forked from ArchipelagoMW/Archipelago
-
Notifications
You must be signed in to change notification settings - Fork 2
/
Copy pathtest_fill.py
860 lines (683 loc) · 39.3 KB
/
test_fill.py
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
from typing import List, Iterable
import unittest
from Options import Accessibility
from test.general import generate_items, generate_locations, generate_test_multiworld
from Fill import FillError, balance_multiworld_progression, fill_restrictive, \
distribute_early_items, distribute_items_restrictive
from BaseClasses import Entrance, LocationProgressType, MultiWorld, Region, Item, Location, \
ItemClassification
from worlds.generic.Rules import CollectionRule, add_item_rule, locality_rules, set_rule
class PlayerDefinition(object):
multiworld: MultiWorld
id: int
menu: Region
locations: List[Location]
prog_items: List[Item]
basic_items: List[Item]
regions: List[Region]
def __init__(self, multiworld: MultiWorld, id: int, menu: Region, locations: List[Location] = [], prog_items: List[Item] = [], basic_items: List[Item] = []):
self.multiworld = multiworld
self.id = id
self.menu = menu
self.locations = locations
self.prog_items = prog_items
self.basic_items = basic_items
self.regions = [menu]
def generate_region(self, parent: Region, size: int, access_rule: CollectionRule = lambda state: True) -> Region:
region_tag = f"_region{len(self.regions)}"
region_name = f"player{self.id}{region_tag}"
region = Region(f"player{self.id}{region_tag}", self.id, self.multiworld)
self.locations += generate_locations(size, self.id, region, None, region_tag)
entrance = Entrance(self.id, f"{region_name}_entrance", parent)
parent.exits.append(entrance)
entrance.connect(region)
entrance.access_rule = access_rule
self.regions.append(region)
self.multiworld.regions.append(region)
return region
def fill_region(multiworld: MultiWorld, region: Region, items: List[Item]) -> List[Item]:
items = items.copy()
while len(items) > 0:
location = region.locations.pop(0)
region.locations.append(location)
if location.item:
return items
item = items.pop(0)
multiworld.push_item(location, item, False)
return items
def region_contains(region: Region, item: Item) -> bool:
for location in region.locations:
if location.item == item:
return True
return False
def generate_player_data(multiworld: MultiWorld, player_id: int, location_count: int = 0, prog_item_count: int = 0, basic_item_count: int = 0) -> PlayerDefinition:
menu = multiworld.get_region("Menu", player_id)
locations = generate_locations(location_count, player_id, menu, None)
prog_items = generate_items(prog_item_count, player_id, True)
multiworld.itempool += prog_items
basic_items = generate_items(basic_item_count, player_id, False)
multiworld.itempool += basic_items
return PlayerDefinition(multiworld, player_id, menu, locations, prog_items, basic_items)
def names(objs: list) -> Iterable[str]:
return map(lambda o: o.name, objs)
class TestFillRestrictive(unittest.TestCase):
def test_basic_fill(self):
"""Tests `fill_restrictive` fills and removes the locations and items from their respective lists"""
multiworld = generate_test_multiworld()
player1 = generate_player_data(multiworld, 1, 2, 2)
item0 = player1.prog_items[0]
item1 = player1.prog_items[1]
loc0 = player1.locations[0]
loc1 = player1.locations[1]
fill_restrictive(multiworld, multiworld.state,
player1.locations, player1.prog_items)
self.assertEqual(loc0.item, item1)
self.assertEqual(loc1.item, item0)
self.assertEqual([], player1.locations)
self.assertEqual([], player1.prog_items)
def test_ordered_fill(self):
"""Tests `fill_restrictive` fulfills set rules"""
multiworld = generate_test_multiworld()
player1 = generate_player_data(multiworld, 1, 2, 2)
items = player1.prog_items
locations = player1.locations
multiworld.completion_condition[player1.id] = lambda state: state.has(
items[0].name, player1.id) and state.has(items[1].name, player1.id)
set_rule(locations[1], lambda state: state.has(
items[0].name, player1.id))
fill_restrictive(multiworld, multiworld.state,
player1.locations.copy(), player1.prog_items.copy())
self.assertEqual(locations[0].item, items[0])
self.assertEqual(locations[1].item, items[1])
def test_partial_fill(self):
"""Tests that `fill_restrictive` returns unfilled locations"""
multiworld = generate_test_multiworld()
player1 = generate_player_data(multiworld, 1, 3, 2)
item0 = player1.prog_items[0]
item1 = player1.prog_items[1]
loc0 = player1.locations[0]
loc1 = player1.locations[1]
loc2 = player1.locations[2]
multiworld.completion_condition[player1.id] = lambda state: state.has(
item0.name, player1.id) and state.has(item1.name, player1.id)
set_rule(loc1, lambda state: state.has(
item0.name, player1.id))
# forces a swap
set_rule(loc2, lambda state: state.has(
item0.name, player1.id))
fill_restrictive(multiworld, multiworld.state,
player1.locations, player1.prog_items)
self.assertEqual(loc0.item, item0)
self.assertEqual(loc1.item, item1)
self.assertEqual(1, len(player1.locations))
self.assertEqual(player1.locations[0], loc2)
def test_minimal_fill(self):
"""Test that fill for minimal player can have unreachable items"""
multiworld = generate_test_multiworld()
player1 = generate_player_data(multiworld, 1, 2, 2)
items = player1.prog_items
locations = player1.locations
multiworld.worlds[player1.id].options.accessibility = Accessibility.from_any(Accessibility.option_minimal)
multiworld.completion_condition[player1.id] = lambda state: state.has(
items[1].name, player1.id)
set_rule(locations[1], lambda state: state.has(
items[0].name, player1.id))
fill_restrictive(multiworld, multiworld.state,
player1.locations.copy(), player1.prog_items.copy())
self.assertEqual(locations[0].item, items[1])
# Unnecessary unreachable Item
self.assertEqual(locations[1].item, items[0])
def test_minimal_mixed_fill(self):
"""
Test that fill for 1 minimal and 1 non-minimal player will correctly place items in a way that lets
the non-minimal player get all items.
"""
multiworld = generate_test_multiworld(2)
player1 = generate_player_data(multiworld, 1, 3, 3)
player2 = generate_player_data(multiworld, 2, 3, 3)
multiworld.worlds[player1.id].options.accessibility.value = Accessibility.option_minimal
multiworld.worlds[player2.id].options.accessibility.value = Accessibility.option_full
multiworld.completion_condition[player1.id] = lambda state: True
multiworld.completion_condition[player2.id] = lambda state: state.has(player2.prog_items[2].name, player2.id)
set_rule(player1.locations[1], lambda state: state.has(player1.prog_items[0].name, player1.id))
set_rule(player1.locations[2], lambda state: state.has(player1.prog_items[1].name, player1.id))
set_rule(player2.locations[1], lambda state: state.has(player2.prog_items[0].name, player2.id))
set_rule(player2.locations[2], lambda state: state.has(player2.prog_items[1].name, player2.id))
# force-place an item that makes it impossible to have all locations accessible
player1.locations[0].place_locked_item(player1.prog_items[2])
# fill remaining locations with remaining items
location_pool = player1.locations[1:] + player2.locations
item_pool = player1.prog_items[:-1] + player2.prog_items
fill_restrictive(multiworld, multiworld.state, location_pool, item_pool)
multiworld.state.sweep_for_advancements() # collect everything
# all of player2's locations and items should be accessible (not all of player1's)
for item in player2.prog_items:
self.assertTrue(multiworld.state.has(item.name, player2.id),
f"{item} is unreachable in {item.location}")
def test_reversed_fill(self):
"""Test a different set of rules can be satisfied"""
multiworld = generate_test_multiworld()
player1 = generate_player_data(multiworld, 1, 2, 2)
item0 = player1.prog_items[0]
item1 = player1.prog_items[1]
loc0 = player1.locations[0]
loc1 = player1.locations[1]
multiworld.completion_condition[player1.id] = lambda state: state.has(
item0.name, player1.id) and state.has(item1.name, player1.id)
set_rule(loc1, lambda state: state.has(item1.name, player1.id))
fill_restrictive(multiworld, multiworld.state,
player1.locations, player1.prog_items)
self.assertEqual(loc0.item, item1)
self.assertEqual(loc1.item, item0)
def test_multi_step_fill(self):
"""Test that fill is able to satisfy multiple spheres"""
multiworld = generate_test_multiworld()
player1 = generate_player_data(multiworld, 1, 4, 4)
items = player1.prog_items
locations = player1.locations
multiworld.completion_condition[player1.id] = lambda state: state.has(
items[2].name, player1.id) and state.has(items[3].name, player1.id)
set_rule(locations[1], lambda state: state.has(
items[0].name, player1.id))
set_rule(locations[2], lambda state: state.has(
items[1].name, player1.id))
set_rule(locations[3], lambda state: state.has(
items[1].name, player1.id))
fill_restrictive(multiworld, multiworld.state,
player1.locations.copy(), player1.prog_items.copy())
self.assertEqual(locations[0].item, items[1])
self.assertEqual(locations[1].item, items[2])
self.assertEqual(locations[2].item, items[0])
self.assertEqual(locations[3].item, items[3])
def test_impossible_fill(self):
"""Test that fill raises an error when it can't place any items"""
multiworld = generate_test_multiworld()
player1 = generate_player_data(multiworld, 1, 2, 2)
items = player1.prog_items
locations = player1.locations
multiworld.completion_condition[player1.id] = lambda state: state.has(
items[0].name, player1.id) and state.has(items[1].name, player1.id)
set_rule(locations[1], lambda state: state.has(
items[1].name, player1.id))
set_rule(locations[0], lambda state: state.has(
items[0].name, player1.id))
self.assertRaises(FillError, fill_restrictive, multiworld, multiworld.state,
player1.locations.copy(), player1.prog_items.copy())
def test_circular_fill(self):
"""Test that fill raises an error when it can't place all items"""
multiworld = generate_test_multiworld()
player1 = generate_player_data(multiworld, 1, 3, 3)
item0 = player1.prog_items[0]
item1 = player1.prog_items[1]
item2 = player1.prog_items[2]
loc0 = player1.locations[0]
loc1 = player1.locations[1]
loc2 = player1.locations[2]
multiworld.completion_condition[player1.id] = lambda state: state.has(
item0.name, player1.id) and state.has(item1.name, player1.id) and state.has(item2.name, player1.id)
set_rule(loc1, lambda state: state.has(item0.name, player1.id))
set_rule(loc2, lambda state: state.has(item1.name, player1.id))
set_rule(loc0, lambda state: state.has(item2.name, player1.id))
self.assertRaises(FillError, fill_restrictive, multiworld, multiworld.state,
player1.locations.copy(), player1.prog_items.copy())
def test_competing_fill(self):
"""Test that fill raises an error when it can't place items in a way to satisfy the conditions"""
multiworld = generate_test_multiworld()
player1 = generate_player_data(multiworld, 1, 2, 2)
item0 = player1.prog_items[0]
item1 = player1.prog_items[1]
loc1 = player1.locations[1]
multiworld.completion_condition[player1.id] = lambda state: state.has(
item0.name, player1.id) and state.has(item0.name, player1.id) and state.has(item1.name, player1.id)
set_rule(loc1, lambda state: state.has(item0.name, player1.id)
and state.has(item1.name, player1.id))
self.assertRaises(FillError, fill_restrictive, multiworld, multiworld.state,
player1.locations.copy(), player1.prog_items.copy())
def test_multiplayer_fill(self):
"""Test that items can be placed across worlds"""
multiworld = generate_test_multiworld(2)
player1 = generate_player_data(multiworld, 1, 2, 2)
player2 = generate_player_data(multiworld, 2, 2, 2)
multiworld.completion_condition[player1.id] = lambda state: state.has(
player1.prog_items[0].name, player1.id) and state.has(
player1.prog_items[1].name, player1.id)
multiworld.completion_condition[player2.id] = lambda state: state.has(
player2.prog_items[0].name, player2.id) and state.has(
player2.prog_items[1].name, player2.id)
fill_restrictive(multiworld, multiworld.state, player1.locations +
player2.locations, player1.prog_items + player2.prog_items)
self.assertEqual(player1.locations[0].item, player1.prog_items[1])
self.assertEqual(player1.locations[1].item, player2.prog_items[1])
self.assertEqual(player2.locations[0].item, player1.prog_items[0])
self.assertEqual(player2.locations[1].item, player2.prog_items[0])
def test_multiplayer_rules_fill(self):
"""Test that fill across worlds satisfies the rules"""
multiworld = generate_test_multiworld(2)
player1 = generate_player_data(multiworld, 1, 2, 2)
player2 = generate_player_data(multiworld, 2, 2, 2)
multiworld.completion_condition[player1.id] = lambda state: state.has(
player1.prog_items[0].name, player1.id) and state.has(
player1.prog_items[1].name, player1.id)
multiworld.completion_condition[player2.id] = lambda state: state.has(
player2.prog_items[0].name, player2.id) and state.has(
player2.prog_items[1].name, player2.id)
set_rule(player2.locations[1], lambda state: state.has(
player2.prog_items[0].name, player2.id))
fill_restrictive(multiworld, multiworld.state, player1.locations +
player2.locations, player1.prog_items + player2.prog_items)
self.assertEqual(player1.locations[0].item, player2.prog_items[0])
self.assertEqual(player1.locations[1].item, player2.prog_items[1])
self.assertEqual(player2.locations[0].item, player1.prog_items[0])
self.assertEqual(player2.locations[1].item, player1.prog_items[1])
def test_restrictive_progress(self):
"""Test that various spheres with different requirements can be filled"""
multiworld = generate_test_multiworld()
player1 = generate_player_data(multiworld, 1, prog_item_count=25)
items = player1.prog_items.copy()
multiworld.completion_condition[player1.id] = lambda state: state.has_all(
names(player1.prog_items), player1.id)
player1.generate_region(player1.menu, 5)
player1.generate_region(player1.menu, 5, lambda state: state.has_all(
names(items[2:7]), player1.id))
player1.generate_region(player1.menu, 5, lambda state: state.has_all(
names(items[7:12]), player1.id))
player1.generate_region(player1.menu, 5, lambda state: state.has_all(
names(items[12:17]), player1.id))
player1.generate_region(player1.menu, 5, lambda state: state.has_all(
names(items[17:22]), player1.id))
locations = multiworld.get_unfilled_locations()
fill_restrictive(multiworld, multiworld.state,
locations, player1.prog_items)
def test_swap_to_earlier_location_with_item_rule(self):
"""Test that item swap happens and works as intended"""
# test for PR#1109
multiworld = generate_test_multiworld(1)
player1 = generate_player_data(multiworld, 1, 4, 4)
locations = player1.locations[:] # copy required
items = player1.prog_items[:] # copy required
# for the test to work, item and location order is relevant: Sphere 1 last, allowed_item not last
for location in locations[:-1]: # Sphere 2
# any one provides access to Sphere 2
set_rule(location, lambda state: any(state.has(item.name, player1.id) for item in items))
# forbid all but 1 item in Sphere 1
sphere1_loc = locations[-1]
allowed_item = items[1]
add_item_rule(sphere1_loc, lambda item_to_place: item_to_place == allowed_item)
# test our rules
self.assertTrue(location.can_fill(None, allowed_item, False), "Test is flawed")
self.assertTrue(location.can_fill(None, items[2], False), "Test is flawed")
self.assertTrue(sphere1_loc.can_fill(None, allowed_item, False), "Test is flawed")
self.assertFalse(sphere1_loc.can_fill(None, items[2], False), "Test is flawed")
# fill has to place items[1] in locations[0] which will result in a swap because of placement order
fill_restrictive(multiworld, multiworld.state, player1.locations, player1.prog_items)
# assert swap happened
self.assertTrue(sphere1_loc.item, "Did not swap required item into Sphere 1")
self.assertEqual(sphere1_loc.item, allowed_item, "Wrong item in Sphere 1")
def test_swap_to_earlier_location_with_item_rule2(self):
"""Test that swap works before all items are placed"""
multiworld = generate_test_multiworld(1)
player1 = generate_player_data(multiworld, 1, 5, 5)
locations = player1.locations[:] # copy required
items = player1.prog_items[:] # copy required
# Two items provide access to sphere 2.
# One of them is forbidden in sphere 1, the other is first placed in sphere 4 because of placement order,
# requiring a swap.
# There are spheres in between, so for the swap to work, it'll have to assume all other items are collected.
one_to_two1 = items[4].name
one_to_two2 = items[3].name
three_to_four = items[2].name
two_to_three1 = items[1].name
two_to_three2 = items[0].name
# Sphere 4
set_rule(locations[0], lambda state: ((state.has(one_to_two1, player1.id) or state.has(one_to_two2, player1.id))
and state.has(two_to_three1, player1.id)
and state.has(two_to_three2, player1.id)
and state.has(three_to_four, player1.id)))
# Sphere 3
set_rule(locations[1], lambda state: ((state.has(one_to_two1, player1.id) or state.has(one_to_two2, player1.id))
and state.has(two_to_three1, player1.id)
and state.has(two_to_three2, player1.id)))
# Sphere 2
set_rule(locations[2], lambda state: state.has(one_to_two1, player1.id) or state.has(one_to_two2, player1.id))
# Sphere 1
sphere1_loc1 = locations[3]
sphere1_loc2 = locations[4]
# forbid one_to_two2 in sphere 1 to make the swap happen as described above
add_item_rule(sphere1_loc1, lambda item_to_place: item_to_place.name != one_to_two2)
add_item_rule(sphere1_loc2, lambda item_to_place: item_to_place.name != one_to_two2)
# Now fill should place one_to_two1 in sphere1_loc1 or sphere1_loc2 via swap,
# which it will attempt before two_to_three and three_to_four are placed, testing the behavior.
fill_restrictive(multiworld, multiworld.state, player1.locations, player1.prog_items)
# assert swap happened
self.assertTrue(sphere1_loc1.item and sphere1_loc2.item, "Did not swap required item into Sphere 1")
self.assertTrue(sphere1_loc1.item.name == one_to_two1 or
sphere1_loc2.item.name == one_to_two1, "Wrong item in Sphere 1")
def test_double_sweep(self):
"""Test that sweep doesn't duplicate Event items when sweeping"""
# test for PR1114
multiworld = generate_test_multiworld(1)
player1 = generate_player_data(multiworld, 1, 1, 1)
location = player1.locations[0]
location.address = None
item = player1.prog_items[0]
item.code = None
location.place_locked_item(item)
multiworld.state.sweep_for_advancements()
multiworld.state.sweep_for_advancements()
self.assertTrue(multiworld.state.prog_items[item.player][item.name], "Sweep did not collect - Test flawed")
self.assertEqual(multiworld.state.prog_items[item.player][item.name], 1, "Sweep collected multiple times")
def test_correct_item_instance_removed_from_pool(self):
"""Test that a placed item gets removed from the submitted pool"""
multiworld = generate_test_multiworld()
player1 = generate_player_data(multiworld, 1, 2, 2)
player1.prog_items[0].name = "Different_item_instance_but_same_item_name"
player1.prog_items[1].name = "Different_item_instance_but_same_item_name"
loc0 = player1.locations[0]
fill_restrictive(multiworld, multiworld.state,
[loc0], player1.prog_items)
self.assertEqual(1, len(player1.prog_items))
self.assertIsNot(loc0.item, player1.prog_items[0], "Filled item was still present in item pool")
class TestDistributeItemsRestrictive(unittest.TestCase):
def test_basic_distribute(self):
"""Test that distribute_items_restrictive is deterministic"""
multiworld = generate_test_multiworld()
player1 = generate_player_data(
multiworld, 1, 4, prog_item_count=2, basic_item_count=2)
locations = player1.locations
prog_items = player1.prog_items
basic_items = player1.basic_items
distribute_items_restrictive(multiworld)
self.assertEqual(locations[0].item, basic_items[1])
self.assertFalse(locations[0].advancement)
self.assertEqual(locations[1].item, prog_items[0])
self.assertTrue(locations[1].advancement)
self.assertEqual(locations[2].item, prog_items[1])
self.assertTrue(locations[2].advancement)
self.assertEqual(locations[3].item, basic_items[0])
self.assertFalse(locations[3].advancement)
def test_excluded_distribute(self):
"""Test that distribute_items_restrictive doesn't put advancement items on excluded locations"""
multiworld = generate_test_multiworld()
player1 = generate_player_data(
multiworld, 1, 4, prog_item_count=2, basic_item_count=2)
locations = player1.locations
locations[1].progress_type = LocationProgressType.EXCLUDED
locations[2].progress_type = LocationProgressType.EXCLUDED
distribute_items_restrictive(multiworld)
self.assertFalse(locations[1].item.advancement)
self.assertFalse(locations[2].item.advancement)
def test_non_excluded_item_distribute(self):
"""Test that useful items aren't placed on excluded locations"""
multiworld = generate_test_multiworld()
player1 = generate_player_data(
multiworld, 1, 4, prog_item_count=2, basic_item_count=2)
locations = player1.locations
basic_items = player1.basic_items
locations[1].progress_type = LocationProgressType.EXCLUDED
basic_items[1].classification = ItemClassification.useful
distribute_items_restrictive(multiworld)
self.assertEqual(locations[1].item, basic_items[0])
def test_too_many_excluded_distribute(self):
"""Test that fill fails if it can't place all progression items due to too many excluded locations"""
multiworld = generate_test_multiworld()
player1 = generate_player_data(
multiworld, 1, 4, prog_item_count=2, basic_item_count=2)
locations = player1.locations
locations[0].progress_type = LocationProgressType.EXCLUDED
locations[1].progress_type = LocationProgressType.EXCLUDED
locations[2].progress_type = LocationProgressType.EXCLUDED
self.assertRaises(FillError, distribute_items_restrictive, multiworld)
def test_non_excluded_item_must_distribute(self):
"""Test that fill fails if it can't place useful items due to too many excluded locations"""
multiworld = generate_test_multiworld()
player1 = generate_player_data(
multiworld, 1, 4, prog_item_count=2, basic_item_count=2)
locations = player1.locations
basic_items = player1.basic_items
locations[1].progress_type = LocationProgressType.EXCLUDED
locations[2].progress_type = LocationProgressType.EXCLUDED
basic_items[0].classification = ItemClassification.useful
basic_items[1].classification = ItemClassification.useful
self.assertRaises(FillError, distribute_items_restrictive, multiworld)
def test_priority_distribute(self):
"""Test that priority locations receive advancement items"""
multiworld = generate_test_multiworld()
player1 = generate_player_data(
multiworld, 1, 4, prog_item_count=2, basic_item_count=2)
locations = player1.locations
locations[0].progress_type = LocationProgressType.PRIORITY
locations[3].progress_type = LocationProgressType.PRIORITY
distribute_items_restrictive(multiworld)
self.assertTrue(locations[0].item.advancement)
self.assertTrue(locations[3].item.advancement)
def test_excess_priority_distribute(self):
"""Test that if there's more priority locations than advancement items, they can still fill"""
multiworld = generate_test_multiworld()
player1 = generate_player_data(
multiworld, 1, 4, prog_item_count=2, basic_item_count=2)
locations = player1.locations
locations[0].progress_type = LocationProgressType.PRIORITY
locations[1].progress_type = LocationProgressType.PRIORITY
locations[2].progress_type = LocationProgressType.PRIORITY
distribute_items_restrictive(multiworld)
self.assertFalse(locations[3].item.advancement)
def test_multiple_world_priority_distribute(self):
"""Test that priority fill can be satisfied for multiple worlds"""
multiworld = generate_test_multiworld(3)
player1 = generate_player_data(
multiworld, 1, 4, prog_item_count=2, basic_item_count=2)
player2 = generate_player_data(
multiworld, 2, 4, prog_item_count=1, basic_item_count=3)
player3 = generate_player_data(
multiworld, 3, 6, prog_item_count=4, basic_item_count=2)
player1.locations[2].progress_type = LocationProgressType.PRIORITY
player1.locations[3].progress_type = LocationProgressType.PRIORITY
player2.locations[1].progress_type = LocationProgressType.PRIORITY
player3.locations[0].progress_type = LocationProgressType.PRIORITY
player3.locations[1].progress_type = LocationProgressType.PRIORITY
player3.locations[2].progress_type = LocationProgressType.PRIORITY
player3.locations[3].progress_type = LocationProgressType.PRIORITY
distribute_items_restrictive(multiworld)
self.assertTrue(player1.locations[2].item.advancement)
self.assertTrue(player1.locations[3].item.advancement)
self.assertTrue(player2.locations[1].item.advancement)
self.assertTrue(player3.locations[0].item.advancement)
self.assertTrue(player3.locations[1].item.advancement)
self.assertTrue(player3.locations[2].item.advancement)
self.assertTrue(player3.locations[3].item.advancement)
def test_can_remove_locations_in_fill_hook(self):
"""Test that distribute_items_restrictive calls the fill hook and allows for item and location removal"""
multiworld = generate_test_multiworld()
player1 = generate_player_data(
multiworld, 1, 4, prog_item_count=2, basic_item_count=2)
removed_item: list[Item] = []
removed_location: list[Location] = []
def fill_hook(progitempool, usefulitempool, filleritempool, fill_locations):
removed_item.append(filleritempool.pop(0))
removed_location.append(fill_locations.pop(0))
multiworld.worlds[player1.id].fill_hook = fill_hook
distribute_items_restrictive(multiworld)
self.assertIsNone(removed_item[0].location)
self.assertIsNone(removed_location[0].item)
def test_seed_robust_to_item_order(self):
"""Test deterministic fill"""
mw1 = generate_test_multiworld()
gen1 = generate_player_data(
mw1, 1, 4, prog_item_count=2, basic_item_count=2)
distribute_items_restrictive(mw1)
mw2 = generate_test_multiworld()
gen2 = generate_player_data(
mw2, 1, 4, prog_item_count=2, basic_item_count=2)
mw2.itempool.append(mw2.itempool.pop(0))
distribute_items_restrictive(mw2)
self.assertEqual(gen1.locations[0].item, gen2.locations[0].item)
self.assertEqual(gen1.locations[1].item, gen2.locations[1].item)
self.assertEqual(gen1.locations[2].item, gen2.locations[2].item)
self.assertEqual(gen1.locations[3].item, gen2.locations[3].item)
def test_seed_robust_to_location_order(self):
"""Test deterministic fill even if locations in a region are reordered"""
mw1 = generate_test_multiworld()
gen1 = generate_player_data(
mw1, 1, 4, prog_item_count=2, basic_item_count=2)
distribute_items_restrictive(mw1)
mw2 = generate_test_multiworld()
gen2 = generate_player_data(
mw2, 1, 4, prog_item_count=2, basic_item_count=2)
reg = mw2.get_region("Menu", gen2.id)
reg.locations.append(reg.locations.pop(0))
distribute_items_restrictive(mw2)
self.assertEqual(gen1.locations[0].item, gen2.locations[0].item)
self.assertEqual(gen1.locations[1].item, gen2.locations[1].item)
self.assertEqual(gen1.locations[2].item, gen2.locations[2].item)
self.assertEqual(gen1.locations[3].item, gen2.locations[3].item)
def test_can_reserve_advancement_items_for_general_fill(self):
"""Test that priority locations fill still satisfies item rules"""
multiworld = generate_test_multiworld()
player1 = generate_player_data(
multiworld, 1, location_count=5, prog_item_count=5)
items = player1.prog_items
multiworld.completion_condition[player1.id] = lambda state: state.has_all(
names(items), player1.id)
location = player1.locations[0]
location.progress_type = LocationProgressType.PRIORITY
location.item_rule = lambda item: item not in items[:4]
distribute_items_restrictive(multiworld)
self.assertEqual(location.item, items[4])
def test_non_excluded_local_items(self):
"""Test that local items get placed locally in a multiworld"""
multiworld = generate_test_multiworld(2)
player1 = generate_player_data(
multiworld, 1, location_count=5, basic_item_count=5)
player2 = generate_player_data(
multiworld, 2, location_count=5, basic_item_count=5)
for item in multiworld.get_items():
item.classification = ItemClassification.useful
multiworld.worlds[player1.id].options.local_items.value = set(names(player1.basic_items))
multiworld.worlds[player2.id].options.local_items.value = set(names(player2.basic_items))
locality_rules(multiworld)
distribute_items_restrictive(multiworld)
for item in multiworld.get_items():
self.assertEqual(item.player, item.location.player)
self.assertFalse(item.location.advancement, False)
def test_early_items(self) -> None:
"""Test that the early items API successfully places items early"""
mw = generate_test_multiworld(2)
player1 = generate_player_data(mw, 1, location_count=5, basic_item_count=5)
player2 = generate_player_data(mw, 2, location_count=5, basic_item_count=5)
mw.early_items[1][player1.basic_items[0].name] = 1
mw.early_items[2][player2.basic_items[2].name] = 1
mw.early_items[2][player2.basic_items[3].name] = 1
early_items = [
player1.basic_items[0],
player2.basic_items[2],
player2.basic_items[3],
]
# copied this code from the beginning of `distribute_items_restrictive`
# before `distribute_early_items` is called
fill_locations = sorted(mw.get_unfilled_locations())
mw.random.shuffle(fill_locations)
itempool = sorted(mw.itempool)
mw.random.shuffle(itempool)
fill_locations, itempool = distribute_early_items(mw, fill_locations, itempool)
remaining_p1 = [item for item in itempool if item.player == 1]
remaining_p2 = [item for item in itempool if item.player == 2]
assert len(itempool) == 7, f"number of items remaining after early_items: {len(itempool)}"
assert len(remaining_p1) == 4, f"number of p1 items after early_items: {len(remaining_p1)}"
assert len(remaining_p2) == 3, f"number of p2 items after early_items: {len(remaining_p1)}"
for i in range(5):
if i != 0:
assert player1.basic_items[i] in itempool, "non-early item to remain in itempool"
if i not in {2, 3}:
assert player2.basic_items[i] in itempool, "non-early item to remain in itempool"
for item in early_items:
assert item not in itempool, "early item to be taken out of itempool"
assert len(fill_locations) == len(mw.get_locations()) - len(early_items), \
f"early location count from {mw.get_locations()} to {len(fill_locations)} " \
f"after {len(early_items)} early items"
items_in_locations = {loc.item for loc in mw.get_locations() if loc.item}
assert len(items_in_locations) == len(early_items), \
f"{len(early_items)} early items in {len(items_in_locations)} locations"
for item in early_items:
assert item in items_in_locations, "early item to be placed in location"
class TestBalanceMultiworldProgression(unittest.TestCase):
def assertRegionContains(self, region: Region, item: Item) -> bool:
for location in region.locations:
if location.item and location.item == item:
return True
self.fail(f"Expected {region.name} to contain {item.name}.\n"
f"Contains{list(map(lambda location: location.item, region.locations))}")
def setUp(self) -> None:
multiworld = generate_test_multiworld(2)
self.multiworld = multiworld
player1 = generate_player_data(
multiworld, 1, prog_item_count=2, basic_item_count=40)
self.player1 = player1
player2 = generate_player_data(
multiworld, 2, prog_item_count=2, basic_item_count=40)
self.player2 = player2
multiworld.completion_condition[player1.id] = lambda state: state.has(
player1.prog_items[0].name, player1.id) and state.has(
player1.prog_items[1].name, player1.id)
multiworld.completion_condition[player2.id] = lambda state: state.has(
player2.prog_items[0].name, player2.id) and state.has(
player2.prog_items[1].name, player2.id)
items = player1.basic_items + player2.basic_items
# Sphere 1
region = player1.generate_region(player1.menu, 20)
items = fill_region(multiworld, region, [
player1.prog_items[0]] + items)
# Sphere 2
region = player1.generate_region(
player1.regions[1], 20, lambda state: state.has(player1.prog_items[0].name, player1.id))
items = fill_region(
multiworld, region, [player1.prog_items[1], player2.prog_items[0]] + items)
# Sphere 3
region = player2.generate_region(
player2.menu, 20, lambda state: state.has(player2.prog_items[0].name, player2.id))
fill_region(multiworld, region, [player2.prog_items[1]] + items)
def test_balances_progression(self) -> None:
"""Tests that progression balancing moves progression items earlier"""
self.multiworld.worlds[self.player1.id].options.progression_balancing.value = 50
self.multiworld.worlds[self.player2.id].options.progression_balancing.value = 50
self.assertRegionContains(
self.player1.regions[2], self.player2.prog_items[0])
balance_multiworld_progression(self.multiworld)
self.assertRegionContains(
self.player1.regions[1], self.player2.prog_items[0])
def test_balances_progression_light(self) -> None:
"""Test that progression balancing still moves items earlier on minimum value"""
self.multiworld.worlds[self.player1.id].options.progression_balancing.value = 1
self.multiworld.worlds[self.player2.id].options.progression_balancing.value = 1
self.assertRegionContains(
self.player1.regions[2], self.player2.prog_items[0])
balance_multiworld_progression(self.multiworld)
# TODO: arrange for a result that's different from the default
self.assertRegionContains(
self.player1.regions[1], self.player2.prog_items[0])
def test_balances_progression_heavy(self) -> None:
"""Test that progression balancing moves items earlier on maximum value"""
self.multiworld.worlds[self.player1.id].options.progression_balancing.value = 99
self.multiworld.worlds[self.player2.id].options.progression_balancing.value = 99
self.assertRegionContains(
self.player1.regions[2], self.player2.prog_items[0])
balance_multiworld_progression(self.multiworld)
# TODO: arrange for a result that's different from the default
self.assertRegionContains(
self.player1.regions[1], self.player2.prog_items[0])
def test_skips_balancing_progression(self) -> None:
"""Test that progression balancing is skipped when players have it disabled"""
self.multiworld.worlds[self.player1.id].options.progression_balancing.value = 0
self.multiworld.worlds[self.player2.id].options.progression_balancing.value = 0
self.assertRegionContains(
self.player1.regions[2], self.player2.prog_items[0])
balance_multiworld_progression(self.multiworld)
self.assertRegionContains(
self.player1.regions[2], self.player2.prog_items[0])
def test_ignores_priority_locations(self) -> None:
"""Test that progression items on priority locations don't get moved by balancing"""
self.multiworld.worlds[self.player1.id].options.progression_balancing.value = 50
self.multiworld.worlds[self.player2.id].options.progression_balancing.value = 50
self.player2.prog_items[0].location.progress_type = LocationProgressType.PRIORITY
balance_multiworld_progression(self.multiworld)
self.assertRegionContains(
self.player1.regions[2], self.player2.prog_items[0])