Spaces:
Running
Running
File size: 11,806 Bytes
e60e568 |
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 |
'''
This file is part of PM4Py (More Info: https://pm4py.fit.fraunhofer.de).
PM4Py 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.
PM4Py 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 PM4Py. If not, see <https://www.gnu.org/licenses/>.
'''
import itertools
import sys
from abc import ABC
from collections import Counter
from typing import Collection, Any, List, Optional, Generic, Dict
from typing import Tuple
from pm4py.algo.discovery.inductive.cuts.abc import Cut
from pm4py.algo.discovery.inductive.cuts.abc import T
from pm4py.algo.discovery.inductive.dtypes.im_dfg import InductiveDFG
from pm4py.algo.discovery.inductive.dtypes.im_ds import IMDataStructureUVCL, IMDataStructureDFG
from pm4py.objects.dfg import util as dfu
from pm4py.objects.dfg.obj import DFG
from pm4py.objects.process_tree.obj import Operator, ProcessTree
class SequenceCut(Cut[T], ABC, Generic[T]):
@classmethod
def operator(cls, parameters: Optional[Dict[str, Any]] = None) -> ProcessTree:
return ProcessTree(operator=Operator.SEQUENCE)
@staticmethod
def check_merge_condition(g1, g2, trans_succ):
for a1 in g1:
for a2 in g2:
if (a2 in trans_succ[a1] and a1 in trans_succ[a2]) or (a2 not in trans_succ[a1] and a1 not in trans_succ[a2]):
return True
return False
@staticmethod
def merge_groups(groups, trans_succ):
i = 0
while i < len(groups):
j = i + 1
while j < len(groups):
if SequenceCut.check_merge_condition(groups[i], groups[j], trans_succ):
groups[i] = groups[i].union(groups[j])
del groups[j]
continue
j = j + 1
i = i + 1
return groups
@classmethod
def holds(cls, obj: T, parameters: Optional[Dict[str, Any]] = None) -> Optional[List[Collection[Any]]]:
'''
This method finds a sequence cut in the dfg.
Implementation follows function sequence on page 188 of
"Robust Process Mining with Guarantees" by Sander J.J. Leemans (ISBN: 978-90-386-4257-4)
Basic Steps:
1. create a group per activity
2. merge pairwise reachable nodes (based on transitive relations)
3. merge pairwise unreachable nodes (based on transitive relations)
4. sort the groups based on their reachability
'''
dfg = obj.dfg
alphabet = dfu.get_vertices(dfg)
transitive_predecessors, transitive_successors = dfu.get_transitive_relations(dfg)
groups = [{a} for a in alphabet]
if len(groups) == 0:
return None
old_size = None
while old_size != len(groups):
old_size = len(groups)
groups = SequenceCut.merge_groups(groups, transitive_successors)
groups = list(sorted(groups, key=lambda g: len(
transitive_predecessors[next(iter(g))]) + (len(alphabet) - len(transitive_successors[next(iter(g))]))))
return groups if len(groups) > 1 else None
class StrictSequenceCut(SequenceCut[T], ABC, Generic[T]):
@classmethod
def _skippable(cls, p: int, dfg: DFG, start: Collection[Any], end: Collection[Any],
groups: List[Collection[Any]], parameters: Optional[Dict[str, Any]] = None) -> bool:
"""
This method implements the function SKIPPABLE as defined on page 233 of
"Robust Process Mining with Guarantees" by Sander J.J. Leemans (ISBN: 978-90-386-4257-4)
The function is used as a helper function for the strict sequence cut detection mechanism, which detects
larger groups of skippable activities.
"""
for i, j in itertools.product(range(0, p), range(p + 1, len(groups))):
for a, b in itertools.product(groups[i], groups[j]):
if (a, b) in dfg.graph:
return True
for i in range(p + 1, len(groups)):
for a in groups[i]:
if a in start:
return True
for i in range(0, p):
for a in groups[i]:
if a in end:
return True
return False
@classmethod
def holds(cls, obj: T, parameters: Optional[Dict[str, Any]] = None) -> Optional[List[Collection[Any]]]:
"""
This method implements the strict sequence cut as defined on page 233 of
"Robust Process Mining with Guarantees" by Sander J.J. Leemans (ISBN: 978-90-386-4257-4)
The function merges groups that together can be skipped.
"""
dfg = obj.dfg
c = SequenceCut.holds(obj)
start = set(dfg.start_activities.keys())
end = set(dfg.end_activities.keys())
if c is not None:
mf = [-1 * sys.maxsize if len(set(G).intersection(start)) > 0 else sys.maxsize for G in c]
mt = [sys.maxsize if len(set(G).intersection(end)) > 0 else -1 * sys.maxsize for G in c]
cmap = cls._construct_alphabet_cluster_map(c)
for (a, b) in dfg.graph:
mf[cmap[b]] = min(mf[cmap[b]], cmap[a])
mt[cmap[a]] = max(mt[cmap[a]], cmap[b])
for p in range(0, len(c)):
if cls._skippable(p, dfg, start, end, c):
q = p - 1
while q >= 0 and mt[q] <= p:
c[p] = c[p].union(c[q])
c[q] = set()
q -= 1
q = p + 1
while q < len(mf) and mf[q] >= p:
c[p] = c[p].union(c[q])
c[q] = set()
q += 1
return list(filter(lambda g: len(g) > 0, c))
return None
@classmethod
def _construct_alphabet_cluster_map(cls, c: List[Collection[Any]], parameters: Optional[Dict[str, Any]] = None):
map = dict()
for i in range(0, len(c)):
for a in c[i]:
map[a] = i
return map
class SequenceCutUVCL(SequenceCut[IMDataStructureUVCL]):
@classmethod
def project(cls, obj: IMDataStructureUVCL, groups: List[Collection[Any]], parameters: Optional[Dict[str, Any]] = None) -> List[IMDataStructureUVCL]:
logs = [Counter() for g in groups]
for t in obj.data_structure:
i = 0
split_point = 0
act_union = set()
while i < len(groups):
new_split_point = cls._find_split_point(
t, groups[i], split_point, act_union)
trace_i = tuple()
j = split_point
while j < new_split_point:
if t[j] in groups[i]:
trace_i = trace_i + (t[j],)
j = j + 1
logs[i].update({trace_i: obj.data_structure[t]})
split_point = new_split_point
act_union = act_union.union(set(groups[i]))
i = i + 1
return list(map(lambda l: IMDataStructureUVCL(l), logs))
@classmethod
def _find_split_point(cls, t: Tuple[Any], group: Collection[Any], start: int, ignore: Collection[Any], parameters: Optional[Dict[str, Any]] = None) -> int:
least_cost = 0
position_with_least_cost = start
cost = 0
i = start
while i < len(t):
if t[i] in group:
cost = cost - 1
elif t[i] not in ignore:
cost = cost + 1
if cost < least_cost:
least_cost = cost
position_with_least_cost = i + 1
i = i + 1
return position_with_least_cost
class StrictSequenceCutUVCL(StrictSequenceCut[IMDataStructureUVCL], SequenceCutUVCL):
@classmethod
def holds(cls, obj: T, parameters: Optional[Dict[str, Any]] = None) -> Optional[List[Collection[Any]]]:
return StrictSequenceCut.holds(obj, parameters)
class SequenceCutDFG(SequenceCut[IMDataStructureDFG]):
@classmethod
def project(cls, obj: IMDataStructureDFG, groups: List[Collection[Any]], parameters: Optional[Dict[str, Any]] = None) -> List[IMDataStructureDFG]:
dfg = obj.dfg
start_activities = []
end_activities = []
activities = []
dfgs = []
skippable = []
for g in groups:
skippable.append(False)
activities_idx = {}
for gind, g in enumerate(groups):
for act in g:
activities_idx[act] = int(gind)
i = 0
while i < len(groups):
to_succ_arcs = Counter()
from_prev_arcs = Counter()
if i < len(groups) - 1:
for (a, b) in dfg.graph:
if a in groups[i] and b in groups[i + 1]:
to_succ_arcs[a] += dfg.graph[(a, b)]
if i > 0:
for (a, b) in dfg.graph:
if a in groups[i - 1] and b in groups[i]:
from_prev_arcs[b] += dfg.graph[(a, b)]
if i == 0:
start_activities.append({})
for a in dfg.start_activities:
if a in groups[i]:
start_activities[i][a] = dfg.start_activities[a]
else:
j = i
while j < activities_idx[a]:
skippable[j] = True
j = j + 1
else:
start_activities.append(from_prev_arcs)
if i == len(groups) - 1:
end_activities.append({})
for a in dfg.end_activities:
if a in groups[i]:
end_activities[i][a] = dfg.end_activities[a]
else:
j = activities_idx[a] + 1
while j <= i:
skippable[j] = True
j = j + 1
else:
end_activities.append(to_succ_arcs)
activities.append({})
act_count = dfu.get_vertex_frequencies(dfg)
for a in groups[i]:
activities[i][a] = act_count[a]
dfgs.append({})
for (a, b) in dfg.graph:
if a in groups[i] and b in groups[i]:
dfgs[i][(a, b)] = dfg.graph[(a, b)]
i = i + 1
i = 0
while i < len(dfgs):
dfi = DFG()
[dfi.graph.update({(a, b): dfgs[i][(a, b)]}) for (a, b) in dfgs[i]]
[dfi.start_activities.update({a: start_activities[i][a]}) for a in start_activities[i]]
[dfi.end_activities.update({a: end_activities[i][a]}) for a in end_activities[i]]
dfgs[i] = dfi
i = i + 1
for (a, b) in dfg.graph:
z = activities_idx[b]
j = activities_idx[a] + 1
while j < z:
skippable[j] = False
j = j + 1
return [IMDataStructureDFG(InductiveDFG(dfg=dfgs[i], skip=skippable[i])) for i in range(len(dfgs))]
class StrictSequenceCutDFG(StrictSequenceCut[IMDataStructureDFG], SequenceCutDFG):
@classmethod
def holds(cls, obj: T, parameters: Optional[Dict[str, Any]] = None) -> Optional[List[Collection[Any]]]:
return StrictSequenceCut.holds(obj, parameters)
|