-
Notifications
You must be signed in to change notification settings - Fork 7
/
stepfilters.py
526 lines (427 loc) · 18.1 KB
/
stepfilters.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
# Contains filters for the various modifiers possible
import random, math, games, copy
NOT_STEPS = ["D", "S", "L", "W", "B", "R"]
# 0 - Normal
# 1 - Mirror
# 2 - Left
# 3 - Right
# -1 - Shuffle, -2 - random
# Map A to B using table C; A[i] -> B[C[i]].
# FIXME: Use direction strings here... They're simpler.
STEP_MAPPINGS = {
"SINGLE": [[0, 1, 2, 3], [3, 2, 1, 0], [1, 3, 0, 2], [2, 0, 3, 1]],
"5PANEL": [[0, 1, 2, 3, 4], [3, 4, 2, 0, 1], [4, 0, 2, 1, 3],
[1, 3, 2, 4, 5]],
"PARAPARA": [[0, 1, 2, 3, 4], [1, 0, 2, 4, 3], [4, 0, 1, 2, 3],
[1, 2, 3, 4, 0]],
"6PANEL": [[0, 1, 2, 3, 4, 5], [4, 5, 3, 2, 0, 1], [2, 0, 5, 1, 3, 4],
[1, 3, 0, 4, 5, 2]],
"8PANEL": [[0, 1, 2, 3, 4, 5, 6, 7], [5, 6, 7, 4, 3, 0, 1, 2],
[3, 0, 1, 7, 2, 4, 5, 6], [1, 2, 4, 0, 5, 6, 7, 3]],
"9PANEL": [[0, 1, 2, 3, 4, 5, 6, 7, 8], [6, 7, 8, 5, 4, 3, 0, 1, 2],
[3, 0, 1, 8, 4, 2, 5, 6, 7], [1, 2, 5, 0, 4, 6, 7, 8, 3]],
"DMX": [[0, 1, 2, 3], [2, 3, 0, 1], [3, 0, 1, 2], [1, 2, 3, 0]],
"EZ2SINGLE": [[0, 1, 2, 3, 4], [4, 3, 2, 1, 0], [4, 3, 2, 1, 0],
[4, 3, 2, 1, 0]],
"EZ2REAL": [[0, 1, 2, 3, 4, 5, 6], [6, 4, 5, 3, 1, 2, 0],
[3, 2, 4, 6, 1, 5, 0], [6, 4, 1, 0, 2, 5, 3]],
}
MAP_EQUIVS = {
"SINGLE": ["DOUBLE", "COUPLE", "VERSUS"],
"5PANEL": ["5DOUBLE", "5COUPLE", "5VERSUS"],
"6PANEL": ["6DOUBLE", "6COUPLE", "6VERSUS"],
"8PANEL": ["8DOUBLE", "8COUPLE", "8VERSUS"],
"9PANEL": ["9DOUBLE", "9COUPLE", "9VERSUS"],
"PARAPARA": ["PARADOUBLE", "PARACOUPLE", "PARAVERSUS"],
"EZ2SINGLE": ["EZ2VERSUS", "EZ2DOUBLE", "EZ2COUPLE"],
"EZ2REAL": ["REALVERSUS", "REALCOUPLE", "REALDOUBLE"],
}
for mode, equivs in MAP_EQUIVS.items():
for eq in equivs: STEP_MAPPINGS[eq] = STEP_MAPPINGS[mode]
# This is a pseudorandom number generator that we're guaranteed will
# always return the same results between different versions of Python,
# given the same seed. This makes sure the same autogenerated steps are
# made across different versions and platforms.
# A) Never change these numbers; they do ensure good randomness.
# B) Never use this module anywhere except pydance; its randomness is bad.
class NonRandom(random.Random):
def __init__(self, seed = 1):
self.seed(seed)
self.m = 16807
self.n = 2147483647
def getstate(self): return self.seed
def setstate(self, state): self.seed(state)
def seed(self, seed = 1): self.seed = seed
def jumpahead(self, n): pass
def random(self):
self.seed = (self.seed * self.m) % self.n
return float(self.seed) / self.n
# General step transformation class. By default, identity transform.
class Transform(object):
def __init__(self, *args): pass
def transform(self, steps):
return [self._update_state(s) or self._transform(s) for s in steps]
def _update_state(self, s): pass
def _transform(self, s): return list(s)
# Compress the steps to remove empty lines. FIXME: Do this in fileparsers.
def compress(steps):
new_steps = []
beat_count = 0
last_event = None
for s in steps:
if not isinstance(s[0], float): # Not a step
if last_event is not None: new_steps.append([beat_count] + last_event)
last_event = None
beat_count = 0
new_steps.append(s)
elif s[1:].count(0) != (len(s) - 1) or last_event == None: # Non-empty
if last_event is not None: new_steps.append([beat_count] + last_event)
last_event = s[1:]
beat_count = s[0]
else: # Empty event
beat_count += s[0]
if last_event is not None: new_steps.append([beat_count] + last_event)
return new_steps
# Rotation, mirroring, shuffle, and random.
class MappingTransform(Transform):
def __init__(self, mode, opt):
self._mapping = STEP_MAPPINGS[mode][opt][:]
def _transform(self, steps):
if steps[0] not in NOT_STEPS:
steps = steps[:]
step = steps[1:]
for j,s in enumerate(step): steps[self._mapping[j] + 1] = s
return steps
else:
return steps[:]
class MirrorTransform(MappingTransform):
def __init__(self, mode): MappingTransform.__init__(self, mode, 1)
class LeftTransform(MappingTransform):
def __init__(self, mode): MappingTransform.__init__(self, mode, 2)
class RightTransform(MappingTransform):
def __init__(self, mode): MappingTransform.__init__(self, mode, 3)
class ShuffleTransform(MappingTransform):
def __init__(self, mode):
MappingTransform.__init__(self, mode, 0)
random.shuffle(self._mapping)
class RandomTransform(ShuffleTransform):
def __init__(self, mode):
ShuffleTransform.__init__(self, mode)
self._holds = []
def _update_state(self, steps):
if steps[0] not in NOT_STEPS:
if len(self._holds) == 0: random.shuffle(self._mapping)
for i,s in enumerate(steps[1:]):
if s & 1 and i in self._holds: self._holds.remove(i)
if s & 2: self._holds.append(i)
rotate = [Transform, MirrorTransform, LeftTransform, RightTransform,
RandomTransform, ShuffleTransform]
# Apply myriad additions/deletions to the step pattern
# FIXME: Return a list rather than in-place modify.
# Shit this is ugly because of that.
def size(steps, opt):
if opt == 1: little(steps, 4) # Tiny
elif opt == 2: little(steps, 2) # Little
elif opt == 3: insert_taps(steps, 4.0, 2.0, False) # Big
elif opt == 4: insert_taps(steps, 2.0, 1.0, False) # Quick
elif opt == 5: insert_taps(steps, 4.0, 3.0, True) # Skippy
# Remove steps that aren't on the beat
def little(steps, mod):
beat = 0.0
# We have to be careful here to end hold arrows at the correct time,
# even if the end falls on an off-beat. Otherwise they can run off into
# infinity.
holds = []
for s in steps:
if s[0] not in NOT_STEPS:
old_s = s[:]
if beat % mod != 0:
s[1:] = [0] * (len(s) - 1)
for i in holds[:]:
if old_s[i] & 1:
s[i] |= 1
holds.remove(i)
beat += s[0]
for i,si in enumerate(s[1:]):
if i not in holds and si & 2: holds.append(i)
elif s[0] == "D": beat += s[1]
# Insert taps if a note falls on a interval-even beat, and the next step
# is interval-away. Insert the new step offset away from the original step
# (and therefore the time for the offset set is interval - offset).
# not_same makes sure the random tap inserted isn't the same as either of
# the surrounding ones.
# Inspired by Stepmania's function of the same name, in src/NoteData.cpp.
def insert_taps(steps, interval, offset, not_same):
new_steps = []
holds = []
beat = 0.0
rand = NonRandom(int(interval * offset * len(steps)))
for i,step in enumerate(steps[:-1]):
if isinstance(step[0], float): # This is a note...
for j,s in enumerate(step[1:]):
if s & 2: holds.append(j)
elif s & 1 and j in holds: holds.remove(j)
if not isinstance(steps[i + 1][0], float):
new_steps.append(step) # Next isn't a note.
elif (step[1:].count(0) == len(step[1:]) or
steps[i + 1][1:].count(0) == len(steps[i + 1][1:])):
# The surrounding notes are both empty.
new_steps.append(step)
elif len(holds) > 1: # Don't add things during two holds
new_steps.append(step)
elif step[0] == interval and beat % interval == 0: # Bingo!
step = steps[i][0] = offset
beat += offset
new_steps.append(step)
if not_same:
start = rand.randint(0, len(steps[i][1:]) -1)
empty = [0] * len(step[1:])
for j in range(len(step[1:])):
checking = (start + j) % len(step[1:])
if not (step[checking + 1] or steps[i + 1][checking + 1]):
empty[checking] = 1
break
new_steps.append([interval - offset] + empty)
else:
new_step = [1] + ([0] * (len(steps[i][1:]) - 2))
rand.shuffle(new_step)
new_steps.append([interval - offset] + new_step)
else: new_steps.append(steps[i])
beat += new_steps[-1][0]
# Stupid inaccurate floating point.
if int(beat + 0.00001) > int(beat): beat = int(beat + 0.00001)
else:
if steps[i][0] == "D": beat += steps[i][1]
new_steps.append(steps[i])
steps[0:-1] = new_steps # Copy into place.
# Pretty obvious.
class RemoveHoldTransform(Transform):
def _transform(self, s):
if s[0] not in NOT_STEPS: return [s[0]] + [i & 13 for i in s[1:]]
else: return s[:]
# Remove secret steps; defined by the 4 bit being on, so 5, 6, or 7.
class RemoveSecret(Transform):
def _transform(self, s):
s = s[:]
if s[0] not in NOT_STEPS:
for i,si in enumerate(s[1:]):
if si & 4: si = s[i] = 0
return s
class RemoveJumps(Transform):
def __init__(self):
self._side = 0
self._holds = []
def _transform(self, s):
if s[0] not in NOT_STEPS:
step = s[1:]
for i,s in enumerate(step):
if s & 2 and i not in self._holds: self._holds.append(i)
if step.count(0) < len(step) - 1:
if self._side and not self._holds: step.reverse()
for i,stepi in enumerate(step):
if stepi:
step[i] = 0
break
if self._side and not self._holds: step.reverse()
self._side ^= 1
for i,s in enumerate(step):
if s & 1 and not s & 2 and i in self._holds:
self._holds.remove(i)
return [s[0]] + step
else: return s[:]
# Add jumps to on-beats with steps.
class WideTransform(Transform):
def __init__(self):
self._beat = 0.0
self._holds = []
def _update_state(self, s):
if s[0] not in NOT_STEPS:
for i,si in enumerate(s[1:]):
if si & 1 and i in self._holds: self._holds.remove(i)
def _transform(self, s):
if s[0] not in NOT_STEPS:
step = s[1:]
if (self._beat % 4 == 0 and s.count(0) == len(step) - 1 and
len(self._holds) == 0):
first = 0
while step[first] == 0: first += 1 # Find the first step
to_add = int(math.sqrt(self._beat)) % len(step)
if step[to_add] != 0: to_add = (to_add + 1) % len(step)
step[to_add] = 1
s = [s[0]] + step
for i,si in enumerate(s[1:]):
if si & 2: self._holds.append(i)
self._beat += s[0]
elif s[0] == "D": beat += s[1]
return s[:]
jumps = [RemoveJumps, Transform, WideTransform]
# Now, here's where stuff gets tricky. We have to randomly but
# deterministically generate fun steps for modes not in the file.
# Transform step patterns from N panel to M panel, M >= N.
class PanelTransform(Transform):
# key is a direction; value is a list of directions that are "fun" to
# map to. Repeating a direction makes it more likely to be chosen. No
# direction should map to itself; making that extra-likely is taken
# care of in the map generation algorithm already.
accept = { "k": "uullc", "u": "kkzzc", "z": "uurrc",
"l": "kkwwc", "c": "lluuddrrkzwg", "r": "zzggc",
"w": "llddc", "d": "wwggc", "g": "ddrrc" }
def __init__(self, orig_panels, new_panels, rand, freq = 0.075):
self.orig_panels = orig_panels
self.new_panels = new_panels
self.freq = freq
self.rand = rand # A NonRandom instance preseeded.
# This is the chance (increasing by 'freq' for each datum we process)
# that a new pseudorandom mapping table (see below) will be generated,
# when a data transform is requested.
self.count = 0
# This is true if we encounter the same (exact) pattern twice in a row.
# We never generate a new mapping table in such a case, and so preserve
# multiple taps on the same arrow.
self.repeating = False
# This tracks the last two patterns transformed.
self.last_processed = [None, None]
# A dictionary mapping directions (characters) to a list of directions
# (strings), that are a) "near" the original direction, and b) in
# self.new_panels.
self.accept = self._generate_accept()
# A list of directions in orig_panels currently behind held.
self.holds = []
self._generate_transform()
# Reduce the acceptable mapping table to only include directions from the
# game modes in question.
def _generate_accept(self):
accept = {}
for dir in PanelTransform.accept:
if dir in self.orig_panels:
accept[dir] = "".join([c for c in PanelTransform.accept[dir]
if c in self.new_panels])
for dir in self.orig_panels:
if not accept.has_key(dir):
accept[dir] = "".join(self.new_panels)
return accept
# transform_mapping (trans) is, essentially, a scrambled list of directions
# from orig_panels, but shoved into an array the size of new_panels.
def _generate_transform(self):
unmapped = list(self.orig_panels)
trans = [None] * len(self.new_panels)
# First, map any hold arrows to the same place they were in for the
# last transformation, so they end at the right time.
for d in self.holds:
if d in self.transform_mapping:
unmapped.remove(d)
trans[self.transform_mapping.index(d)] = d
# Next, give a high chance that any direction will be mapped
# to itself, assuming it exists in the new panels. Or, always map
# it to its old self, if no alternatives exist.
# Originally, the "high chance" was hard coded around 0.5 to 0.75.
# Lower values meant steps got shuffled more than desired on small
# mappings, but higher ones resulted in small->large mappings not
# using many of the available arrows. So, now the chance is the
# ratio between the two sizes.
chance = float(len(self.orig_panels)) / float(len(self.new_panels))
unmapped_original = list(unmapped) # Don't modify the iterating list
for d in unmapped_original:
if d in self.new_panels:
if self.rand.random() < chance or len(self.accept[d]) == 0:
trans[self.new_panels.index(d)] = d
unmapped.remove(d)
# Next, go through whatever is left, and map it acceptably.
for d in unmapped:
accept = [a for a in self.accept[d] if
trans[self.new_panels.index(a)] is None]
if len(accept) != 0:
dir = self.rand.choice(accept)
trans[self.new_panels.index(dir)] = d
else:
# If possible, see if we can map to the original direction again
# (all its neighbors may have gotten overwritten between the first
# check and now).
if d in self.new_panels and trans[self.new_panels.index(d)] is None:
trans[self.new_panels.index(d)] = d
# And finally --
# Arrows in this direction won't get mapped to anything.
# However, in order to avoid accidentally dropping many events
# through poorly generated mappings, generate a new one more
# quickly if this situation occurs.
else: self.count = self.freq * 5
self.transform_mapping = trans
# Update our internal state based on the data passed in.
def _update_state(self, steps):
# This is a delay / BPM change / etc
if not isinstance(steps[0], float): return
self.count += self.freq * steps[0]
# Check the last two events to see if this is a repeat.
self.repeating = False
if steps[1:] in self.last_processed:
self.repeating = True
self.last_processed.append(steps[1:])
self.last_processed.pop(0)
# If we're not repeating, see if it's time to generate a new
# transformation table.
if not self.repeating:
if self.count > self.rand.random():
self.count = 0
self._generate_transform()
# Note any holds, so they don't get remapped in the middle.
for i,s in enumerate(steps[1:]):
if s & 2:
self.holds.append(self.orig_panels[i])
elif s & 1 and self.orig_panels[i] in self.holds:
self.holds.remove(self.orig_panels[i])
# Actually perform the transformation on the data.
def _transform(self, steps):
# This is not a note, it's a BPM change / delay / etc
if not isinstance(steps[0], float): return list(steps)
new_steps = [0] * (len(self.transform_mapping) + 1)
new_steps[0] = steps[0]
for i,s in enumerate(steps[1:]):
if s and self.orig_panels[i] in self.transform_mapping:
new_steps[self.transform_mapping.index(self.orig_panels[i]) + 1] = s
return new_steps
# This lets us cover the common case of DDR steps for KSF files, and avoid
# losing steps in the downscaling.
class FiveToFourTransform(Transform):
def __init__(self, orig_panels, new_panels, rand):
self.holding_center = False
self.rand = rand # A NonRandom instance preseeded.
def _update_state(self, steps):
if not isinstance(steps[0], float): return
if steps[3] & 2: holding_center = True
elif steps[3] & 1: holding_center = False
def _transform(self, steps):
if not isinstance(steps[0], float): return list(steps)
new_steps = [steps[0], steps[1], steps[2], steps[4], steps[5]]
if 0 in new_steps:
possible = [i for i,new_step in enumerate(new_steps) if new_step == 0]
new_steps[self.rand.choice(possible)] |= steps[3]
return new_steps
# Transform a song's steps from one mode and difficulty to a target mode.
def generate_mode(song, difficulty, target_mode, pid):
equiv = {"SINGLE": "5PANEL", "VERSUS": "5VERSUS",
"COUPLE": "5COUPLE", "DOUBLE": "5DOUBLE" }
if target_mode in games.SINGLE: mode = "SINGLE"
elif target_mode in games.VERSUS: mode = "VERSUS"
elif target_mode in games.ONLY_COUPLE: mode = "COUPLE"
elif target_mode in games.DOUBLE: mode = "DOUBLE"
if song.steps.has_key(mode):
# Dance ManiaX can be found as DWI files, with exact visual mappings.
# Cheat and use that.
if target_mode[:3] == "DMX":
if target_mode in games.COUPLE:
return song.steps[mode][difficulty][pid]
else: return song.steps[mode][difficulty]
steps = song.steps[mode][difficulty]
T = PanelTransform
elif song.steps.has_key(equiv[mode]):
steps = song.steps[equiv[mode]][difficulty]
mode = equiv[mode]
if len(games.GAMES[target_mode].dirs) == 4: T = FiveToFourTransform
else: T = PanelTransform
else: print "This shouldn't happen! Email pyddr-devel@icculus.org."
if mode in games.COUPLE: steps = steps[pid]
seed = song.info["bpm"]
if song.info["gap"] != 0: seed *= song.info["gap"]
trans = T(games.GAMES[mode].dirs, games.GAMES[target_mode].dirs,
NonRandom(int(song.info["bpm"])))
return trans.transform(steps)