summaryrefslogtreecommitdiff
path: root/extras/insert_texts.py
blob: d5a801503a882beddce74357438b73b29cdee72c (plain)
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
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
#author: Bryan Bishop <kanzure@gmail.com>
#date: 2012-01-07, 2012-01-17, 2012-01-27
#insert TX_FAR targets into pokered.asm
#and other insertion tasks
import extract_maps
from analyze_texts import analyze_texts, text_pretty_printer_at, scan_rom_for_tx_fars
from pretty_map_headers import map_name_cleaner, make_text_label, map_constants, find_all_tx_fars, tx_far_pretty_printer, tx_far_label_maker
import pretty_map_headers
from analyze_incbins import asm, offset_to_pointer, find_incbin_to_replace_for, split_incbin_line_into_three, generate_diff_insert, load_asm, isolate_incbins, process_incbins, reset_incbins, apply_diff
import analyze_incbins
from gbz80disasm import text_asm_pretty_printer, output_bank_opcodes, load_labels, find_label
import os, sys
import subprocess
spacing = "	"
tx_fars = None
failed_attempts = {}

pokemons = ["BULBASAUR", "IVYSAUR", "VENUSAUR", "CHARMANDER", "CHARMELEON", "CHARIZARD", "SQUIRTLE", "WARTORTLE", "BLASTOISE", "CATERPIE", "METAPOD", "BUTTERFREE", "WEEDLE", "KAKUNA", "BEEDRILL", "PIDGEY", "PIDGEOTTO", "PIDGEOT", "RATTATA", "RATICATE", "SPEAROW", "FEAROW", "EKANS", "ARBOK", "PIKACHU", "RAICHU", "SANDSHREW", "SANDSLASH", "NIDORANF", "NIDORINA", "NIDOQUEEN", "NIDORANM", "NIDORINO", "NIDOKING", "CLEFAIRY", "CLEFABLE", "VULPIX", "NINETALES", "JIGGLYPUFF", "WIGGLYTUFF", "ZUBAT", "GOLBAT", "ODDISH", "GLOOM", "VILEPLUME", "PARAS", "PARASECT", "VENONAT", "VENOMOTH", "DIGLETT", "DUGTRIO", "MEOWTH", "PERSIAN", "PSYDUCK", "GOLDUCK", "MANKEY", "PRIMEAPE", "GROWLITHE", "ARCANINE", "POLIWAG", "POLIWHIRL", "POLIWRATH", "ABRA", "KADABRA", "ALAKAZAM", "MACHOP", "MACHOKE", "MACHAMP", "BELLSPROUT", "WEEPINBELL", "VICTREEBEL", "TENTACOOL", "TENTACRUEL", "GEODUDE", "GRAVELER", "GOLEM", "PONYTA", "RAPIDASH", "SLOWPOKE", "SLOWBRO", "MAGNEMITE", "MAGNETON", "FARFETCH_D", "DODUO", "DODRIO", "SEEL", "DEWGONG", "GRIMER", "MUK", "SHELLDER", "CLOYSTER", "GASTLY", "HAUNTER", "GENGAR", "ONIX", "DROWZEE", "HYPNO", "KRABBY", "KINGLER", "VOLTORB", "ELECTRODE", "EXEGGCUTE", "EXEGGUTOR", "CUBONE", "MAROWAK", "HITMONLEE", "HITMONCHAN", "LICKITUNG", "KOFFING", "WEEZING", "RHYHORN", "RHYDON", "CHANSEY", "TANGELA", "KANGASKHAN", "HORSEA", "SEADRA", "GOLDEEN", "SEAKING", "STARYU", "STARMIE", "MR_MIME", "SCYTHER", "JYNX", "ELECTABUZZ", "MAGMAR", "PINSIR", "TAUROS", "MAGIKARP", "GYARADOS", "LAPRAS", "DITTO", "EEVEE", "VAPOREON", "JOLTEON", "FLAREON", "PORYGON", "OMANYTE", "OMASTAR", "KABUTO", "KABUTOPS", "AERODACTYL", "SNORLAX", "ARTICUNO", "ZAPDOS", "MOLTRES", "DRATINI", "DRAGONAIR", "DRAGONITE", "MEWTWO", "MEW"]

moves = [["POUND", 0x01], ["KARATE_CHOP", 0x02], ["DOUBLESLAP", 0x03], ["COMET_PUNCH", 0x04], ["MEGA_PUNCH", 0x05], ["PAY_DAY", 0x06], ["FIRE_PUNCH", 0x07], ["ICE_PUNCH", 0x08], ["THUNDERPUNCH", 0x09], ["SCRATCH", 0x0A], ["VICEGRIP", 0x0B], ["GUILLOTINE", 0x0C], ["RAZOR_WIND", 0x0D], ["SWORDS_DANCE", 0x0E], ["CUT", 0x0F], ["GUST", 0x10], ["WING_ATTACK", 0x11], ["WHIRLWIND", 0x12], ["FLY", 0x13], ["BIND", 0x14], ["SLAM", 0x15], ["VINE_WHIP", 0x16], ["STOMP", 0x17], ["DOUBLE_KICK", 0x18], ["MEGA_KICK", 0x19], ["JUMP_KICK", 0x1A], ["ROLLING_KICK", 0x1B], ["SAND_ATTACK", 0x1C], ["HEADBUTT", 0x1D], ["HORN_ATTACK", 0x1E], ["FURY_ATTACK", 0x1F], ["HORN_DRILL", 0x20], ["TACKLE", 0x21], ["BODY_SLAM", 0x22], ["WRAP", 0x23], ["TAKE_DOWN", 0x24], ["THRASH", 0x25], ["DOUBLE_EDGE", 0x26], ["TAIL_WHIP", 0x27], ["POISON_STING", 0x28], ["TWINEEDLE", 0x29], ["PIN_MISSILE", 0x2A], ["LEER", 0x2B], ["BITE", 0x2C], ["GROWL", 0x2D], ["ROAR", 0x2E], ["SING", 0x2F], ["SUPERSONIC", 0x30], ["SONICBOOM", 0x31], ["DISABLE", 0x32], ["ACID", 0x33], ["EMBER", 0x34], ["FLAMETHROWER", 0x35], ["MIST", 0x36], ["WATER_GUN", 0x37], ["HYDRO_PUMP", 0x38], ["SURF", 0x39], ["ICE_BEAM", 0x3A], ["BLIZZARD", 0x3B], ["PSYBEAM", 0x3C], ["BUBBLEBEAM", 0x3D], ["AURORA_BEAM", 0x3E], ["HYPER_BEAM", 0x3F], ["PECK", 0x40], ["DRILL_PECK", 0x41], ["SUBMISSION", 0x42], ["LOW_KICK", 0x43], ["COUNTER", 0x44], ["SEISMIC_TOSS", 0x45], ["STRENGTH", 0x46], ["ABSORB", 0x47], ["MEGA_DRAIN", 0x48], ["LEECH_SEED", 0x49], ["GROWTH", 0x4A], ["RAZOR_LEAF", 0x4B], ["SOLARBEAM", 0x4C], ["POISONPOWDER", 0x4D], ["STUN_SPORE", 0x4E], ["SLEEP_POWDER", 0x4F], ["PETAL_DANCE", 0x50], ["STRING_SHOT", 0x51], ["DRAGON_RAGE", 0x52], ["FIRE_SPIN", 0x53], ["THUNDERSHOCK", 0x54], ["THUNDERBOLT", 0x55], ["THUNDER_WAVE", 0x56], ["THUNDER", 0x57], ["ROCK_THROW", 0x58], ["EARTHQUAKE", 0x59], ["FISSURE", 0x5A], ["DIG", 0x5B], ["TOXIC", 0x5C], ["CONFUSION", 0x5D], ["PSYCHIC_M", 0x5E], ["HYPNOSIS", 0x5F], ["MEDITATE", 0x60], ["AGILITY", 0x61], ["QUICK_ATTACK", 0x62], ["RAGE", 0x63], ["TELEPORT", 0x64], ["NIGHT_SHADE", 0x65], ["MIMIC", 0x66], ["SCREECH", 0x67], ["DOUBLE_TEAM", 0x68], ["RECOVER", 0x69], ["HARDEN", 0x6A], ["MINIMIZE", 0x6B], ["SMOKESCREEN", 0x6C], ["CONFUSE_RAY", 0x6D], ["WITHDRAW", 0x6E], ["DEFENSE_CURL", 0x6F], ["BARRIER", 0x70], ["LIGHT_SCREEN", 0x71], ["HAZE", 0x72], ["REFLECT", 0x73], ["FOCUS_ENERGY", 0x74], ["BIDE", 0x75], ["METRONOME", 0x76], ["MIRROR_MOVE", 0x77], ["SELFDESTRUCT", 0x78], ["EGG_BOMB", 0x79], ["LICK", 0x7A], ["SMOG", 0x7B], ["SLUDGE", 0x7C], ["BONE_CLUB", 0x7D], ["FIRE_BLAST", 0x7E], ["WATERFALL", 0x7F], ["CLAMP", 0x80], ["SWIFT", 0x81], ["SKULL_BASH", 0x82], ["SPIKE_CANNON", 0x83], ["CONSTRICT", 0x84], ["AMNESIA", 0x85], ["KINESIS", 0x86], ["SOFTBOILED", 0x87], ["HI_JUMP_KICK", 0x88], ["GLARE", 0x89], ["DREAM_EATER", 0x8A], ["POISON_GAS", 0x8B], ["BARRAGE", 0x8C], ["LEECH_LIFE", 0x8D], ["LOVELY_KISS", 0x8E], ["SKY_ATTACK", 0x8F], ["TRANSFORM", 0x90], ["BUBBLE", 0x91], ["DIZZY_PUNCH", 0x92], ["SPORE", 0x93], ["FLASH", 0x94], ["PSYWAVE", 0x95], ["SPLASH", 0x96], ["ACID_ARMOR", 0x97], ["CRABHAMMER", 0x98], ["EXPLOSION", 0x99], ["FURY_SWIPES", 0x9A], ["BONEMERANG", 0x9B], ["REST", 0x9C], ["ROCK_SLIDE", 0x9D], ["HYPER_FANG", 0x9E], ["SHARPEN", 0x9F], ["CONVERSION", 0xA0], ["TRI_ATTACK", 0xA1], ["SUPER_FANG", 0xA2], ["SLASH", 0xA3], ["SUBSTITUTE", 0xA4], ["STRUGGLE", 0xA5]]

elemental_types = [
["NORMAL", "EQU", 0x00],
["FIGHTING", "EQU", 0x01],
["FLYING", "EQU", 0x02],
["POISON", "EQU", 0x03],
["GROUND", "EQU", 0x04],
["ROCK", "EQU", 0x05],
["BUG", "EQU", 0x07],
["GHOST", "EQU", 0x08],
["FIRE", "EQU", 0x14],
["WATER", "EQU", 0x15],
["GRASS", "EQU", 0x16],
["ELECTRIC", "EQU", 0x17],
["PSYCHIC", "EQU", 0x18],
["ICE", "EQU", 0x19],
["DRAGON", "EQU", 0x1A]]

def local_reset_incbins():
    asm = None
    incbin_lines = []
    processed_incbins = {}
    analyze_incbins.asm = None
    analyze_incbins.incbin_lines = []
    analyze_incbins.processed_incbins = {}
    
    #reload
    load_asm()
    isolate_incbins()
    process_incbins()

def find_tx_far_entry(map_id, text_id):
    for tx_far_line in tx_fars:
        if tx_far_line[0] == map_id and tx_far_line[1] == text_id:
            return tx_far_line

def insert_tx_far(map_id, text_id, tx_far_line=None):
    "inserts a tx_far"
    global tx_fars
    if tx_far_line == None:
        tx_far_line = find_tx_far_entry(map_id, text_id)
    text_pointer = tx_far_line[2]
    start_address = tx_far_line[3]
    tx_far_object = tx_far_line[4]
    end_address = tx_far_object[1]["end_address"] + 1 #the end byte; +1 because of a bug somewhere :(

    line_number = find_incbin_to_replace_for(start_address)
    if line_number == None:
        print "skipping tx_far for map_id=" + str(map_id) + " text_id=" + str(text_id) + " text_pointer=" + hex(text_pointer) + " tx_far_start_address=" + hex(start_address)
        return

    #also do a name check
    label = tx_far_label_maker(extract_maps.map_headers[map_id]["name"], text_id)
    if (label + ":") in "\n".join(analyze_incbins.asm):
        print "skipping tx_far for map_id=" + str(map_id) + " text_id=" + str(text_id) + " text_pointer=" + hex(text_pointer) + " tx_far_start_address=" + hex(start_address)
        return
    
    newlines = split_incbin_line_into_three(line_number, start_address, end_address - start_address)
    tx_far_asm = tx_far_pretty_printer(tx_far_line)

    newlines = newlines.split("\n")
    if len(newlines) == 2: index = 0 #replace the 1st line with new content
    elif len(newlines) == 3: index = 1 #replace the 2nd line with new content
    
    newlines[index] = tx_far_asm

    if len(newlines) == 3 and newlines[2][-2:] == "$0":
        #get rid of the last incbin line if it is only including 0 bytes
        del newlines[2]
        #note that this has to be done after adding in the new asm
    newlines = "\n".join(line for line in newlines)
    newlines = newlines.replace("$x", "$") #where does this keep coming from??

    #signs are dumb; cluster the labels please
    if "\"needs fulfilled!\", $55" in newlines:
        newlines = "\n" + label + ": "
        line_number += 1
    if ("STRENGTH to move!" in newlines) or ("it the way it is." in newlines):
        newlines = "\n" + label + ": "
        line_number += 1
    if "@\"" in newlines and not "@@\"" in newlines:
        newlines = newlines.replace("@", "@@")

    #Char52 doesn't work yet? oh well
    newlines = newlines.replace("Char52", "$52")

    diff = generate_diff_insert(line_number, newlines)
    print "working on map_id=" + str(map_id) + " text_id=" + str(text_id)
    print diff
    apply_diff(diff)

def insert_all_tx_far_targets():
    for tx_far in tx_fars:
        map_id = tx_far[0]
        text_id = tx_far[1]
        #if map_id <= 185: continue #i'm just trying to get it going faster

        insert_tx_far(map_id, text_id, tx_far_line=tx_far)

        reset_incbins()
        analyze_incbins.reset_incbins()
        asm = None
        incbin_lines = []
        processed_incbins = {}
        analyze_incbins.asm = None
        analyze_incbins.incbin_lines = []
        analyze_incbins.processed_incbins = {}

        load_asm()
        isolate_incbins()
        process_incbins()

def all_texts_are_tx_fars(map_id):
    map2 = extract_maps.map_headers[map_id]
    for text_id in map2["texts"]:
        txt = map2["texts"][text_id]
        if not "TX_FAR" in txt[0].keys(): return False
    return True

def texts_label_pretty_printer(map_id):
    "output a texts label for map if all texts are TX_FARs and in the asm already"
    #extract_maps.map_headers[map_id]["texts"][text_id][0]["TX_FAR"]
    #if not all_texts_are_tx_fars(map_id): return None
    map2 = extract_maps.map_headers[map_id]

    #pointer to the list of texts
    texts_list_pointer = int(map2["texts_pointer"], 16)

    #get the label for this texts list
    base_label = map_name_cleaner(map2["name"], None)[:-2]
    label = base_label + "Texts"

    #make up a label for each text
    text_labels = []
    text_id = 1
    for text in map2["texts"].keys():
        text_label = base_label + "Text" + str(text_id)
        text_labels.append(text_label)
        text_id += 1
    
    output = label + ": ; " + hex(texts_list_pointer)
    output += "\n"
    output += spacing + "dw "

    first = True
    for labela in text_labels:
        if not first:
            output += ", " + labela
        else:
            output += labela
            first = False

    return output

def insert_texts_label(map_id):
    #if not all_texts_are_tx_fars(map_id): return None
    map2 = extract_maps.map_headers[map_id]
    
    base_label = map_name_cleaner(map2["name"], None)[:-2]
    label = base_label + "Texts"
    texts_pointer = int(map2["texts_pointer"], 16)

    insert_asm = texts_label_pretty_printer(map_id)

    line_number = find_incbin_to_replace_for(texts_pointer)
    if line_number == None:
        print "skipping texts label for map_id=" + str(map_id) + " texts_pointer=" + hex(texts_pointer) + " because the address is taken"
        return

    #also do a name check
    if (label + ":") in "\n".join(analyze_incbins.asm):
        print "skipping texts label for map_id=" + str(map_id) + " texts_pointer=" + hex(texts_pointer) + " because the label is already used"
        return
    
    newlines = split_incbin_line_into_three(line_number, texts_pointer, len(map2["referenced_texts"])*2 )

    newlines = newlines.split("\n")
    if len(newlines) == 2: index = 0 #replace the 1st line with new content
    elif len(newlines) == 3: index = 1 #replace the 2nd line with new content
    
    newlines[index] = insert_asm

    if len(newlines) == 3 and newlines[2][-2:] == "$0":
        #get rid of the last incbin line if it is only including 0 bytes
        del newlines[2]
        #note that this has to be done after adding in the new asm
    newlines = "\n".join(line for line in newlines)
    newlines = newlines.replace("$x", "$")

    diff = generate_diff_insert(line_number, newlines)
    print "working on map_id=" + str(map_id) + " texts_pointer=" + hex(texts_pointer)
    print diff
    apply_diff(diff)

#untested as of 2012-01-07
def insert_all_texts_labels():
    for map_id in extract_maps.map_headers.keys():
        if map_id not in extract_maps.bad_maps:
            if len(extract_maps.map_headers[map_id]["referenced_texts"]) > 0:
                insert_texts_label(map_id)
        
                reset_incbins()
                analyze_incbins.reset_incbins()
                asm = None
                incbin_lines = []
                processed_incbins = {}
                analyze_incbins.asm = None
                analyze_incbins.incbin_lines = []
                analyze_incbins.processed_incbins = {}
        
                load_asm()
                isolate_incbins()
                process_incbins()

def txt_to_tx_far_pretty_printer(address, label, target_label, include_byte=False):
    output = "\n" + label + ": ; " + hex(address) + "\n"
    output += spacing + "TX_FAR " + target_label + "\n"
    if include_byte:
        output += spacing + "db $50\n"
    return output

def insert_text_label_tx_far(map_id, text_id):
    if map_id in extract_maps.bad_maps:
        print "bad map id=" + str(map_id)
        return
    map2 = extract_maps.map_headers[map_id]
    if map2["texts"][text_id] == {0: {}}: return None

    base_label = map_name_cleaner(map2["name"], None)[:-2]
    label = base_label + "Text" + str(text_id)
    target_label = "_" + label
    start_address = map2["texts"][text_id][0]["start_address"]
    if 0x4000 <= start_address <= 0x7fff:
        start_address = extract_maps.calculate_pointer(start_address, int(map2["bank"],16))
    include_byte = False
    print map2["texts"][text_id]
    if "type" in map2["texts"][text_id][1].keys():
        if map2["texts"][text_id][1]["type"] == 0x50:
            include_byte = True
    tx_far_asm = txt_to_tx_far_pretty_printer(start_address, label, target_label, include_byte=include_byte)
    
    line_number = find_incbin_to_replace_for(start_address)
    if line_number == None:
        print "skipping text label that calls TX_FAR for map_id=" + str(map_id) + " text_id=" + str(text_id) + " because the address is taken " + hex(start_address)
        return

    #also do a name check
    if 1 < ("\n".join(analyze_incbins.asm)).count("\n" + label + ":"):
        print "skipping text label that calls TX_FAR for map_id=" + str(map_id) + " text_id" + str(text_id) + " because the label is already used (" + label + ":)"
        return
    
    extra = 0
    if include_byte: extra += 1
    newlines = split_incbin_line_into_three(line_number, start_address, 4 + extra )

    newlines = newlines.split("\n")
    if len(newlines) == 2: index = 0 #replace the 1st line with new content
    elif len(newlines) == 3: index = 1 #replace the 2nd line with new content
    
    newlines[index] = tx_far_asm

    if len(newlines) == 3 and newlines[2][-2:] == "$0":
        #get rid of the last incbin line if it is only including 0 bytes
        del newlines[2]
        #note that this has to be done after adding in the new asm
    newlines = "\n".join(line for line in newlines)

    newlines = newlines.replace("$x", "$")

    diff = generate_diff_insert(line_number, newlines)
    print "working on map_id=" + str(map_id) + " text_id=" + str(text_id)
    print diff
    apply_diff(diff)

def insert_all_text_labels():
    for map_id in extract_maps.map_headers.keys():
        if map_id <= 100: continue #skip
        if map_id not in extract_maps.bad_maps:
            for text_id in extract_maps.map_headers[map_id]["referenced_texts"]:
                insert_text_label_tx_far(map_id, text_id)
        
                reset_incbins()
                analyze_incbins.reset_incbins()
                asm = None
                incbin_lines = []
                processed_incbins = {}
                analyze_incbins.asm = None
                analyze_incbins.incbin_lines = []
                analyze_incbins.processed_incbins = {}
        
                load_asm()
                isolate_incbins()
                process_incbins()

#TODO: if line_id !=0 then don't include the label?
def insert_08_asm(map_id, text_id, line_id=0):
    map2 = extract_maps.map_headers[map_id]
    base_label = map_name_cleaner(map2["name"], None)[:-2]
    label = base_label + "Text" + str(text_id)

    start_address = all_texts[map_id][text_id][line_id]["start_address"]

    (text_asm, end_address) = text_asm_pretty_printer(label, start_address)
    print "end address is: " + hex(end_address)

    #find where to insert the assembly
    line_number = find_incbin_to_replace_for(start_address)
    if line_number == None:
        print "skipping text label for a $08 on map_id=" + str(map_id) + " text_id=" + str(text_id) + " because the address is taken"
        return

    #also do a name check
    if 1 <= ("\n".join(analyze_incbins.asm)).count("\n" + label + ":"):
        print "skipping text label for a $08 on map_id=" + str(map_id) + " text_id=" + str(text_id) + " because the label is already taken (" + label + ":)"
        return
    
    newlines = split_incbin_line_into_three(line_number, start_address, end_address - start_address )
    
    newlines = newlines.split("\n")
    if len(newlines) == 2: index = 0 #replace the 1st line with new content
    elif len(newlines) == 3: index = 1 #replace the 2nd line with new content
    
    newlines[index] = text_asm

    if len(newlines) == 3 and newlines[2][-2:] == "$0":
        #get rid of the last incbin line if it is only including 0 bytes
        del newlines[2]
        #note that this has to be done after adding in the new asm
    newlines = "\n".join(line for line in newlines)

    newlines = newlines.replace("$x", "$")

    diff = generate_diff_insert(line_number, newlines)
    print "working on map_id=" + str(map_id) + " text_id=" + str(text_id)
    print diff
    result = apply_diff(diff)

    if result == False:
        failed_attempts[len(failed_attempts.keys())] = {"map_id": map_id, "text_id": text_id}

def find_all_08s():
    all_08s = []
    for map_id in all_texts:
        for text_id in all_texts[map_id].keys():
            if 0 in all_texts[map_id][text_id].keys():
                for line_id in all_texts[map_id][text_id].keys():
                    if all_texts[map_id][text_id][line_id]["type"] == 0x8:
                        all_08s.append([map_id, text_id, line_id])
    return all_08s

def insert_all_08s():
    all_08s = find_all_08s()
    for the_08_line in all_08s:
        map_id = the_08_line[0]
        if map_id <= 86: continue #speed things up
        text_id = the_08_line[1]
        line_id = the_08_line[2]

        print "processing map_id=" + str(map_id) + " text_id=" + str(text_id)
        insert_08_asm(map_id, text_id, line_id)
        
        #reset everything
        analyze_incbins.reset_incbins()
        asm = None
        incbin_lines = []
        processed_incbins = {}
        analyze_incbins.asm = None
        analyze_incbins.incbin_lines = []
        analyze_incbins.processed_incbins = {}

        #reload
        load_asm()
        isolate_incbins()
        process_incbins()

def insert_asm(start_address, label, text_asm=None, end_address=None):
    if text_asm == None and end_address == None:
        (text_asm, end_address) = text_asm_pretty_printer(label, start_address, include_08=False)
        print "end address is: " + hex(end_address)

    #find where to insert the assembly
    line_number = find_incbin_to_replace_for(start_address)
    if line_number == None:
        print "skipping asm because the address is taken"
        return False

    #name check
    if (label + ":") in "\n".join(analyze_incbins.asm):
        print "skipping asm because the label is taken"
        return False

    newlines = split_incbin_line_into_three(line_number, start_address, end_address - start_address )
    
    newlines = newlines.split("\n")
    if len(newlines) == 2: index = 0 #replace the 1st line with new content
    elif len(newlines) == 3: index = 1 #replace the 2nd line with new content
    
    newlines[index] = text_asm

    if len(newlines) == 3 and newlines[2][-2:] == "$0":
        #get rid of the last incbin line if it is only including 0 bytes
        del newlines[2]
        #note that this has to be done after adding in the new asm
    newlines = "\n".join(line for line in newlines)

    newlines = newlines.replace("$x", "$")

    diff = generate_diff_insert(line_number, newlines)
    print diff
    result = apply_diff(diff, try_fixing=True)
    return True

def insert_text(address, label, apply=False, try_fixing=True):
    "inserts a text script (but not $8s)"
    start_address = address

    line_number = find_incbin_to_replace_for(start_address)
    if line_number == None:
        print "skipping text at " + hex(start_address) + " with address " + label
        return "skip"

    #another reason to skip is if the interval is 0
    processed_incbin = analyze_incbins.processed_incbins[line_number]
    if processed_incbin["interval"] == 0:
        print "skipping text at " + hex(start_address) + " with address " + label + " because the interval is 0"
        return "skip"

    text_asm, byte_count = text_pretty_printer_at(start_address, label)
    end_address = start_address + byte_count
    newlines = split_incbin_line_into_three(line_number, start_address, byte_count)

    newlines = newlines.split("\n")
    if len(newlines) == 2: index = 0 #replace the 1st line with new content
    elif len(newlines) == 3: index = 1 #replace the 2nd line with new content
    
    newlines[index] = text_asm

    if len(newlines) == 3 and newlines[2][-2:] == "$0":
        #get rid of the last incbin line if it is only including 0 bytes
        del newlines[2]
        #note that this has to be done after adding in the new asm
    newlines = "\n".join(line for line in newlines)
    newlines = newlines.replace("$x", "$") #where does this keep coming from??

    #Char52 doesn't work yet
    newlines = newlines.replace("Char52", "$52")

    diff = generate_diff_insert(line_number, newlines)
    print diff
    if apply:
        return apply_diff(diff, try_fixing=try_fixing)
    else: #simulate a successful insertion
        return True

#move this into another file?
def scan_for_map_scripts_pointer():
    for map_id in extract_maps.map_headers.keys(): #skip id=0 (Pallet Town) because the naming conventions are wonky
        map2 = extract_maps.map_headers[map_id]
        if map_id in extract_maps.bad_maps or map_id in [0, 39, 37, 38]: continue #skip
        script_pointer = int(map2["script_pointer"], 16)

        main_asm_output, offset, last_hl_address, last_a_address, used_3d97 = output_bank_opcodes(script_pointer)
        hl_pointer = "None"
        
        first_script_text = ""
        if last_hl_address != None and last_hl_address != "None" and used_3d97==True:
            if last_hl_address > 0x3fff:
                hl_pointer = extract_maps.calculate_pointer(last_hl_address, int(map2["bank"], 16))
            else:
                hl_pointer = last_hl_address
            byte1 = ord(extract_maps.rom[hl_pointer])
            byte2 = ord(extract_maps.rom[hl_pointer+1])
            address = byte1 + (byte2 << 8)

            if address > 0x3fff:
                first_script_pointer = extract_maps.calculate_pointer(address, int(map2["bank"], 16))
            else:
                first_script_pointer = address

            #for later output
            first_script_text = " first_script=" + hex(first_script_pointer)

            #go ahead and insert this script pointer
            insert_asm(first_script_pointer, map_name_cleaner(map2["name"], None)[:-2] + "Script0")
            
            #reset everything
            #analyze_incbins.reset_incbins()
            asm = None
            incbin_lines = []
            processed_incbins = {}
            analyze_incbins.asm = None
            analyze_incbins.incbin_lines = []
            analyze_incbins.processed_incbins = {}
    
            #reload
            load_asm()
            isolate_incbins()
            process_incbins()

            a_numbers = [0]
            last_a_id = 0
            script_pointers = [hex(first_script_pointer)]
            latest_script_pointer = first_script_pointer
            while last_a_id == (max(a_numbers)) or last_a_id==0:
                asm_output, offset, last_hl_address2, last_a_id, byte1, byte2, address = None, None, None, None, None, None, None
                asm_output, offset, last_hl_address2, last_a_id, used_3d97_2 = output_bank_opcodes(latest_script_pointer)
                
                if last_a_id == (max(a_numbers) + 1):
                    a_numbers.append(last_a_id)
                else:
                    break
                
                byte1 = ord(extract_maps.rom[hl_pointer + (2*last_a_id)])
                byte2 = ord(extract_maps.rom[hl_pointer + (2*last_a_id) + 1])
                address2 = byte1 + (byte2 << 8)
                if address2 > 0x3fff:
                    latest_script_pointer = extract_maps.calculate_pointer(address2, int(map2["bank"], 16))
                else:
                    latest_script_pointer = address2

                script_pointers.append(hex(latest_script_pointer))
                #print "latest script pointer (part 1): " + hex(address2)
                #print "latest script pointer: " + hex(latest_script_pointer)

                #go ahead and insert the asm for this script
                result = insert_asm(latest_script_pointer, map_name_cleaner(map2["name"], None)[:-2] + "Script" + str(len(script_pointers) - 1))
                
                if result:
                    #reset everything
                    #analyze_incbins.reset_incbins()
                    asm = None
                    incbin_lines = []
                    processed_incbins = {}
                    analyze_incbins.asm = None
                    analyze_incbins.incbin_lines = []
                    analyze_incbins.processed_incbins = {}
            
                    #reload
                    load_asm()
                    isolate_incbins()
                    process_incbins()

            print "map_id=" + str(map_id) + " scripts are: " + str(script_pointers)
        
        if last_hl_address == None: last_hl_address = "None"
        else: last_hl_address = hex(last_hl_address)

        if hl_pointer != None and hl_pointer != "None": hl_pointer = hex(hl_pointer)

        print "map_id=" + str(map_id) + " " + map2["name"] + " script_pointer=" + hex(script_pointer) + " script_pointers=" + hl_pointer + first_script_text
        print main_asm_output
        print "\n\n"

        #insert asm for the main script
        result = insert_asm(script_pointer, map_name_cleaner(map2["name"], None)[:-2] + "Script")
        
        if result:
            #reset everything
            #analyze_incbins.reset_incbins()
            asm = None
            incbin_lines = []
            processed_incbins = {}
            analyze_incbins.asm = None
            analyze_incbins.incbin_lines = []
            analyze_incbins.processed_incbins = {}
    
            #reload
            load_asm()
            isolate_incbins()
            process_incbins()

        #insert script pointer list asm if there's anything of value
        if hl_pointer != None and hl_pointer != "None" and used_3d97==True:
            start_address = int(hl_pointer, 16) #where to insert this list
            total_size = len(a_numbers) * 2
            
            script_label = map_name_cleaner(map2["name"], None)[:-2] + "Script"
            scripts_label = script_label  + "s"
            script_asm = scripts_label + ": ; " + hex(start_address) + "\n"
            script_asm += spacing + "dw"

            first = True
            for id in a_numbers:
                if first:
                    script_asm += " "
                    first = False
                else:
                    script_asm += ", "
                script_asm += script_label + str(id)
            script_asm += "\n" #extra newline?

            result = insert_asm(start_address, scripts_label, text_asm=script_asm, end_address=start_address + total_size)
            if result:
                #reset everything
                #analyze_incbins.reset_incbins()
                asm = None
                incbin_lines = []
                processed_incbins = {}
                analyze_incbins.asm = None
                analyze_incbins.incbin_lines = []
                analyze_incbins.processed_incbins = {}
        
                #reload
                load_asm()
                isolate_incbins()
                process_incbins()
            else:
                print "trouble inserting map script pointer list"
                print script_asm
                sys.exit(0)

def scan_rom_for_tx_fars_and_insert():
    """calls analyze_texts.scan_rom_for_tx_fars()
    looks through INCBIN'd addresses from main.asm,
    finds TX_FARs that aren't included yet.
    """
    x = 0
    address_bundles = scan_rom_for_tx_fars(printer=True)
    for address_bundle in address_bundles:
        tx_far_address = address_bundle[1]
        tx_far_target_address = address_bundle[0]
        if tx_far_address in [0xeff2]: continue #skip
        #if tx_far_address < 0x7627b: continue #because it stopped a few times for errors

        tx_far_label = "UnnamedText_%.2x" % (tx_far_address)
        tx_far_target_label = "_" + tx_far_label

        #let's also do a quick check if it might be in the file already
        if not (": ; " + hex(tx_far_address) in analyze_incbins.asm):
            print "inserting text at " + hex(tx_far_address)
            result = insert_text(tx_far_target_address, tx_far_target_label, apply=True)
        else:
            #we can't just pretend like it worked, because we don't know what label was used
            #so, figure out the label
            for line in analyze_incbins.asm_lines:
                if ": ; " + hex(tx_far_address) in line:
                    tx_far_target_label = line.split(":")[0]
            result = "skip"

        if result == True or result == None:
            local_reset_incbins()
            result2 = insert_text(tx_far_address, tx_far_label, apply=True)
            local_reset_incbins()
        elif result == "skip":
            print "skipping " + hex(tx_far_address)
        #    result2 = insert_text(tx_far_address, tx_far_label, apply=True)
        #    local_reset_incbins()

        #just skip these for now
        #if not result or not result2:
        #    sys.exit(0)

def get_mon_name(id):
    return pokemons[id]

def get_type_label(id):
    for line in elemental_types:
        if line[2] == id: return line[0]
    return None

def get_attack_label(id):
    for move in moves:
        if move[1] == id: return move[0]
    return "0" #no move

def get_pointer_target_at(address):
    rom = extract_maps.rom
    byte1 = ord(rom[address])
    byte2 = ord(rom[address+1])
    pointer = (byte1 + (byte2 << 8))
    return pointer

def get_frontsprite_label(id):
    return get_mon_name(id).title() + "PicFront"
def get_backsprite_label(id):
    return get_mon_name(id).title() + "PicBack"

def base_data_pretty_printer(id):
    """returns beautified asm for this pokemon

    uses 28 bytes

    pokedex number, base hp, base attack, base defense, base speed, base special
    type 1 (label), type 2 (label), catch rate, base experience yield
    dimensions of frontsprite (byte)
    frontsprite label pointer
    backsprite label pointer
    attacks known at level 1 (4 bytes, 4 constants)
    growth rate (byte)
    incbin - tm/hm flags (7 bytes)
    padding (0)
    """
    output = ""
    rom = extract_maps.rom
    base_address = 0x383de + (28 * (id))
    
    pokedex_number = id
    mon_name = get_mon_name(id)
    base_hp = ord(rom[base_address + 1])
    base_attack = ord(rom[base_address + 2])
    base_defense = ord(rom[base_address + 3])
    base_speed = ord(rom[base_address + 4])
    base_special = ord(rom[base_address + 5])
    
    type1_id = ord(rom[base_address + 6])
    type2_id = ord(rom[base_address + 7])
    type1 = get_type_label(type1_id)
    type2 = get_type_label(type2_id)

    catch_rate = ord(rom[base_address + 8])
    base_exp_yield = ord(rom[base_address + 9])
    frontsprite_dimensions = ord(rom[base_address + 10])

    frontsprite = get_frontsprite_label(id)
    backsprite = get_backsprite_label(id)

    #attacks known at level 0
    attack1 = get_attack_label(ord(rom[base_address + 15]))
    attack2 = get_attack_label(ord(rom[base_address + 16]))
    attack3 = get_attack_label(ord(rom[base_address + 17]))
    attack4 = get_attack_label(ord(rom[base_address + 18]))

    growth_rate = ord(rom[base_address + 19])

    incbin_start_address = base_address + 20
    incbin_end_address = base_address + 27

    output = mon_name.title() + ("BaseStats: ; 0x%.x" % (base_address)) + "\n"
    output += spacing + "db DEX_" + mon_name.upper() + " ; pokedex id\n"
    output += spacing + ("db " + str(base_hp)) + " ; base hp\n"
    output += spacing + "db " + str(base_attack) + " ; base attack\n"
    output += spacing + "db " + str(base_defense) + " ; base defense\n"
    output += spacing + "db " + str(base_speed) + " ; base speed\n"
    output += spacing + "db " + str(base_special) + " ; base special\n\n"
    output += spacing + "db " + type1 + " ; species type 1\n"
    output += spacing + "db " + type2 + " ; species type 2\n\n"
    output += spacing + "db " + str(catch_rate) + " ; catch rate\n"
    output += spacing + "db " + str(base_exp_yield) + " ; base exp yield\n"
    output += spacing + ("db $%.2x" % (frontsprite_dimensions)) + " ; sprite dimensions\n\n"
    output += spacing + "dw " + frontsprite + "\n"
    output += spacing + "dw " + backsprite + "\n"
    output += spacing + "\n" + spacing + "; attacks known at lvl 0\n"
    output += spacing + "db " + attack1 + "\n"
    output += spacing + "db " + attack2 + "\n"
    output += spacing + "db " + attack3 + "\n"
    output += spacing + "db " + attack4 + "\n\n"
    output += spacing + "db " + str(growth_rate) + " ; growth rate\n"
    output += spacing + "\n" + spacing + "; learnset\n"
    
    #learnset crap
    output += spacing + "db %" + bin(ord(rom[base_address + 20]))[2:].zfill(8) + "\n"
    output += spacing + "db %" + bin(ord(rom[base_address + 21]))[2:].zfill(8) + "\n"
    output += spacing + "db %" + bin(ord(rom[base_address + 22]))[2:].zfill(8) + "\n"
    output += spacing + "db %" + bin(ord(rom[base_address + 23]))[2:].zfill(8) + "\n"
    output += spacing + "db %" + bin(ord(rom[base_address + 24]))[2:].zfill(8) + "\n"
    output += spacing + "db %" + bin(ord(rom[base_address + 25]))[2:].zfill(8) + "\n"
    output += spacing + "db %" + bin(ord(rom[base_address + 26]))[2:].zfill(8) + "\n\n"

    output += spacing + "db 0 ; padding\n"

    return output

def insert_base_stats(id):
    insert_asm = base_data_pretty_printer(id)

    address = 0x383de + (28 * (id))
    line_number = find_incbin_to_replace_for(address)
    label = get_mon_name(id).title() + "BaseStats"
    if line_number == None:
        print "skipping, already inserted at " + hex(address)
        return

    #also do a name check
    if (label + ":") in "\n".join(analyze_incbins.asm):
        print "skipping " + label + " because it is already in use.."
        return
    
    newlines = split_incbin_line_into_three(line_number, address, 28 )

    newlines = newlines.split("\n")
    if len(newlines) == 2: index = 0 #replace the 1st line with new content
    elif len(newlines) == 3: index = 1 #replace the 2nd line with new content
    
    newlines[index] = insert_asm

    if len(newlines) == 3 and newlines[2][-2:] == "$0":
        #get rid of the last incbin line if it is only including 0 bytes
        del newlines[2]
        #note that this has to be done after adding in the new asm
    newlines = "\n".join(line for line in newlines)
    newlines = newlines.replace("$x", "$")

    diff = generate_diff_insert(line_number, newlines)
    print diff
    apply_diff(diff, try_fixing=False, do_compile=False)

def insert_all_base_stats():
    for id in range(0, 151):
        #if id < 62: continue #skip
        insert_base_stats(id)
        
        #reset everything
        reset_incbins()
        analyze_incbins.reset_incbins()
        asm = None
        incbin_lines = []
        processed_incbins = {}
        analyze_incbins.asm = None
        analyze_incbins.incbin_lines = []
        analyze_incbins.processed_incbins = {}

        #reload
        load_asm()
        isolate_incbins()
        process_incbins()

if __name__ == "__main__":
    #load map headers and object data
    extract_maps.load_rom()
    extract_maps.load_map_pointers()
    extract_maps.read_all_map_headers()
    load_labels()
    #print base_data_pretty_printer(0)
    load_asm()
    isolate_incbins()
    process_incbins()
    #insert_base_stats(1)
    insert_all_base_stats()

    #load texts (these two have different formats)
    #all_texts = pretty_map_headers.analyze_texts.analyze_texts()
    #pretty_map_headers.all_texts = all_texts
    #tx_fars = pretty_map_headers.find_all_tx_fars()

    #load incbins
    #reset_incbins()

    #scan_for_map_scripts_pointer()
    #scan_rom_for_tx_fars_and_insert()
    #insert_text(0xa586b, "_VermilionCityText14")

    #insert _ViridianCityText10
    #insert_tx_far(1, 10)

    #just me testing a pokemart sign duplicate
    #insert_tx_far(3, 14)

    #this is the big one
    #insert_all_tx_far_targets()
    
    #for map_id in extract_maps.map_headers.keys():
    #    if map_id not in extract_maps.bad_maps:
    #        if len(extract_maps.map_headers[map_id]["referenced_texts"]) > 0:
    #            texts_label_pretty_printer(map_id)
    
    #insert_texts_label(240)
    #insert_all_texts_labels()

    #insert_text_label_tx_far(240, 1)
    #insert_all_text_labels()

    #insert_08_asm(83, 1)
    #insert_all_08s()

    #insert_asm(0x1da56, "NameRaterText1")
    #insert_text_label_tx_far(91, 1)

    #insert_text(0x44276, "ViridianPokeCenterText4")
    #insert_texts_label(4)
    #insert_all_texts_labels()