forked from harvardnlp/data2text
-
Notifications
You must be signed in to change notification settings - Fork 0
/
box_preprocess.lua
522 lines (427 loc) · 16.8 KB
/
box_preprocess.lua
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
require('onmt.init')
local tds = require('tds')
local path = require('pl.path')
local cjson = require('cjson')
local stringx = require('pl.stringx')
local cmd = torch.CmdLine()
cmd:text("")
cmd:text("box_preprocess.lua")
cmd:text("")
cmd:text("**Preprocess Options**")
cmd:text("")
cmd:text("")
cmd:option('-config', '', [[Read options from this file]])
cmd:option('-json_data_dir', '', [[Path to directory containing json data]])
cmd:option('-save_data', '', [[Output file for the prepared data]])
cmd:option('-src_vocab_size', 50000, [[Size of the source vocabulary]])
cmd:option('-tgt_vocab_size', 50000, [[Size of the target vocabulary]])
cmd:option('-src_vocab', '', [[Path to an existing source vocabulary]])
cmd:option('-tgt_vocab', '', [[Path to an existing target vocabulary]])
cmd:option('-features_vocabs_prefix', '', [[Path prefix to existing features vocabularies]])
cmd:option('-ptr_fi', '', [[Path to pointer file (for conditional copy attn)]])
cmd:option('-src_seq_length', 5000000, [[Maximum source sequence length]])
cmd:option('-tgt_seq_length', 5000000, [[Maximum target sequence length]])
cmd:option('-shuffle', 1, [[Shuffle data]])
cmd:option('-seed', 3435, [[Random seed]])
cmd:option('-players_per_team', 13, [[Max players per team]])
cmd:option('-report_every', 100000, [[Report status every this many sentences]])
local opt = cmd:parse(arg)
local function hasFeatures(filename)
local reader = onmt.utils.FileReader.new(filename)
local _, _, numFeatures = onmt.utils.Features.extract(reader:next())
reader:close()
return numFeatures > 0
end
local bs_keys = {"PLAYER_NAME", "START_POSITION", "MIN", "PTS", "FGM", "FGA",
"FG_PCT", "FG3M", "FG3A", "FG3_PCT", "FTM", "FTA", "FT_PCT", "OREB",
"DREB", "REB", "AST", "TO", "STL", "BLK", "PF", "FIRST_NAME",
"SECOND_NAME"}
local ls_keys = {"TEAM-PTS_QTR1", "TEAM-PTS_QTR2", "TEAM-PTS_QTR3", "TEAM-PTS_QTR4", "TEAM-PTS",
"TEAM-FG_PCT", "TEAM-FG3_PCT", "TEAM-FT_PCT", "TEAM-REB", "TEAM-AST", "TEAM-TOV", "TEAM-WINS", "TEAM-LOSSES",
"TEAM-CITY", "TEAM-NAME"}
-- this will make vocab for every word in summary or in a table cell or header
local function makeVocabulary(jsondat, size)
local wordVocab = onmt.utils.Dict.new({onmt.Constants.PAD_WORD, onmt.Constants.UNK_WORD,
onmt.Constants.BOS_WORD, onmt.Constants.EOS_WORD})
local featuresVocabs = {}
local colVocab = onmt.utils.Dict.new({onmt.Constants.PAD_WORD, onmt.Constants.UNK_WORD}) -- UNK not really necessary
for i = 1, #bs_keys do
colVocab:add(bs_keys[i])
end
for i = 1, #ls_keys do
colVocab:add(ls_keys[i])
end
local rowVocab = onmt.utils.Dict.new({onmt.Constants.PAD_WORD, onmt.Constants.UNK_WORD})
local cellVocab = onmt.utils.Dict.new({onmt.Constants.PAD_WORD, onmt.Constants.UNK_WORD})
cellVocab:add("N/A")
for i = 1, #jsondat do
local game = jsondat[i]
-- add all the words in the summary
for j = 1, #game.summary do
wordVocab:add(game.summary[j])
end
-- add all the box score poop
for t = 1, #bs_keys do
local k = bs_keys[t]
local tbl = game.box_score[k]
for idx, val in pairs(tbl) do
wordVocab:add(val)
cellVocab:add(val)
end
end
-- add all the linescore stuff
for t = 1, #ls_keys do
local k = ls_keys[t]
local v = game.home_line[k]
wordVocab:add(v)
cellVocab:add(v)
wordVocab:add(game.vis_line[k])
cellVocab:add(game.vis_line[k])
end
for k, v in pairs(game.box_score.PLAYER_NAME) do
rowVocab:add(v)
end
rowVocab:add(game.home_line["TEAM-NAME"])
rowVocab:add(game.vis_line["TEAM-NAME"])
end
local originalSize = wordVocab:size()
wordVocab = wordVocab:prune(size)
print('Created dictionary of size ' .. wordVocab:size() .. ' (pruned from ' .. originalSize .. ')')
return wordVocab, featuresVocabs, colVocab, rowVocab, cellVocab
end
local function initVocabulary(name, jsondat, vocabFile, vocabSize, featuresVocabsFiles)
local wordVocab, colVocab, rowVocab, cellVocab
local featuresVocabs = {}
if vocabFile:len() > 0 then
-- If given, load existing word dictionary.
print('Reading ' .. name .. ' vocabulary from \'' .. vocabFile .. '\'...')
wordVocab = onmt.utils.Dict.new()
wordVocab:loadFile(vocabFile)
print('Loaded ' .. wordVocab:size() .. ' ' .. name .. ' words')
end
if featuresVocabsFiles:len() > 0 then
-- If given, discover existing features dictionaries.
local j = 1
while true do
local file = featuresVocabsFiles .. '.' .. name .. '_feature_' .. j .. '.dict'
if not path.exists(file) then
break
end
print('Reading ' .. name .. ' feature ' .. j .. ' vocabulary from \'' .. file .. '\'...')
featuresVocabs[j] = onmt.utils.Dict.new()
featuresVocabs[j]:loadFile(file)
print('Loaded ' .. featuresVocabs[j]:size() .. ' labels')
j = j + 1
end
end
if wordVocab == nil or (#featuresVocabs == 0 and hasFeatures(dataFile)) then
-- If a dictionary is still missing, generate it.
print('Building ' .. name .. ' vocabulary...')
local genWordVocab, genFeaturesVocabs, genColVocab, genRowVocab, genCellVocab = makeVocabulary(jsondat, vocabSize)
if wordVocab == nil then
wordVocab = genWordVocab
colVocab = genColVocab
rowVocab = genRowVocab
cellVocab = genCellVocab
end
if #featuresVocabs == 0 then
featuresVocabs = genFeaturesVocabs
end
end
print('')
return {
words = wordVocab,
features = featuresVocabs,
cols = colVocab,
rows = rowVocab,
cells = cellVocab
}
end
local function saveVocabulary(name, vocab, file)
print('Saving ' .. name .. ' vocabulary to \'' .. file .. '\'...')
vocab:writeFile(file)
end
local function saveFeaturesVocabularies(name, vocabs, prefix)
for j = 1, #vocabs do
local file = prefix .. '.' .. name .. '_feature_' .. j .. '.dict'
print('Saving ' .. name .. ' feature ' .. j .. ' vocabulary to \'' .. file .. '\'...')
vocabs[j]:writeFile(file)
end
end
local function vecToTensor(vec)
local t = torch.Tensor(#vec)
for i, v in pairs(vec) do
t[i] = v
end
return t
end
local function get_player_idxs(game, max_per_team)
local home_players, vis_players = {}, {}
-- count total number of players
local nplayers = 0
for k,v in pairs(game.box_score['PTS']) do
nplayers = nplayers + 1
end
local num_home, num_vis = 0, 0
for i = 1, nplayers do
local player_city = game.box_score.TEAM_CITY[tostring(i-1)]
if player_city == game.home_city then
if #home_players < max_per_team then
table.insert(home_players, tostring(i-1))
num_home = num_home + 1
end
else
if #vis_players < max_per_team then
table.insert(vis_players, tostring(i-1))
num_vis = num_vis + 1
end
end
end
--print("adding", num_home, num_vis, "players")
return home_players, vis_players
end
local function makeData(jsondat, srcDicts, tgtDicts, shuffle)
local players_per_team = opt.players_per_team
-- make a src for each row
local srcs = {}
for i = 1, 2*players_per_team+2 do -- 2 teams
table.insert(srcs, tds.Vec())
end
local srcFeatures = tds.Vec()
local srcTriples = tds.Vec()
local tgt = tds.Vec()
local tgtFeatures = tds.Vec()
local sizes = tds.Vec() -- will be target sizes...
local count = 0
local ignored = 0
for i = 1, #jsondat do
local game = jsondat[i]
-- get player_idxs for each team, since there're not always 13 of each
local home_players, vis_players = get_player_idxs(game, players_per_team)
local tgtTokens = game.summary
-- row, col, val
-- leave out PLAYER_NAME, FIRST_NAME, SECOND_NAME in bs_keys, and TEAM-NAME, TEAM-CITY in ls_keys
local gameTriples = torch.IntTensor(2*players_per_team*(#bs_keys-3) + 2*(#ls_keys-2), 3):fill(1)
if #tgtTokens > 0 and #tgtTokens <= opt.tgt_seq_length then
local tgtWords = tgtTokens
local tripleIdx = 1
for ii, player_list in ipairs({home_players, vis_players}) do
for j = 1, players_per_team do
local src_j = {}
local player_key = player_list[j] -- can be nil if not enough
local playerIdx = srcDicts.rows:lookup(game.box_score.PLAYER_NAME[player_key])
if not playerIdx and not shuffle then -- validation
playerIdx = 2 -- UNK
end
assert(playerIdx or not player_key)
for k, key in ipairs(bs_keys) do
local val = game.box_score[key][player_key]
assert(val or (not player_key))
table.insert(src_j, val or "N/A")
if player_key and key ~= "PLAYER_NAME" and key ~= "FIRST_NAME" and key ~= "SECOND_NAME" then
local colIdx = srcDicts.cols:lookup(key)
assert(colIdx)
local valIdx = srcDicts.cells:lookup(val)
assert(valIdx)
gameTriples[tripleIdx][1] = playerIdx
gameTriples[tripleIdx][2] = colIdx
gameTriples[tripleIdx][3] = valIdx
tripleIdx = tripleIdx + 1
end
end
local idxs = srcDicts.words:convertToIdx(src_j, onmt.Constants.UNK_WORD)
assert(idxs:dim() > 0)
srcs[(ii-1)*players_per_team+j]:insert(idxs)
end
end
-- make line scores the same size as box scores by pre-padding
local home_src, vis_src = {}, {}
for j = 1, (#bs_keys - #ls_keys) do
table.insert(home_src, onmt.Constants.PAD_WORD)
table.insert(vis_src, onmt.Constants.PAD_WORD)
end
-- add rest of the stuff
local homeIdx = srcDicts.rows:lookup(game.home_line["TEAM-NAME"])
assert(homeIdx)
local visIdx = srcDicts.rows:lookup(game.vis_line["TEAM-NAME"])
assert(visIdx)
for k, key in ipairs(ls_keys) do
local colIdx = srcDicts.cols:lookup(key)
assert(colIdx)
table.insert(home_src, game.home_line[key])
local homeValIdx = srcDicts.cells:lookup(game.home_line[key])
assert(homeValIdx)
table.insert(vis_src, game.vis_line[key])
local visValIdx = srcDicts.cells:lookup(game.vis_line[key])
if not visValIdx and not shuffle then --Validation
visValIdx = 2
end
assert(visValIdx)
if key ~= "TEAM-NAME" and key ~= "TEAM-CITY" then
gameTriples[tripleIdx][1] = homeIdx
gameTriples[tripleIdx][2] = colIdx
gameTriples[tripleIdx][3] = homeValIdx
tripleIdx = tripleIdx + 1
gameTriples[tripleIdx][1] = visIdx
gameTriples[tripleIdx][2] = colIdx
gameTriples[tripleIdx][3] = visValIdx
tripleIdx = tripleIdx + 1
end
end
assert(#home_src == srcs[1][1]:size(1))
assert(#vis_src == srcs[1][1]:size(1))
local idxs = srcDicts.words:convertToIdx(home_src, onmt.Constants.UNK_WORD)
assert(idxs:dim() > 0)
srcs[2*players_per_team+1]:insert(idxs)
idxs = srcDicts.words:convertToIdx(vis_src, onmt.Constants.UNK_WORD)
assert(idxs:dim() > 0)
srcs[2*players_per_team+2]:insert(idxs)
srcTriples:insert(gameTriples)
--src:insert(srcDicts.words:convertToIdx(srcWords, onmt.Constants.UNK_WORD))
tgt:insert(tgtDicts.words:convertToIdx(tgtWords,
onmt.Constants.UNK_WORD,
onmt.Constants.BOS_WORD,
onmt.Constants.EOS_WORD))
if #srcDicts.features > 0 then
srcFeatures:insert(onmt.utils.Features.generateSource(srcDicts.features, srcFeats, true))
end
if #tgtDicts.features > 0 then
tgtFeatures:insert(onmt.utils.Features.generateTarget(tgtDicts.features, tgtFeats, true))
end
sizes:insert(#tgtWords)
else
ignored = ignored + 1
end
count = count + 1
if count % opt.report_every == 0 then
print('... ' .. count .. ' sentences prepared')
end
end -- end for i = 1, #jsondat
local function reorderData(perm)
tgt = onmt.utils.Table.reorder(tgt, perm, true)
for j = 1, #srcs do
srcs[j] = onmt.utils.Table.reorder(srcs[j], perm, true)
end
srcTriples = onmt.utils.Table.reorder(srcTriples, perm, true)
if opt.ptr_fi:len() > 0 then
g_ptrStuff = onmt.utils.Table.reorder(g_ptrStuff, perm, true)
end
if #srcDicts.features > 0 then
srcFeatures = onmt.utils.Table.reorder(srcFeatures, perm, true)
end
if #tgtDicts.features > 0 then
tgtFeatures = onmt.utils.Table.reorder(tgtFeatures, perm, true)
end
end
if opt.ptr_fi:len() > 0 then
assert(not shuffle or #g_ptrStuff == #srcs[1])
end
if shuffle then
print('... shuffling sentences')
local perm = torch.randperm(#tgt)
sizes = onmt.utils.Table.reorder(sizes, perm, true)
reorderData(perm)
end
if shuffle then
print('... sorting sentences by size')
local _, perm = torch.sort(vecToTensor(sizes))
reorderData(perm)
end
print('Prepared ' .. #tgt .. ' sentences (' .. ignored
.. ' ignored due to source length > ' .. opt.src_seq_length
.. ' or target length > ' .. opt.tgt_seq_length .. ')')
local srcData = {
words = srcs,
features = srcFeatures,
triples = srcTriples
}
local tgtData = {
words = tgt,
features = tgtFeatures,
pointers = shuffle and g_ptrStuff
}
return srcData, tgtData
end
local function main()
local requiredOptions = {
"json_data_dir",
"save_data"
}
onmt.utils.Opt.init(opt, requiredOptions)
local jsondir = opt.json_data_dir
if jsondir:sub(jsondir:len(), jsondir:len()) ~= '/' then
jsondir = jsondir .. '/'
end
local f = io.open(jsondir .. "train.json")
local jsondat_train = cjson.decode(f:read("*all"))
f:close()
local f = io.open(jsondir .. "valid.json")
local jsondat_valid = cjson.decode(f:read("*all"))
f:close()
local f = io.open(jsondir .. "test.json")
local jsondat_test = cjson.decode(f:read("*all"))
f:close()
if opt.ptr_fi:len() > 0 then -- mapping from target values to table values
g_ptrStuff = tds.Vec()
local fi = assert(io.open(opt.ptr_fi, "r"))
while true do
local line = fi:read()
if line == nil then
break
end
local pieces = stringx.split(line)
local lineTuples = {}
local maxTupleLen = 0
for j = 1, #pieces do
local tuple = stringx.split(pieces[j], ',')
table.insert(lineTuples, tuple)
if #tuple > maxTupleLen then
maxTupleLen = #tuple
end
end
assert(#pieces == #lineTuples)
-- put these in a tensor
local tupleTensor = torch.IntTensor(#lineTuples, maxTupleLen+1) -- last idx will have length
for j = 1, #lineTuples do
tupleTensor[j][maxTupleLen+1] = #lineTuples[j]-1 -- number of labels/ptrsrcs
tupleTensor[j][1] = tonumber(lineTuples[j][1])+1 -- make 1-indexed
for k = 2, #lineTuples[j] do
tupleTensor[j][k] = tonumber(lineTuples[j][k])+1 -- make 1-indexed
end
end
g_ptrStuff:insert(tupleTensor)
end
fi:close()
end
local data = {}
data.dicts = {}
data.dicts.src = initVocabulary('source', jsondat_train, opt.src_vocab,
opt.src_vocab_size, opt.features_vocabs_prefix)
--for k,v in pairs(data.dicts.src) do print(k) end
data.dicts.tgt = data.dicts.src
print('Preparing training data...')
data.train = {}
data.train.src, data.train.tgt = makeData(jsondat_train,
data.dicts.src, data.dicts.tgt, true)
print('')
print('Preparing validation data...')
data.valid = {}
data.valid.src, data.valid.tgt = makeData(jsondat_valid,
data.dicts.src, data.dicts.tgt, false)
data.test = {}
data.test.src, data.test.tgt = makeData(jsondat_test, data.dicts.src, data.dicts.tgt, false)
print('')
if opt.src_vocab:len() == 0 then
saveVocabulary('source', data.dicts.src.words, opt.save_data .. '.src.dict')
end
if opt.tgt_vocab:len() == 0 then
saveVocabulary('target', data.dicts.tgt.words, opt.save_data .. '.tgt.dict')
end
if opt.features_vocabs_prefix:len() == 0 then
saveFeaturesVocabularies('source', data.dicts.src.features, opt.save_data)
saveFeaturesVocabularies('target', data.dicts.tgt.features, opt.save_data)
end
print('Saving data to \'' .. opt.save_data .. '-train.t7\'...')
torch.save(opt.save_data .. '-train.t7', data, 'binary', false)
end
main()