in this scheme.
                                  note: subschemes are considered as Node.
                                  So if any defined it should be given as a
                                  Node
    :param typing.Iterable links: set of Link
    :type: typing.Iterable

58
    """
59
60
61
62
63
    _JSON_DESCRIPTION = 'description'
    _JSON_TITLE = 'title'
    _JSON_NODES = 'nodes'
    _JSON_LINKS = 'links'

64
    def __init__(self, nodes: typing.Iterable=None, links: typing.Iterable=None,
65
                 description: str=None, title: str=None):
66
67
68
69
        self.__rnodes = None
        self.__rlinks = None
        self.__rsub_schemes = None

70
71
72
        self._reset(nodes=nodes, links=links, description=description,
                    title=title)

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
    @property
    def rnodes(self):
        """All the nodes recursively. So this mean contained in this scheme or
        in a subscheme"""
        if self.__rnodes is None:
            self.__rnodes = list(self.nodes)
            for subscheme in self.sub_schemes:
                self.__rnodes.extend(subscheme.rnodes)
        return self.__rnodes

    @property
    def rlinks(self):
        """all the links recursively. So this mean contained in this scheme or
        in a subscheme"""
        if self.__rlinks is None:
            self.__rlinks = self.links
            for subscheme in self.sub_schemes:
                self.__rlinks.update(subscheme.rlinks)
        return self.__rlinks

    @property
    def sub_schemes(self):
        """list of sub schemes contained by this scheme"""
        return self.__sub_schemes

    @property
    def rsub_schemes(self):
        """list of all sub schemes contained in this scheme recursively
        """
        if self.__rsub_schemes is None:
            self.__rsub_schemes = self.sub_schemes
            for sub_scheme in self.sub_schemes:
                self.__rsub_schemes.extend(sub_scheme.rsub_schemes)
        return self.__rsub_schemes

108
109
110
    def _reset(self, nodes, links, description, title):
        self.title = title or ''
        self.description = description or ''
111
112
113
114
115
116
        # clear structure
        self.links = {}  # keys are link ID, values are Link
        self.__sub_schemes = []
        self.__rnodes = None
        self.__rlinks = None

117
118
119
120
121
122
123
        if links is not None:
            for link in links:
                self.links[link.id] = link
        self.nodes = nodes or []
        """list of nodes"""
        self.nodes_dict = {}
        """dict with node id as key and node as value"""
124
125

        # register all sub scheme
126
        for node in self.nodes:
127
128
129
130
131
            if isinstance(node, SubScheme):
                self.__sub_schemes.append(node)
                self.nodes_dict[node.id] = node
        # update nodes for sub_schemes
        for node in self.rnodes:
132
            self.nodes_dict[node.id] = node
133

134
135
136
        if links is not None:
            self._update_nodes_from_links()

137
    def final_nodes(self):
138
139
140
141
142
143
        """

        :return: list of final nodes (with no output) and which hasn't any
                 control node upstream
        """
        res = []
144
        for node in self.rnodes:
145
146
147
148
149
150
            assert isinstance(node, Node)
            if node.isfinal():
                res.append(node)
        return res

    def start_nodes(self):
151
152
153
154
155
156
        """

        :return: list of nodes starting the workflow. Those does not require
        any input_data
        :rtype: list
        """
157
        res = []
158
        for node in self.rnodes:
159
160
161
162
163
164
            assert isinstance(node, Node)
            if node.isstart():
                res.append(node)
        return res

    def endlessNodes(self):
165
166
167
168
169
        """

        :return: list of final nodes.
        :rtype: list
        """
170
        res = []
171
        for node in self.rnodes:
172
173
174
175
176
            assert isinstance(node, Node)
            if node.endless is True:
                res.append(node)
        return res

177
    def save_to(self, output_file: str):
178
179
        """
        Save the scheme as an xml formated file to `stream`
180

181
        :param output_file: name of the output file.
182
        :type: str
183
        """
Olof Svensson's avatar
Olof Svensson committed
184
        if output_file.lower().endswith('.json'):
185
186
187
188
            self.save_as_json(output_file)
        else:
            self.save_as_xml(output_file)

189
    def save_as_xml(self, output_file: str):
190
191
192
193
194
        """
        save current scheme to a default xml format

        :param str output_file: file path
        """
195
196
197
198
199
        tree = self.scheme_to_etree(data_format="literal")
        indent(tree.getroot(), 0)

        tree.write(output_file)

200
    def save_as_json(self, output_file: str):
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
        """
        save current scheme to a default json format

        :param str output_file: file path
        """
        with open(output_file, 'w') as json_file:
            json.dump(self.to_json(), json_file)

    def nodes_to_json(self):
        """

        :return: nodes to json compatible format
        :rtype: list
        """
        res = []
        for node in self.nodes:
            res.append(node.to_json())
        return res

220
221
    @staticmethod
    def nodes_from_json(json_data):
222
223
224
        """

        :param json_data: data containing the json definition
225
226
227
        :return: nodes, sub_schemes
        :rtype: tuple
        :raises: ValueError if the json is not a valid Node description
228
        """
229
230
231
232
233
234
235
236
        nodes = []
        sub_schemes = []
        for node_json_data in json_data:
            # if contains a node description then this is a subscheme
            if Scheme._JSON_NODES in node_json_data:
                sub_scheme = SubScheme.load_from_json(node_json_data)
                sub_schemes.append(sub_scheme)
            else:
237
                nodes.append(Node.from_json(node_json_data))
238
        return nodes, sub_schemes
239
240
241
242
243
244
245
246
247
248
249
250

    def links_to_json(self):
        """

        :return: links to json compatible format
        :rtype: list
        """
        res = []
        for link in self.links.values():
            res.append(link.to_json())
        return res

251
252
    @staticmethod
    def links_from_json(json_data):
253
254
255
256
257
258
        """

        :param json_data: data containing the json definition
        :return: list of Link defined by the json data
        :rtype: list
        """
259
260
261
262
        links = []
        for link_json_data in json_data:
            links.append(Link.from_json(link_json_data))
        return links
263
264

    def to_json(self):
265
266
267
268
269
270
        """
        Convert scheme to json

        :return: json dict
        :rtype: dict
        """
271
272
273
274
275
276
277
278
        return {
            self._JSON_DESCRIPTION: self.description,
            self._JSON_TITLE: self.title,
            self._JSON_NODES: self.nodes_to_json(),
            self._JSON_LINKS: self.links_to_json(),
        }

    @staticmethod
279
280
281
282
283
284
285
286
287
    def from_json_file(json_file_path: str):
        """
        Create and load Scheme from a json file

        :param str json_file_path: json file
        :return: Scheme fitting the json description
        :rtype: Scheme
        :raises: ValueError if file not found or invalid
        """
288
289
290
291
292
293
294
295
296
        scheme = Scheme()
        try:
            scheme.load_from_json_file(json_file_path)
        except ValueError as e:
            _logger.error(e)
            return None
        else:
            return scheme

297
    def load_from_json_file(self, json_file_path: str):
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
        """

        :param str json_file_path: path to the json file containing the scheme
                                   description
        :return: Scheme fitting the json description contains if the file.
                 If description is incomplete, return None
        :rtype:Union[Scheme,None]
        """
        try:
            with open(json_file_path, 'r') as json_file:
                json_data = json.load(json_file)
        except IOError as e:
            _logger.error('fail to read json file', str(e))
        else:
            self.load_from_json(json_data=json_data)

314
315
    @staticmethod
    def load_scheme_info_from_json(json_data):
316
        """
317
        load fom json stream the Scheme Information
318
319

        :param json_data: scheme description
320
321
        :return: nodes, links, sub-schemes, title, description
        :rtype: tuple
322
323
        """
        # load title
324
        if Scheme._JSON_TITLE not in json_data:
325
326
327
            _logger.warning('no title found in the json')
            title = None
        else:
328
            title = json_data[Scheme._JSON_TITLE]
329
        # load description
330
        if Scheme._JSON_DESCRIPTION not in json_data:
331
332
333
            _logger.warning('no description found in the json')
            description = None
        else:
334
            description = json_data[Scheme._JSON_DESCRIPTION]
335
        # load links
336
        if Scheme._JSON_LINKS not in json_data:
337
338
339
340
            _logger.error('no link found in the json')
            links = None
        else:
            try:
341
                links = Scheme.links_from_json(json_data=json_data[Scheme._JSON_LINKS])
342
343
344
345
            except ValueError as e:
                _logger.error(e)
                links = None
        # load nodes
346
347
        nodes, sub_schemes = None, None
        if Scheme._JSON_NODES not in json_data:
348
349
350
            _logger.error('no nodes found in the json')
        else:
            try:
351
                nodes, sub_schemes = Scheme.nodes_from_json(json_data=json_data[Scheme._JSON_NODES])
352
353
354
            except ValueError as e:
                _logger.error(e)
                nodes = None
355
356
357
358
        return nodes, links, sub_schemes, title, description

    def load_from_json(self, json_data):
        """
359

360
361
362
363
364
        :param json_data: scheme description
        :raise ValueError: if sink or source channel missing or if link id
                           missing or if sink or source node missing
        """
        nodes, links, sub_schemes, title, description = self.load_scheme_info_from_json(json_data)
365
366
367
368
369
        # create scheme if possible
        if nodes is None or links is None:
            raise ValueError('unable to load scheme from json description.'
                             'Information missing')
        else:
370
371
            if sub_schemes is not None:
                nodes.extend(sub_schemes)
372
373
374
            self._reset(nodes=nodes, links=links, description=description,
                        title=title)

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
    def scheme_to_etree(self, data_format="literal", pickle_fallback=False):
        """
        Return an `xml.etree.ElementTree` representation of the `scheme.
        """
        builder = TreeBuilder(element_factory=Element)
        builder.start("scheme", {"version": "2.0",
                                 "title": self.title or "",
                                 "description": self.description or ""})

        # Nodes
        node_ids = defaultdict(count().__next__)
        builder.start("nodes", {})
        for node in self.nodes:  # type: SchemeNode
            attrs = {"id": node.id,
                     "qualified_name": node._qualified_name,
                     }

            if type(node) is not Node:
                attrs["scheme_node_type"] = "%s.%s" % (type(node).__name__,
                                                       type(node).__module__)
            builder.start("node", attrs)
            builder.end("node")

        builder.end("nodes")

        # Links
        link_ids = defaultdict(count().__next__)
        builder.start("links", {})
        for link in self.links:  # type: SchemeLink
            source = link.source_node_id
            sink = link.sink_node_id
            source_id = node_ids[source]
            sink_id = node_ids[sink]
            attrs = {"id": str(link_ids[link]),
                     "source_node_id": str(source_id),
                     "sink_node_id": str(sink_id),
                     "source_channel": link.source_channel,
                     "sink_channel": link.sink_channel,
                     "enabled": "true" if link.enabled else "false",
                     }
            builder.start("link", attrs)
            builder.end("link")

        builder.end("links")

        # Annotations
        annotation_ids = defaultdict(count().__next__)
        builder.start("thumbnail", {})
        builder.end("thumbnail")

        # Node properties/settings
        builder.start("node_properties", {})
        for node in self.nodes:
            data = None
            if node.properties:
                try:
                    data, format = dumps(node.properties, format=data_format,
                                         pickle_fallback=pickle_fallback)
                except Exception:
                    _logger.error("Error serializing properties for node %r",
                              node.title, exc_info=True)
                if data is not None:
                    builder.start("properties",
                                  {"node_id": str(node_ids[node]),
                                   "format": format})
                    builder.data(data)
                    builder.end("properties")

        builder.end("node_properties")

        builder.end("scheme")
        root = builder.close()
        tree = ElementTree(root)
        return tree

    def _update_nodes_from_links(self):
        """
        Update upstream and downstream nodes from links definition
        """
        self._clear_nodes_connections()
        for link_id, link in self.links.items():
            source_node = self.nodes_dict[link.source_node_id]
            sink_node = self.nodes_dict[link.sink_node_id]
            source_node.downstream_nodes.add(self.nodes_dict[link.sink_node_id])
            sink_node.upstream_nodes.add(self.nodes_dict[link.source_node_id])

    def _clear_nodes_connections(self):
        """
        clear for all nodes downstream and upstream nodes
        """
        for node in self.nodes:
            assert isinstance(node, Node)
            node.downstream_nodes = set()
            node.upstream_nodes = set()

    def has_final_join(self):
471
472
473
474
475
        """
        :return: True if we need to send a 'end' signal before closing the
                 workflow. This is needed in the 'acquisition workflow' like
                 tomwer and the DataWatcher process for example.
        :rtype: bool
476
477
478
479
480
481
482
483
        """
        for node in self.nodes:
            if node.need_stop_join:
                return True
        return False

    @staticmethod
    def from_desc(desc):
484
485
486
487
488
489
        """

        :param desc:
        :return: instance of Scheme from it description.
        :rtype: :class:`Scheme`
        """
490
491
492
493
        nodes = []
        nodes_dict = {}

        for node_d in desc.nodes:
494
            node = Node(id=node_d.id, processing_pt=node_d.qualified_name)
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
            nodes.append(node)
            nodes_dict[node.id] = node
            data = node_d.data
            if data:
                properties = loads(data.data, data.format)
                node.properties = properties
            else:
                node.properties = {}
            node.qualified_name = node_d.qualified_name

        for link_d in desc.links:
            upstream_node = nodes_dict[link_d.source_node_id]
            assert isinstance(upstream_node, Node)
            downstream_node = nodes_dict[link_d.sink_node_id]
            upstream_node.downstream_nodes.add(downstream_node)
            downstream_node.upstream_nodes.add(upstream_node)

        scheme = Scheme(nodes=nodes, links=desc.links)
        scheme.title = desc.title
        scheme.description = desc.description

        return scheme

518
    def load_handlers(self):
519
520
521
        """
        load all nodes handlers.
        """
522
523
524
        for node in self.nodes:
            node.load_handlers()

525

526
class SubScheme(Scheme, Node):
527
528
529
530
531
532
533
534
535
536
537
538
539
    """
    Define a sub-scheme of the workflow (or subworkflow). SubScheme are
    as Scheme expect that they are not 'root'.

    :param nodes: set of Node
    :type: typing.Iterable
    :param links: set of Node
    :type: typing.Iterable
    :param description: description of the subscheme
    :type: str
    :param error_handler: ErrorHandler
    """
    def __init__(self, nodes: typing.Iterable, links: typing.Iterable,
540
541
542
543
                 description: str=None, error_handler=None, title='', id=None):
        Node.__init__(self, processing_pt=None, error_handler=error_handler,
                      id=id)
        Scheme.__init__(self, nodes=nodes, links=links, title=title,
544
545
                        description=description)

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
    def to_json(self):
        """
        Convert sub scheme to json

        :return: json dict
        :rtype: dict
        """
        desc = Scheme.to_json(self)
        desc.update(Node.to_json(self))
        return desc

    @staticmethod
    def load_from_json(json_data):
        """

        :param json_data: scheme description
        :raise ValueError: if sink or source channel missing or if link id
                           missing or if sink or source node missing
        """
        nodes, links, sub_schemes, title, description = Scheme.load_scheme_info_from_json(json_data)
        _id, _properties, _process_pt = Node.load_node_info_from_json(json_data)
        if _properties is not None:
            _logger.warning('SubScheme properties are not managed')
        if _process_pt is not None:
            _logger.warning('SubScheme pointer to ptocess is not managed')
        # create scheme if possible
        if nodes is None or links is None:
            raise ValueError('unable to load scheme from json description.'
                             'Information missing')
        sub_scheme = SubScheme(nodes=nodes, links=links,
                               description=description, title=title, id=_id)
        return sub_scheme

579

580
581
582
583
584
585
586
def contains_control_nodes(nodes_list: typing.Iterable):
    """
    Return the list of the 'control' nodes.

    :param typing.Iterable nodes_list:
    :return:
    """
587
588
589
590
591
592
    for _node in nodes_list:
        if _node.endless or contains_control_nodes(_node.upstream_nodes):
            return True
    return False


593
594
595
596
597
598
599
600
def loads(string: str, format: str) -> object:
    """load stream from the given format.

    :param str string: stream to load
    :param str format: format
    :return: object contained in the stream
    :rtype: object
    """
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
    if format == "literal":
        return literal_eval(string)
    elif format == "json":
        return json.loads(string)
    elif format == "pickle":
        return pickle.loads(base64.decodebytes(string.encode('ascii')))
    else:
        raise ValueError("Unknown format")


def indent(element, level=0, indent="\t"):
    """
    Indent an instance of a :class:`Element`. Based on
    (http://effbot.org/zone/element-lib.htm#prettyprint).

    """
    def empty(text):
        return not text or not text.strip()

    def indent_(element, level, last):
        child_count = len(element)

        if child_count:
            if empty(element.text):
                element.text = "\n" + indent * (level + 1)

            if empty(element.tail):
                element.tail = "\n" + indent * (level + (-1 if last else 0))

            for i, child in enumerate(element):
                indent_(child, level + 1, i == child_count - 1)

        else:
            if empty(element.tail):
                element.tail = "\n" + indent * (level + (-1 if last else 0))

    return indent_(element, level, True)


def dumps(obj, format="literal", prettyprint=False, pickle_fallback=False):
    """
    Serialize `obj` using `format` ('json' or 'literal') and return its
    string representation and the used serialization format ('literal',
    'json' or 'pickle').

    If `pickle_fallback` is True and the serialization with `format`
    fails object's pickle representation will be returned

    """
    if format == "literal":
        try:
            return (literal_dumps(obj, prettyprint=prettyprint, indent=1),
                    "literal")
        except (ValueError, TypeError) as ex:
            if not pickle_fallback:
                raise

            _logger.debug("Could not serialize to a literal string")

    elif format == "json":
        try:
            return (json.dumps(obj, indent=1 if prettyprint else None),
                    "json")
        except (ValueError, TypeError):
            if not pickle_fallback:
                raise

            _logger.debug("Could not serialize to a json string")

    elif format == "pickle":
        return base64.encodebytes(pickle.dumps(obj)).decode('ascii'), "pickle"

    else:
        raise ValueError("Unsupported format %r" % format)

    if pickle_fallback:
        _logger.warning("Using pickle fallback")
        return base64.encodebytes(pickle.dumps(obj)).decode('ascii'), "pickle"
    else:
        raise Exception("Something strange happened.")


# This is a subset of PyON serialization.
def literal_dumps(obj, prettyprint=False, indent=4):
    """
    Write obj into a string as a python literal.
    """
    memo = {}
    NoneType = type(None)

    def check(obj):
        if type(obj) in [int, float, bool, NoneType, str, bytes]:
            return True

        if id(obj) in memo:
            raise ValueError("{0} is a recursive structure".format(obj))

        memo[id(obj)] = obj

        if type(obj) in [list, tuple]:
            return all(map(check, obj))
        elif type(obj) is dict:
            return all(map(check, chain(iter(obj.keys()), iter(obj.values()))))
        else:
            raise TypeError("{0} can not be serialized as a python "
                             "literal".format(type(obj)))

    check(obj)

    if prettyprint:
        return pprint.pformat(obj, indent=indent)
    else:
        return repr(obj)