File size: 30,665 Bytes
d758c99
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
import itertools
import operator

import numpy as np
import torch
from torch import nn
import torchtext

from seq2struct.models import variational_lstm
from seq2struct.models import transformer
from seq2struct.utils import batched_sequence


def clamp(value, abs_max):
    value = max(-abs_max, value)
    value = min(abs_max, value)
    return value


def get_attn_mask(seq_lengths):
    # Given seq_lengths like [3, 1, 2], this will produce
    # [[[1, 1, 1],
    #   [1, 1, 1],
    #   [1, 1, 1]],
    #  [[1, 0, 0],
    #   [0, 0, 0],
    #   [0, 0, 0]],
    #  [[1, 1, 0],
    #   [1, 1, 0],
    #   [0, 0, 0]]]
    # int(max(...)) so that it has type 'int instead of numpy.int64
    max_length, batch_size = int(max(seq_lengths)), len(seq_lengths)
    attn_mask = torch.LongTensor(batch_size, max_length, max_length).fill_(0)
    for batch_idx, seq_length in enumerate(seq_lengths):
      attn_mask[batch_idx, :seq_length, :seq_length] = 1
    return attn_mask


class LookupEmbeddings(torch.nn.Module):
    def __init__(self, device, vocab, embedder, emb_size, learnable_words=[]):
        super().__init__()
        self._device = device
        self.vocab = vocab
        self.embedder = embedder
        self.emb_size = emb_size

        self.embedding = torch.nn.Embedding(
                num_embeddings=len(self.vocab),
                embedding_dim=emb_size)
        if self.embedder:
            assert emb_size == self.embedder.dim

        # init embedding
        self.learnable_words = learnable_words
        init_embed_list = []
        for i, word in enumerate(self.vocab):
            if self.embedder.contains(word):
                init_embed_list.append( \
                   self.embedder.lookup(word))
            else:
                init_embed_list.append(self.embedding.weight[i])
        init_embed_weight = torch.stack(init_embed_list, 0)
        self.embedding.weight = nn.Parameter(init_embed_weight)

    def forward_unbatched(self, token_lists):
        # token_lists: list of list of lists
        # [batch, num descs, desc length]
        # - each list contains tokens
        # - each list corresponds to a column name, table name, etc.

        embs = []
        for tokens in token_lists:
            # token_indices shape: batch (=1) x length
            token_indices = torch.tensor(
                self.vocab.indices(tokens), device=self._device).unsqueeze(0)

            # emb shape: batch (=1) x length x word_emb_size
            emb = self.embedding(token_indices)

            # emb shape: desc length x batch (=1) x word_emb_size
            emb = emb.transpose(0, 1)
            embs.append(emb)

        # all_embs shape: sum of desc lengths x batch (=1) x word_emb_size
        all_embs = torch.cat(embs, dim=0)

        # boundaries shape: num of descs + 1
        # If desc lengths are [2, 3, 4],
        # then boundaries is [0, 2, 5, 9]
        boundaries = np.cumsum([0] + [emb.shape[0] for emb in embs])

        return all_embs, boundaries
    
    def _compute_boundaries(self, token_lists):
        # token_lists: list of list of lists
        # [batch, num descs, desc length]
        # - each list contains tokens
        # - each list corresponds to a column name, table name, etc.
        boundaries = [
            np.cumsum([0] + [len(token_list) for token_list in token_lists_for_item])
            for token_lists_for_item in token_lists]

        return boundaries
    
    def _embed_token(self, token, batch_idx):
        if token in self.learnable_words or not self.embedder.contains(token):
            return self.embedding.weight[self.vocab.index(token)]
        else:
            emb = self.embedder.lookup(token)
            return emb.to(self._device)

    def forward(self, token_lists):
        # token_lists: list of list of lists
        # [batch, num descs, desc length]
        # - each list contains tokens
        # - each list corresponds to a column name, table name, etc.
        # PackedSequencePlus, with shape: [batch, sum of desc lengths, emb_size]
        all_embs = batched_sequence.PackedSequencePlus.from_lists(
            lists=[
                [
                    token
                    for token_list in token_lists_for_item
                    for token in token_list
                ]
                for token_lists_for_item in token_lists
            ],
            item_shape=(self.emb_size,),
            device=self._device,
            item_to_tensor=self._embed_token)
        all_embs = all_embs.apply(lambda d: d.to(self._device))
        
        return all_embs, self._compute_boundaries(token_lists)

    def _embed_words_learned(self, token_lists):
        # token_lists: list of list of lists
        # [batch, num descs, desc length]
        # - each list contains tokens
        # - each list corresponds to a column name, table name, etc.

        # PackedSequencePlus, with shape: [batch, num descs * desc length (sum of desc lengths)]
        indices = batched_sequence.PackedSequencePlus.from_lists(
            lists=[
                [
                    token
                    for token_list in token_lists_for_item
                    for token in token_list
                ]
                for token_lists_for_item in token_lists
            ],
            item_shape=(1,),  # For compatibility with old PyTorch versions
            tensor_type=torch.LongTensor,
            item_to_tensor=lambda token, batch_idx, out: out.fill_(self.vocab.index(token))
        )
        indices = indices.apply(lambda d: d.to(self._device))
        # PackedSequencePlus, with shape: [batch, sum of desc lengths, emb_size]
        all_embs = indices.apply(lambda x: self.embedding(x.squeeze(-1)))

        return all_embs, self._compute_boundaries(token_lists)


class EmbLinear(torch.nn.Module):
    def __init__(self, input_size, output_size):
        super().__init__()
        self.linear = torch.nn.Linear(input_size, output_size)
    
    def forward(self, input_):
        all_embs, boundaries = input_
        all_embs = all_embs.apply(lambda d: self.linear(d))
        return all_embs, boundaries


class BiLSTM(torch.nn.Module):
    def __init__(self, input_size, output_size, dropout, summarize, use_native=False):
        # input_size: dimensionality of input
        # output_size: dimensionality of output
        # dropout
        # summarize:
        # - True: return Tensor of 1 x batch x emb size 
        # - False: return Tensor of seq len x batch x emb size 
        super().__init__()

        if use_native:
            self.lstm = torch.nn.LSTM(
                    input_size=input_size,
                    hidden_size=output_size // 2,
                    bidirectional=True,
                    dropout=dropout)
            self.dropout = torch.nn.Dropout(dropout)
        else:
            self.lstm = variational_lstm.LSTM(
                    input_size=input_size,
                    hidden_size=int(output_size // 2),
                    bidirectional=True,
                    dropout=dropout)
        self.summarize = summarize
        self.use_native = use_native

    def forward_unbatched(self, input_):
        # all_embs shape: sum of desc lengths x batch (=1) x input_size
        all_embs, boundaries = input_

        new_boundaries = [0]
        outputs = []
        for left, right in zip(boundaries, boundaries[1:]):
            # state shape:
            # - h: num_layers (=1) * num_directions (=2) x batch (=1) x recurrent_size / 2
            # - c: num_layers (=1) * num_directions (=2) x batch (=1) x recurrent_size / 2
            # output shape: seq len x batch size x output_size
            if self.use_native:
                inp = self.dropout(all_embs[left:right])
                output, (h, c) = self.lstm(inp)
            else:
                output, (h, c) = self.lstm(all_embs[left:right])
            if self.summarize:
                seq_emb = torch.cat((h[0], h[1]), dim=-1).unsqueeze(0)
                new_boundaries.append(new_boundaries[-1] + 1)
            else:
                seq_emb = output
                new_boundaries.append(new_boundaries[-1] + output.shape[0])
            outputs.append(seq_emb)

        return torch.cat(outputs, dim=0), new_boundaries

    def forward(self, input_):
        # all_embs shape: PackedSequencePlus with shape [batch, sum of desc lengths, input_size]
        # boundaries: list of lists with shape [batch, num descs + 1]
        all_embs, boundaries = input_

        # List of the following:
        # (batch_idx, desc_idx, length)
        desc_lengths = []
        batch_desc_to_flat_map = {}
        for batch_idx, boundaries_for_item in enumerate(boundaries):
            for desc_idx, (left, right) in enumerate(zip(boundaries_for_item, boundaries_for_item[1:])):
                desc_lengths.append((batch_idx, desc_idx, right - left))
                batch_desc_to_flat_map[batch_idx, desc_idx] = len(batch_desc_to_flat_map)

        # Recreate PackedSequencePlus into shape
        # [batch * num descs, desc length, input_size]
        # with name `rearranged_all_embs`
        remapped_ps_indices = []
        def rearranged_all_embs_map_index(desc_lengths_idx, seq_idx):
            batch_idx, desc_idx, _ = desc_lengths[desc_lengths_idx]
            return batch_idx, boundaries[batch_idx][desc_idx] + seq_idx
        def rearranged_all_embs_gather_from_indices(indices):
            batch_indices, seq_indices = zip(*indices)
            remapped_ps_indices[:] =  all_embs.raw_index(batch_indices, seq_indices)
            return all_embs.ps.data[torch.LongTensor(remapped_ps_indices)]
        rearranged_all_embs = batched_sequence.PackedSequencePlus.from_gather(
            lengths=[length for _, _, length in desc_lengths],
            map_index=rearranged_all_embs_map_index,
            gather_from_indices=rearranged_all_embs_gather_from_indices)
        rev_remapped_ps_indices = tuple(
            x[0] for x in sorted(
                enumerate(remapped_ps_indices), key=operator.itemgetter(1)))

        # output shape: PackedSequence, [batch * num_descs, desc length, output_size]
        # state shape:
        # - h: [num_layers (=1) * num_directions (=2), batch, output_size / 2]
        # - c: [num_layers (=1) * num_directions (=2), batch, output_size / 2]
        if self.use_native:
            rearranged_all_embs = rearranged_all_embs.apply(self.dropout)
        output, (h, c) = self.lstm(rearranged_all_embs.ps)
        if self.summarize:
            # h shape: [batch * num descs, output_size]
            h = torch.cat((h[0], h[1]), dim=-1)

            # new_all_embs: PackedSequencePlus, [batch, num descs, input_size]
            new_all_embs = batched_sequence.PackedSequencePlus.from_gather(
                lengths=[len(boundaries_for_item) - 1 for boundaries_for_item in boundaries],
                map_index=lambda batch_idx, desc_idx: rearranged_all_embs.sort_to_orig[batch_desc_to_flat_map[batch_idx, desc_idx]],
                gather_from_indices=lambda indices: h[torch.LongTensor(indices)])

            new_boundaries = [
                list(range(len(boundaries_for_item)))
                for boundaries_for_item in boundaries
            ]
        else:
            new_all_embs = all_embs.apply(
                lambda _: output.data[torch.LongTensor(rev_remapped_ps_indices)])
            new_boundaries = boundaries

        return new_all_embs, new_boundaries


class RelationalTransformerUpdate(torch.nn.Module):

    def __init__(self, device, num_layers, num_heads, hidden_size, 
            ff_size=None,
            dropout=0.1,
            merge_types=False,
            tie_layers=False,
            qq_max_dist=2,
            #qc_token_match=True,
            #qt_token_match=True,
            #cq_token_match=True,
            cc_foreign_key=True,
            cc_table_match=True,
            cc_max_dist=2,
            ct_foreign_key=True,
            ct_table_match=True,
            #tq_token_match=True,
            tc_table_match=True,
            tc_foreign_key=True,
            tt_max_dist=2,
            tt_foreign_key=True,
            sc_link=False,
            cv_link=False,
            ):
        super().__init__()
        self._device = device
        self.num_heads = num_heads

        self.qq_max_dist    = qq_max_dist
        #self.qc_token_match = qc_token_match
        #self.qt_token_match = qt_token_match
        #self.cq_token_match = cq_token_match
        self.cc_foreign_key = cc_foreign_key
        self.cc_table_match = cc_table_match
        self.cc_max_dist    = cc_max_dist
        self.ct_foreign_key = ct_foreign_key
        self.ct_table_match = ct_table_match
        #self.tq_token_match = tq_token_match
        self.tc_table_match = tc_table_match
        self.tc_foreign_key = tc_foreign_key
        self.tt_max_dist    = tt_max_dist
        self.tt_foreign_key = tt_foreign_key

        self.relation_ids = {}
        def add_relation(name):
            self.relation_ids[name] = len(self.relation_ids)
        def add_rel_dist(name, max_dist):
            for i in range(-max_dist, max_dist + 1):
                add_relation((name, i))

        add_rel_dist('qq_dist', qq_max_dist)

        add_relation('qc_default')
        #if qc_token_match:
        #    add_relation('qc_token_match')

        add_relation('qt_default')
        #if qt_token_match:
        #    add_relation('qt_token_match')

        add_relation('cq_default')
        #if cq_token_match:
        #    add_relation('cq_token_match')

        add_relation('cc_default')
        if cc_foreign_key:
            add_relation('cc_foreign_key_forward')
            add_relation('cc_foreign_key_backward')
        if cc_table_match:
            add_relation('cc_table_match')
        add_rel_dist('cc_dist', cc_max_dist)

        add_relation('ct_default')
        if ct_foreign_key:
            add_relation('ct_foreign_key')
        if ct_table_match:
            add_relation('ct_primary_key')
            add_relation('ct_table_match')
            add_relation('ct_any_table')

        add_relation('tq_default')
        #if cq_token_match:
        #    add_relation('tq_token_match')

        add_relation('tc_default')
        if tc_table_match:
            add_relation('tc_primary_key')
            add_relation('tc_table_match')
            add_relation('tc_any_table')
        if tc_foreign_key:
            add_relation('tc_foreign_key')

        add_relation('tt_default')
        if tt_foreign_key:
            add_relation('tt_foreign_key_forward')
            add_relation('tt_foreign_key_backward')
            add_relation('tt_foreign_key_both')
        add_rel_dist('tt_dist', tt_max_dist)

        # schema linking relations
        # forward_backward
        if sc_link:
            add_relation('qcCEM')
            add_relation('cqCEM')
            add_relation('qtTEM')
            add_relation('tqTEM')
            add_relation('qcCPM')
            add_relation('cqCPM')
            add_relation('qtTPM')
            add_relation('tqTPM')
        
        if cv_link:
            add_relation("qcNUMBER")
            add_relation("cqNUMBER")
            add_relation("qcTIME")
            add_relation("cqTIME")
            add_relation("qcCELLMATCH")
            add_relation("cqCELLMATCH")

        if merge_types:
            assert not cc_foreign_key
            assert not cc_table_match
            assert not ct_foreign_key
            assert not ct_table_match
            assert not tc_foreign_key
            assert not tc_table_match
            assert not tt_foreign_key

            assert cc_max_dist == qq_max_dist
            assert tt_max_dist == qq_max_dist

            add_relation('xx_default')
            self.relation_ids['qc_default'] = self.relation_ids['xx_default']
            self.relation_ids['qt_default'] = self.relation_ids['xx_default']
            self.relation_ids['cq_default'] = self.relation_ids['xx_default']
            self.relation_ids['cc_default'] = self.relation_ids['xx_default']
            self.relation_ids['ct_default'] = self.relation_ids['xx_default']
            self.relation_ids['tq_default'] = self.relation_ids['xx_default']
            self.relation_ids['tc_default'] = self.relation_ids['xx_default']
            self.relation_ids['tt_default'] = self.relation_ids['xx_default']

            if sc_link:
                self.relation_ids['qcCEM'] = self.relation_ids['xx_default']
                self.relation_ids['qcCPM'] = self.relation_ids['xx_default']
                self.relation_ids['qtTEM'] = self.relation_ids['xx_default']
                self.relation_ids['qtTPM'] = self.relation_ids['xx_default']
                self.relation_ids['cqCEM'] = self.relation_ids['xx_default']
                self.relation_ids['cqCPM'] = self.relation_ids['xx_default']
                self.relation_ids['tqTEM'] = self.relation_ids['xx_default']
                self.relation_ids['tqTPM'] = self.relation_ids['xx_default']
            if cv_link:
                self.relation_ids["qcNUMBER"] = self.relation_ids['xx_default']
                self.relation_ids["cqNUMBER"] = self.relation_ids['xx_default']
                self.relation_ids["qcTIME"] = self.relation_ids['xx_default']
                self.relation_ids["cqTIME"] = self.relation_ids['xx_default']
                self.relation_ids["qcCELLMATCH"] = self.relation_ids['xx_default']
                self.relation_ids["cqCELLMATCH"] = self.relation_ids['xx_default']

            for i in range(-qq_max_dist, qq_max_dist + 1):
                self.relation_ids['cc_dist', i] = self.relation_ids['qq_dist', i]
                self.relation_ids['tt_dist', i] = self.relation_ids['tt_dist', i]

        if ff_size is None:
            ff_size = hidden_size * 4
        self.encoder = transformer.Encoder(
            lambda: transformer.EncoderLayer(
                hidden_size, 
                transformer.MultiHeadedAttentionWithRelations(
                    num_heads,
                    hidden_size,
                    dropout),
                transformer.PositionwiseFeedForward(
                    hidden_size,
                    ff_size,
                    dropout),
                len(self.relation_ids),
                dropout),
            hidden_size,
            num_layers,
            tie_layers)
        
        self.align_attn = transformer.PointerWithRelations(hidden_size, 
            len(self.relation_ids), dropout)
    
    def create_align_mask(self, num_head, q_length, c_length, t_length):
        # mask with size num_heads * all_len * all * len
        all_length = q_length + c_length + t_length
        mask_1 = torch.ones(num_head - 1, all_length, all_length, device=self._device)
        mask_2 = torch.zeros(1, all_length, all_length, device=self._device)
        for i in range(q_length):
            for j in range(q_length, q_length + c_length):
                mask_2[0, i, j] = 1
                mask_2[0, j, i] = 1
        mask = torch.cat([mask_1, mask_2], 0)
        return mask

    
    def forward_unbatched(self, desc, q_enc, c_enc, c_boundaries, t_enc, t_boundaries):
        # enc shape: total len x batch (=1) x recurrent size
        enc = torch.cat((q_enc, c_enc, t_enc), dim=0)

        # enc shape: batch (=1) x total len x recurrent size
        enc = enc.transpose(0, 1)

        # Catalogue which things are where
        relations = self.compute_relations(
                desc,
                enc_length=enc.shape[1],
                q_enc_length=q_enc.shape[0],
                c_enc_length=c_enc.shape[0],
                c_boundaries=c_boundaries,
                t_boundaries=t_boundaries)

        relations_t = torch.LongTensor(relations).to(self._device)
        enc_new = self.encoder(enc, relations_t, mask=None)

        # Split updated_enc again
        c_base = q_enc.shape[0]
        t_base = q_enc.shape[0] + c_enc.shape[0]
        q_enc_new = enc_new[:, :c_base]
        c_enc_new = enc_new[:, c_base:t_base]
        t_enc_new = enc_new[:, t_base:]

        m2c_align_mat = self.align_attn(enc_new, enc_new[:, c_base:t_base], \
            enc_new[:, c_base:t_base], relations_t[:, c_base:t_base])
        m2t_align_mat = self.align_attn(enc_new, enc_new[:, t_base:], \
            enc_new[:, t_base:], relations_t[:, t_base:])
        return q_enc_new, c_enc_new, t_enc_new, (m2c_align_mat, m2t_align_mat)

    def forward(self, descs, q_enc, c_enc, c_boundaries, t_enc, t_boundaries):
        # TODO: Update to also compute m2c_align_mat and m2t_align_mat
        # enc: PackedSequencePlus with shape [batch, total len, recurrent size]
        enc = batched_sequence.PackedSequencePlus.cat_seqs((q_enc, c_enc, t_enc))

        q_enc_lengths = list(q_enc.orig_lengths())
        c_enc_lengths = list(c_enc.orig_lengths())
        t_enc_lengths = list(t_enc.orig_lengths())
        enc_lengths = list(enc.orig_lengths())
        max_enc_length = max(enc_lengths)

        all_relations = []
        for batch_idx, desc in enumerate(descs):
            enc_length = enc_lengths[batch_idx]
            relations_for_item = self.compute_relations(
                desc,
                enc_length,
                q_enc_lengths[batch_idx],
                c_enc_lengths[batch_idx],
                c_boundaries[batch_idx],
                t_boundaries[batch_idx])
            all_relations.append(np.pad(relations_for_item, ((0, max_enc_length - enc_length),), 'constant'))
        relations_t = torch.from_numpy(np.stack(all_relations)).to(self._device)

        # mask shape: [batch, total len, total len]
        mask = get_attn_mask(enc_lengths).to(self._device)
        # enc_new: shape [batch, total len, recurrent size]
        enc_padded, _ = enc.pad(batch_first=True)
        enc_new = self.encoder(enc_padded, relations_t, mask=mask)

        # Split enc_new again
        def gather_from_enc_new(indices):
            batch_indices, seq_indices = zip(*indices)
            return enc_new[torch.LongTensor(batch_indices), torch.LongTensor(seq_indices)]

        q_enc_new = batched_sequence.PackedSequencePlus.from_gather(
            lengths=q_enc_lengths,
            map_index=lambda batch_idx, seq_idx: (batch_idx, seq_idx),
            gather_from_indices=gather_from_enc_new)
        c_enc_new = batched_sequence.PackedSequencePlus.from_gather(
            lengths=c_enc_lengths,
            map_index=lambda batch_idx, seq_idx: (batch_idx, q_enc_lengths[batch_idx] + seq_idx),
            gather_from_indices=gather_from_enc_new)
        t_enc_new = batched_sequence.PackedSequencePlus.from_gather(
            lengths=t_enc_lengths,
            map_index=lambda batch_idx, seq_idx: (batch_idx, q_enc_lengths[batch_idx] + c_enc_lengths[batch_idx] + seq_idx),
            gather_from_indices=gather_from_enc_new)
        return q_enc_new, c_enc_new, t_enc_new

    def compute_relations(self, desc, enc_length, q_enc_length, c_enc_length, c_boundaries, t_boundaries):
        sc_link = desc.get('sc_link', {'q_col_match': {}, 'q_tab_match': {}})
        cv_link = desc.get('cv_link', {'num_date_match': {}, 'cell_match': {}}) 

        # Catalogue which things are where
        loc_types = {}
        for i in range(q_enc_length):
            loc_types[i] = ('question',)

        c_base = q_enc_length
        for c_id, (c_start, c_end) in enumerate(zip(c_boundaries, c_boundaries[1:])):
            for i in range(c_start + c_base, c_end + c_base):
                loc_types[i] = ('column', c_id)
        t_base = q_enc_length + c_enc_length
        for t_id, (t_start, t_end) in enumerate(zip(t_boundaries, t_boundaries[1:])):
            for i in range(t_start + t_base, t_end + t_base):
                loc_types[i] = ('table', t_id)
        
        relations = np.empty((enc_length, enc_length), dtype=np.int64)

        for i, j in itertools.product(range(enc_length),repeat=2):
            def set_relation(name):
                relations[i, j] = self.relation_ids[name]

            i_type, j_type = loc_types[i], loc_types[j]
            if i_type[0] == 'question':
                if j_type[0] == 'question':
                    set_relation(('qq_dist', clamp(j - i, self.qq_max_dist)))
                elif j_type[0] == 'column':
                    # set_relation('qc_default')
                    j_real = j - c_base
                    if f"{i},{j_real}" in sc_link["q_col_match"]:
                        set_relation("qc" + sc_link["q_col_match"][f"{i},{j_real}"])
                    elif f"{i},{j_real}" in cv_link["cell_match"]:
                        set_relation("qc" + cv_link["cell_match"][f"{i},{j_real}"])
                    elif f"{i},{j_real}" in cv_link["num_date_match"]:
                        set_relation("qc" + cv_link["num_date_match"][f"{i},{j_real}"])
                    else:
                        set_relation('qc_default')
                elif j_type[0] == 'table':
                    # set_relation('qt_default')
                    j_real = j - t_base
                    if f"{i},{j_real}" in sc_link["q_tab_match"]:
                        set_relation("qt" + sc_link["q_tab_match"][f"{i},{j_real}"])
                    else:
                        set_relation('qt_default')

            elif i_type[0] == 'column':
                if j_type[0] == 'question':
                    # set_relation('cq_default')
                    i_real = i - c_base
                    if f"{j},{i_real}" in sc_link["q_col_match"]:
                        set_relation("cq" + sc_link["q_col_match"][f"{j},{i_real}"])
                    elif f"{j},{i_real}" in cv_link["cell_match"]:
                        set_relation("cq" + cv_link["cell_match"][f"{j},{i_real}"])
                    elif f"{j},{i_real}" in cv_link["num_date_match"]:
                        set_relation("cq" + cv_link["num_date_match"][f"{j},{i_real}"])
                    else:
                        set_relation('cq_default')
                elif j_type[0] == 'column':
                    col1, col2 = i_type[1], j_type[1]
                    if col1 == col2:
                        set_relation(('cc_dist', clamp(j - i, self.cc_max_dist)))
                    else:
                        set_relation('cc_default')
                        if self.cc_foreign_key:
                            if desc['foreign_keys'].get(str(col1)) == col2:
                                set_relation('cc_foreign_key_forward')
                            if desc['foreign_keys'].get(str(col2)) == col1:
                                set_relation('cc_foreign_key_backward')
                        if (self.cc_table_match and 
                            desc['column_to_table'][str(col1)] == desc['column_to_table'][str(col2)]):
                            set_relation('cc_table_match')

                elif j_type[0] == 'table':
                    col, table = i_type[1], j_type[1]
                    set_relation('ct_default')
                    if self.ct_foreign_key and self.match_foreign_key(desc, col, table):
                        set_relation('ct_foreign_key')
                    if self.ct_table_match:
                        col_table = desc['column_to_table'][str(col)] 
                        if col_table == table:
                            if col in desc['primary_keys']:
                                set_relation('ct_primary_key')
                            else:
                                set_relation('ct_table_match')
                        elif col_table is None:
                            set_relation('ct_any_table')

            elif i_type[0] == 'table':
                if j_type[0] == 'question':
                    # set_relation('tq_default')
                    i_real = i - t_base
                    if f"{j},{i_real}" in sc_link["q_tab_match"]:
                        set_relation("tq" + sc_link["q_tab_match"][f"{j},{i_real}"])
                    else:
                        set_relation('tq_default')
                elif j_type[0] == 'column':
                    table, col = i_type[1], j_type[1]
                    set_relation('tc_default')

                    if self.tc_foreign_key and self.match_foreign_key(desc, col, table):
                        set_relation('tc_foreign_key')
                    if self.tc_table_match:
                        col_table = desc['column_to_table'][str(col)] 
                        if col_table == table:
                            if col in desc['primary_keys']:
                                set_relation('tc_primary_key')
                            else:
                                set_relation('tc_table_match')
                        elif col_table is None:
                            set_relation('tc_any_table')
                elif j_type[0] == 'table':
                    table1, table2 = i_type[1], j_type[1]
                    if table1 == table2:
                        set_relation(('tt_dist', clamp(j - i, self.tt_max_dist)))
                    else:
                        set_relation('tt_default')
                        if self.tt_foreign_key:
                            forward = table2 in desc['foreign_keys_tables'].get(str(table1), ())
                            backward = table1 in desc['foreign_keys_tables'].get(str(table2), ())
                            if forward and backward:
                                set_relation('tt_foreign_key_both')
                            elif forward:
                                set_relation('tt_foreign_key_forward')
                            elif backward:
                                set_relation('tt_foreign_key_backward')
        return relations

    @classmethod
    def match_foreign_key(cls, desc, col, table):
        foreign_key_for = desc['foreign_keys'].get(str(col))
        if foreign_key_for is None:
            return False

        foreign_table = desc['column_to_table'][str(foreign_key_for)]
        return desc['column_to_table'][str(col)] == foreign_table


class NoOpUpdate:
    def __init__(self, device, hidden_size):
        pass

    def __call__(self, desc, q_enc, c_enc, c_boundaries, t_enc, t_boundaries):
        #return q_enc.transpose(0, 1), c_enc.transpose(0, 1), t_enc.transpose(0, 1)
        return q_enc, c_enc, t_enc
    
    def forward_unbatched(self, desc, q_enc, c_enc, c_boundaries, t_enc, t_boundaries):
        """
        The same interface with RAT
        return: encodings with size: length * embed_size, alignment matrix
        """
        return q_enc.transpose(0,1), c_enc.transpose(0,1), t_enc.transpose(0,1), (None, None)