workflow.py 15 KB
Newer Older
Jerome Mariette's avatar
Jerome Mariette committed
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
#
# Copyright (C) 2012 INRA
# 
# This program is free software: you can redistribute it and/or modify
# it under the terms of the GNU General Public License as published by
# the Free Software Foundation, either version 3 of the License, or
# (at your option) any later version.
# 
# This program 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 General Public License for more details.
# 
# You should have received a copy of the GNU General Public License
# along with this program.  If not, see <http://www.gnu.org/licenses/>.
#

import inspect
import os
import sys
import uuid
import pkgutil
import tempfile
import pickle
import time
import threading

import jflow.components
import jflow.utils as utils
from jflow.config_reader import WorkflowConfigReader
from jflow.parameter import Parameter
from jflow.workflows_manager import WorkflowsManager, JFlowConfigReader

from weaver.script import ABSTRACTIONS
from weaver.script import DATASETS
from weaver.script import FUNCTIONS
from weaver.script import NESTS
from weaver.script import OPTIONS
from weaver.script import STACKS
from weaver.nest import Nest
from weaver.options import Options
from cctools.makeflow import MakeflowLog
from cctools.makeflow.log import Node

class Workflow(threading.Thread):
    """
    """

    PROPERTIES_FILE_NAME = "workflow.properties"
    MAKEFLOW_LOG_FILE_NAME = "Makeflow.makeflowlog"
    DUMP_FILE_NAME = "workflow.dump"
    WORKING = ".working"
    STATUS_STARTED = "started"
    STATUS_COMPLETED = "completed"
    STATUS_FAILED = "failed"
    STATUS_ABORTED = "aborted"
    
    def __init__(self, args={}, id=None):
        """ 
        """
        # define as a thread
        threading.Thread.__init__(self)
        self.config_reader = WorkflowConfigReader(self._get_property_path())
        self.jflow_config_reader = JFlowConfigReader()
        self.manager = WorkflowsManager()
        self.components = []
        self.makes = {}
        self.globals = {}
        self.options = Options()
        self.status = self.STATUS_STARTED
        self.start_time = None
        self.end_time = None
        self.comp_pckg = self._import_components()
        self.engine_arguments = ''
        self.component_nameids_is_init = False
        self.component_nameids = {}
        self.args = args
        # try to parse engine arguments
        try:
            type, options = self.jflow_config_reader.get_batch()
            if type: self.engine_arguments += ' -T ' + type
            if options : self.engine_arguments += ' -B "' + options + '"'
        except: self.engine_arguments = None
        try:
            # Add name and description values
            self.name = self.config_reader.get_name()
            self.description = self.config_reader.get_description()
            # Add parameter values
            self.parameters = self.config_reader.get_parameters()
        except:
            raise IOError(self.__class__.__name__ + " workflow property file not found or invalid.")
        self.id = id
        if self.id is not None:
            self.directory = self.manager.get_workflow_directory(self.name, self.id)
            if not os.path.isdir(self.directory):
                os.makedirs(self.directory, 0751)
            self._serialize()

    def __setstate__(self, state):
        self.__dict__ = state.copy()
        threading.Thread.__init__(self, name=self.name)
        
    def __getstate__(self):
        """
        Threading uses Lock Object, do not consider these objects when serializing a workflow
        """
        odict = self.__dict__.copy()
        del odict['_Thread__started']
        del odict['_Thread__block']
        del odict['_Thread__stderr']
        return odict
    
    def add_component(self, component_name, args=[], kwargs={}, component_prefix="default"):
        # first build and check if this component is OK
        my_pckge = __import__(self.comp_pckg[component_name], globals(), locals(), [component_name], -1)
        # build the object and define required field
        cmpt_object = getattr(my_pckge, component_name)()
        cmpt_object.output_directory = self.get_component_output_directory(component_name, component_prefix)
        cmpt_object.prefix = component_prefix
        if kwargs: cmpt_object.define_parameters(**kwargs)
        else: cmpt_object.define_parameters(*args)
        
        # if already init, add the component to the list and check if weaver should be executed
        if self.component_nameids_is_init:
            # add the component
            self.components.append(cmpt_object)
            # if this one require a dynamic pipeline, execute the first part
            if cmpt_object.is_dynamic:
                self._execute_weaver()
        else: 
            if self._component_is_duplicated(cmpt_object):
                raise ValueError("Component " + cmpt_object.__class__.__name__ + " with prefix " + 
                                 cmpt_object.prefix + " already exist in this pipeline!")
            self.component_nameids[cmpt_object.get_nameid()] = None
        
        return cmpt_object
    
    def pre_process(self):
        pass
    
    def process(self):
        """ 
        Run the workflow, has to be implemented by subclasses
        """
        raise NotImplementedError
    
    def post_process(self):
        pass
    
    def get_temporary_file(self, suffix=".txt"):
        # first check if tmp directory exists
        if not os.path.isdir(self.jflow_config_reader.get_tmp_directory()):
            os.makedirs(self.jflow_config_reader.get_tmp_directory(), 0751)
        tempfile_name = os.path.basename(tempfile.NamedTemporaryFile(suffix=suffix).name)
        return os.path.join(self.jflow_config_reader.get_tmp_directory(), tempfile_name)

    def get_component_output_directory(self, component_name, component_prefix):
        return os.path.join(self.directory, component_name + "_" + component_prefix)
    
    def get_components_nameid(self):
        return self.component_nameids.keys()
    
    def run(self):
        """
        Only require for Threading
        """
        self.start_time = time.time()
        self.status = self.STATUS_STARTED
        self.end_time = None  
        # serialize the workflow before pre processing
        self._serialize()
        self.pre_process()
        # serialize the workflow before running
        self._serialize()
        # if not done yet, init the component nameids list before to run
        if not self.component_nameids_is_init:
            self.process()
            self.component_nameids_is_init = True
            self._serialize()
        self.process()
        if len(self.components) > 0:
            self._execute_weaver()
        self._serialize()
        self.post_process()
        if self.status == self.STATUS_STARTED: self.status = self.STATUS_COMPLETED
        self.end_time = time.time()
        self._serialize()
    
    def _update_status_from_log(self):
        # first update the status from weaver folders
        # TODO update self.end_time
        try:
            working_directory = os.path.join(self.directory, self.WORKING)
            make_states = []
            for wdir in os.listdir(working_directory):
                log_path = os.path.join(working_directory, wdir, self.MAKEFLOW_LOG_FILE_NAME)
                log = MakeflowLog(log_path)
                log.parse()
                make_states.append(log.state)
            if self.STATUS_ABORTED in make_states:
                self.status = self.STATUS_ABORTED
            elif self.STATUS_FAILED in make_states:
                self.status = self.STATUS_FAILED
            elif self.status != self.STATUS_COMPLETED:
                self.status = self.STATUS_STARTED
        except: pass
        self._serialize()

    def get_status(self):
        self._update_status_from_log()
        return self.status
    
    def get_component_status(self, component_nameid):
        status = {"time": 0.0,
                  "tasks": 0,
                  "waiting": 0,
                  "running": 0,
                  "failed":0,
                  "aborted": 0,
                  "completed": 0}
        try:
            log = MakeflowLog(self.component_nameids[component_nameid])
            log.parse()
            symbols = set(n.symbol for n in log.nodes if n.symbol)
            if not symbols: return None
            for n in log.nodes:
                if not n.symbol: continue
                if n.symbol == component_nameid:
                    status["tasks"] += 1
                    status["time"]  += n.elapsed_time
                    if n.state == Node.WAITING:
                        status["waiting"] += 1
                    elif n.state == Node.RUNNING:
                        status["running"] += 1
                    elif n.state == Node.FAILED:
                        status["failed"] += 1
                    elif n.state == Node.ABORTED:
                        status["aborted"] += 1
                    elif n.state == Node.COMPLETED:
                        status["completed"] += 1
        except: pass
        return status        
    
    def makeflow_pretty_print_node(self, dag, node):
        sys.stdout.write('{0:>10} {1} {2}\n'.format('NODE', node.id, node.symbol))
        
        for output_file in sorted(node.output_files):
            sys.stdout.write('{0:>10} {1:>10} {2}\n'.format('', 'OUTPUT', output_file))
    
        for input_file in sorted(node.input_files):
            sys.stdout.write('{0:>10} {1:>10} {2}\n'.format('', 'INPUT', input_file))
            
        sys.stdout.write('{0:>10} {1:>10} {2}\n'.format('', 'COMMAND', node.command))

    def _execute_weaver(self, engine_wrapper=None):
        # Add nest path and path to script to Python module path to allow
        # for importing modules outside of $PYTHONPATH
        make_directory, new_make = self._get_current_make()
        current_working_directory = os.path.join(os.path.join(self.directory, self.WORKING), make_directory)
        sys.path.insert(0, os.path.abspath(os.path.dirname(current_working_directory)))

        # Load built-ins
        self._import('abstraction', ABSTRACTIONS)
        self._import('dataset', DATASETS)
        self._import('function', FUNCTIONS)
        self._import('nest', NESTS)
        self._import('options', OPTIONS)
        self._import('stack', STACKS)
                
        # Execute nest
        with Nest(current_working_directory, wrapper=engine_wrapper, path=self.jflow_config_reader.get_makeflow_path()) as nest:
            with self.options:
                if new_make:
                    try:
                        for component in self.components:
                            nest.symbol = component.get_nameid()
                            self.component_nameids[component.get_nameid()] = os.path.join(current_working_directory, self.MAKEFLOW_LOG_FILE_NAME)
                            component.execute()
                        # create the DAG
                        nest.compile()
                    except Exception as e:
                        sys.exit('Error compiling script: {0}'.format(e))
                        self.status = self.STATUS_FAILED
                        self._serialize()
                self.components = []
                # Once a weaver script is compiled, serialize the workflow
                self._serialize()
                try:
                    nest.execute(self.engine_arguments, exit_on_failure=True)
                except:
                    self.status = self.STATUS_FAILED
                    self.end_time = time.time()
                    self._serialize()
                    sys.exit()

    def _get_current_make(self):
        current_component, make_directory, new_make = [], None, False
        for component in self.components:
            current_component.append(component.get_nameid())
        for make in self.makes:
            if set(current_component) == set(self.makes[make]):
                make_directory = make
        # If the components in the queue have already been compiled
        if make_directory is None:
            make_directory = uuid.uuid4().hex[:10]
            self.makes[make_directory] = current_component
            new_make = True
        return [make_directory, new_make]
    
    def _serialize(self):
        self.dump_path = os.path.join(self.directory, self.DUMP_FILE_NAME)
        workflow_dump = open(self.dump_path, "wb")
        pickle.dump(self, workflow_dump)
        workflow_dump.close()

    def _component_is_duplicated(self, component):
        if component.get_nameid() in self.component_nameids.keys():
            return True
        return False

    def _get_property_path(self):
        """ 
        Return(type:string):  the path to the workflow properties file, None if does not exist
        """
        property_file = os.path.join(os.path.dirname(inspect.getfile(self.__class__)), self.PROPERTIES_FILE_NAME)
        return property_file if os.path.isfile(property_file) else None

    def _import_components(self):
        pckge = {}
        # first import global packages
        for importer, modname, ispkg in pkgutil.iter_modules(jflow.components.__path__, jflow.components.__name__ + "."):
            m = __import__(modname)
            for class_name, obj in inspect.getmembers(sys.modules[modname], inspect.isclass):
                if issubclass(obj, jflow.component.Component) and obj.__name__ != jflow.component.Component.__name__:
                    pckge[class_name] = modname
        # then import pipeline packages
        pipeline_dir = os.path.dirname(inspect.getfile(self.__class__))
        for importer, modname, ispkg in pkgutil.iter_modules([os.path.join(pipeline_dir, "components")], "workflows." + 
                                                             os.path.basename(pipeline_dir) + ".components."):
            m = __import__(modname)
            for class_name, obj in inspect.getmembers(sys.modules[modname], inspect.isclass):
                if issubclass(obj, jflow.component.Component) and obj.__name__ != jflow.component.Component.__name__:
                    pckge[class_name] = modname
        # finally import workflows shared packages
        workflows_dir = os.path.dirname(os.path.dirname(inspect.getfile(self.__class__)))
        for importer, modname, ispkg in pkgutil.iter_modules([os.path.join(workflows_dir, "components")], "workflows.components."):
            m = __import__(modname)
            for class_name, obj in inspect.getmembers(sys.modules[modname], inspect.isclass):
                if issubclass(obj, jflow.component.Component) and obj.__name__ != jflow.component.Component.__name__:
                    pckge[class_name] = modname
        return pckge

    def _import(self, module, symbols):
        """ Import ``symbols`` from ``module`` into global namespace. """
        # Import module
        m = 'weaver.{0}'.format(module)
        m = __import__(m, self.globals, self.globals, symbols, -1)

        # Import symbols from module into global namespace, which we store as
        # an attribute for later use (i.e. during compile)
        for symbol in symbols:
            self.globals[symbol] = getattr(m, symbol)