Upload 10 files
Browse files- csvs/X_test.csv +0 -0
- csvs/X_train.csv +0 -0
- csvs/y_test.csv +694 -0
- csvs/y_train.csv +2772 -0
- dataset/task1_data.pkl +3 -0
- src/deployment_utils.py +607 -0
- src/plotting.py +230 -0
- src/preprocessing.py +591 -0
- src/style.css +94 -0
- src/utils.py +389 -0
csvs/X_test.csv
ADDED
The diff for this file is too large to render.
See raw diff
|
|
csvs/X_train.csv
ADDED
The diff for this file is too large to render.
See raw diff
|
|
csvs/y_test.csv
ADDED
@@ -0,0 +1,694 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
,winner_index
|
2 |
+
397,1
|
3 |
+
559,1
|
4 |
+
401,0
|
5 |
+
923,0
|
6 |
+
334,1
|
7 |
+
1726,0
|
8 |
+
1653,0
|
9 |
+
1905,0
|
10 |
+
1899,0
|
11 |
+
1455,0
|
12 |
+
855,1
|
13 |
+
548,1
|
14 |
+
1666,0
|
15 |
+
1116,1
|
16 |
+
344,0
|
17 |
+
1277,1
|
18 |
+
410,0
|
19 |
+
1174,1
|
20 |
+
1308,0
|
21 |
+
2188,0
|
22 |
+
2324,0
|
23 |
+
222,1
|
24 |
+
1783,0
|
25 |
+
1395,0
|
26 |
+
888,1
|
27 |
+
1275,1
|
28 |
+
999,0
|
29 |
+
978,0
|
30 |
+
175,1
|
31 |
+
94,0
|
32 |
+
542,0
|
33 |
+
341,0
|
34 |
+
1833,0
|
35 |
+
1027,0
|
36 |
+
679,1
|
37 |
+
711,1
|
38 |
+
306,0
|
39 |
+
393,1
|
40 |
+
1764,0
|
41 |
+
1204,1
|
42 |
+
1310,1
|
43 |
+
326,0
|
44 |
+
414,0
|
45 |
+
180,1
|
46 |
+
208,0
|
47 |
+
1009,0
|
48 |
+
753,1
|
49 |
+
589,0
|
50 |
+
312,0
|
51 |
+
336,0
|
52 |
+
219,0
|
53 |
+
477,0
|
54 |
+
355,0
|
55 |
+
644,0
|
56 |
+
1059,0
|
57 |
+
757,1
|
58 |
+
450,0
|
59 |
+
1279,0
|
60 |
+
1304,1
|
61 |
+
2250,0
|
62 |
+
1124,0
|
63 |
+
133,1
|
64 |
+
1031,0
|
65 |
+
1024,1
|
66 |
+
1818,0
|
67 |
+
1305,0
|
68 |
+
150,1
|
69 |
+
177,0
|
70 |
+
122,0
|
71 |
+
990,1
|
72 |
+
489,0
|
73 |
+
1692,0
|
74 |
+
1106,1
|
75 |
+
1623,0
|
76 |
+
680,1
|
77 |
+
1155,0
|
78 |
+
103,0
|
79 |
+
494,0
|
80 |
+
964,1
|
81 |
+
2338,0
|
82 |
+
603,0
|
83 |
+
1646,0
|
84 |
+
1415,0
|
85 |
+
800,0
|
86 |
+
2161,0
|
87 |
+
1159,1
|
88 |
+
584,0
|
89 |
+
1228,0
|
90 |
+
659,0
|
91 |
+
1645,0
|
92 |
+
612,0
|
93 |
+
510,1
|
94 |
+
1838,0
|
95 |
+
885,1
|
96 |
+
1374,0
|
97 |
+
621,1
|
98 |
+
83,1
|
99 |
+
541,0
|
100 |
+
844,1
|
101 |
+
2237,0
|
102 |
+
598,1
|
103 |
+
58,1
|
104 |
+
522,0
|
105 |
+
1307,0
|
106 |
+
2045,0
|
107 |
+
463,0
|
108 |
+
1048,0
|
109 |
+
1003,1
|
110 |
+
149,1
|
111 |
+
1177,0
|
112 |
+
556,1
|
113 |
+
1758,0
|
114 |
+
965,0
|
115 |
+
361,0
|
116 |
+
2179,0
|
117 |
+
2361,0
|
118 |
+
10,0
|
119 |
+
301,1
|
120 |
+
1857,0
|
121 |
+
230,0
|
122 |
+
113,0
|
123 |
+
1516,0
|
124 |
+
450,1
|
125 |
+
435,0
|
126 |
+
1738,0
|
127 |
+
697,1
|
128 |
+
1007,1
|
129 |
+
488,1
|
130 |
+
474,1
|
131 |
+
2344,0
|
132 |
+
975,0
|
133 |
+
318,1
|
134 |
+
266,0
|
135 |
+
17,1
|
136 |
+
1328,1
|
137 |
+
483,0
|
138 |
+
661,0
|
139 |
+
683,1
|
140 |
+
27,0
|
141 |
+
2317,0
|
142 |
+
1605,0
|
143 |
+
142,0
|
144 |
+
666,0
|
145 |
+
2038,0
|
146 |
+
592,0
|
147 |
+
648,0
|
148 |
+
1409,0
|
149 |
+
634,1
|
150 |
+
359,1
|
151 |
+
549,0
|
152 |
+
626,1
|
153 |
+
1824,0
|
154 |
+
889,0
|
155 |
+
746,1
|
156 |
+
62,1
|
157 |
+
97,1
|
158 |
+
2352,0
|
159 |
+
391,1
|
160 |
+
1503,0
|
161 |
+
1517,0
|
162 |
+
886,0
|
163 |
+
2068,0
|
164 |
+
828,1
|
165 |
+
1377,0
|
166 |
+
145,1
|
167 |
+
1449,0
|
168 |
+
1802,0
|
169 |
+
167,0
|
170 |
+
537,1
|
171 |
+
843,0
|
172 |
+
825,1
|
173 |
+
1909,0
|
174 |
+
1250,1
|
175 |
+
1797,0
|
176 |
+
536,1
|
177 |
+
91,1
|
178 |
+
1106,0
|
179 |
+
996,1
|
180 |
+
840,0
|
181 |
+
716,1
|
182 |
+
1198,0
|
183 |
+
227,0
|
184 |
+
1386,0
|
185 |
+
801,1
|
186 |
+
430,0
|
187 |
+
1695,0
|
188 |
+
328,1
|
189 |
+
1337,1
|
190 |
+
849,0
|
191 |
+
1220,1
|
192 |
+
1815,0
|
193 |
+
913,1
|
194 |
+
650,1
|
195 |
+
1072,0
|
196 |
+
1023,0
|
197 |
+
636,1
|
198 |
+
106,0
|
199 |
+
473,0
|
200 |
+
660,1
|
201 |
+
1398,0
|
202 |
+
510,0
|
203 |
+
181,0
|
204 |
+
882,0
|
205 |
+
1385,0
|
206 |
+
1298,1
|
207 |
+
796,0
|
208 |
+
387,0
|
209 |
+
431,1
|
210 |
+
1271,1
|
211 |
+
1332,1
|
212 |
+
639,0
|
213 |
+
472,0
|
214 |
+
545,1
|
215 |
+
65,1
|
216 |
+
2232,0
|
217 |
+
397,0
|
218 |
+
214,1
|
219 |
+
688,1
|
220 |
+
356,0
|
221 |
+
405,1
|
222 |
+
1348,1
|
223 |
+
18,1
|
224 |
+
1830,0
|
225 |
+
739,1
|
226 |
+
1788,0
|
227 |
+
1435,0
|
228 |
+
544,0
|
229 |
+
1750,0
|
230 |
+
663,1
|
231 |
+
892,0
|
232 |
+
476,1
|
233 |
+
770,0
|
234 |
+
2249,0
|
235 |
+
2139,0
|
236 |
+
1123,1
|
237 |
+
1326,1
|
238 |
+
1107,0
|
239 |
+
1349,1
|
240 |
+
816,0
|
241 |
+
265,0
|
242 |
+
1640,0
|
243 |
+
1234,0
|
244 |
+
553,1
|
245 |
+
1933,0
|
246 |
+
1921,0
|
247 |
+
909,1
|
248 |
+
691,0
|
249 |
+
772,0
|
250 |
+
872,0
|
251 |
+
675,1
|
252 |
+
1160,1
|
253 |
+
585,0
|
254 |
+
1664,0
|
255 |
+
150,0
|
256 |
+
1053,1
|
257 |
+
700,1
|
258 |
+
617,0
|
259 |
+
1892,0
|
260 |
+
1249,1
|
261 |
+
205,0
|
262 |
+
470,0
|
263 |
+
348,1
|
264 |
+
582,1
|
265 |
+
1002,0
|
266 |
+
1554,0
|
267 |
+
290,0
|
268 |
+
502,1
|
269 |
+
130,0
|
270 |
+
2046,0
|
271 |
+
471,1
|
272 |
+
635,0
|
273 |
+
987,0
|
274 |
+
493,0
|
275 |
+
1242,1
|
276 |
+
38,1
|
277 |
+
910,1
|
278 |
+
1609,0
|
279 |
+
85,1
|
280 |
+
760,1
|
281 |
+
244,0
|
282 |
+
1762,0
|
283 |
+
622,0
|
284 |
+
420,0
|
285 |
+
420,1
|
286 |
+
1212,0
|
287 |
+
2111,0
|
288 |
+
1082,0
|
289 |
+
1059,1
|
290 |
+
1081,1
|
291 |
+
533,1
|
292 |
+
2365,0
|
293 |
+
466,0
|
294 |
+
898,1
|
295 |
+
705,1
|
296 |
+
631,1
|
297 |
+
1300,1
|
298 |
+
1443,0
|
299 |
+
800,1
|
300 |
+
984,0
|
301 |
+
1615,0
|
302 |
+
1260,0
|
303 |
+
1672,0
|
304 |
+
1636,0
|
305 |
+
582,0
|
306 |
+
1330,1
|
307 |
+
61,0
|
308 |
+
1010,1
|
309 |
+
2299,0
|
310 |
+
1796,0
|
311 |
+
907,1
|
312 |
+
586,1
|
313 |
+
1032,1
|
314 |
+
1613,0
|
315 |
+
950,0
|
316 |
+
249,1
|
317 |
+
1948,0
|
318 |
+
132,1
|
319 |
+
1011,0
|
320 |
+
1224,1
|
321 |
+
1352,0
|
322 |
+
308,1
|
323 |
+
1751,0
|
324 |
+
724,0
|
325 |
+
766,1
|
326 |
+
1316,0
|
327 |
+
821,0
|
328 |
+
721,0
|
329 |
+
982,1
|
330 |
+
941,1
|
331 |
+
1025,0
|
332 |
+
1603,0
|
333 |
+
1014,1
|
334 |
+
989,0
|
335 |
+
123,0
|
336 |
+
2222,0
|
337 |
+
390,1
|
338 |
+
565,1
|
339 |
+
293,0
|
340 |
+
333,1
|
341 |
+
2115,0
|
342 |
+
216,0
|
343 |
+
778,1
|
344 |
+
1073,1
|
345 |
+
812,0
|
346 |
+
1264,0
|
347 |
+
758,1
|
348 |
+
1223,0
|
349 |
+
1877,0
|
350 |
+
577,1
|
351 |
+
104,1
|
352 |
+
91,0
|
353 |
+
1233,0
|
354 |
+
245,1
|
355 |
+
1262,1
|
356 |
+
1188,1
|
357 |
+
1208,1
|
358 |
+
956,1
|
359 |
+
1644,0
|
360 |
+
802,0
|
361 |
+
1383,0
|
362 |
+
1649,0
|
363 |
+
780,0
|
364 |
+
1050,0
|
365 |
+
1343,1
|
366 |
+
1424,0
|
367 |
+
116,1
|
368 |
+
1835,0
|
369 |
+
191,1
|
370 |
+
475,0
|
371 |
+
1224,0
|
372 |
+
1820,0
|
373 |
+
23,0
|
374 |
+
1626,0
|
375 |
+
2274,0
|
376 |
+
381,0
|
377 |
+
757,0
|
378 |
+
752,0
|
379 |
+
335,0
|
380 |
+
475,1
|
381 |
+
608,1
|
382 |
+
895,1
|
383 |
+
1241,1
|
384 |
+
192,0
|
385 |
+
925,0
|
386 |
+
696,1
|
387 |
+
764,1
|
388 |
+
432,1
|
389 |
+
899,1
|
390 |
+
1154,0
|
391 |
+
677,0
|
392 |
+
507,1
|
393 |
+
2180,0
|
394 |
+
1390,0
|
395 |
+
1648,0
|
396 |
+
335,1
|
397 |
+
24,0
|
398 |
+
1722,0
|
399 |
+
600,1
|
400 |
+
295,0
|
401 |
+
1437,0
|
402 |
+
283,0
|
403 |
+
592,1
|
404 |
+
1963,0
|
405 |
+
546,1
|
406 |
+
2112,0
|
407 |
+
942,1
|
408 |
+
543,0
|
409 |
+
2351,0
|
410 |
+
60,1
|
411 |
+
924,1
|
412 |
+
195,0
|
413 |
+
1075,0
|
414 |
+
404,1
|
415 |
+
161,0
|
416 |
+
1915,0
|
417 |
+
1241,0
|
418 |
+
2138,0
|
419 |
+
39,1
|
420 |
+
376,0
|
421 |
+
462,1
|
422 |
+
520,1
|
423 |
+
166,1
|
424 |
+
1199,0
|
425 |
+
1902,0
|
426 |
+
918,0
|
427 |
+
1425,0
|
428 |
+
1019,1
|
429 |
+
570,1
|
430 |
+
347,0
|
431 |
+
842,1
|
432 |
+
751,0
|
433 |
+
1885,0
|
434 |
+
664,0
|
435 |
+
1454,0
|
436 |
+
707,0
|
437 |
+
683,0
|
438 |
+
1289,1
|
439 |
+
47,1
|
440 |
+
198,0
|
441 |
+
44,0
|
442 |
+
1427,0
|
443 |
+
1043,0
|
444 |
+
24,1
|
445 |
+
973,0
|
446 |
+
1684,0
|
447 |
+
828,0
|
448 |
+
197,0
|
449 |
+
1812,0
|
450 |
+
2193,0
|
451 |
+
1366,0
|
452 |
+
623,0
|
453 |
+
1141,1
|
454 |
+
995,0
|
455 |
+
2113,0
|
456 |
+
1157,1
|
457 |
+
784,0
|
458 |
+
1755,0
|
459 |
+
444,1
|
460 |
+
2205,0
|
461 |
+
723,1
|
462 |
+
916,0
|
463 |
+
1391,0
|
464 |
+
417,0
|
465 |
+
377,0
|
466 |
+
550,1
|
467 |
+
1048,1
|
468 |
+
261,1
|
469 |
+
1679,0
|
470 |
+
197,1
|
471 |
+
1209,1
|
472 |
+
86,0
|
473 |
+
1102,1
|
474 |
+
94,1
|
475 |
+
110,0
|
476 |
+
144,0
|
477 |
+
803,0
|
478 |
+
193,1
|
479 |
+
337,0
|
480 |
+
316,0
|
481 |
+
1924,0
|
482 |
+
1290,1
|
483 |
+
370,1
|
484 |
+
853,0
|
485 |
+
933,1
|
486 |
+
1898,0
|
487 |
+
265,1
|
488 |
+
662,1
|
489 |
+
839,1
|
490 |
+
1451,0
|
491 |
+
671,0
|
492 |
+
552,0
|
493 |
+
1506,0
|
494 |
+
1715,0
|
495 |
+
110,1
|
496 |
+
1988,0
|
497 |
+
2320,0
|
498 |
+
945,0
|
499 |
+
1232,0
|
500 |
+
681,0
|
501 |
+
1292,0
|
502 |
+
2006,0
|
503 |
+
643,1
|
504 |
+
1638,0
|
505 |
+
1096,1
|
506 |
+
917,1
|
507 |
+
2177,0
|
508 |
+
1011,1
|
509 |
+
355,1
|
510 |
+
1,1
|
511 |
+
1057,0
|
512 |
+
418,0
|
513 |
+
1149,0
|
514 |
+
563,0
|
515 |
+
899,0
|
516 |
+
1151,1
|
517 |
+
799,1
|
518 |
+
462,0
|
519 |
+
327,1
|
520 |
+
1142,1
|
521 |
+
1358,0
|
522 |
+
948,0
|
523 |
+
1064,1
|
524 |
+
131,0
|
525 |
+
1279,1
|
526 |
+
1564,0
|
527 |
+
112,1
|
528 |
+
1005,0
|
529 |
+
1682,0
|
530 |
+
918,1
|
531 |
+
793,0
|
532 |
+
32,1
|
533 |
+
647,1
|
534 |
+
424,1
|
535 |
+
1070,0
|
536 |
+
1996,0
|
537 |
+
1159,0
|
538 |
+
275,1
|
539 |
+
490,1
|
540 |
+
2163,0
|
541 |
+
736,1
|
542 |
+
973,1
|
543 |
+
2321,0
|
544 |
+
601,1
|
545 |
+
901,0
|
546 |
+
2085,0
|
547 |
+
293,1
|
548 |
+
1459,0
|
549 |
+
1244,1
|
550 |
+
905,0
|
551 |
+
935,1
|
552 |
+
887,1
|
553 |
+
95,1
|
554 |
+
108,1
|
555 |
+
1044,0
|
556 |
+
479,0
|
557 |
+
370,0
|
558 |
+
1126,0
|
559 |
+
255,1
|
560 |
+
1254,0
|
561 |
+
169,1
|
562 |
+
28,0
|
563 |
+
216,1
|
564 |
+
863,0
|
565 |
+
1965,0
|
566 |
+
581,0
|
567 |
+
486,0
|
568 |
+
163,0
|
569 |
+
1038,0
|
570 |
+
2042,0
|
571 |
+
130,1
|
572 |
+
1229,1
|
573 |
+
1851,0
|
574 |
+
891,1
|
575 |
+
761,1
|
576 |
+
136,1
|
577 |
+
868,1
|
578 |
+
1195,1
|
579 |
+
1978,0
|
580 |
+
497,0
|
581 |
+
817,0
|
582 |
+
517,0
|
583 |
+
262,0
|
584 |
+
1147,0
|
585 |
+
446,0
|
586 |
+
964,0
|
587 |
+
314,0
|
588 |
+
1181,1
|
589 |
+
250,1
|
590 |
+
1207,0
|
591 |
+
823,1
|
592 |
+
187,1
|
593 |
+
726,1
|
594 |
+
886,1
|
595 |
+
1180,0
|
596 |
+
542,1
|
597 |
+
1122,1
|
598 |
+
228,1
|
599 |
+
1346,1
|
600 |
+
2013,0
|
601 |
+
569,0
|
602 |
+
336,1
|
603 |
+
2167,0
|
604 |
+
1987,0
|
605 |
+
354,1
|
606 |
+
607,1
|
607 |
+
550,0
|
608 |
+
2131,0
|
609 |
+
678,0
|
610 |
+
2063,0
|
611 |
+
170,0
|
612 |
+
1201,1
|
613 |
+
332,0
|
614 |
+
806,1
|
615 |
+
1476,0
|
616 |
+
195,1
|
617 |
+
818,0
|
618 |
+
1276,1
|
619 |
+
2175,0
|
620 |
+
622,1
|
621 |
+
141,1
|
622 |
+
140,0
|
623 |
+
549,1
|
624 |
+
576,0
|
625 |
+
30,0
|
626 |
+
718,0
|
627 |
+
1577,0
|
628 |
+
1100,0
|
629 |
+
1592,0
|
630 |
+
2101,0
|
631 |
+
2089,0
|
632 |
+
753,0
|
633 |
+
921,0
|
634 |
+
1145,0
|
635 |
+
612,1
|
636 |
+
2062,0
|
637 |
+
1217,0
|
638 |
+
534,0
|
639 |
+
400,1
|
640 |
+
766,0
|
641 |
+
491,0
|
642 |
+
1128,1
|
643 |
+
1512,0
|
644 |
+
1163,1
|
645 |
+
485,1
|
646 |
+
1022,0
|
647 |
+
70,1
|
648 |
+
1373,0
|
649 |
+
1060,0
|
650 |
+
867,1
|
651 |
+
893,0
|
652 |
+
2255,0
|
653 |
+
776,0
|
654 |
+
1826,0
|
655 |
+
271,1
|
656 |
+
2316,0
|
657 |
+
1556,0
|
658 |
+
878,0
|
659 |
+
2066,0
|
660 |
+
1148,1
|
661 |
+
1662,0
|
662 |
+
174,1
|
663 |
+
1051,1
|
664 |
+
548,0
|
665 |
+
455,0
|
666 |
+
628,0
|
667 |
+
685,1
|
668 |
+
833,1
|
669 |
+
730,1
|
670 |
+
2226,0
|
671 |
+
433,1
|
672 |
+
992,1
|
673 |
+
1055,1
|
674 |
+
1487,0
|
675 |
+
1321,0
|
676 |
+
1823,0
|
677 |
+
1678,0
|
678 |
+
1509,0
|
679 |
+
131,1
|
680 |
+
323,1
|
681 |
+
2069,0
|
682 |
+
1282,0
|
683 |
+
457,0
|
684 |
+
1008,0
|
685 |
+
129,0
|
686 |
+
1121,0
|
687 |
+
42,1
|
688 |
+
203,0
|
689 |
+
156,1
|
690 |
+
1054,0
|
691 |
+
539,1
|
692 |
+
1119,0
|
693 |
+
732,0
|
694 |
+
1819,0
|
csvs/y_train.csv
ADDED
@@ -0,0 +1,2772 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
,winner_index
|
2 |
+
182,0
|
3 |
+
1245,0
|
4 |
+
880,0
|
5 |
+
170,1
|
6 |
+
1171,1
|
7 |
+
940,0
|
8 |
+
16,0
|
9 |
+
605,0
|
10 |
+
2256,0
|
11 |
+
2367,0
|
12 |
+
1744,0
|
13 |
+
1560,0
|
14 |
+
985,1
|
15 |
+
1945,0
|
16 |
+
926,1
|
17 |
+
595,0
|
18 |
+
346,1
|
19 |
+
855,0
|
20 |
+
102,0
|
21 |
+
1472,0
|
22 |
+
488,0
|
23 |
+
786,0
|
24 |
+
2105,0
|
25 |
+
1514,0
|
26 |
+
380,1
|
27 |
+
2130,0
|
28 |
+
432,0
|
29 |
+
1273,0
|
30 |
+
852,0
|
31 |
+
386,1
|
32 |
+
272,1
|
33 |
+
181,1
|
34 |
+
392,1
|
35 |
+
851,1
|
36 |
+
2238,0
|
37 |
+
427,0
|
38 |
+
1119,1
|
39 |
+
572,1
|
40 |
+
806,0
|
41 |
+
530,1
|
42 |
+
405,0
|
43 |
+
1416,0
|
44 |
+
241,1
|
45 |
+
1031,1
|
46 |
+
2363,0
|
47 |
+
689,0
|
48 |
+
304,1
|
49 |
+
1005,1
|
50 |
+
25,0
|
51 |
+
2332,0
|
52 |
+
896,0
|
53 |
+
179,1
|
54 |
+
1242,0
|
55 |
+
2372,0
|
56 |
+
1711,0
|
57 |
+
1128,0
|
58 |
+
344,1
|
59 |
+
715,1
|
60 |
+
1521,0
|
61 |
+
2356,0
|
62 |
+
2257,0
|
63 |
+
1298,0
|
64 |
+
1290,0
|
65 |
+
175,0
|
66 |
+
456,1
|
67 |
+
495,1
|
68 |
+
613,0
|
69 |
+
1326,0
|
70 |
+
183,0
|
71 |
+
627,1
|
72 |
+
2199,0
|
73 |
+
606,0
|
74 |
+
1445,0
|
75 |
+
77,1
|
76 |
+
1544,0
|
77 |
+
1110,1
|
78 |
+
1862,0
|
79 |
+
1716,0
|
80 |
+
378,0
|
81 |
+
1082,1
|
82 |
+
1158,0
|
83 |
+
1094,1
|
84 |
+
1868,0
|
85 |
+
937,1
|
86 |
+
2306,0
|
87 |
+
890,1
|
88 |
+
426,0
|
89 |
+
1276,0
|
90 |
+
1072,1
|
91 |
+
2296,0
|
92 |
+
1074,1
|
93 |
+
1701,0
|
94 |
+
211,1
|
95 |
+
93,0
|
96 |
+
604,0
|
97 |
+
111,1
|
98 |
+
1342,1
|
99 |
+
48,1
|
100 |
+
690,1
|
101 |
+
1177,1
|
102 |
+
254,1
|
103 |
+
1798,0
|
104 |
+
2230,0
|
105 |
+
619,0
|
106 |
+
545,0
|
107 |
+
1479,0
|
108 |
+
1185,0
|
109 |
+
771,1
|
110 |
+
1175,0
|
111 |
+
1651,0
|
112 |
+
214,0
|
113 |
+
1907,0
|
114 |
+
2044,0
|
115 |
+
1465,0
|
116 |
+
812,1
|
117 |
+
1077,1
|
118 |
+
791,0
|
119 |
+
200,1
|
120 |
+
847,0
|
121 |
+
1339,1
|
122 |
+
1101,1
|
123 |
+
192,1
|
124 |
+
1872,0
|
125 |
+
429,0
|
126 |
+
271,0
|
127 |
+
2135,0
|
128 |
+
1187,0
|
129 |
+
1247,0
|
130 |
+
299,0
|
131 |
+
976,1
|
132 |
+
831,0
|
133 |
+
652,1
|
134 |
+
865,1
|
135 |
+
190,1
|
136 |
+
515,1
|
137 |
+
371,1
|
138 |
+
2005,0
|
139 |
+
632,1
|
140 |
+
1317,1
|
141 |
+
425,1
|
142 |
+
76,0
|
143 |
+
299,1
|
144 |
+
1036,0
|
145 |
+
1084,0
|
146 |
+
281,0
|
147 |
+
783,1
|
148 |
+
1207,1
|
149 |
+
1220,0
|
150 |
+
237,0
|
151 |
+
762,0
|
152 |
+
1922,0
|
153 |
+
670,1
|
154 |
+
1853,0
|
155 |
+
126,1
|
156 |
+
2221,0
|
157 |
+
229,1
|
158 |
+
1080,1
|
159 |
+
2265,0
|
160 |
+
813,1
|
161 |
+
1211,1
|
162 |
+
1113,0
|
163 |
+
1105,0
|
164 |
+
303,1
|
165 |
+
834,1
|
166 |
+
1328,0
|
167 |
+
748,0
|
168 |
+
151,1
|
169 |
+
1315,1
|
170 |
+
47,0
|
171 |
+
290,1
|
172 |
+
229,0
|
173 |
+
672,0
|
174 |
+
714,1
|
175 |
+
1375,0
|
176 |
+
1297,1
|
177 |
+
623,1
|
178 |
+
352,1
|
179 |
+
460,1
|
180 |
+
1144,1
|
181 |
+
1143,0
|
182 |
+
566,0
|
183 |
+
539,0
|
184 |
+
1295,1
|
185 |
+
1655,0
|
186 |
+
939,1
|
187 |
+
1482,0
|
188 |
+
9,1
|
189 |
+
374,0
|
190 |
+
1076,1
|
191 |
+
1208,0
|
192 |
+
1357,0
|
193 |
+
1389,0
|
194 |
+
88,0
|
195 |
+
870,1
|
196 |
+
2247,0
|
197 |
+
698,1
|
198 |
+
903,1
|
199 |
+
915,0
|
200 |
+
1874,0
|
201 |
+
1292,1
|
202 |
+
59,0
|
203 |
+
526,0
|
204 |
+
1263,0
|
205 |
+
638,1
|
206 |
+
669,1
|
207 |
+
1026,1
|
208 |
+
890,0
|
209 |
+
2110,0
|
210 |
+
668,0
|
211 |
+
1941,0
|
212 |
+
1105,1
|
213 |
+
738,1
|
214 |
+
2092,0
|
215 |
+
931,0
|
216 |
+
1401,0
|
217 |
+
1345,0
|
218 |
+
929,0
|
219 |
+
933,0
|
220 |
+
2165,0
|
221 |
+
781,1
|
222 |
+
136,0
|
223 |
+
11,0
|
224 |
+
2362,0
|
225 |
+
1500,0
|
226 |
+
802,1
|
227 |
+
1402,0
|
228 |
+
407,1
|
229 |
+
445,0
|
230 |
+
1189,1
|
231 |
+
1675,0
|
232 |
+
848,0
|
233 |
+
1203,0
|
234 |
+
785,0
|
235 |
+
1066,0
|
236 |
+
749,0
|
237 |
+
1318,0
|
238 |
+
980,1
|
239 |
+
1234,1
|
240 |
+
698,0
|
241 |
+
2326,0
|
242 |
+
1502,0
|
243 |
+
505,0
|
244 |
+
1840,0
|
245 |
+
2,1
|
246 |
+
844,0
|
247 |
+
167,1
|
248 |
+
528,1
|
249 |
+
11,1
|
250 |
+
762,1
|
251 |
+
805,0
|
252 |
+
1747,0
|
253 |
+
430,1
|
254 |
+
411,1
|
255 |
+
2169,0
|
256 |
+
113,1
|
257 |
+
219,1
|
258 |
+
1404,0
|
259 |
+
647,0
|
260 |
+
657,0
|
261 |
+
798,0
|
262 |
+
1100,1
|
263 |
+
620,1
|
264 |
+
1121,1
|
265 |
+
2305,0
|
266 |
+
1896,0
|
267 |
+
1746,0
|
268 |
+
1710,0
|
269 |
+
1569,0
|
270 |
+
740,0
|
271 |
+
1227,1
|
272 |
+
1089,1
|
273 |
+
525,1
|
274 |
+
55,1
|
275 |
+
630,0
|
276 |
+
562,0
|
277 |
+
1299,0
|
278 |
+
1611,0
|
279 |
+
747,1
|
280 |
+
36,0
|
281 |
+
2330,0
|
282 |
+
653,1
|
283 |
+
2129,0
|
284 |
+
765,0
|
285 |
+
56,1
|
286 |
+
991,1
|
287 |
+
452,1
|
288 |
+
1221,1
|
289 |
+
362,0
|
290 |
+
1083,1
|
291 |
+
460,0
|
292 |
+
1900,0
|
293 |
+
2301,0
|
294 |
+
1414,0
|
295 |
+
1017,1
|
296 |
+
2314,0
|
297 |
+
308,0
|
298 |
+
1870,0
|
299 |
+
316,1
|
300 |
+
518,0
|
301 |
+
349,0
|
302 |
+
1136,0
|
303 |
+
217,1
|
304 |
+
2244,0
|
305 |
+
614,1
|
306 |
+
1844,0
|
307 |
+
1299,1
|
308 |
+
343,1
|
309 |
+
959,0
|
310 |
+
1541,0
|
311 |
+
930,1
|
312 |
+
1343,0
|
313 |
+
120,1
|
314 |
+
2280,0
|
315 |
+
83,0
|
316 |
+
296,1
|
317 |
+
841,0
|
318 |
+
7,1
|
319 |
+
365,0
|
320 |
+
21,0
|
321 |
+
1205,1
|
322 |
+
1096,0
|
323 |
+
115,0
|
324 |
+
936,1
|
325 |
+
87,0
|
326 |
+
2026,0
|
327 |
+
442,0
|
328 |
+
160,1
|
329 |
+
307,1
|
330 |
+
1156,0
|
331 |
+
1773,0
|
332 |
+
534,1
|
333 |
+
570,0
|
334 |
+
599,0
|
335 |
+
803,1
|
336 |
+
1494,0
|
337 |
+
578,0
|
338 |
+
883,0
|
339 |
+
950,1
|
340 |
+
2036,0
|
341 |
+
203,1
|
342 |
+
2100,0
|
343 |
+
13,1
|
344 |
+
1702,0
|
345 |
+
1230,0
|
346 |
+
1349,0
|
347 |
+
274,0
|
348 |
+
673,1
|
349 |
+
569,1
|
350 |
+
248,0
|
351 |
+
2349,0
|
352 |
+
69,1
|
353 |
+
1432,0
|
354 |
+
459,0
|
355 |
+
836,1
|
356 |
+
385,0
|
357 |
+
1323,1
|
358 |
+
1302,0
|
359 |
+
1951,0
|
360 |
+
1341,1
|
361 |
+
291,1
|
362 |
+
1335,1
|
363 |
+
161,1
|
364 |
+
2166,0
|
365 |
+
1114,1
|
366 |
+
252,1
|
367 |
+
423,0
|
368 |
+
1452,0
|
369 |
+
1186,0
|
370 |
+
583,1
|
371 |
+
1889,0
|
372 |
+
579,1
|
373 |
+
324,0
|
374 |
+
139,1
|
375 |
+
1691,0
|
376 |
+
1034,1
|
377 |
+
342,1
|
378 |
+
374,1
|
379 |
+
790,1
|
380 |
+
147,1
|
381 |
+
748,1
|
382 |
+
600,0
|
383 |
+
1098,1
|
384 |
+
1768,0
|
385 |
+
317,0
|
386 |
+
81,0
|
387 |
+
1340,0
|
388 |
+
472,1
|
389 |
+
934,1
|
390 |
+
822,0
|
391 |
+
1331,1
|
392 |
+
1035,1
|
393 |
+
1319,0
|
394 |
+
1243,0
|
395 |
+
1260,1
|
396 |
+
0,1
|
397 |
+
1674,0
|
398 |
+
257,1
|
399 |
+
2334,0
|
400 |
+
1338,1
|
401 |
+
1405,0
|
402 |
+
1928,0
|
403 |
+
1620,0
|
404 |
+
715,0
|
405 |
+
2182,0
|
406 |
+
1595,0
|
407 |
+
1216,1
|
408 |
+
453,1
|
409 |
+
1308,1
|
410 |
+
624,0
|
411 |
+
1051,0
|
412 |
+
71,0
|
413 |
+
22,1
|
414 |
+
1527,0
|
415 |
+
1867,0
|
416 |
+
282,0
|
417 |
+
338,1
|
418 |
+
320,1
|
419 |
+
10,1
|
420 |
+
41,1
|
421 |
+
961,1
|
422 |
+
1728,0
|
423 |
+
561,0
|
424 |
+
744,1
|
425 |
+
535,0
|
426 |
+
1047,1
|
427 |
+
575,1
|
428 |
+
1303,0
|
429 |
+
168,1
|
430 |
+
1420,0
|
431 |
+
1665,0
|
432 |
+
1515,0
|
433 |
+
1311,1
|
434 |
+
2359,0
|
435 |
+
1351,0
|
436 |
+
788,1
|
437 |
+
1274,1
|
438 |
+
684,1
|
439 |
+
1736,0
|
440 |
+
674,0
|
441 |
+
1686,0
|
442 |
+
164,0
|
443 |
+
1132,0
|
444 |
+
621,0
|
445 |
+
109,0
|
446 |
+
6,1
|
447 |
+
373,1
|
448 |
+
1008,1
|
449 |
+
1929,0
|
450 |
+
1287,0
|
451 |
+
568,0
|
452 |
+
652,0
|
453 |
+
4,0
|
454 |
+
172,1
|
455 |
+
312,1
|
456 |
+
2376,0
|
457 |
+
85,0
|
458 |
+
300,0
|
459 |
+
1075,1
|
460 |
+
787,0
|
461 |
+
14,1
|
462 |
+
489,1
|
463 |
+
407,0
|
464 |
+
369,1
|
465 |
+
275,0
|
466 |
+
2055,0
|
467 |
+
1103,0
|
468 |
+
314,1
|
469 |
+
487,1
|
470 |
+
135,1
|
471 |
+
134,1
|
472 |
+
1122,0
|
473 |
+
1286,0
|
474 |
+
2033,0
|
475 |
+
866,1
|
476 |
+
478,1
|
477 |
+
722,1
|
478 |
+
792,0
|
479 |
+
95,0
|
480 |
+
454,0
|
481 |
+
313,1
|
482 |
+
597,0
|
483 |
+
343,0
|
484 |
+
541,1
|
485 |
+
2346,0
|
486 |
+
2212,0
|
487 |
+
349,1
|
488 |
+
1323,0
|
489 |
+
104,0
|
490 |
+
2070,0
|
491 |
+
808,0
|
492 |
+
815,1
|
493 |
+
1990,0
|
494 |
+
1583,0
|
495 |
+
201,1
|
496 |
+
1050,1
|
497 |
+
834,0
|
498 |
+
218,0
|
499 |
+
602,1
|
500 |
+
2194,0
|
501 |
+
580,0
|
502 |
+
100,1
|
503 |
+
1488,0
|
504 |
+
1265,1
|
505 |
+
850,1
|
506 |
+
46,1
|
507 |
+
231,0
|
508 |
+
2076,0
|
509 |
+
2151,0
|
510 |
+
125,1
|
511 |
+
1071,0
|
512 |
+
419,1
|
513 |
+
1334,0
|
514 |
+
1344,1
|
515 |
+
775,1
|
516 |
+
1152,0
|
517 |
+
709,1
|
518 |
+
1204,0
|
519 |
+
1164,0
|
520 |
+
609,1
|
521 |
+
419,0
|
522 |
+
944,0
|
523 |
+
2303,0
|
524 |
+
1297,0
|
525 |
+
1799,0
|
526 |
+
526,1
|
527 |
+
353,0
|
528 |
+
2302,0
|
529 |
+
889,1
|
530 |
+
1980,0
|
531 |
+
1238,0
|
532 |
+
225,1
|
533 |
+
468,1
|
534 |
+
1060,1
|
535 |
+
287,1
|
536 |
+
1765,0
|
537 |
+
1629,0
|
538 |
+
559,0
|
539 |
+
642,0
|
540 |
+
2001,0
|
541 |
+
740,1
|
542 |
+
463,1
|
543 |
+
1198,1
|
544 |
+
1680,0
|
545 |
+
210,0
|
546 |
+
1117,0
|
547 |
+
1992,0
|
548 |
+
1189,0
|
549 |
+
1269,0
|
550 |
+
728,1
|
551 |
+
297,1
|
552 |
+
239,1
|
553 |
+
2174,0
|
554 |
+
159,1
|
555 |
+
879,1
|
556 |
+
2127,0
|
557 |
+
1384,0
|
558 |
+
1800,0
|
559 |
+
72,1
|
560 |
+
975,1
|
561 |
+
272,0
|
562 |
+
339,1
|
563 |
+
1372,0
|
564 |
+
1307,1
|
565 |
+
1133,0
|
566 |
+
691,1
|
567 |
+
731,0
|
568 |
+
74,1
|
569 |
+
2266,0
|
570 |
+
158,0
|
571 |
+
790,0
|
572 |
+
938,1
|
573 |
+
1559,0
|
574 |
+
491,1
|
575 |
+
473,1
|
576 |
+
1045,1
|
577 |
+
732,1
|
578 |
+
1091,1
|
579 |
+
536,0
|
580 |
+
1301,0
|
581 |
+
2019,0
|
582 |
+
1182,0
|
583 |
+
661,1
|
584 |
+
1071,1
|
585 |
+
189,0
|
586 |
+
501,0
|
587 |
+
745,1
|
588 |
+
1125,1
|
589 |
+
1704,0
|
590 |
+
341,1
|
591 |
+
815,0
|
592 |
+
233,1
|
593 |
+
840,1
|
594 |
+
809,0
|
595 |
+
1120,1
|
596 |
+
589,1
|
597 |
+
1192,0
|
598 |
+
1673,0
|
599 |
+
2197,0
|
600 |
+
369,0
|
601 |
+
767,0
|
602 |
+
687,0
|
603 |
+
241,0
|
604 |
+
402,0
|
605 |
+
651,1
|
606 |
+
618,0
|
607 |
+
963,1
|
608 |
+
551,0
|
609 |
+
671,1
|
610 |
+
915,1
|
611 |
+
597,1
|
612 |
+
1314,1
|
613 |
+
587,0
|
614 |
+
1203,1
|
615 |
+
152,1
|
616 |
+
994,1
|
617 |
+
767,1
|
618 |
+
2243,0
|
619 |
+
295,1
|
620 |
+
364,1
|
621 |
+
2297,0
|
622 |
+
1995,0
|
623 |
+
702,1
|
624 |
+
1832,0
|
625 |
+
317,1
|
626 |
+
1120,0
|
627 |
+
1086,0
|
628 |
+
673,0
|
629 |
+
777,0
|
630 |
+
1423,0
|
631 |
+
1473,0
|
632 |
+
1137,1
|
633 |
+
1376,0
|
634 |
+
754,1
|
635 |
+
284,1
|
636 |
+
1312,1
|
637 |
+
846,0
|
638 |
+
669,0
|
639 |
+
832,0
|
640 |
+
884,1
|
641 |
+
1316,1
|
642 |
+
34,0
|
643 |
+
1249,0
|
644 |
+
2210,0
|
645 |
+
321,1
|
646 |
+
1251,1
|
647 |
+
268,0
|
648 |
+
41,0
|
649 |
+
464,0
|
650 |
+
69,0
|
651 |
+
992,0
|
652 |
+
306,1
|
653 |
+
213,0
|
654 |
+
108,0
|
655 |
+
215,0
|
656 |
+
512,1
|
657 |
+
874,1
|
658 |
+
846,1
|
659 |
+
28,1
|
660 |
+
1604,0
|
661 |
+
1520,0
|
662 |
+
499,0
|
663 |
+
449,0
|
664 |
+
116,0
|
665 |
+
1040,1
|
666 |
+
1092,1
|
667 |
+
1761,0
|
668 |
+
1484,0
|
669 |
+
760,0
|
670 |
+
1140,1
|
671 |
+
395,1
|
672 |
+
1142,0
|
673 |
+
1586,0
|
674 |
+
1092,0
|
675 |
+
300,1
|
676 |
+
2192,0
|
677 |
+
830,1
|
678 |
+
1748,0
|
679 |
+
269,1
|
680 |
+
1294,1
|
681 |
+
1231,0
|
682 |
+
701,1
|
683 |
+
326,1
|
684 |
+
1439,0
|
685 |
+
928,1
|
686 |
+
398,0
|
687 |
+
145,0
|
688 |
+
1670,0
|
689 |
+
428,1
|
690 |
+
2253,0
|
691 |
+
1009,1
|
692 |
+
986,0
|
693 |
+
270,0
|
694 |
+
1382,0
|
695 |
+
924,0
|
696 |
+
204,1
|
697 |
+
540,0
|
698 |
+
1491,0
|
699 |
+
1522,0
|
700 |
+
1113,1
|
701 |
+
199,1
|
702 |
+
298,0
|
703 |
+
838,0
|
704 |
+
2366,0
|
705 |
+
302,0
|
706 |
+
982,0
|
707 |
+
1123,0
|
708 |
+
1183,1
|
709 |
+
60,0
|
710 |
+
1133,1
|
711 |
+
1061,0
|
712 |
+
664,1
|
713 |
+
712,1
|
714 |
+
1813,0
|
715 |
+
820,0
|
716 |
+
1743,0
|
717 |
+
977,0
|
718 |
+
909,0
|
719 |
+
689,1
|
720 |
+
759,0
|
721 |
+
2375,0
|
722 |
+
761,0
|
723 |
+
2128,0
|
724 |
+
309,1
|
725 |
+
120,0
|
726 |
+
1314,0
|
727 |
+
988,0
|
728 |
+
1669,0
|
729 |
+
1162,1
|
730 |
+
1272,1
|
731 |
+
1007,0
|
732 |
+
1959,0
|
733 |
+
1057,1
|
734 |
+
1707,0
|
735 |
+
580,1
|
736 |
+
1387,0
|
737 |
+
1066,1
|
738 |
+
557,1
|
739 |
+
499,1
|
740 |
+
1440,0
|
741 |
+
854,1
|
742 |
+
252,0
|
743 |
+
1135,1
|
744 |
+
372,0
|
745 |
+
2189,0
|
746 |
+
1184,1
|
747 |
+
794,0
|
748 |
+
198,1
|
749 |
+
1296,0
|
750 |
+
1821,0
|
751 |
+
330,1
|
752 |
+
1218,1
|
753 |
+
826,0
|
754 |
+
1336,0
|
755 |
+
123,1
|
756 |
+
2010,0
|
757 |
+
1270,0
|
758 |
+
1413,0
|
759 |
+
358,0
|
760 |
+
665,1
|
761 |
+
396,0
|
762 |
+
367,1
|
763 |
+
752,1
|
764 |
+
1214,0
|
765 |
+
416,1
|
766 |
+
1789,0
|
767 |
+
2067,0
|
768 |
+
1193,1
|
769 |
+
322,0
|
770 |
+
1016,1
|
771 |
+
2154,0
|
772 |
+
1767,0
|
773 |
+
494,1
|
774 |
+
19,1
|
775 |
+
332,1
|
776 |
+
35,0
|
777 |
+
1632,0
|
778 |
+
376,1
|
779 |
+
66,1
|
780 |
+
144,1
|
781 |
+
72,0
|
782 |
+
959,1
|
783 |
+
649,1
|
784 |
+
658,0
|
785 |
+
263,0
|
786 |
+
2261,0
|
787 |
+
946,1
|
788 |
+
1330,0
|
789 |
+
1042,1
|
790 |
+
1588,0
|
791 |
+
1518,0
|
792 |
+
2248,0
|
793 |
+
1230,1
|
794 |
+
1849,0
|
795 |
+
1817,0
|
796 |
+
1699,0
|
797 |
+
2358,0
|
798 |
+
1185,1
|
799 |
+
769,0
|
800 |
+
1232,1
|
801 |
+
484,1
|
802 |
+
40,1
|
803 |
+
1277,0
|
804 |
+
829,1
|
805 |
+
637,1
|
806 |
+
1190,0
|
807 |
+
483,1
|
808 |
+
1667,0
|
809 |
+
9,0
|
810 |
+
1759,0
|
811 |
+
368,0
|
812 |
+
2288,0
|
813 |
+
2118,0
|
814 |
+
721,1
|
815 |
+
138,0
|
816 |
+
117,1
|
817 |
+
103,1
|
818 |
+
1035,0
|
819 |
+
1852,0
|
820 |
+
373,0
|
821 |
+
2381,0
|
822 |
+
2204,0
|
823 |
+
1477,0
|
824 |
+
48,0
|
825 |
+
533,0
|
826 |
+
1194,0
|
827 |
+
1055,0
|
828 |
+
124,0
|
829 |
+
1288,1
|
830 |
+
1829,0
|
831 |
+
400,0
|
832 |
+
1197,0
|
833 |
+
8,0
|
834 |
+
59,1
|
835 |
+
1319,1
|
836 |
+
717,1
|
837 |
+
994,0
|
838 |
+
410,1
|
839 |
+
255,0
|
840 |
+
1860,0
|
841 |
+
506,0
|
842 |
+
2152,0
|
843 |
+
1769,0
|
844 |
+
598,0
|
845 |
+
5,0
|
846 |
+
2245,0
|
847 |
+
755,0
|
848 |
+
1331,0
|
849 |
+
188,1
|
850 |
+
1447,0
|
851 |
+
2201,0
|
852 |
+
1069,1
|
853 |
+
289,1
|
854 |
+
1339,0
|
855 |
+
206,1
|
856 |
+
1890,0
|
857 |
+
1078,1
|
858 |
+
504,1
|
859 |
+
337,1
|
860 |
+
1348,0
|
861 |
+
1049,1
|
862 |
+
603,1
|
863 |
+
1912,0
|
864 |
+
979,0
|
865 |
+
477,1
|
866 |
+
222,0
|
867 |
+
972,1
|
868 |
+
981,1
|
869 |
+
795,1
|
870 |
+
171,0
|
871 |
+
16,1
|
872 |
+
594,1
|
873 |
+
121,1
|
874 |
+
202,1
|
875 |
+
2336,0
|
876 |
+
1070,1
|
877 |
+
1265,0
|
878 |
+
604,1
|
879 |
+
646,1
|
880 |
+
2347,0
|
881 |
+
2094,0
|
882 |
+
560,1
|
883 |
+
496,0
|
884 |
+
239,0
|
885 |
+
1456,0
|
886 |
+
242,0
|
887 |
+
897,0
|
888 |
+
448,0
|
889 |
+
2061,0
|
890 |
+
679,0
|
891 |
+
532,0
|
892 |
+
1727,0
|
893 |
+
1286,1
|
894 |
+
653,0
|
895 |
+
1850,0
|
896 |
+
1193,0
|
897 |
+
854,0
|
898 |
+
771,0
|
899 |
+
606,1
|
900 |
+
1396,0
|
901 |
+
641,0
|
902 |
+
493,1
|
903 |
+
960,0
|
904 |
+
2034,0
|
905 |
+
651,0
|
906 |
+
451,0
|
907 |
+
2383,0
|
908 |
+
1263,1
|
909 |
+
1165,1
|
910 |
+
1153,1
|
911 |
+
84,0
|
912 |
+
996,0
|
913 |
+
1236,1
|
914 |
+
945,1
|
915 |
+
976,0
|
916 |
+
1094,0
|
917 |
+
877,0
|
918 |
+
2025,0
|
919 |
+
1039,1
|
920 |
+
857,0
|
921 |
+
1267,1
|
922 |
+
1280,1
|
923 |
+
1088,0
|
924 |
+
1116,0
|
925 |
+
157,1
|
926 |
+
555,1
|
927 |
+
962,0
|
928 |
+
285,0
|
929 |
+
1197,1
|
930 |
+
246,0
|
931 |
+
263,1
|
932 |
+
983,1
|
933 |
+
619,1
|
934 |
+
968,1
|
935 |
+
1309,0
|
936 |
+
484,0
|
937 |
+
869,0
|
938 |
+
643,0
|
939 |
+
1322,1
|
940 |
+
979,1
|
941 |
+
756,1
|
942 |
+
1252,0
|
943 |
+
734,0
|
944 |
+
870,0
|
945 |
+
133,0
|
946 |
+
1010,0
|
947 |
+
223,1
|
948 |
+
1671,0
|
949 |
+
93,1
|
950 |
+
1942,0
|
951 |
+
969,0
|
952 |
+
687,1
|
953 |
+
1099,1
|
954 |
+
1548,0
|
955 |
+
860,0
|
956 |
+
511,1
|
957 |
+
554,1
|
958 |
+
447,1
|
959 |
+
620,0
|
960 |
+
640,1
|
961 |
+
587,1
|
962 |
+
642,1
|
963 |
+
1169,1
|
964 |
+
258,0
|
965 |
+
1471,0
|
966 |
+
882,1
|
967 |
+
930,0
|
968 |
+
2185,0
|
969 |
+
843,1
|
970 |
+
413,1
|
971 |
+
632,0
|
972 |
+
640,0
|
973 |
+
1906,0
|
974 |
+
8,1
|
975 |
+
1327,1
|
976 |
+
1199,1
|
977 |
+
50,1
|
978 |
+
234,0
|
979 |
+
584,1
|
980 |
+
1213,0
|
981 |
+
328,0
|
982 |
+
1076,0
|
983 |
+
1991,0
|
984 |
+
794,1
|
985 |
+
568,1
|
986 |
+
565,0
|
987 |
+
1555,0
|
988 |
+
92,1
|
989 |
+
920,1
|
990 |
+
1805,0
|
991 |
+
923,1
|
992 |
+
324,1
|
993 |
+
346,0
|
994 |
+
856,0
|
995 |
+
12,0
|
996 |
+
1079,1
|
997 |
+
274,1
|
998 |
+
2208,0
|
999 |
+
814,1
|
1000 |
+
1792,0
|
1001 |
+
1054,1
|
1002 |
+
2155,0
|
1003 |
+
2339,0
|
1004 |
+
881,1
|
1005 |
+
1033,0
|
1006 |
+
363,0
|
1007 |
+
393,0
|
1008 |
+
151,0
|
1009 |
+
1621,0
|
1010 |
+
82,0
|
1011 |
+
440,1
|
1012 |
+
871,1
|
1013 |
+
1244,0
|
1014 |
+
787,1
|
1015 |
+
452,0
|
1016 |
+
2126,0
|
1017 |
+
1700,0
|
1018 |
+
743,0
|
1019 |
+
100,0
|
1020 |
+
1315,0
|
1021 |
+
381,1
|
1022 |
+
954,1
|
1023 |
+
395,0
|
1024 |
+
1790,0
|
1025 |
+
19,0
|
1026 |
+
1139,1
|
1027 |
+
861,1
|
1028 |
+
294,1
|
1029 |
+
1780,0
|
1030 |
+
119,0
|
1031 |
+
1154,1
|
1032 |
+
137,1
|
1033 |
+
2254,0
|
1034 |
+
1172,1
|
1035 |
+
207,1
|
1036 |
+
563,1
|
1037 |
+
43,0
|
1038 |
+
1140,0
|
1039 |
+
112,0
|
1040 |
+
2279,0
|
1041 |
+
1178,1
|
1042 |
+
881,0
|
1043 |
+
1597,0
|
1044 |
+
596,0
|
1045 |
+
283,1
|
1046 |
+
680,0
|
1047 |
+
322,1
|
1048 |
+
537,0
|
1049 |
+
1097,1
|
1050 |
+
564,1
|
1051 |
+
558,1
|
1052 |
+
146,0
|
1053 |
+
210,1
|
1054 |
+
968,0
|
1055 |
+
282,1
|
1056 |
+
281,1
|
1057 |
+
29,1
|
1058 |
+
525,0
|
1059 |
+
538,1
|
1060 |
+
339,0
|
1061 |
+
2145,0
|
1062 |
+
758,0
|
1063 |
+
518,1
|
1064 |
+
1062,0
|
1065 |
+
591,1
|
1066 |
+
177,1
|
1067 |
+
1723,0
|
1068 |
+
152,0
|
1069 |
+
1001,1
|
1070 |
+
2158,0
|
1071 |
+
345,1
|
1072 |
+
659,1
|
1073 |
+
81,1
|
1074 |
+
325,0
|
1075 |
+
628,1
|
1076 |
+
402,1
|
1077 |
+
1356,0
|
1078 |
+
98,0
|
1079 |
+
1056,1
|
1080 |
+
1210,1
|
1081 |
+
1148,0
|
1082 |
+
3,1
|
1083 |
+
686,1
|
1084 |
+
1782,0
|
1085 |
+
56,0
|
1086 |
+
522,1
|
1087 |
+
1030,0
|
1088 |
+
1791,0
|
1089 |
+
971,1
|
1090 |
+
32,0
|
1091 |
+
978,1
|
1092 |
+
588,1
|
1093 |
+
678,1
|
1094 |
+
2292,0
|
1095 |
+
386,0
|
1096 |
+
737,0
|
1097 |
+
1925,0
|
1098 |
+
609,0
|
1099 |
+
364,0
|
1100 |
+
763,0
|
1101 |
+
645,0
|
1102 |
+
1847,0
|
1103 |
+
1104,1
|
1104 |
+
1461,0
|
1105 |
+
137,0
|
1106 |
+
240,0
|
1107 |
+
674,1
|
1108 |
+
601,0
|
1109 |
+
2088,0
|
1110 |
+
155,1
|
1111 |
+
938,0
|
1112 |
+
907,0
|
1113 |
+
74,0
|
1114 |
+
911,1
|
1115 |
+
974,0
|
1116 |
+
742,1
|
1117 |
+
983,0
|
1118 |
+
940,1
|
1119 |
+
347,1
|
1120 |
+
492,1
|
1121 |
+
80,0
|
1122 |
+
1231,1
|
1123 |
+
1724,0
|
1124 |
+
782,1
|
1125 |
+
1495,0
|
1126 |
+
2200,0
|
1127 |
+
51,0
|
1128 |
+
751,1
|
1129 |
+
191,0
|
1130 |
+
1422,0
|
1131 |
+
496,1
|
1132 |
+
974,1
|
1133 |
+
805,1
|
1134 |
+
1657,0
|
1135 |
+
433,0
|
1136 |
+
438,0
|
1137 |
+
285,1
|
1138 |
+
1897,0
|
1139 |
+
1175,1
|
1140 |
+
1320,0
|
1141 |
+
2331,0
|
1142 |
+
1546,0
|
1143 |
+
780,1
|
1144 |
+
820,1
|
1145 |
+
1266,0
|
1146 |
+
977,1
|
1147 |
+
1496,0
|
1148 |
+
1763,0
|
1149 |
+
1842,0
|
1150 |
+
1144,0
|
1151 |
+
1547,0
|
1152 |
+
228,0
|
1153 |
+
1708,0
|
1154 |
+
577,0
|
1155 |
+
401,1
|
1156 |
+
418,1
|
1157 |
+
511,0
|
1158 |
+
590,1
|
1159 |
+
1013,1
|
1160 |
+
567,0
|
1161 |
+
470,1
|
1162 |
+
227,1
|
1163 |
+
436,1
|
1164 |
+
1047,0
|
1165 |
+
822,1
|
1166 |
+
117,0
|
1167 |
+
29,0
|
1168 |
+
997,0
|
1169 |
+
1093,0
|
1170 |
+
1090,1
|
1171 |
+
624,1
|
1172 |
+
164,1
|
1173 |
+
2378,0
|
1174 |
+
1084,1
|
1175 |
+
75,1
|
1176 |
+
793,1
|
1177 |
+
910,0
|
1178 |
+
676,0
|
1179 |
+
908,0
|
1180 |
+
2029,0
|
1181 |
+
92,0
|
1182 |
+
38,0
|
1183 |
+
505,1
|
1184 |
+
409,1
|
1185 |
+
1388,0
|
1186 |
+
2122,0
|
1187 |
+
1997,0
|
1188 |
+
441,0
|
1189 |
+
134,0
|
1190 |
+
107,0
|
1191 |
+
1026,0
|
1192 |
+
236,1
|
1193 |
+
261,0
|
1194 |
+
1602,0
|
1195 |
+
2195,0
|
1196 |
+
1937,0
|
1197 |
+
2121,0
|
1198 |
+
627,0
|
1199 |
+
697,0
|
1200 |
+
1340,1
|
1201 |
+
516,1
|
1202 |
+
467,1
|
1203 |
+
1173,0
|
1204 |
+
1448,0
|
1205 |
+
2065,0
|
1206 |
+
456,0
|
1207 |
+
1474,0
|
1208 |
+
2322,0
|
1209 |
+
1274,0
|
1210 |
+
1784,0
|
1211 |
+
773,1
|
1212 |
+
818,1
|
1213 |
+
153,1
|
1214 |
+
53,0
|
1215 |
+
1109,1
|
1216 |
+
2350,0
|
1217 |
+
1578,0
|
1218 |
+
1858,0
|
1219 |
+
321,0
|
1220 |
+
986,1
|
1221 |
+
180,0
|
1222 |
+
1115,0
|
1223 |
+
1873,0
|
1224 |
+
1869,0
|
1225 |
+
416,0
|
1226 |
+
187,0
|
1227 |
+
819,1
|
1228 |
+
220,0
|
1229 |
+
156,0
|
1230 |
+
1827,0
|
1231 |
+
685,0
|
1232 |
+
1255,1
|
1233 |
+
952,1
|
1234 |
+
2341,0
|
1235 |
+
595,1
|
1236 |
+
509,0
|
1237 |
+
251,0
|
1238 |
+
2172,0
|
1239 |
+
1934,0
|
1240 |
+
118,1
|
1241 |
+
521,0
|
1242 |
+
1237,1
|
1243 |
+
807,0
|
1244 |
+
345,0
|
1245 |
+
457,1
|
1246 |
+
955,1
|
1247 |
+
633,0
|
1248 |
+
330,0
|
1249 |
+
423,1
|
1250 |
+
21,1
|
1251 |
+
875,1
|
1252 |
+
280,1
|
1253 |
+
1165,0
|
1254 |
+
70,0
|
1255 |
+
404,0
|
1256 |
+
389,1
|
1257 |
+
1012,1
|
1258 |
+
929,1
|
1259 |
+
1378,0
|
1260 |
+
1025,1
|
1261 |
+
1453,0
|
1262 |
+
730,0
|
1263 |
+
1342,0
|
1264 |
+
486,1
|
1265 |
+
1540,0
|
1266 |
+
528,0
|
1267 |
+
1146,0
|
1268 |
+
1291,1
|
1269 |
+
148,1
|
1270 |
+
778,0
|
1271 |
+
1151,0
|
1272 |
+
675,0
|
1273 |
+
823,0
|
1274 |
+
2187,0
|
1275 |
+
498,0
|
1276 |
+
2134,0
|
1277 |
+
1004,0
|
1278 |
+
1248,1
|
1279 |
+
1273,1
|
1280 |
+
912,0
|
1281 |
+
984,1
|
1282 |
+
615,1
|
1283 |
+
1013,0
|
1284 |
+
1760,0
|
1285 |
+
469,1
|
1286 |
+
2325,0
|
1287 |
+
1311,0
|
1288 |
+
958,0
|
1289 |
+
194,0
|
1290 |
+
811,1
|
1291 |
+
736,0
|
1292 |
+
1734,0
|
1293 |
+
842,0
|
1294 |
+
426,1
|
1295 |
+
1163,0
|
1296 |
+
845,0
|
1297 |
+
1660,0
|
1298 |
+
1866,0
|
1299 |
+
629,1
|
1300 |
+
1238,1
|
1301 |
+
849,1
|
1302 |
+
467,0
|
1303 |
+
741,0
|
1304 |
+
1324,1
|
1305 |
+
749,1
|
1306 |
+
247,1
|
1307 |
+
1668,0
|
1308 |
+
25,1
|
1309 |
+
185,0
|
1310 |
+
1584,0
|
1311 |
+
1338,0
|
1312 |
+
223,0
|
1313 |
+
657,1
|
1314 |
+
440,0
|
1315 |
+
558,0
|
1316 |
+
1063,0
|
1317 |
+
792,1
|
1318 |
+
1772,0
|
1319 |
+
121,0
|
1320 |
+
727,0
|
1321 |
+
527,1
|
1322 |
+
1164,1
|
1323 |
+
764,0
|
1324 |
+
278,0
|
1325 |
+
773,0
|
1326 |
+
1168,1
|
1327 |
+
1859,0
|
1328 |
+
1046,1
|
1329 |
+
926,0
|
1330 |
+
1019,0
|
1331 |
+
52,0
|
1332 |
+
1044,1
|
1333 |
+
1136,1
|
1334 |
+
1910,0
|
1335 |
+
277,0
|
1336 |
+
305,0
|
1337 |
+
523,0
|
1338 |
+
922,1
|
1339 |
+
2304,0
|
1340 |
+
63,0
|
1341 |
+
211,0
|
1342 |
+
1549,0
|
1343 |
+
2020,0
|
1344 |
+
1394,0
|
1345 |
+
396,1
|
1346 |
+
919,0
|
1347 |
+
50,0
|
1348 |
+
1450,0
|
1349 |
+
831,1
|
1350 |
+
240,1
|
1351 |
+
1310,0
|
1352 |
+
958,1
|
1353 |
+
58,0
|
1354 |
+
856,1
|
1355 |
+
478,0
|
1356 |
+
636,0
|
1357 |
+
1137,0
|
1358 |
+
357,1
|
1359 |
+
1081,0
|
1360 |
+
953,1
|
1361 |
+
1696,0
|
1362 |
+
1806,0
|
1363 |
+
279,1
|
1364 |
+
713,1
|
1365 |
+
2364,0
|
1366 |
+
114,0
|
1367 |
+
190,0
|
1368 |
+
371,0
|
1369 |
+
221,0
|
1370 |
+
720,1
|
1371 |
+
1508,0
|
1372 |
+
310,0
|
1373 |
+
903,0
|
1374 |
+
795,0
|
1375 |
+
1102,0
|
1376 |
+
2272,0
|
1377 |
+
479,1
|
1378 |
+
1186,1
|
1379 |
+
585,1
|
1380 |
+
64,1
|
1381 |
+
967,1
|
1382 |
+
1434,0
|
1383 |
+
1579,0
|
1384 |
+
1111,1
|
1385 |
+
616,1
|
1386 |
+
1777,0
|
1387 |
+
993,0
|
1388 |
+
873,0
|
1389 |
+
774,1
|
1390 |
+
1157,0
|
1391 |
+
824,1
|
1392 |
+
1919,0
|
1393 |
+
608,0
|
1394 |
+
917,0
|
1395 |
+
1037,0
|
1396 |
+
1324,0
|
1397 |
+
379,0
|
1398 |
+
1368,0
|
1399 |
+
635,1
|
1400 |
+
30,1
|
1401 |
+
1718,0
|
1402 |
+
231,1
|
1403 |
+
1103,1
|
1404 |
+
4,1
|
1405 |
+
417,1
|
1406 |
+
2043,0
|
1407 |
+
68,1
|
1408 |
+
1347,1
|
1409 |
+
1836,0
|
1410 |
+
481,1
|
1411 |
+
1596,0
|
1412 |
+
1468,0
|
1413 |
+
807,1
|
1414 |
+
315,1
|
1415 |
+
1807,0
|
1416 |
+
599,1
|
1417 |
+
1257,1
|
1418 |
+
797,1
|
1419 |
+
412,0
|
1420 |
+
246,1
|
1421 |
+
943,1
|
1422 |
+
1262,0
|
1423 |
+
124,1
|
1424 |
+
724,1
|
1425 |
+
1463,0
|
1426 |
+
260,0
|
1427 |
+
562,1
|
1428 |
+
799,0
|
1429 |
+
768,1
|
1430 |
+
487,0
|
1431 |
+
637,0
|
1432 |
+
193,0
|
1433 |
+
1018,1
|
1434 |
+
1865,0
|
1435 |
+
273,0
|
1436 |
+
579,0
|
1437 |
+
209,0
|
1438 |
+
704,0
|
1439 |
+
1064,0
|
1440 |
+
808,1
|
1441 |
+
1083,0
|
1442 |
+
226,0
|
1443 |
+
398,1
|
1444 |
+
524,1
|
1445 |
+
2337,0
|
1446 |
+
891,0
|
1447 |
+
1272,0
|
1448 |
+
894,0
|
1449 |
+
517,1
|
1450 |
+
1293,0
|
1451 |
+
705,0
|
1452 |
+
630,1
|
1453 |
+
862,0
|
1454 |
+
1162,0
|
1455 |
+
33,1
|
1456 |
+
183,1
|
1457 |
+
2176,0
|
1458 |
+
277,1
|
1459 |
+
1705,0
|
1460 |
+
1141,0
|
1461 |
+
1616,0
|
1462 |
+
914,1
|
1463 |
+
1221,0
|
1464 |
+
1936,0
|
1465 |
+
435,1
|
1466 |
+
865,0
|
1467 |
+
276,1
|
1468 |
+
1891,0
|
1469 |
+
1268,1
|
1470 |
+
1226,1
|
1471 |
+
304,0
|
1472 |
+
40,0
|
1473 |
+
377,1
|
1474 |
+
1091,0
|
1475 |
+
57,0
|
1476 |
+
200,0
|
1477 |
+
1267,0
|
1478 |
+
789,1
|
1479 |
+
206,0
|
1480 |
+
34,1
|
1481 |
+
972,0
|
1482 |
+
2353,0
|
1483 |
+
232,1
|
1484 |
+
1161,1
|
1485 |
+
535,1
|
1486 |
+
1637,0
|
1487 |
+
1127,0
|
1488 |
+
2223,0
|
1489 |
+
242,1
|
1490 |
+
46,0
|
1491 |
+
1253,1
|
1492 |
+
22,0
|
1493 |
+
357,0
|
1494 |
+
439,0
|
1495 |
+
1030,1
|
1496 |
+
876,0
|
1497 |
+
1462,0
|
1498 |
+
153,0
|
1499 |
+
235,0
|
1500 |
+
1253,0
|
1501 |
+
2214,0
|
1502 |
+
251,1
|
1503 |
+
1561,0
|
1504 |
+
714,0
|
1505 |
+
677,1
|
1506 |
+
311,0
|
1507 |
+
1256,0
|
1508 |
+
922,0
|
1509 |
+
1087,1
|
1510 |
+
353,1
|
1511 |
+
519,0
|
1512 |
+
1489,0
|
1513 |
+
1015,1
|
1514 |
+
1334,1
|
1515 |
+
1278,1
|
1516 |
+
1721,0
|
1517 |
+
2098,0
|
1518 |
+
221,1
|
1519 |
+
931,1
|
1520 |
+
1557,0
|
1521 |
+
2258,0
|
1522 |
+
1893,0
|
1523 |
+
905,1
|
1524 |
+
504,0
|
1525 |
+
1181,0
|
1526 |
+
770,1
|
1527 |
+
80,1
|
1528 |
+
2096,0
|
1529 |
+
1725,0
|
1530 |
+
1022,1
|
1531 |
+
1127,1
|
1532 |
+
1379,0
|
1533 |
+
508,1
|
1534 |
+
1598,0
|
1535 |
+
2109,0
|
1536 |
+
1720,0
|
1537 |
+
1730,0
|
1538 |
+
531,0
|
1539 |
+
15,1
|
1540 |
+
1418,0
|
1541 |
+
273,1
|
1542 |
+
1861,0
|
1543 |
+
135,0
|
1544 |
+
1475,0
|
1545 |
+
699,0
|
1546 |
+
553,0
|
1547 |
+
852,1
|
1548 |
+
725,1
|
1549 |
+
500,1
|
1550 |
+
1828,0
|
1551 |
+
692,0
|
1552 |
+
1261,1
|
1553 |
+
524,0
|
1554 |
+
1179,0
|
1555 |
+
1940,0
|
1556 |
+
540,1
|
1557 |
+
1275,0
|
1558 |
+
2227,0
|
1559 |
+
90,1
|
1560 |
+
320,0
|
1561 |
+
706,1
|
1562 |
+
362,1
|
1563 |
+
173,1
|
1564 |
+
634,0
|
1565 |
+
1932,0
|
1566 |
+
965,1
|
1567 |
+
365,1
|
1568 |
+
302,1
|
1569 |
+
458,0
|
1570 |
+
105,1
|
1571 |
+
98,1
|
1572 |
+
1571,0
|
1573 |
+
127,0
|
1574 |
+
2183,0
|
1575 |
+
1694,0
|
1576 |
+
682,0
|
1577 |
+
387,1
|
1578 |
+
264,1
|
1579 |
+
482,1
|
1580 |
+
551,1
|
1581 |
+
1108,0
|
1582 |
+
745,0
|
1583 |
+
937,0
|
1584 |
+
2116,0
|
1585 |
+
313,0
|
1586 |
+
168,0
|
1587 |
+
719,0
|
1588 |
+
1753,0
|
1589 |
+
804,1
|
1590 |
+
1947,0
|
1591 |
+
1000,0
|
1592 |
+
1131,1
|
1593 |
+
935,0
|
1594 |
+
1863,0
|
1595 |
+
1337,0
|
1596 |
+
530,0
|
1597 |
+
20,1
|
1598 |
+
1313,1
|
1599 |
+
660,0
|
1600 |
+
6,0
|
1601 |
+
114,1
|
1602 |
+
443,0
|
1603 |
+
2354,0
|
1604 |
+
1085,0
|
1605 |
+
1717,0
|
1606 |
+
692,1
|
1607 |
+
1303,1
|
1608 |
+
906,0
|
1609 |
+
1261,0
|
1610 |
+
552,1
|
1611 |
+
394,0
|
1612 |
+
1171,0
|
1613 |
+
2072,0
|
1614 |
+
613,1
|
1615 |
+
832,1
|
1616 |
+
1962,0
|
1617 |
+
1531,0
|
1618 |
+
1179,1
|
1619 |
+
1888,0
|
1620 |
+
663,0
|
1621 |
+
1871,0
|
1622 |
+
267,0
|
1623 |
+
329,0
|
1624 |
+
1430,0
|
1625 |
+
2164,0
|
1626 |
+
1733,0
|
1627 |
+
1069,0
|
1628 |
+
465,1
|
1629 |
+
1095,1
|
1630 |
+
2269,0
|
1631 |
+
658,1
|
1632 |
+
1110,0
|
1633 |
+
1642,0
|
1634 |
+
813,0
|
1635 |
+
106,1
|
1636 |
+
111,0
|
1637 |
+
471,0
|
1638 |
+
292,1
|
1639 |
+
358,1
|
1640 |
+
189,1
|
1641 |
+
655,1
|
1642 |
+
2310,0
|
1643 |
+
1300,0
|
1644 |
+
414,1
|
1645 |
+
1098,0
|
1646 |
+
403,1
|
1647 |
+
1572,0
|
1648 |
+
1317,0
|
1649 |
+
35,1
|
1650 |
+
876,1
|
1651 |
+
1880,0
|
1652 |
+
1970,0
|
1653 |
+
896,1
|
1654 |
+
2015,0
|
1655 |
+
160,0
|
1656 |
+
1243,1
|
1657 |
+
1714,0
|
1658 |
+
2294,0
|
1659 |
+
694,0
|
1660 |
+
1228,1
|
1661 |
+
97,0
|
1662 |
+
165,0
|
1663 |
+
722,0
|
1664 |
+
521,1
|
1665 |
+
1222,0
|
1666 |
+
469,0
|
1667 |
+
811,0
|
1668 |
+
971,0
|
1669 |
+
115,1
|
1670 |
+
248,1
|
1671 |
+
363,1
|
1672 |
+
79,0
|
1673 |
+
1259,0
|
1674 |
+
1505,0
|
1675 |
+
835,1
|
1676 |
+
512,0
|
1677 |
+
243,0
|
1678 |
+
564,0
|
1679 |
+
196,1
|
1680 |
+
830,0
|
1681 |
+
1235,0
|
1682 |
+
1894,0
|
1683 |
+
508,0
|
1684 |
+
15,0
|
1685 |
+
89,1
|
1686 |
+
5,1
|
1687 |
+
1344,0
|
1688 |
+
1169,0
|
1689 |
+
1226,0
|
1690 |
+
948,1
|
1691 |
+
1438,0
|
1692 |
+
383,0
|
1693 |
+
2379,0
|
1694 |
+
1345,1
|
1695 |
+
1271,0
|
1696 |
+
1538,0
|
1697 |
+
2267,0
|
1698 |
+
729,1
|
1699 |
+
185,1
|
1700 |
+
1729,0
|
1701 |
+
763,1
|
1702 |
+
701,0
|
1703 |
+
633,1
|
1704 |
+
1381,0
|
1705 |
+
932,1
|
1706 |
+
1576,0
|
1707 |
+
1166,0
|
1708 |
+
162,1
|
1709 |
+
234,1
|
1710 |
+
949,0
|
1711 |
+
610,0
|
1712 |
+
2097,0
|
1713 |
+
1499,0
|
1714 |
+
1464,0
|
1715 |
+
1028,0
|
1716 |
+
389,0
|
1717 |
+
904,1
|
1718 |
+
1467,0
|
1719 |
+
1406,0
|
1720 |
+
1073,0
|
1721 |
+
1511,0
|
1722 |
+
1436,0
|
1723 |
+
999,1
|
1724 |
+
503,1
|
1725 |
+
574,1
|
1726 |
+
706,0
|
1727 |
+
182,1
|
1728 |
+
1138,1
|
1729 |
+
708,1
|
1730 |
+
503,0
|
1731 |
+
946,0
|
1732 |
+
2309,0
|
1733 |
+
372,1
|
1734 |
+
224,1
|
1735 |
+
872,1
|
1736 |
+
566,1
|
1737 |
+
1021,1
|
1738 |
+
814,0
|
1739 |
+
571,1
|
1740 |
+
785,1
|
1741 |
+
73,1
|
1742 |
+
719,1
|
1743 |
+
1843,0
|
1744 |
+
288,1
|
1745 |
+
1614,0
|
1746 |
+
1112,0
|
1747 |
+
1017,0
|
1748 |
+
1093,1
|
1749 |
+
2251,0
|
1750 |
+
54,0
|
1751 |
+
631,0
|
1752 |
+
662,0
|
1753 |
+
385,1
|
1754 |
+
742,0
|
1755 |
+
178,1
|
1756 |
+
796,1
|
1757 |
+
1681,0
|
1758 |
+
449,1
|
1759 |
+
2233,0
|
1760 |
+
1795,0
|
1761 |
+
2114,0
|
1762 |
+
801,0
|
1763 |
+
887,0
|
1764 |
+
142,1
|
1765 |
+
906,1
|
1766 |
+
66,0
|
1767 |
+
406,1
|
1768 |
+
1690,0
|
1769 |
+
829,0
|
1770 |
+
73,0
|
1771 |
+
238,1
|
1772 |
+
1856,0
|
1773 |
+
176,0
|
1774 |
+
149,0
|
1775 |
+
1216,0
|
1776 |
+
908,1
|
1777 |
+
951,1
|
1778 |
+
927,0
|
1779 |
+
1513,0
|
1780 |
+
1712,0
|
1781 |
+
702,0
|
1782 |
+
1129,0
|
1783 |
+
1256,1
|
1784 |
+
919,1
|
1785 |
+
1884,0
|
1786 |
+
311,1
|
1787 |
+
36,1
|
1788 |
+
1794,0
|
1789 |
+
2093,0
|
1790 |
+
82,1
|
1791 |
+
1018,0
|
1792 |
+
2027,0
|
1793 |
+
1065,1
|
1794 |
+
1639,0
|
1795 |
+
809,1
|
1796 |
+
1318,1
|
1797 |
+
90,0
|
1798 |
+
1074,0
|
1799 |
+
2342,0
|
1800 |
+
1305,1
|
1801 |
+
947,1
|
1802 |
+
1168,0
|
1803 |
+
67,1
|
1804 |
+
270,1
|
1805 |
+
388,0
|
1806 |
+
1306,1
|
1807 |
+
1040,0
|
1808 |
+
352,0
|
1809 |
+
453,0
|
1810 |
+
1545,0
|
1811 |
+
2207,0
|
1812 |
+
645,1
|
1813 |
+
1046,0
|
1814 |
+
857,1
|
1815 |
+
644,1
|
1816 |
+
1519,0
|
1817 |
+
513,0
|
1818 |
+
2173,0
|
1819 |
+
268,1
|
1820 |
+
718,1
|
1821 |
+
109,1
|
1822 |
+
693,0
|
1823 |
+
759,1
|
1824 |
+
997,1
|
1825 |
+
1285,1
|
1826 |
+
1068,1
|
1827 |
+
2217,0
|
1828 |
+
1289,0
|
1829 |
+
1079,0
|
1830 |
+
789,0
|
1831 |
+
45,1
|
1832 |
+
867,0
|
1833 |
+
1446,0
|
1834 |
+
1581,0
|
1835 |
+
529,1
|
1836 |
+
2106,0
|
1837 |
+
1735,0
|
1838 |
+
2078,0
|
1839 |
+
879,0
|
1840 |
+
501,1
|
1841 |
+
1240,0
|
1842 |
+
947,0
|
1843 |
+
1052,0
|
1844 |
+
2278,0
|
1845 |
+
2368,0
|
1846 |
+
1787,0
|
1847 |
+
1202,0
|
1848 |
+
746,0
|
1849 |
+
1039,0
|
1850 |
+
781,0
|
1851 |
+
303,0
|
1852 |
+
1214,1
|
1853 |
+
573,1
|
1854 |
+
391,0
|
1855 |
+
735,0
|
1856 |
+
625,0
|
1857 |
+
1215,1
|
1858 |
+
1042,0
|
1859 |
+
147,0
|
1860 |
+
14,0
|
1861 |
+
1294,0
|
1862 |
+
1628,0
|
1863 |
+
256,0
|
1864 |
+
1080,0
|
1865 |
+
1312,0
|
1866 |
+
694,1
|
1867 |
+
1213,1
|
1868 |
+
78,0
|
1869 |
+
286,0
|
1870 |
+
827,1
|
1871 |
+
894,1
|
1872 |
+
141,0
|
1873 |
+
1283,0
|
1874 |
+
348,0
|
1875 |
+
1115,1
|
1876 |
+
1052,1
|
1877 |
+
557,0
|
1878 |
+
1770,0
|
1879 |
+
1259,1
|
1880 |
+
447,0
|
1881 |
+
53,1
|
1882 |
+
862,1
|
1883 |
+
1037,1
|
1884 |
+
775,0
|
1885 |
+
1268,0
|
1886 |
+
704,1
|
1887 |
+
1347,0
|
1888 |
+
860,1
|
1889 |
+
2017,0
|
1890 |
+
233,0
|
1891 |
+
1781,0
|
1892 |
+
1329,1
|
1893 |
+
86,1
|
1894 |
+
354,0
|
1895 |
+
824,0
|
1896 |
+
1683,0
|
1897 |
+
31,1
|
1898 |
+
1063,1
|
1899 |
+
388,1
|
1900 |
+
1886,0
|
1901 |
+
331,1
|
1902 |
+
359,0
|
1903 |
+
37,1
|
1904 |
+
1134,1
|
1905 |
+
279,0
|
1906 |
+
1745,0
|
1907 |
+
375,1
|
1908 |
+
962,1
|
1909 |
+
422,1
|
1910 |
+
51,1
|
1911 |
+
2328,0
|
1912 |
+
1269,1
|
1913 |
+
1421,0
|
1914 |
+
171,1
|
1915 |
+
895,0
|
1916 |
+
126,0
|
1917 |
+
1804,0
|
1918 |
+
1709,0
|
1919 |
+
1689,0
|
1920 |
+
309,0
|
1921 |
+
425,0
|
1922 |
+
1528,0
|
1923 |
+
1341,0
|
1924 |
+
237,1
|
1925 |
+
399,0
|
1926 |
+
1001,0
|
1927 |
+
838,1
|
1928 |
+
1779,0
|
1929 |
+
1650,0
|
1930 |
+
755,1
|
1931 |
+
27,1
|
1932 |
+
179,0
|
1933 |
+
1000,1
|
1934 |
+
2246,0
|
1935 |
+
939,0
|
1936 |
+
514,1
|
1937 |
+
1219,1
|
1938 |
+
868,0
|
1939 |
+
1219,0
|
1940 |
+
444,0
|
1941 |
+
667,0
|
1942 |
+
2240,0
|
1943 |
+
538,0
|
1944 |
+
738,0
|
1945 |
+
774,0
|
1946 |
+
1333,1
|
1947 |
+
1192,1
|
1948 |
+
1132,1
|
1949 |
+
378,1
|
1950 |
+
817,1
|
1951 |
+
1108,1
|
1952 |
+
243,1
|
1953 |
+
2,0
|
1954 |
+
709,0
|
1955 |
+
626,0
|
1956 |
+
1191,0
|
1957 |
+
654,0
|
1958 |
+
641,1
|
1959 |
+
988,1
|
1960 |
+
783,0
|
1961 |
+
101,0
|
1962 |
+
3,0
|
1963 |
+
991,0
|
1964 |
+
415,0
|
1965 |
+
128,1
|
1966 |
+
2360,0
|
1967 |
+
96,1
|
1968 |
+
476,0
|
1969 |
+
1336,1
|
1970 |
+
2198,0
|
1971 |
+
1460,0
|
1972 |
+
207,0
|
1973 |
+
1215,0
|
1974 |
+
1062,1
|
1975 |
+
399,1
|
1976 |
+
519,1
|
1977 |
+
1776,0
|
1978 |
+
966,1
|
1979 |
+
253,1
|
1980 |
+
1251,0
|
1981 |
+
1156,1
|
1982 |
+
1532,0
|
1983 |
+
1139,0
|
1984 |
+
1756,0
|
1985 |
+
765,1
|
1986 |
+
458,1
|
1987 |
+
2073,0
|
1988 |
+
1935,0
|
1989 |
+
2235,0
|
1990 |
+
208,1
|
1991 |
+
514,0
|
1992 |
+
436,0
|
1993 |
+
607,0
|
1994 |
+
199,0
|
1995 |
+
379,1
|
1996 |
+
43,1
|
1997 |
+
756,0
|
1998 |
+
57,1
|
1999 |
+
583,0
|
2000 |
+
184,1
|
2001 |
+
885,0
|
2002 |
+
88,1
|
2003 |
+
1183,0
|
2004 |
+
656,0
|
2005 |
+
1147,1
|
2006 |
+
52,1
|
2007 |
+
638,0
|
2008 |
+
733,0
|
2009 |
+
1523,0
|
2010 |
+
438,1
|
2011 |
+
1,0
|
2012 |
+
727,1
|
2013 |
+
1411,0
|
2014 |
+
1876,0
|
2015 |
+
750,1
|
2016 |
+
2079,0
|
2017 |
+
826,1
|
2018 |
+
1652,0
|
2019 |
+
1786,0
|
2020 |
+
970,1
|
2021 |
+
1301,1
|
2022 |
+
520,0
|
2023 |
+
654,1
|
2024 |
+
864,0
|
2025 |
+
368,1
|
2026 |
+
1400,0
|
2027 |
+
1403,0
|
2028 |
+
2150,0
|
2029 |
+
1255,0
|
2030 |
+
1367,0
|
2031 |
+
837,0
|
2032 |
+
485,0
|
2033 |
+
1043,1
|
2034 |
+
1118,1
|
2035 |
+
1254,1
|
2036 |
+
490,0
|
2037 |
+
575,0
|
2038 |
+
1808,0
|
2039 |
+
695,0
|
2040 |
+
712,0
|
2041 |
+
2119,0
|
2042 |
+
735,1
|
2043 |
+
2148,0
|
2044 |
+
264,0
|
2045 |
+
1107,1
|
2046 |
+
901,1
|
2047 |
+
2124,0
|
2048 |
+
893,1
|
2049 |
+
646,0
|
2050 |
+
611,0
|
2051 |
+
779,1
|
2052 |
+
1170,0
|
2053 |
+
816,1
|
2054 |
+
1067,1
|
2055 |
+
31,0
|
2056 |
+
523,1
|
2057 |
+
648,1
|
2058 |
+
315,0
|
2059 |
+
639,1
|
2060 |
+
1600,0
|
2061 |
+
42,0
|
2062 |
+
1217,1
|
2063 |
+
71,1
|
2064 |
+
1841,0
|
2065 |
+
278,1
|
2066 |
+
1155,1
|
2067 |
+
148,0
|
2068 |
+
1485,0
|
2069 |
+
1248,0
|
2070 |
+
544,1
|
2071 |
+
1200,0
|
2072 |
+
176,1
|
2073 |
+
1056,0
|
2074 |
+
682,1
|
2075 |
+
1355,0
|
2076 |
+
629,0
|
2077 |
+
1053,0
|
2078 |
+
739,0
|
2079 |
+
162,0
|
2080 |
+
713,0
|
2081 |
+
1350,0
|
2082 |
+
672,1
|
2083 |
+
230,1
|
2084 |
+
2335,0
|
2085 |
+
1239,1
|
2086 |
+
2216,0
|
2087 |
+
1247,1
|
2088 |
+
119,1
|
2089 |
+
1809,0
|
2090 |
+
1166,1
|
2091 |
+
37,0
|
2092 |
+
1280,0
|
2093 |
+
1361,0
|
2094 |
+
726,0
|
2095 |
+
26,0
|
2096 |
+
782,0
|
2097 |
+
361,1
|
2098 |
+
209,1
|
2099 |
+
186,1
|
2100 |
+
107,1
|
2101 |
+
841,1
|
2102 |
+
429,1
|
2103 |
+
1195,0
|
2104 |
+
1167,0
|
2105 |
+
301,0
|
2106 |
+
1911,0
|
2107 |
+
340,1
|
2108 |
+
1304,0
|
2109 |
+
819,0
|
2110 |
+
850,0
|
2111 |
+
1601,0
|
2112 |
+
625,1
|
2113 |
+
1101,0
|
2114 |
+
89,0
|
2115 |
+
2081,0
|
2116 |
+
235,1
|
2117 |
+
1201,0
|
2118 |
+
75,0
|
2119 |
+
710,0
|
2120 |
+
434,1
|
2121 |
+
334,0
|
2122 |
+
1737,0
|
2123 |
+
1065,0
|
2124 |
+
1284,0
|
2125 |
+
276,0
|
2126 |
+
1239,0
|
2127 |
+
138,1
|
2128 |
+
1176,0
|
2129 |
+
392,0
|
2130 |
+
250,0
|
2131 |
+
836,0
|
2132 |
+
260,1
|
2133 |
+
443,1
|
2134 |
+
827,0
|
2135 |
+
725,0
|
2136 |
+
409,0
|
2137 |
+
605,1
|
2138 |
+
1854,0
|
2139 |
+
1364,0
|
2140 |
+
78,1
|
2141 |
+
717,0
|
2142 |
+
329,1
|
2143 |
+
105,0
|
2144 |
+
262,1
|
2145 |
+
912,1
|
2146 |
+
173,0
|
2147 |
+
1752,0
|
2148 |
+
96,0
|
2149 |
+
1146,1
|
2150 |
+
427,1
|
2151 |
+
350,1
|
2152 |
+
2307,0
|
2153 |
+
1481,0
|
2154 |
+
1058,0
|
2155 |
+
547,1
|
2156 |
+
731,1
|
2157 |
+
1321,1
|
2158 |
+
943,0
|
2159 |
+
1624,0
|
2160 |
+
259,0
|
2161 |
+
1209,0
|
2162 |
+
1130,1
|
2163 |
+
1302,1
|
2164 |
+
1117,1
|
2165 |
+
1875,0
|
2166 |
+
555,0
|
2167 |
+
1478,0
|
2168 |
+
422,0
|
2169 |
+
1145,1
|
2170 |
+
1740,0
|
2171 |
+
556,0
|
2172 |
+
424,0
|
2173 |
+
1433,0
|
2174 |
+
2048,0
|
2175 |
+
2051,0
|
2176 |
+
421,1
|
2177 |
+
1282,1
|
2178 |
+
1966,0
|
2179 |
+
1058,1
|
2180 |
+
1233,1
|
2181 |
+
474,0
|
2182 |
+
54,1
|
2183 |
+
169,0
|
2184 |
+
498,1
|
2185 |
+
1196,0
|
2186 |
+
1617,0
|
2187 |
+
143,0
|
2188 |
+
1089,0
|
2189 |
+
847,1
|
2190 |
+
1566,0
|
2191 |
+
681,1
|
2192 |
+
737,1
|
2193 |
+
383,1
|
2194 |
+
1293,1
|
2195 |
+
703,1
|
2196 |
+
1878,0
|
2197 |
+
350,0
|
2198 |
+
1237,0
|
2199 |
+
247,0
|
2200 |
+
966,0
|
2201 |
+
129,1
|
2202 |
+
459,1
|
2203 |
+
699,1
|
2204 |
+
617,1
|
2205 |
+
573,0
|
2206 |
+
188,0
|
2207 |
+
1469,0
|
2208 |
+
952,0
|
2209 |
+
99,0
|
2210 |
+
869,1
|
2211 |
+
1125,0
|
2212 |
+
356,1
|
2213 |
+
1426,0
|
2214 |
+
1731,0
|
2215 |
+
810,1
|
2216 |
+
1090,0
|
2217 |
+
981,0
|
2218 |
+
921,1
|
2219 |
+
289,0
|
2220 |
+
284,0
|
2221 |
+
690,0
|
2222 |
+
1006,1
|
2223 |
+
955,0
|
2224 |
+
1114,0
|
2225 |
+
2084,0
|
2226 |
+
408,0
|
2227 |
+
1258,1
|
2228 |
+
1086,1
|
2229 |
+
825,0
|
2230 |
+
492,0
|
2231 |
+
615,0
|
2232 |
+
1831,0
|
2233 |
+
1882,0
|
2234 |
+
1172,0
|
2235 |
+
961,0
|
2236 |
+
581,1
|
2237 |
+
554,0
|
2238 |
+
1839,0
|
2239 |
+
1174,0
|
2240 |
+
772,1
|
2241 |
+
1619,0
|
2242 |
+
2086,0
|
2243 |
+
754,0
|
2244 |
+
2140,0
|
2245 |
+
1176,1
|
2246 |
+
244,1
|
2247 |
+
26,1
|
2248 |
+
729,0
|
2249 |
+
1138,0
|
2250 |
+
985,0
|
2251 |
+
382,1
|
2252 |
+
1417,0
|
2253 |
+
1641,0
|
2254 |
+
1585,0
|
2255 |
+
960,1
|
2256 |
+
411,0
|
2257 |
+
437,0
|
2258 |
+
1848,0
|
2259 |
+
351,1
|
2260 |
+
217,0
|
2261 |
+
2108,0
|
2262 |
+
1134,0
|
2263 |
+
710,1
|
2264 |
+
695,1
|
2265 |
+
1283,1
|
2266 |
+
2149,0
|
2267 |
+
441,1
|
2268 |
+
1130,0
|
2269 |
+
1457,0
|
2270 |
+
578,1
|
2271 |
+
655,0
|
2272 |
+
1281,0
|
2273 |
+
561,1
|
2274 |
+
506,1
|
2275 |
+
1322,0
|
2276 |
+
527,0
|
2277 |
+
1562,0
|
2278 |
+
927,1
|
2279 |
+
1143,1
|
2280 |
+
256,1
|
2281 |
+
1676,0
|
2282 |
+
500,0
|
2283 |
+
2181,0
|
2284 |
+
1739,0
|
2285 |
+
970,0
|
2286 |
+
360,0
|
2287 |
+
76,1
|
2288 |
+
898,0
|
2289 |
+
1325,0
|
2290 |
+
157,0
|
2291 |
+
741,1
|
2292 |
+
616,0
|
2293 |
+
166,0
|
2294 |
+
925,1
|
2295 |
+
2060,0
|
2296 |
+
226,1
|
2297 |
+
1006,0
|
2298 |
+
1150,1
|
2299 |
+
1225,0
|
2300 |
+
1023,1
|
2301 |
+
1419,0
|
2302 |
+
12,1
|
2303 |
+
1410,0
|
2304 |
+
904,0
|
2305 |
+
1498,0
|
2306 |
+
1741,0
|
2307 |
+
408,1
|
2308 |
+
1029,1
|
2309 |
+
1846,0
|
2310 |
+
1524,0
|
2311 |
+
224,0
|
2312 |
+
225,0
|
2313 |
+
366,1
|
2314 |
+
454,1
|
2315 |
+
333,0
|
2316 |
+
1020,0
|
2317 |
+
1041,0
|
2318 |
+
723,0
|
2319 |
+
1654,0
|
2320 |
+
254,0
|
2321 |
+
49,1
|
2322 |
+
139,0
|
2323 |
+
63,1
|
2324 |
+
415,1
|
2325 |
+
128,0
|
2326 |
+
384,1
|
2327 |
+
1360,0
|
2328 |
+
1775,0
|
2329 |
+
543,1
|
2330 |
+
204,0
|
2331 |
+
351,0
|
2332 |
+
1033,1
|
2333 |
+
1480,0
|
2334 |
+
2117,0
|
2335 |
+
1917,0
|
2336 |
+
920,0
|
2337 |
+
236,0
|
2338 |
+
464,1
|
2339 |
+
998,1
|
2340 |
+
1210,0
|
2341 |
+
1504,0
|
2342 |
+
708,0
|
2343 |
+
1291,0
|
2344 |
+
804,0
|
2345 |
+
1466,0
|
2346 |
+
786,1
|
2347 |
+
461,0
|
2348 |
+
1635,0
|
2349 |
+
913,0
|
2350 |
+
1160,0
|
2351 |
+
750,0
|
2352 |
+
611,1
|
2353 |
+
861,0
|
2354 |
+
1270,1
|
2355 |
+
2050,0
|
2356 |
+
186,0
|
2357 |
+
547,0
|
2358 |
+
911,0
|
2359 |
+
859,1
|
2360 |
+
297,0
|
2361 |
+
45,0
|
2362 |
+
1266,1
|
2363 |
+
1003,0
|
2364 |
+
158,1
|
2365 |
+
1246,1
|
2366 |
+
269,0
|
2367 |
+
874,0
|
2368 |
+
1264,1
|
2369 |
+
2080,0
|
2370 |
+
934,0
|
2371 |
+
1112,1
|
2372 |
+
1067,0
|
2373 |
+
1822,0
|
2374 |
+
213,1
|
2375 |
+
67,0
|
2376 |
+
1693,0
|
2377 |
+
1014,0
|
2378 |
+
1002,1
|
2379 |
+
439,1
|
2380 |
+
588,0
|
2381 |
+
532,1
|
2382 |
+
784,1
|
2383 |
+
1354,0
|
2384 |
+
839,0
|
2385 |
+
442,1
|
2386 |
+
413,0
|
2387 |
+
892,1
|
2388 |
+
1346,0
|
2389 |
+
883,1
|
2390 |
+
319,0
|
2391 |
+
1587,0
|
2392 |
+
259,1
|
2393 |
+
2091,0
|
2394 |
+
602,0
|
2395 |
+
791,1
|
2396 |
+
184,0
|
2397 |
+
1399,0
|
2398 |
+
178,0
|
2399 |
+
707,1
|
2400 |
+
375,0
|
2401 |
+
1565,0
|
2402 |
+
461,1
|
2403 |
+
99,1
|
2404 |
+
1184,0
|
2405 |
+
2270,0
|
2406 |
+
1327,0
|
2407 |
+
676,1
|
2408 |
+
212,0
|
2409 |
+
437,1
|
2410 |
+
858,1
|
2411 |
+
1281,1
|
2412 |
+
1016,0
|
2413 |
+
1223,1
|
2414 |
+
610,1
|
2415 |
+
49,0
|
2416 |
+
1313,0
|
2417 |
+
574,0
|
2418 |
+
928,0
|
2419 |
+
1973,0
|
2420 |
+
665,0
|
2421 |
+
502,0
|
2422 |
+
1296,1
|
2423 |
+
1887,0
|
2424 |
+
1149,1
|
2425 |
+
866,0
|
2426 |
+
65,0
|
2427 |
+
342,0
|
2428 |
+
967,0
|
2429 |
+
412,1
|
2430 |
+
1061,1
|
2431 |
+
1803,0
|
2432 |
+
666,1
|
2433 |
+
325,1
|
2434 |
+
1024,0
|
2435 |
+
596,1
|
2436 |
+
1816,0
|
2437 |
+
1814,0
|
2438 |
+
1020,1
|
2439 |
+
2357,0
|
2440 |
+
1287,1
|
2441 |
+
1881,0
|
2442 |
+
163,1
|
2443 |
+
743,1
|
2444 |
+
2058,0
|
2445 |
+
1393,0
|
2446 |
+
668,1
|
2447 |
+
693,1
|
2448 |
+
837,1
|
2449 |
+
1397,0
|
2450 |
+
294,0
|
2451 |
+
1658,0
|
2452 |
+
902,1
|
2453 |
+
747,0
|
2454 |
+
884,0
|
2455 |
+
296,0
|
2456 |
+
859,0
|
2457 |
+
1021,0
|
2458 |
+
481,0
|
2459 |
+
1634,0
|
2460 |
+
2011,0
|
2461 |
+
1252,1
|
2462 |
+
989,1
|
2463 |
+
258,1
|
2464 |
+
916,1
|
2465 |
+
2168,0
|
2466 |
+
1246,0
|
2467 |
+
1392,0
|
2468 |
+
788,0
|
2469 |
+
465,0
|
2470 |
+
288,0
|
2471 |
+
969,1
|
2472 |
+
776,1
|
2473 |
+
319,1
|
2474 |
+
1837,0
|
2475 |
+
1362,0
|
2476 |
+
875,0
|
2477 |
+
593,1
|
2478 |
+
2024,0
|
2479 |
+
1537,0
|
2480 |
+
323,0
|
2481 |
+
995,1
|
2482 |
+
567,1
|
2483 |
+
1235,1
|
2484 |
+
531,1
|
2485 |
+
2008,0
|
2486 |
+
146,1
|
2487 |
+
331,0
|
2488 |
+
215,1
|
2489 |
+
218,1
|
2490 |
+
1938,0
|
2491 |
+
1698,0
|
2492 |
+
993,1
|
2493 |
+
1202,1
|
2494 |
+
1309,1
|
2495 |
+
266,1
|
2496 |
+
390,0
|
2497 |
+
1697,0
|
2498 |
+
1153,0
|
2499 |
+
863,1
|
2500 |
+
716,0
|
2501 |
+
403,0
|
2502 |
+
125,0
|
2503 |
+
1845,0
|
2504 |
+
1152,1
|
2505 |
+
1212,1
|
2506 |
+
446,1
|
2507 |
+
2037,0
|
2508 |
+
1194,1
|
2509 |
+
1236,0
|
2510 |
+
1004,1
|
2511 |
+
2298,0
|
2512 |
+
797,0
|
2513 |
+
1486,0
|
2514 |
+
1353,0
|
2515 |
+
1359,0
|
2516 |
+
1335,0
|
2517 |
+
667,1
|
2518 |
+
1182,1
|
2519 |
+
1329,0
|
2520 |
+
1068,0
|
2521 |
+
594,0
|
2522 |
+
1109,0
|
2523 |
+
87,1
|
2524 |
+
1492,0
|
2525 |
+
873,1
|
2526 |
+
507,0
|
2527 |
+
650,0
|
2528 |
+
286,1
|
2529 |
+
1104,0
|
2530 |
+
39,0
|
2531 |
+
858,0
|
2532 |
+
1663,0
|
2533 |
+
298,1
|
2534 |
+
777,1
|
2535 |
+
703,0
|
2536 |
+
769,1
|
2537 |
+
318,0
|
2538 |
+
1245,1
|
2539 |
+
1412,0
|
2540 |
+
590,0
|
2541 |
+
618,1
|
2542 |
+
576,1
|
2543 |
+
428,0
|
2544 |
+
1190,1
|
2545 |
+
2206,0
|
2546 |
+
1369,0
|
2547 |
+
340,0
|
2548 |
+
1483,0
|
2549 |
+
61,1
|
2550 |
+
1325,1
|
2551 |
+
572,0
|
2552 |
+
1187,1
|
2553 |
+
2056,0
|
2554 |
+
1968,0
|
2555 |
+
1225,1
|
2556 |
+
851,0
|
2557 |
+
1825,0
|
2558 |
+
591,0
|
2559 |
+
810,0
|
2560 |
+
845,1
|
2561 |
+
1099,0
|
2562 |
+
744,0
|
2563 |
+
380,0
|
2564 |
+
2157,0
|
2565 |
+
1507,0
|
2566 |
+
1173,1
|
2567 |
+
670,0
|
2568 |
+
154,1
|
2569 |
+
212,1
|
2570 |
+
1931,0
|
2571 |
+
1732,0
|
2572 |
+
7,0
|
2573 |
+
990,0
|
2574 |
+
1028,1
|
2575 |
+
1087,0
|
2576 |
+
84,1
|
2577 |
+
1793,0
|
2578 |
+
1129,1
|
2579 |
+
1126,1
|
2580 |
+
220,1
|
2581 |
+
733,1
|
2582 |
+
384,0
|
2583 |
+
1429,0
|
2584 |
+
122,1
|
2585 |
+
394,1
|
2586 |
+
2103,0
|
2587 |
+
1771,0
|
2588 |
+
174,0
|
2589 |
+
33,0
|
2590 |
+
1200,1
|
2591 |
+
310,1
|
2592 |
+
1036,1
|
2593 |
+
688,0
|
2594 |
+
2133,0
|
2595 |
+
902,0
|
2596 |
+
779,0
|
2597 |
+
1879,0
|
2598 |
+
1332,0
|
2599 |
+
17,0
|
2600 |
+
1170,1
|
2601 |
+
1295,0
|
2602 |
+
159,0
|
2603 |
+
900,1
|
2604 |
+
1883,0
|
2605 |
+
434,0
|
2606 |
+
878,1
|
2607 |
+
2035,0
|
2608 |
+
2377,0
|
2609 |
+
1643,0
|
2610 |
+
2315,0
|
2611 |
+
686,0
|
2612 |
+
448,1
|
2613 |
+
62,0
|
2614 |
+
102,1
|
2615 |
+
68,0
|
2616 |
+
2277,0
|
2617 |
+
0,0
|
2618 |
+
238,0
|
2619 |
+
1188,0
|
2620 |
+
431,0
|
2621 |
+
194,1
|
2622 |
+
202,0
|
2623 |
+
44,1
|
2624 |
+
421,0
|
2625 |
+
466,1
|
2626 |
+
1493,0
|
2627 |
+
546,0
|
2628 |
+
1754,0
|
2629 |
+
1627,0
|
2630 |
+
64,0
|
2631 |
+
482,0
|
2632 |
+
253,0
|
2633 |
+
2123,0
|
2634 |
+
1470,0
|
2635 |
+
720,0
|
2636 |
+
2219,0
|
2637 |
+
649,0
|
2638 |
+
1180,1
|
2639 |
+
305,1
|
2640 |
+
516,0
|
2641 |
+
132,0
|
2642 |
+
864,1
|
2643 |
+
140,1
|
2644 |
+
1774,0
|
2645 |
+
2159,0
|
2646 |
+
165,1
|
2647 |
+
13,0
|
2648 |
+
1196,1
|
2649 |
+
451,1
|
2650 |
+
980,0
|
2651 |
+
954,0
|
2652 |
+
1778,0
|
2653 |
+
480,0
|
2654 |
+
1012,0
|
2655 |
+
280,0
|
2656 |
+
205,1
|
2657 |
+
366,0
|
2658 |
+
529,0
|
2659 |
+
1029,0
|
2660 |
+
560,0
|
2661 |
+
1045,0
|
2662 |
+
1038,1
|
2663 |
+
1218,0
|
2664 |
+
1284,1
|
2665 |
+
292,0
|
2666 |
+
1749,0
|
2667 |
+
1811,0
|
2668 |
+
914,0
|
2669 |
+
2064,0
|
2670 |
+
232,0
|
2671 |
+
1085,1
|
2672 |
+
480,1
|
2673 |
+
1135,0
|
2674 |
+
656,1
|
2675 |
+
307,0
|
2676 |
+
1536,0
|
2677 |
+
728,0
|
2678 |
+
711,0
|
2679 |
+
897,1
|
2680 |
+
513,1
|
2681 |
+
1229,0
|
2682 |
+
1111,0
|
2683 |
+
291,0
|
2684 |
+
1431,0
|
2685 |
+
118,0
|
2686 |
+
172,0
|
2687 |
+
1161,0
|
2688 |
+
497,1
|
2689 |
+
2271,0
|
2690 |
+
257,0
|
2691 |
+
1258,0
|
2692 |
+
998,0
|
2693 |
+
1097,0
|
2694 |
+
1206,1
|
2695 |
+
1550,0
|
2696 |
+
445,1
|
2697 |
+
509,1
|
2698 |
+
287,0
|
2699 |
+
79,1
|
2700 |
+
833,0
|
2701 |
+
360,1
|
2702 |
+
101,1
|
2703 |
+
23,1
|
2704 |
+
1370,0
|
2705 |
+
143,1
|
2706 |
+
1027,1
|
2707 |
+
871,0
|
2708 |
+
1785,0
|
2709 |
+
55,0
|
2710 |
+
944,1
|
2711 |
+
880,1
|
2712 |
+
1158,1
|
2713 |
+
942,0
|
2714 |
+
2178,0
|
2715 |
+
1320,1
|
2716 |
+
515,0
|
2717 |
+
1088,1
|
2718 |
+
1371,0
|
2719 |
+
957,1
|
2720 |
+
1191,1
|
2721 |
+
1407,0
|
2722 |
+
835,0
|
2723 |
+
155,0
|
2724 |
+
1049,0
|
2725 |
+
987,1
|
2726 |
+
696,0
|
2727 |
+
2313,0
|
2728 |
+
468,0
|
2729 |
+
382,0
|
2730 |
+
821,1
|
2731 |
+
455,1
|
2732 |
+
1124,1
|
2733 |
+
1250,0
|
2734 |
+
949,1
|
2735 |
+
798,1
|
2736 |
+
957,0
|
2737 |
+
1041,1
|
2738 |
+
1706,0
|
2739 |
+
267,1
|
2740 |
+
1444,0
|
2741 |
+
1118,0
|
2742 |
+
1408,0
|
2743 |
+
77,0
|
2744 |
+
1757,0
|
2745 |
+
2132,0
|
2746 |
+
495,0
|
2747 |
+
1510,0
|
2748 |
+
853,1
|
2749 |
+
1222,1
|
2750 |
+
586,0
|
2751 |
+
614,0
|
2752 |
+
1742,0
|
2753 |
+
734,1
|
2754 |
+
2142,0
|
2755 |
+
245,0
|
2756 |
+
877,1
|
2757 |
+
888,0
|
2758 |
+
1257,0
|
2759 |
+
700,0
|
2760 |
+
249,0
|
2761 |
+
2370,0
|
2762 |
+
1497,0
|
2763 |
+
1240,1
|
2764 |
+
1015,0
|
2765 |
+
127,1
|
2766 |
+
848,1
|
2767 |
+
1167,1
|
2768 |
+
1206,0
|
2769 |
+
1864,0
|
2770 |
+
941,0
|
2771 |
+
2141,0
|
2772 |
+
2283,0
|
dataset/task1_data.pkl
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:046121ecaa796a0d453ce75820b5b6d53d468a03b7352074029504c9f96e3c32
|
3 |
+
size 4611306
|
src/deployment_utils.py
ADDED
@@ -0,0 +1,607 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
# global
|
2 |
+
from typing import Tuple, List
|
3 |
+
import re
|
4 |
+
import numpy as np
|
5 |
+
import pandas as pd
|
6 |
+
|
7 |
+
import tensorflow as tf
|
8 |
+
from tensorflow import keras
|
9 |
+
from keras.utils import pad_sequences
|
10 |
+
from keras.preprocessing.text import Tokenizer
|
11 |
+
from gensim.models.doc2vec import Doc2Vec
|
12 |
+
|
13 |
+
import transformers
|
14 |
+
from transformers import pipeline, BertTokenizer
|
15 |
+
|
16 |
+
import fasttext
|
17 |
+
|
18 |
+
# local
|
19 |
+
from preprocessing import Preprocessor
|
20 |
+
from utils import read_data
|
21 |
+
|
22 |
+
|
23 |
+
# read data
|
24 |
+
X_train, X_test, y_train, y_test = read_data()
|
25 |
+
|
26 |
+
# instantiate preprocessor object
|
27 |
+
preprocessor = Preprocessor()
|
28 |
+
|
29 |
+
# load models
|
30 |
+
doc2vec_model_embeddings = Doc2Vec.load(
|
31 |
+
"F:/Graduation Project/Project/models/best_doc2vec_embeddings")
|
32 |
+
doc2vec_model = keras.models.load_model(
|
33 |
+
"F:/Graduation Project/Project/models/best_doc2vec_model.h5")
|
34 |
+
tfidf_model = keras.models.load_model(
|
35 |
+
"F:/Graduation Project/Project/models/best_tfidf_model.h5")
|
36 |
+
cnn_model = keras.models.load_model(
|
37 |
+
"F:/Graduation Project/Project/models/best_cnn_model.h5")
|
38 |
+
glove_model = keras.models.load_model(
|
39 |
+
"F:/Graduation Project/Project/models/best_glove_model.h5")
|
40 |
+
lstm_model = keras.models.load_model(
|
41 |
+
"F:/Graduation Project/Project/models/best_lstm_model.h5")
|
42 |
+
bert_model = keras.models.load_model(
|
43 |
+
"F:/Graduation Project/Project/models/best_bert_model.h5", custom_objects={"TFBertModel": transformers.TFBertModel})
|
44 |
+
fasttext_model = fasttext.load_model(
|
45 |
+
"F:/Graduation Project/Project/models/best_fasttext_model.bin")
|
46 |
+
summarization_model = pipeline(
|
47 |
+
"summarization", model="facebook/bart-large-cnn")
|
48 |
+
|
49 |
+
|
50 |
+
# TODO: Add Docstrings
|
51 |
+
def extract_case_information(case_content: str):
|
52 |
+
content_list = case_content.split("\n")
|
53 |
+
petitioner = re.findall(r"petitioner:(.+)", content_list[0])[0]
|
54 |
+
respondent = re.findall(r"respondent:(.+)", content_list[1])[0]
|
55 |
+
facts = re.findall(r"facts:(.+)", content_list[2])[0]
|
56 |
+
|
57 |
+
return petitioner, respondent, facts
|
58 |
+
|
59 |
+
|
60 |
+
def generate_random_sample() -> Tuple[str, str, str, int]:
|
61 |
+
"""
|
62 |
+
Randomly fetch a random case from `X_test` to test it.
|
63 |
+
|
64 |
+
Returns:
|
65 |
+
--------
|
66 |
+
A tuple contains the following:
|
67 |
+
- petitioner : str
|
68 |
+
Contains petitioner name.
|
69 |
+
- respondent : str
|
70 |
+
Contains respondent name.
|
71 |
+
- facts : str
|
72 |
+
Contains case facts.
|
73 |
+
- label : int
|
74 |
+
Represents the winning index(0 = petitioner, 1 = respondent).
|
75 |
+
"""
|
76 |
+
|
77 |
+
random_idx = np.random.randint(low=0, high=len(X_test))
|
78 |
+
|
79 |
+
petitioner = X_test["first_party"].iloc[random_idx]
|
80 |
+
respondent = X_test["second_party"].iloc[random_idx]
|
81 |
+
facts = X_test["Facts"].iloc[random_idx]
|
82 |
+
label = y_test.iloc[random_idx][0]
|
83 |
+
|
84 |
+
return petitioner, respondent, facts, label
|
85 |
+
|
86 |
+
|
87 |
+
def generate_highlighted_words(facts: str, petitioner_words: List[str], respondent_words: List[str]):
|
88 |
+
"""
|
89 |
+
Highlight `petitioner_words` and `respondent_words` for model
|
90 |
+
interpretation.
|
91 |
+
|
92 |
+
Parameters:
|
93 |
+
-----------
|
94 |
+
- facts : str
|
95 |
+
Facts of a specific case.
|
96 |
+
- petitioner_words : List[str]
|
97 |
+
Contains all words that model pays attention
|
98 |
+
to be a petetioner words.
|
99 |
+
- respondent_words : List[str]
|
100 |
+
Contains all words that model pays attention
|
101 |
+
to be a respondent words.
|
102 |
+
|
103 |
+
Returns:
|
104 |
+
--------
|
105 |
+
- rendered_text : str
|
106 |
+
Contains the `facts` but with adding
|
107 |
+
highlighting mechanism to visualize it using CSS in HTML format.
|
108 |
+
|
109 |
+
Example:
|
110 |
+
--------
|
111 |
+
>>> facts_ = 'Mohammed shot Aly after a hot negotiation happened between
|
112 |
+
... them about the profits of their company'
|
113 |
+
>>> petitioner_words_ = ['shot', 'hot']
|
114 |
+
>>> respondent_words_ = ['profits']
|
115 |
+
>>> generate_highlighted_words(facts, petitioner_words_, respondent_words_)
|
116 |
+
|
117 |
+
>>> output:
|
118 |
+
<div class='text-facts'> Mohammed <span class='highlight-petitioner'>shot</span>
|
119 |
+
Aly after a <span class='highlight-petitioner'>hot</span> negotiation happened
|
120 |
+
between them about <span class='highlight-respondent'>profits</span> of their
|
121 |
+
company </div>
|
122 |
+
"""
|
123 |
+
|
124 |
+
rendered_text = '<div class="text-facts"> '
|
125 |
+
|
126 |
+
for word in facts.split():
|
127 |
+
if word in petitioner_words:
|
128 |
+
highlight_word = ' <span class="highlight-petitioner"> ' + word + " </span> "
|
129 |
+
rendered_text += highlight_word
|
130 |
+
|
131 |
+
elif word in respondent_words:
|
132 |
+
highlight_word = ' <span class="highlight-respondent"> ' + word + " </span> "
|
133 |
+
rendered_text += highlight_word
|
134 |
+
|
135 |
+
else:
|
136 |
+
rendered_text += " " + word
|
137 |
+
|
138 |
+
rendered_text += " </div>"
|
139 |
+
|
140 |
+
return rendered_text
|
141 |
+
|
142 |
+
|
143 |
+
class VectorizerGenerator:
|
144 |
+
"""Responsible for creation and generation of tokenizers and text
|
145 |
+
vectorizers for JudgerAIs' models"""
|
146 |
+
|
147 |
+
def __init__(self) -> None:
|
148 |
+
pass
|
149 |
+
|
150 |
+
def generate_tf_idf_vectorizer(self) -> keras.layers.TextVectorization:
|
151 |
+
"""
|
152 |
+
Generating best text vectroizer of the tf-idf model (3rd combination).
|
153 |
+
|
154 |
+
Returns:
|
155 |
+
-------
|
156 |
+
- text_vectorizer : keras.layers.TextVectorization
|
157 |
+
Represents the case facts' vectorizer that converts case facts to
|
158 |
+
numerical tensors.
|
159 |
+
"""
|
160 |
+
|
161 |
+
first_party_names = X_train["first_party"]
|
162 |
+
second_party_names = X_train["second_party"]
|
163 |
+
facts = X_train["Facts"]
|
164 |
+
|
165 |
+
anonymized_facts = preprocessor.anonymize_data(
|
166 |
+
first_party_names, second_party_names, facts)
|
167 |
+
|
168 |
+
text_vectorizer, _ = preprocessor.convert_text_to_vectors_tf_idf(
|
169 |
+
anonymized_facts)
|
170 |
+
|
171 |
+
return text_vectorizer
|
172 |
+
|
173 |
+
def generate_cnn_vectorizer(self) -> keras.layers.TextVectorization:
|
174 |
+
"""
|
175 |
+
Generating best text vectroizer of the cnn model (2nd combination).
|
176 |
+
|
177 |
+
Returns:
|
178 |
+
-------
|
179 |
+
- text_vectorizer : keras.layers.TextVectorization
|
180 |
+
Represents the case facts' vectorizer that converts case facts to
|
181 |
+
numerical tensors.
|
182 |
+
"""
|
183 |
+
|
184 |
+
balanced_df = preprocessor.balance_data(X_train["Facts"], y_train)
|
185 |
+
X_train_balanced = balanced_df["Facts"]
|
186 |
+
|
187 |
+
text_vectorizer, _ = preprocessor.convert_text_to_vectors_cnn(
|
188 |
+
X_train_balanced)
|
189 |
+
|
190 |
+
return text_vectorizer
|
191 |
+
|
192 |
+
def generate_glove_tokenizer(self) -> keras.preprocessing.text.Tokenizer:
|
193 |
+
"""
|
194 |
+
Generating best glove tokenizer of the GloVe model (2nd combination).
|
195 |
+
|
196 |
+
Returns:
|
197 |
+
-------
|
198 |
+
- glove_tokenizer : keras.preprocessing.text.Tokenizer
|
199 |
+
Represents the case facts' tokenizer that converts case facts to
|
200 |
+
numerical tensors.
|
201 |
+
"""
|
202 |
+
|
203 |
+
balanced_df = preprocessor.balance_data(X_train["Facts"], y_train)
|
204 |
+
X_train_balanced = balanced_df["Facts"]
|
205 |
+
|
206 |
+
glove_tokenizer, _ = preprocessor.convert_text_to_vectors_glove(
|
207 |
+
X_train_balanced)
|
208 |
+
|
209 |
+
return glove_tokenizer
|
210 |
+
|
211 |
+
def generate_lstm_tokenizer(self) -> keras.preprocessing.text.Tokenizer:
|
212 |
+
"""
|
213 |
+
Generating best text tokenizer of the LSTM model (1st combination).
|
214 |
+
|
215 |
+
Returns:
|
216 |
+
-------
|
217 |
+
- lstm_tokenizer : keras.preprocessing.text.Tokenizer
|
218 |
+
Represents the case facts' tokenizer that converts case facts to
|
219 |
+
numerical tensors.
|
220 |
+
"""
|
221 |
+
|
222 |
+
lstm_tokenizer = Tokenizer(num_words=18430)
|
223 |
+
lstm_tokenizer.fit_on_texts(X_train)
|
224 |
+
|
225 |
+
return lstm_tokenizer
|
226 |
+
|
227 |
+
def generate_bert_tokenizer(self) -> transformers.BertTokenizer:
|
228 |
+
"""
|
229 |
+
Generating best bert tokenizer of the BERT model (1st combination).
|
230 |
+
|
231 |
+
Returns:
|
232 |
+
-------
|
233 |
+
- bert_tokenizer : transformers.BertTokenizer
|
234 |
+
Represents the case facts' tokenizer that converts case facts to
|
235 |
+
input ids tensors.
|
236 |
+
"""
|
237 |
+
|
238 |
+
bert_tokenizer = BertTokenizer.from_pretrained('bert-base-cased')
|
239 |
+
return bert_tokenizer
|
240 |
+
|
241 |
+
|
242 |
+
class DataPreparator:
|
243 |
+
"""Responsible for preparing the case facts aka converting case facts to
|
244 |
+
numerical vectors using `VectorizerGenerator` object."""
|
245 |
+
|
246 |
+
def __init__(self) -> None:
|
247 |
+
self.vectorizer_generator = VectorizerGenerator()
|
248 |
+
|
249 |
+
def prepare_doc2vec(self, facts: str) -> pd.DataFrame:
|
250 |
+
"""
|
251 |
+
Responsible for converting `facts` string to numerical vector
|
252 |
+
using `doc2vec_model_embeddings`.
|
253 |
+
|
254 |
+
Parameters:
|
255 |
+
----------
|
256 |
+
- facts : str
|
257 |
+
Represents the case facts.
|
258 |
+
|
259 |
+
Returns:
|
260 |
+
-------
|
261 |
+
- facts_vector : pd.DataFrame
|
262 |
+
A row DataFrame represents the 50-d vector of the `facts`.
|
263 |
+
"""
|
264 |
+
|
265 |
+
facts = pd.Series(facts)
|
266 |
+
facts_processed = preprocessor.preprocess_data(facts)
|
267 |
+
facts_vectors = preprocessor.convert_text_to_vectors_doc2vec(
|
268 |
+
facts_processed, train=False, embeddings_doc2vec=doc2vec_model_embeddings)
|
269 |
+
|
270 |
+
return facts_vectors
|
271 |
+
|
272 |
+
def _anonymize_facts(self, first_party_name: str, second_party_name: str, facts: str) -> str:
|
273 |
+
"""
|
274 |
+
Anonymize case `facts` by replacing `first_party_name` & `second_party_name` with
|
275 |
+
generic tag "__PARTY__".
|
276 |
+
|
277 |
+
Parameters:
|
278 |
+
-----------
|
279 |
+
- first_party_name : str
|
280 |
+
Represents the petitioner name.
|
281 |
+
- second_party_name : str
|
282 |
+
Represents the respondent name.
|
283 |
+
- facts : str
|
284 |
+
Represents the case facts.
|
285 |
+
|
286 |
+
Returns:
|
287 |
+
-------
|
288 |
+
- anonymized_facts : str
|
289 |
+
Represents `facts` after anonymization.
|
290 |
+
"""
|
291 |
+
|
292 |
+
anonymized_facts = preprocessor._anonymize_case_facts(
|
293 |
+
first_party_name, second_party_name, facts)
|
294 |
+
|
295 |
+
return anonymized_facts
|
296 |
+
|
297 |
+
def prepare_tf_idf(self, anonymized_facts: str) -> tf.Tensor:
|
298 |
+
"""
|
299 |
+
Responsible for converting `facts` string to numerical vector
|
300 |
+
using tf-idf `vectorizer_generator` in the 3rd combination.
|
301 |
+
|
302 |
+
Parameters:
|
303 |
+
-----------
|
304 |
+
- anonymized_facts : str
|
305 |
+
Represents the case facts after anonymization.
|
306 |
+
|
307 |
+
Returns:
|
308 |
+
-------
|
309 |
+
- facts_vector : tf.Tensor
|
310 |
+
A Tensor of 10000-d represents `facts`.
|
311 |
+
"""
|
312 |
+
|
313 |
+
anonymized_facts = pd.Series(anonymized_facts)
|
314 |
+
tf_idf_vectorizer = self.vectorizer_generator.generate_tf_idf_vectorizer()
|
315 |
+
|
316 |
+
facts_vector = preprocessor.convert_text_to_vectors_tf_idf(
|
317 |
+
anonymized_facts, train=False, text_vectorizer=tf_idf_vectorizer)
|
318 |
+
|
319 |
+
return facts_vector
|
320 |
+
|
321 |
+
def prepare_cnn(self, facts: str) -> tf.Tensor:
|
322 |
+
"""
|
323 |
+
Responsible for converting `facts` string to numerical vector
|
324 |
+
using cnn `vectorizer_generator` in the 2nd combination.
|
325 |
+
|
326 |
+
Parameters:
|
327 |
+
-----------
|
328 |
+
- facts : str
|
329 |
+
Represents the case facts.
|
330 |
+
|
331 |
+
Returns:
|
332 |
+
-------
|
333 |
+
- facts_vector : tf.Tensor
|
334 |
+
A Tensor of 2000-d represents `facts`.
|
335 |
+
"""
|
336 |
+
facts = pd.Series(facts)
|
337 |
+
|
338 |
+
cnn_vectorizer = self.vectorizer_generator.generate_cnn_vectorizer()
|
339 |
+
|
340 |
+
facts_vector = preprocessor.convert_text_to_vectors_cnn(
|
341 |
+
facts, train=False, text_vectorizer=cnn_vectorizer)
|
342 |
+
|
343 |
+
return facts_vector
|
344 |
+
|
345 |
+
def prepare_glove(self, facts: str) -> np.ndarray:
|
346 |
+
"""
|
347 |
+
Responsible for converting `facts` string to numerical vector
|
348 |
+
using glove `vectorizer_generator` in the 2nd combination.
|
349 |
+
|
350 |
+
Parameters:
|
351 |
+
-----------
|
352 |
+
- facts : str
|
353 |
+
Represents the case facts.
|
354 |
+
|
355 |
+
Returns:
|
356 |
+
-------
|
357 |
+
- facts_vector : np.ndarray
|
358 |
+
A nd.ndarray of 50-d represents `facts`.
|
359 |
+
"""
|
360 |
+
|
361 |
+
facts = pd.Series(facts)
|
362 |
+
|
363 |
+
glove_tokneizer = self.vectorizer_generator.generate_glove_tokenizer()
|
364 |
+
|
365 |
+
facts_vector = preprocessor.convert_text_to_vectors_glove(
|
366 |
+
facts, train=False, glove_tokenizer=glove_tokneizer)
|
367 |
+
|
368 |
+
return facts_vector
|
369 |
+
|
370 |
+
def prepare_lstm(self, facts: str) -> np.ndarray:
|
371 |
+
"""
|
372 |
+
Responsible for converting `facts` string to numerical vector
|
373 |
+
using lstm `vectorizer_generator` in the 1st combination.
|
374 |
+
|
375 |
+
Parameters:
|
376 |
+
-----------
|
377 |
+
- facts : str
|
378 |
+
Represents the case facts.
|
379 |
+
|
380 |
+
Returns:
|
381 |
+
-------
|
382 |
+
- facts_vector_padded : np.ndarray
|
383 |
+
A nd.ndarray of 974-d represents `facts`.
|
384 |
+
"""
|
385 |
+
|
386 |
+
facts = pd.Series(facts)
|
387 |
+
lstm_tokenizer = self.vectorizer_generator.generate_lstm_tokenizer()
|
388 |
+
facts_vector = lstm_tokenizer.texts_to_sequences(facts)
|
389 |
+
facts_vector_padded = pad_sequences(facts_vector, 974)
|
390 |
+
|
391 |
+
return facts_vector_padded
|
392 |
+
|
393 |
+
def prepare_bert(self, facts: str) -> tf.Tensor:
|
394 |
+
"""
|
395 |
+
Responsible for converting `facts` string to numerical vector
|
396 |
+
using bert `vectorizer_generator` in the 1st combination.
|
397 |
+
|
398 |
+
Parameters:
|
399 |
+
-----------
|
400 |
+
- facts : str
|
401 |
+
Represents the case facts.
|
402 |
+
|
403 |
+
Returns:
|
404 |
+
-------
|
405 |
+
- tf.Tensor
|
406 |
+
A tf.Tensor of 256-d represents `facts` input ids.
|
407 |
+
"""
|
408 |
+
|
409 |
+
bert_tokenizer = self.vectorizer_generator.generate_bert_tokenizer()
|
410 |
+
facts_vector_dict = bert_tokenizer.encode_plus(
|
411 |
+
facts,
|
412 |
+
max_length=256,
|
413 |
+
truncation=True,
|
414 |
+
padding='max_length',
|
415 |
+
add_special_tokens=True,
|
416 |
+
return_tensors='tf'
|
417 |
+
)
|
418 |
+
|
419 |
+
return facts_vector_dict["input_ids"]
|
420 |
+
|
421 |
+
|
422 |
+
class Predictor:
|
423 |
+
"""Responsible for get predictions of JudgerAIs' models"""
|
424 |
+
|
425 |
+
def __init__(self) -> None:
|
426 |
+
self.data_preparator = DataPreparator()
|
427 |
+
|
428 |
+
def predict_doc2vec(self, facts: str) -> np.ndarray:
|
429 |
+
"""
|
430 |
+
Get prediction of `facts` using `doc2vec_model`.
|
431 |
+
|
432 |
+
Parameters:
|
433 |
+
----------
|
434 |
+
- facts : str
|
435 |
+
Represents the case facts.
|
436 |
+
|
437 |
+
Returns:
|
438 |
+
--------
|
439 |
+
- pet_res_scores : np.ndarray
|
440 |
+
An array contains 2 elements, one for probability of petitioner winning
|
441 |
+
and the second for the probability of respondent winning.
|
442 |
+
"""
|
443 |
+
|
444 |
+
facts_vector = self.data_preparator.prepare_doc2vec(facts)
|
445 |
+
predictions = doc2vec_model.predict(facts_vector)
|
446 |
+
|
447 |
+
pet_res_scores = []
|
448 |
+
for i in predictions:
|
449 |
+
temp = i[0]
|
450 |
+
pet_res_scores.append(np.array([1 - temp, temp]))
|
451 |
+
|
452 |
+
return np.array(pet_res_scores)
|
453 |
+
|
454 |
+
def predict_tf_idf(self, anonymized_facts: str) -> np.ndarray:
|
455 |
+
"""
|
456 |
+
Get prediction of `facts` using `tfidf_model`.
|
457 |
+
|
458 |
+
Parameters:
|
459 |
+
-----------
|
460 |
+
- anonymized_facts : str
|
461 |
+
Represents the case facts after anonymization.
|
462 |
+
|
463 |
+
Returns:
|
464 |
+
--------
|
465 |
+
- pet_res_scores : np.ndarray
|
466 |
+
An array contains 2 elements, one for probability of petitioner winning
|
467 |
+
and the second for the probability of respondent winning.
|
468 |
+
"""
|
469 |
+
|
470 |
+
facts_vector = self.data_preparator.prepare_tf_idf(anonymized_facts)
|
471 |
+
predictions = tfidf_model.predict(facts_vector)
|
472 |
+
|
473 |
+
pet_res_scores = []
|
474 |
+
for i in predictions:
|
475 |
+
temp = i[0]
|
476 |
+
pet_res_scores.append(np.array([1 - temp, temp]))
|
477 |
+
|
478 |
+
return np.array(pet_res_scores)
|
479 |
+
|
480 |
+
def predict_cnn(self, facts: str) -> np.ndarray:
|
481 |
+
"""
|
482 |
+
Get prediction of `facts` using `cnn_model`.
|
483 |
+
|
484 |
+
Parameters:
|
485 |
+
----------
|
486 |
+
- facts : str
|
487 |
+
Represents the case facts.
|
488 |
+
|
489 |
+
Returns:
|
490 |
+
--------
|
491 |
+
- pet_res_scores : np.ndarray
|
492 |
+
An array contains 2 elements, one for probability of petitioner winning
|
493 |
+
and the second for the probability of respondent winning.
|
494 |
+
"""
|
495 |
+
|
496 |
+
facts_vector = self.data_preparator.prepare_cnn(facts)
|
497 |
+
predictions = cnn_model.predict(facts_vector)
|
498 |
+
|
499 |
+
pet_res_scores = []
|
500 |
+
for i in predictions:
|
501 |
+
temp = i[0]
|
502 |
+
pet_res_scores.append(np.array([1 - temp, temp]))
|
503 |
+
|
504 |
+
return np.array(pet_res_scores)
|
505 |
+
|
506 |
+
def predict_glove(self, facts: str) -> np.ndarray:
|
507 |
+
"""
|
508 |
+
Get prediction of `facts` using `glove_model`.
|
509 |
+
|
510 |
+
Parameters:
|
511 |
+
----------
|
512 |
+
- facts : str
|
513 |
+
Represents the case facts.
|
514 |
+
|
515 |
+
Returns:
|
516 |
+
--------
|
517 |
+
- pet_res_scores : np.ndarray
|
518 |
+
An array contains 2 elements, one for probability of petitioner winning
|
519 |
+
and the second for the probability of respondent winning.
|
520 |
+
"""
|
521 |
+
|
522 |
+
facts_vector = self.data_preparator.prepare_glove(facts)
|
523 |
+
predictions = glove_model.predict(facts_vector)
|
524 |
+
|
525 |
+
pet_res_scores = []
|
526 |
+
for i in predictions:
|
527 |
+
temp = i[0]
|
528 |
+
pet_res_scores.append(np.array([1 - temp, temp]))
|
529 |
+
|
530 |
+
return np.array(pet_res_scores)
|
531 |
+
|
532 |
+
def predict_lstm(self, facts: str) -> np.ndarray:
|
533 |
+
"""
|
534 |
+
Get prediction of `facts` using `lstm_model`.
|
535 |
+
|
536 |
+
Parameters:
|
537 |
+
----------
|
538 |
+
- facts : str
|
539 |
+
Represents the case facts.
|
540 |
+
|
541 |
+
Returns:
|
542 |
+
--------
|
543 |
+
- pet_res_scores : np.ndarray
|
544 |
+
An array contains 2 elements, one for probability of petitioner winning
|
545 |
+
and the second for the probability of respondent winning.
|
546 |
+
"""
|
547 |
+
|
548 |
+
facts_vector = self.data_preparator.prepare_lstm(facts)
|
549 |
+
predictions = lstm_model.predict(facts_vector)
|
550 |
+
|
551 |
+
pet_res_scores = []
|
552 |
+
for i in predictions:
|
553 |
+
temp = i[0]
|
554 |
+
pet_res_scores.append(np.array([1 - temp, temp]))
|
555 |
+
|
556 |
+
return np.array(pet_res_scores)
|
557 |
+
|
558 |
+
def predict_bert(self, facts: str) -> np.ndarray:
|
559 |
+
"""
|
560 |
+
Get prediction of `facts` using `bert_model`.
|
561 |
+
|
562 |
+
Parameters:
|
563 |
+
----------
|
564 |
+
- facts : str
|
565 |
+
Represents the case facts.
|
566 |
+
|
567 |
+
Returns:
|
568 |
+
--------
|
569 |
+
- predictions : np.ndarray
|
570 |
+
An array contains 2 elements, one for probability of petitioner winning
|
571 |
+
and the second for the probability of respondent winning.
|
572 |
+
"""
|
573 |
+
|
574 |
+
facts_vector = self.data_preparator.prepare_bert(facts)
|
575 |
+
predictions = bert_model.predict(facts_vector)
|
576 |
+
|
577 |
+
return predictions
|
578 |
+
|
579 |
+
def predict_fasttext(self, facts: str) -> np.ndarray:
|
580 |
+
"""
|
581 |
+
Get prediction of `facts` using `fasttext`.
|
582 |
+
|
583 |
+
Parameters:
|
584 |
+
----------
|
585 |
+
- facts : str
|
586 |
+
Represents the case facts.
|
587 |
+
|
588 |
+
Returns:
|
589 |
+
--------
|
590 |
+
- pet_res_scores : np.ndarray
|
591 |
+
An array contains 2 elements, one for probability of petitioner winning
|
592 |
+
and the second for the probability of respondent winning.
|
593 |
+
"""
|
594 |
+
|
595 |
+
prediction = fasttext_model.predict(facts)[1]
|
596 |
+
prediction = np.array([prediction])
|
597 |
+
|
598 |
+
pet_res_scores = []
|
599 |
+
for i in prediction:
|
600 |
+
temp = i[0]
|
601 |
+
pet_res_scores.append(np.array([1 - temp, temp]))
|
602 |
+
|
603 |
+
return np.array(pet_res_scores)
|
604 |
+
|
605 |
+
def summarize_facts(self, facts: str) -> str:
|
606 |
+
summarized_case_facts = summarization_model(facts)[0]['summary_text']
|
607 |
+
return summarized_case_facts
|
src/plotting.py
ADDED
@@ -0,0 +1,230 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
from typing import List
|
2 |
+
|
3 |
+
import numpy as np
|
4 |
+
import pandas as pd
|
5 |
+
import matplotlib.pyplot as plt
|
6 |
+
import seaborn as sn
|
7 |
+
|
8 |
+
from sklearn.metrics import auc
|
9 |
+
from sklearn.metrics import roc_curve
|
10 |
+
from sklearn.metrics import classification_report
|
11 |
+
from sklearn.metrics import confusion_matrix
|
12 |
+
|
13 |
+
from tensorflow import keras
|
14 |
+
|
15 |
+
|
16 |
+
class PlottingManager:
|
17 |
+
"""Responsible for providing plots & visualization for the models."""
|
18 |
+
|
19 |
+
def __init__(self) -> None:
|
20 |
+
"""Define style for visualizations."""
|
21 |
+
plt.style.use("seaborn")
|
22 |
+
|
23 |
+
def plot_subplots_curve(
|
24 |
+
self,
|
25 |
+
training_measure: List[List[float]],
|
26 |
+
validation_measure: List[List[float]],
|
27 |
+
title: str,
|
28 |
+
train_color: str = "orangered",
|
29 |
+
validation_color: str = "dodgerblue",
|
30 |
+
) -> None:
|
31 |
+
"""
|
32 |
+
Plotting subplots of the elements of `training_measure` vs. `validation_measure`.
|
33 |
+
|
34 |
+
Parameters:
|
35 |
+
------------
|
36 |
+
- training_measure : List[List[float]]
|
37 |
+
A `k` by `num_epochs` list contains the trained measure whether it's loss or
|
38 |
+
accuracy for each fold.
|
39 |
+
- validation_measure : List[List[float]]
|
40 |
+
A `k` by `num_epochs` list contains the validation measure whether it's loss
|
41 |
+
or accuracy for each fold.
|
42 |
+
- title : str
|
43 |
+
Represents the title of the plot.
|
44 |
+
- train_color : str, optional
|
45 |
+
Represents the graph color for the `training_measure`. (Default is "orangered").
|
46 |
+
- validation_color : str, optional
|
47 |
+
Represents the graph color for the `validation_measure`. (Default is "dodgerblue").
|
48 |
+
"""
|
49 |
+
|
50 |
+
plt.figure(figsize=(12, 8))
|
51 |
+
|
52 |
+
for i in range(len(training_measure)):
|
53 |
+
plt.subplot(2, 2, i + 1)
|
54 |
+
plt.plot(training_measure[i], c=train_color)
|
55 |
+
plt.plot(validation_measure[i], c=validation_color)
|
56 |
+
plt.title("Fold " + str(i + 1))
|
57 |
+
|
58 |
+
plt.suptitle(title)
|
59 |
+
plt.show()
|
60 |
+
|
61 |
+
def plot_heatmap(
|
62 |
+
self, measure: List[List[float]], title: str, cmap: str = "coolwarm"
|
63 |
+
) -> None:
|
64 |
+
"""
|
65 |
+
Plotting a heatmap of the values in `measure`.
|
66 |
+
|
67 |
+
Parameters:
|
68 |
+
------------
|
69 |
+
- measure : List[List[float]]
|
70 |
+
A `k` by `num_epochs` list contains the measure whether it's loss
|
71 |
+
or accuracy for each fold.
|
72 |
+
- title : str
|
73 |
+
Title of the plot.
|
74 |
+
- cmap : str, optional
|
75 |
+
Color map of the plot (default is "coolwarm").
|
76 |
+
"""
|
77 |
+
|
78 |
+
# transpose the array to make it `num_epochs` by `k`
|
79 |
+
values_array = np.array(measure).T
|
80 |
+
df_cm = pd.DataFrame(
|
81 |
+
values_array,
|
82 |
+
range(1, values_array.shape[0] + 1),
|
83 |
+
["fold " + str(i + 1) for i in range(4)],
|
84 |
+
)
|
85 |
+
|
86 |
+
plt.figure(figsize=(10, 8))
|
87 |
+
plt.title(
|
88 |
+
title + " Throughout " + str(values_array.shape[1]) + " Folds", pad=20
|
89 |
+
)
|
90 |
+
sn.heatmap(df_cm, annot=True, cmap=cmap, annot_kws={"size": 10})
|
91 |
+
plt.show()
|
92 |
+
|
93 |
+
def plot_average_curves(
|
94 |
+
self,
|
95 |
+
title: str,
|
96 |
+
x: List[float],
|
97 |
+
y: List[float],
|
98 |
+
x_label: str,
|
99 |
+
y_label: str,
|
100 |
+
train_color: str = "orangered",
|
101 |
+
validation_color: str = "dodgerblue",
|
102 |
+
) -> None:
|
103 |
+
"""
|
104 |
+
Plotting the curves of `x` against `y`, where x and y are training and validation
|
105 |
+
measures (loss or accuracy).
|
106 |
+
|
107 |
+
Parameters:
|
108 |
+
------------
|
109 |
+
- title : str
|
110 |
+
Title of the plot.
|
111 |
+
- x : List[float]
|
112 |
+
Training measure of the models (loss or accuracy).
|
113 |
+
- y : List[float]
|
114 |
+
Validation measure of the models (loss or accuracy).
|
115 |
+
- x_label : str
|
116 |
+
Label of the training measure to put it in plot legend.
|
117 |
+
- y_label : str
|
118 |
+
Label of the validation measure to put it in plot legend.
|
119 |
+
- train_color : str, optional
|
120 |
+
Color of the training plot (default is "orangered").
|
121 |
+
- validation_color : str, optional
|
122 |
+
Color of the validation plot (default is "dodgerblue").
|
123 |
+
"""
|
124 |
+
|
125 |
+
plt.title(title, pad=20)
|
126 |
+
plt.plot(x, c=train_color, label=x_label)
|
127 |
+
plt.plot(y, c=validation_color, label=y_label)
|
128 |
+
plt.legend()
|
129 |
+
plt.show()
|
130 |
+
|
131 |
+
def plot_roc_curve(
|
132 |
+
self,
|
133 |
+
all_models: List[keras.models.Sequential],
|
134 |
+
X_test: pd.DataFrame,
|
135 |
+
y_test: pd.Series,
|
136 |
+
) -> None:
|
137 |
+
"""
|
138 |
+
Plotting the AUC-ROC curve of all the passed models in `all_models`.
|
139 |
+
|
140 |
+
Parameters:
|
141 |
+
------------
|
142 |
+
- all_models : List[keras.models.Sequential]
|
143 |
+
Contains all trained models, number of models equals number of
|
144 |
+
`k` fold cross-validation.
|
145 |
+
- X_test : pd.DataFrame
|
146 |
+
Contains the testing vectors.
|
147 |
+
- y_test : pd.Series
|
148 |
+
Contains the testing labels.
|
149 |
+
"""
|
150 |
+
|
151 |
+
plt.figure(figsize=(12, 8))
|
152 |
+
for i, model in enumerate(all_models):
|
153 |
+
y_pred = model.predict(X_test).ravel()
|
154 |
+
fpr, tpr, _ = roc_curve(y_test, y_pred)
|
155 |
+
auc_curve = auc(fpr, tpr)
|
156 |
+
plt.subplot(2, 2, i + 1)
|
157 |
+
plt.plot([0, 1], [0, 1], color="dodgerblue", linestyle="--")
|
158 |
+
plt.plot(
|
159 |
+
fpr,
|
160 |
+
tpr,
|
161 |
+
color="orangered",
|
162 |
+
label=f"Fold {str(i+1)} (area = {auc_curve:.3f})",
|
163 |
+
)
|
164 |
+
plt.legend(loc="best")
|
165 |
+
plt.title(f"Fold {str(i+1)}")
|
166 |
+
|
167 |
+
plt.suptitle("AUC-ROC curves")
|
168 |
+
plt.show()
|
169 |
+
|
170 |
+
def plot_classification_report(
|
171 |
+
self, model: keras.models.Sequential, X_test: pd.DataFrame, y_test: pd.Series
|
172 |
+
) -> str | dict:
|
173 |
+
"""
|
174 |
+
Plotting the classification report of the passed `model`.
|
175 |
+
|
176 |
+
Parameters:
|
177 |
+
------------
|
178 |
+
- model : keras.models.Sequential
|
179 |
+
The trained model that will be evaluated.
|
180 |
+
- X_test : pd.DataFrame
|
181 |
+
Contains the testing vectors.
|
182 |
+
- y_test : pd.Series
|
183 |
+
Contains the testing labels.
|
184 |
+
|
185 |
+
Returns:
|
186 |
+
--------
|
187 |
+
- str | dict: The classification report for the given model and testing data.
|
188 |
+
It returns a string if `output_format` is set to 'str', and returns
|
189 |
+
a dictionary if `output_format` is set to 'dict'.
|
190 |
+
"""
|
191 |
+
|
192 |
+
y_pred = model.predict(X_test).ravel()
|
193 |
+
preds = np.where(y_pred > 0.5, 1, 0)
|
194 |
+
cls_report = classification_report(y_test, preds)
|
195 |
+
|
196 |
+
return cls_report
|
197 |
+
|
198 |
+
def plot_confusion_matrix(
|
199 |
+
self,
|
200 |
+
all_models: List[keras.models.Sequential],
|
201 |
+
X_test: pd.DataFrame,
|
202 |
+
y_test: pd.Series,
|
203 |
+
) -> None:
|
204 |
+
"""
|
205 |
+
Plotting the confusion matrix of each model in `all_models`.
|
206 |
+
|
207 |
+
Parameters:
|
208 |
+
------------
|
209 |
+
- all_models: list[keras.models.Sequential]
|
210 |
+
Contains all trained models, number of models equals
|
211 |
+
number of `k` fold cross-validation.
|
212 |
+
- X_test: pd.DataFrame
|
213 |
+
Contains the testing vectors.
|
214 |
+
- y_test: pd.Series
|
215 |
+
Contains the testing labels.
|
216 |
+
"""
|
217 |
+
|
218 |
+
_, axes = plt.subplots(nrows=2, ncols=2, figsize=(12, 8))
|
219 |
+
|
220 |
+
for i, (model, ax) in enumerate(zip(all_models, axes.flatten())):
|
221 |
+
y_pred = model.predict(X_test).ravel()
|
222 |
+
preds = np.where(y_pred > 0.5, 1, 0)
|
223 |
+
|
224 |
+
conf_matrix = confusion_matrix(y_test, preds)
|
225 |
+
sn.heatmap(conf_matrix, annot=True, ax=ax)
|
226 |
+
ax.set_title(f"Fold {i+1}")
|
227 |
+
|
228 |
+
plt.suptitle("Confusion Matrices")
|
229 |
+
plt.tight_layout()
|
230 |
+
plt.show()
|
src/preprocessing.py
ADDED
@@ -0,0 +1,591 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
# global
|
2 |
+
import string
|
3 |
+
from typing import List, Tuple
|
4 |
+
|
5 |
+
import numpy as np
|
6 |
+
import pandas as pd
|
7 |
+
|
8 |
+
import re
|
9 |
+
import nltk
|
10 |
+
|
11 |
+
from sklearn.utils import resample
|
12 |
+
|
13 |
+
from gensim.models.doc2vec import Doc2Vec, TaggedDocument
|
14 |
+
from nltk.tokenize import RegexpTokenizer
|
15 |
+
|
16 |
+
import tensorflow as tf
|
17 |
+
from keras.layers import TextVectorization
|
18 |
+
from keras.preprocessing.text import Tokenizer
|
19 |
+
from keras.utils import pad_sequences
|
20 |
+
|
21 |
+
# local
|
22 |
+
from utils import Doc2VecModel
|
23 |
+
|
24 |
+
|
25 |
+
punct = string.punctuation
|
26 |
+
stemmer = nltk.stem.PorterStemmer()
|
27 |
+
eng_stopwords = nltk.corpus.stopwords.words("english")
|
28 |
+
|
29 |
+
|
30 |
+
class Preprocessor:
|
31 |
+
"""Responsible for preprocessing case facts."""
|
32 |
+
|
33 |
+
def __init__(self) -> None:
|
34 |
+
pass
|
35 |
+
|
36 |
+
def _nltk_tokenizer(self, text: str) -> List[str]:
|
37 |
+
"""
|
38 |
+
Tokenize a given `text` using the RegexpTokenizer from the nltk library.
|
39 |
+
|
40 |
+
Parameters:
|
41 |
+
-----------
|
42 |
+
- text : str
|
43 |
+
A string containing the text to be tokenized.
|
44 |
+
|
45 |
+
Returns:
|
46 |
+
--------
|
47 |
+
- tokens : List[str]
|
48 |
+
A list of tokens generated by the tokenizer.
|
49 |
+
"""
|
50 |
+
|
51 |
+
tokenizer = RegexpTokenizer(r"\w+")
|
52 |
+
tokens = tokenizer.tokenize(text)
|
53 |
+
|
54 |
+
return tokens
|
55 |
+
|
56 |
+
def _tokenize_text(self, text_column: pd.Series) -> pd.Series:
|
57 |
+
"""Splitting `text_column` into tokens.
|
58 |
+
|
59 |
+
Parameters:
|
60 |
+
------------
|
61 |
+
- text_column : pd.Series
|
62 |
+
Contains text that needs to be tokenized.
|
63 |
+
|
64 |
+
Returns:
|
65 |
+
--------
|
66 |
+
- tokenized_text : pd.Series
|
67 |
+
Contains tokenized version of `text_column`.
|
68 |
+
"""
|
69 |
+
|
70 |
+
tokenized_text = text_column.apply(self._nltk_tokenizer)
|
71 |
+
return tokenized_text
|
72 |
+
|
73 |
+
def _convert_to_tagged_document(
|
74 |
+
self, text_column: pd.Series
|
75 |
+
) -> Tuple[List[str], List[TaggedDocument]]:
|
76 |
+
"""
|
77 |
+
Convert `text_column` of specific to TaggedDocuments.
|
78 |
+
|
79 |
+
Parameters:
|
80 |
+
------------
|
81 |
+
- column : pd.Series
|
82 |
+
Contains the list of tokens of each fact.
|
83 |
+
|
84 |
+
Returns:
|
85 |
+
--------
|
86 |
+
A tuble containing the following items:
|
87 |
+
- tokens_list : list[str]
|
88 |
+
Contains all tokens of each case in the `text_column`.
|
89 |
+
- tagged_docs : list[TaggedDocument]
|
90 |
+
Contains TaggedDocument object for each case.
|
91 |
+
"""
|
92 |
+
|
93 |
+
tokens_list = text_column.to_list()
|
94 |
+
tagged_docs = [TaggedDocument(t, [str(i)])
|
95 |
+
for i, t in enumerate(tokens_list)]
|
96 |
+
|
97 |
+
return tokens_list, tagged_docs
|
98 |
+
|
99 |
+
def _vectorize_text(
|
100 |
+
self, doc2vec_model: Doc2Vec, df: pd.Series, tokens_list: List[str]
|
101 |
+
) -> pd.DataFrame:
|
102 |
+
"""
|
103 |
+
Convert values of `tokens_list` to a vector.
|
104 |
+
|
105 |
+
Parameters:
|
106 |
+
-----------
|
107 |
+
- doc2vec_model : Doc2Vev
|
108 |
+
Trained Doc2Vec model.
|
109 |
+
- df : pd.Series
|
110 |
+
This will use only to get its indicies for the new generated dataframe.
|
111 |
+
- tokens_list : List[str]
|
112 |
+
Contains all tokens of each case.
|
113 |
+
|
114 |
+
Returns:
|
115 |
+
--------
|
116 |
+
- text_vectors_df : pd.DataFrame
|
117 |
+
Contains the vector representaion for each case.
|
118 |
+
"""
|
119 |
+
|
120 |
+
text_vectors = [doc2vec_model.infer_vector(doc) for doc in tokens_list]
|
121 |
+
text_vectors_df = pd.DataFrame(text_vectors, index=df.index)
|
122 |
+
|
123 |
+
return text_vectors_df
|
124 |
+
|
125 |
+
def _anonymize_case_facts(
|
126 |
+
self, first_party_name: str, second_party_name: str, facts: str
|
127 |
+
) -> str:
|
128 |
+
"""
|
129 |
+
Anonymize case facts by replacing its party names with "_PARTY_" tag.
|
130 |
+
|
131 |
+
Parameters:
|
132 |
+
------------
|
133 |
+
- first_party_name : str
|
134 |
+
Represents first party name or petitioner name.
|
135 |
+
- second_party_name : str
|
136 |
+
Represents second party name or respondent name.
|
137 |
+
- facts : str
|
138 |
+
Represents case facts.
|
139 |
+
|
140 |
+
Returns:
|
141 |
+
--------
|
142 |
+
- anonymized_facts : str
|
143 |
+
An anonymized version of `facts`.
|
144 |
+
"""
|
145 |
+
|
146 |
+
# remove any commas and any non alphabet characters
|
147 |
+
first_party_name = re.sub(r"[\,+]", " ", first_party_name)
|
148 |
+
first_party_name = re.sub(r"[^a-zA-Z]", " ", first_party_name)
|
149 |
+
|
150 |
+
second_party_name = re.sub(r"[\,+]", " ", second_party_name)
|
151 |
+
second_party_name = re.sub(r"[^a-zA-Z]", " ", second_party_name)
|
152 |
+
|
153 |
+
for name in first_party_name.split():
|
154 |
+
facts = re.sub(name, " _PARTY_ ", facts)
|
155 |
+
|
156 |
+
for name in second_party_name.split():
|
157 |
+
facts = re.sub(name, " _PARTY_ ", facts)
|
158 |
+
|
159 |
+
# replace any consecutive _PARTY_ tags with only one _PARTY_ tag.
|
160 |
+
regex_continous_tags = r"(_PARTY_\s+){2,}"
|
161 |
+
anonymized_facts = re.sub(regex_continous_tags, " _PARTY_ ", facts)
|
162 |
+
# remove ant consecutive spaces
|
163 |
+
anonymized_facts = re.sub(r"\s+", " ", anonymized_facts)
|
164 |
+
|
165 |
+
return anonymized_facts
|
166 |
+
|
167 |
+
def _preprocess_text(self, text: str) -> str:
|
168 |
+
"""
|
169 |
+
Preprocessing & cleaning `text` including:
|
170 |
+
- lowercasing
|
171 |
+
- removing quotation marks
|
172 |
+
- removing digits
|
173 |
+
- removing punctuation
|
174 |
+
- removing brackets, braces, and paranthesis
|
175 |
+
- removeing stopwords
|
176 |
+
- stemming tokens
|
177 |
+
|
178 |
+
Parameters:
|
179 |
+
------------
|
180 |
+
- text : str
|
181 |
+
Text need to be processed (cleaned).
|
182 |
+
|
183 |
+
Returns:
|
184 |
+
--------
|
185 |
+
- processed_text : str
|
186 |
+
A preprocessed version of `text`.
|
187 |
+
"""
|
188 |
+
|
189 |
+
text = text.lower()
|
190 |
+
# remove quotation marks
|
191 |
+
text = re.sub(r"\'", "", text)
|
192 |
+
# remove digits
|
193 |
+
text = re.sub(r"\d+", "", text)
|
194 |
+
# remove punctuation but with keeping '_' letter
|
195 |
+
text = "".join([ch for ch in text if (ch == "_") or (ch not in punct)])
|
196 |
+
# remove brackets, braces, and parantheses
|
197 |
+
text = re.sub(r"[\[\]\(\)\{\}]+", " ", text)
|
198 |
+
tokens = nltk.word_tokenize(text)
|
199 |
+
# remove stopwords and stemming tokens
|
200 |
+
tokens = [stemmer.stem(token)
|
201 |
+
for token in tokens if token not in eng_stopwords]
|
202 |
+
# convert tokens back to string
|
203 |
+
processed_text = " ".join(tokens)
|
204 |
+
|
205 |
+
return processed_text
|
206 |
+
|
207 |
+
def convert_text_to_vectors_doc2vec(
|
208 |
+
self,
|
209 |
+
text_column: pd.Series,
|
210 |
+
train: bool = True,
|
211 |
+
embeddings_doc2vec: Doc2Vec = None,
|
212 |
+
) -> Tuple[Doc2Vec, pd.DataFrame] | pd.DataFrame:
|
213 |
+
"""
|
214 |
+
Converting `text_column` to vectors using `Doc2Vec` model
|
215 |
+
|
216 |
+
Parameters:
|
217 |
+
------------
|
218 |
+
- text_column : pd.Series
|
219 |
+
Contains the case facts.
|
220 |
+
- train : bool, optional
|
221 |
+
Defines whether the model will be trained or not. (if True, Doc2Vec will be trained |
|
222 |
+
else, Doc2Vec will used the passed `embeddings_Doc2Vec`). (Default is True).
|
223 |
+
- embeddings_doc2vec : Doc2Vec, optional
|
224 |
+
Trained Doc2Vec model will be used for generating embeddings of `text_column` if
|
225 |
+
`train` is False. (Default is None).
|
226 |
+
|
227 |
+
Returns:
|
228 |
+
--------
|
229 |
+
1. A tuple contains the following:
|
230 |
+
- embeddings_doc2vec : Doc2Vec
|
231 |
+
Trained Doc2Vec model.
|
232 |
+
- text_vectors_df : pd.DataFrame
|
233 |
+
A DataFrame contains `text_column` vectors if `train` is True.
|
234 |
+
|
235 |
+
2. text_vectors_df : pd.DataFrame
|
236 |
+
A DataFrame contains `text_column` vectors if `train` is False.
|
237 |
+
|
238 |
+
Raises:
|
239 |
+
-------
|
240 |
+
- AssertionError
|
241 |
+
If train is False and `embeddings_doc2vec` is None.
|
242 |
+
- AssertionError
|
243 |
+
If train is False and `embedding_doc2vec` is not an instance of Doc2Vec
|
244 |
+
"""
|
245 |
+
|
246 |
+
tokenized_text = self._tokenize_text(text_column)
|
247 |
+
tokens_list, tagged_docs = self._convert_to_tagged_document(
|
248 |
+
tokenized_text)
|
249 |
+
|
250 |
+
if train:
|
251 |
+
doc2vec_model = Doc2VecModel()
|
252 |
+
embeddings_doc2vec = doc2vec_model.train_doc2vec_embeddings_model(
|
253 |
+
tagged_docs
|
254 |
+
)
|
255 |
+
text_vectors_df = self._vectorize_text(
|
256 |
+
embeddings_doc2vec, text_column, tokens_list
|
257 |
+
)
|
258 |
+
return embeddings_doc2vec, text_vectors_df
|
259 |
+
|
260 |
+
assert (
|
261 |
+
embeddings_doc2vec is not None
|
262 |
+
), "`embedding_doc2vec` argument must be not None."
|
263 |
+
assert isinstance(
|
264 |
+
embeddings_doc2vec, Doc2Vec
|
265 |
+
), "`embedding_doc2vec` argument must be an instance of Doc2Vec to infer vectors."
|
266 |
+
text_vectors_df = self._vectorize_text(
|
267 |
+
embeddings_doc2vec, text_column, tokens_list
|
268 |
+
)
|
269 |
+
|
270 |
+
return text_vectors_df
|
271 |
+
|
272 |
+
def convert_text_to_vectors_tf_idf(
|
273 |
+
self,
|
274 |
+
text_column: pd.Series,
|
275 |
+
ngrams: int = 2,
|
276 |
+
max_tokens: int = 10000,
|
277 |
+
output_mode: str = "tf-idf",
|
278 |
+
train: bool = True,
|
279 |
+
text_vectorizer: TextVectorization = None,
|
280 |
+
) -> Tuple[TextVectorization, tf.Tensor] | tf.Tensor:
|
281 |
+
"""
|
282 |
+
Converting `text_column` to vectors using `TextVectorization` layer.
|
283 |
+
|
284 |
+
Parameters:
|
285 |
+
------------
|
286 |
+
- text_column : pd.Series
|
287 |
+
Contains the case facts.
|
288 |
+
- ngrams : int, optional
|
289 |
+
Defines the number of n-gram (Default is 2).
|
290 |
+
- max_tokens : int, optional
|
291 |
+
Defines the number of max_tokens of `text_vectorizer` (Default is 10,000).
|
292 |
+
- output_mode : str, optional
|
293 |
+
Represents the output vectors type whether it is "tfi-df" or "binary" or "count"
|
294 |
+
(Default is "tf-idf").
|
295 |
+
- train : bool, optional
|
296 |
+
Defines whether the model will be trained or not. (if True, TextVectorization
|
297 |
+
will be trained, else, TextVectorization will used the passed `text_vectorizer`).
|
298 |
+
(Default is True).
|
299 |
+
- text_vectorizer : TextVectorization, optional
|
300 |
+
Trained TextVectorization layer will be used for generating embeddings of
|
301 |
+
`text_column` if `train` is False. (Default is None).
|
302 |
+
|
303 |
+
Returns:
|
304 |
+
--------
|
305 |
+
- if `train` == True:
|
306 |
+
A tuple contains the following:
|
307 |
+
- text_vectorizer : TextVectorization
|
308 |
+
Trained TextVectorization layer.
|
309 |
+
- text_vectors : tf.Tensor
|
310 |
+
A Tensor contains `text_column` training vectors.
|
311 |
+
- otherwise:
|
312 |
+
text_vectors : tf.Tensor
|
313 |
+
A Tensor contains `text_column` testing vectors.
|
314 |
+
|
315 |
+
Raises:
|
316 |
+
-------
|
317 |
+
- AssertionError
|
318 |
+
If train is False and `text_vectorizer` is None.
|
319 |
+
- AssertionError
|
320 |
+
If train is False and `text_vectorizer` is not an instance of TextVectorization.
|
321 |
+
"""
|
322 |
+
|
323 |
+
if train:
|
324 |
+
text_vectorizer = TextVectorization(
|
325 |
+
ngrams=ngrams, max_tokens=max_tokens, output_mode=output_mode
|
326 |
+
)
|
327 |
+
text_vectorizer.adapt(text_column)
|
328 |
+
text_vectors = text_vectorizer(text_column)
|
329 |
+
|
330 |
+
return text_vectorizer, text_vectors
|
331 |
+
|
332 |
+
assert (
|
333 |
+
text_vectorizer is not None
|
334 |
+
), "`text_vectorizer` argument must be not None."
|
335 |
+
assert isinstance(
|
336 |
+
text_vectorizer, TextVectorization
|
337 |
+
), "`text_vectorizer` argument must be an instance of TextVectorization to infer vectors."
|
338 |
+
text_vectors = text_vectorizer(text_column)
|
339 |
+
|
340 |
+
return text_vectors
|
341 |
+
|
342 |
+
def convert_text_to_vectors_cnn(
|
343 |
+
self,
|
344 |
+
text_column: pd.Series,
|
345 |
+
max_tokens: int = 2000,
|
346 |
+
output_sequence_length: int = 500,
|
347 |
+
output_mode: str = "int",
|
348 |
+
train: bool = True,
|
349 |
+
text_vectorizer: TextVectorization = None,
|
350 |
+
) -> Tuple[TextVectorization, tf.Tensor] | tf.Tensor:
|
351 |
+
"""
|
352 |
+
Converting `text_column` to vectors using `TextVectorization` layer.
|
353 |
+
|
354 |
+
Parameters:
|
355 |
+
------------
|
356 |
+
- text_column : pd.Series
|
357 |
+
Contains the case facts.
|
358 |
+
- max_tokens : int, optional
|
359 |
+
Defines the number of max_tokens of `text_vectorizer` (Default is 2000).
|
360 |
+
- output_sequence_length : int, optional
|
361 |
+
Represents the dimensions of the output vector (Default is 500).
|
362 |
+
- output_mode : str, optional
|
363 |
+
Represents the output vectors type whether it is "int" or "binary" or "tfi-df".
|
364 |
+
- train : bool, optional
|
365 |
+
Defines whether the model will be trained or not. (if True,
|
366 |
+
TextVectorization will be trained | else, TextVectorization will used the
|
367 |
+
passed `text_vectorizer`). (Default is True).
|
368 |
+
- text_vectorizer : TextVectorization, optional
|
369 |
+
Trained TextVectorization layer will be used for generating embeddings of
|
370 |
+
`text_column` if `train` is False. (Default is None).
|
371 |
+
|
372 |
+
Returns:
|
373 |
+
--------
|
374 |
+
- if `train` == True:
|
375 |
+
A tuple contains the following:
|
376 |
+
- text_vectorizer : TextVectorization
|
377 |
+
Trained TextVectorization layer.
|
378 |
+
- text_vectors : tf.Tensor
|
379 |
+
A Tensor contains `text_column` training vectors.
|
380 |
+
- otherwise:
|
381 |
+
text_vectors : tf.Tensor
|
382 |
+
A Tensor contains `text_column` testing vectors.
|
383 |
+
|
384 |
+
Raises:
|
385 |
+
-------
|
386 |
+
- AssertionError
|
387 |
+
If train is False and `text_vectorizer` is None.
|
388 |
+
- AssertionError
|
389 |
+
If train is False and `text_vectorizer` is not an instance of TextVectorization.
|
390 |
+
"""
|
391 |
+
|
392 |
+
if train:
|
393 |
+
text_vectorizer = TextVectorization(
|
394 |
+
max_tokens=max_tokens,
|
395 |
+
output_mode=output_mode,
|
396 |
+
output_sequence_length=output_sequence_length,
|
397 |
+
)
|
398 |
+
text_vectorizer.adapt(text_column)
|
399 |
+
text_vectors = text_vectorizer(text_column)
|
400 |
+
return text_vectorizer, text_vectors
|
401 |
+
|
402 |
+
assert (
|
403 |
+
text_vectorizer is not None
|
404 |
+
), "`text_vectorizer` argument must be not None."
|
405 |
+
assert isinstance(
|
406 |
+
text_vectorizer, TextVectorization
|
407 |
+
), "`text_vectorizer` argument must be an instance of TextVectorization to infer vectors."
|
408 |
+
text_vectors = text_vectorizer(text_column)
|
409 |
+
|
410 |
+
return text_vectors
|
411 |
+
|
412 |
+
def convert_text_to_vectors_glove(
|
413 |
+
self,
|
414 |
+
text_column: pd.Series,
|
415 |
+
train: bool = True,
|
416 |
+
glove_tokenizer: Tokenizer = None,
|
417 |
+
vocab_size: int = 1000,
|
418 |
+
oov_token: str = "<OOV>",
|
419 |
+
max_length: int = 50,
|
420 |
+
padding_type: str = "post",
|
421 |
+
truncation_type: str = "post",
|
422 |
+
) -> Tuple[Tokenizer, np.ndarray] | np.ndarray:
|
423 |
+
"""
|
424 |
+
Converting `text_column` to vectors using `glove_tokenizer`.
|
425 |
+
|
426 |
+
Parameters:
|
427 |
+
------------
|
428 |
+
- text_column : pd.Series
|
429 |
+
Contains the case facts.
|
430 |
+
- train : bool, optional
|
431 |
+
Defines whether the model will be trained or not. (if True,
|
432 |
+
Tokenizer will be trained | else, Tokenizer will used the
|
433 |
+
passed `glove_tokenizer`). (Default is True).
|
434 |
+
- glove_tokenizer : Tokenizer, optional
|
435 |
+
Trained Tokenizer layer will be used for generating embeddings of
|
436 |
+
`text_column` if `train` is False. (Default is None).
|
437 |
+
- vocab_size : int, optional
|
438 |
+
Represents the number of supported vocabulary of the Tokenizer,
|
439 |
+
any token not in this vocabulary will be treated as an out-of-vocabulary
|
440 |
+
token(OOV). (Default is 1000).
|
441 |
+
- oov_tokens : str, optional
|
442 |
+
Represents the token of an out-of-vocabulary token (Default is "<OOV>").
|
443 |
+
- max_length : int, optional
|
444 |
+
Defins the output vector's dimension. (Default is 50).
|
445 |
+
- padding_type : str, optional
|
446 |
+
Defines the padding type of the vectors, if the vector size is less than
|
447 |
+
`max_length`, the rest of the `max_length` will be padded with 0 (Default is "post").
|
448 |
+
- truncation_type : str, optional
|
449 |
+
Defines the truncation type of the vectors, if the vector size is more than
|
450 |
+
`max_length`, the extra of the `max_length` will be truncated (Default is "post").
|
451 |
+
|
452 |
+
Returns:
|
453 |
+
--------
|
454 |
+
- if `train` == True:
|
455 |
+
A tuple contains the following:
|
456 |
+
- glove_tokenizer : Tokenizer
|
457 |
+
Trained Tokenizer layer.
|
458 |
+
- text_padded : np.ndarray
|
459 |
+
An array contains `text_column` vectors.
|
460 |
+
- otherwise:
|
461 |
+
text_padded : np.ndarray
|
462 |
+
An array contains `text_column` vectors.
|
463 |
+
|
464 |
+
Raises:
|
465 |
+
-------
|
466 |
+
- AssertionError
|
467 |
+
If train is False and `glove_tokenizer` is None.
|
468 |
+
- AssertionError
|
469 |
+
If train is False and `glove_tokenizer` is not instance of Tokenizer.
|
470 |
+
"""
|
471 |
+
|
472 |
+
if train:
|
473 |
+
glove_tokenizer = Tokenizer(
|
474 |
+
num_words=vocab_size, oov_token=oov_token)
|
475 |
+
glove_tokenizer.fit_on_texts(text_column)
|
476 |
+
text_sequences = glove_tokenizer.texts_to_sequences(text_column)
|
477 |
+
text_padded = pad_sequences(
|
478 |
+
text_sequences,
|
479 |
+
maxlen=max_length,
|
480 |
+
padding=padding_type,
|
481 |
+
truncating=truncation_type,
|
482 |
+
)
|
483 |
+
|
484 |
+
return glove_tokenizer, text_padded
|
485 |
+
|
486 |
+
assert (
|
487 |
+
glove_tokenizer is not None
|
488 |
+
), "`glove_tokenizer` argument must be not None."
|
489 |
+
assert isinstance(
|
490 |
+
glove_tokenizer, Tokenizer
|
491 |
+
), "`glove_tokenizer` argument must be an instance of Tokenizer."
|
492 |
+
text_sequences = glove_tokenizer.texts_to_sequences(text_column)
|
493 |
+
text_padded = pad_sequences(
|
494 |
+
text_sequences,
|
495 |
+
maxlen=max_length,
|
496 |
+
padding=padding_type,
|
497 |
+
truncating=truncation_type,
|
498 |
+
)
|
499 |
+
|
500 |
+
return text_padded
|
501 |
+
|
502 |
+
def balance_data(self, X_train: pd.Series, y_train: pd.Series) -> pd.DataFrame:
|
503 |
+
"""
|
504 |
+
Balancing `X_train` and `y_train` to distribute the targets in `y_train` equally.
|
505 |
+
|
506 |
+
Parameters:
|
507 |
+
------------
|
508 |
+
- text_column : pd.Series
|
509 |
+
Contains the case facts.
|
510 |
+
- y_train : pd.Series
|
511 |
+
Contains the training targets.
|
512 |
+
|
513 |
+
Returns:
|
514 |
+
--------
|
515 |
+
- shuffled_balanced_df : pd.DataFrame
|
516 |
+
Contains the new balanced dataframe with shuffling indicies.
|
517 |
+
"""
|
518 |
+
|
519 |
+
df = pd.concat([X_train, y_train], axis=1)
|
520 |
+
|
521 |
+
first_party = df[df["winner_index"] == 0]
|
522 |
+
second_party = df[df["winner_index"] == 1]
|
523 |
+
|
524 |
+
upsample_second_party = resample(
|
525 |
+
second_party, replace=True, n_samples=len(first_party), random_state=42
|
526 |
+
)
|
527 |
+
|
528 |
+
upsample_df = pd.concat([upsample_second_party, first_party])
|
529 |
+
|
530 |
+
shuffled_indices = np.arange(upsample_df.shape[0])
|
531 |
+
np.random.shuffle(shuffled_indices)
|
532 |
+
|
533 |
+
shuffled_balanced_df = upsample_df.iloc[shuffled_indices, :]
|
534 |
+
|
535 |
+
return shuffled_balanced_df
|
536 |
+
|
537 |
+
def anonymize_data(
|
538 |
+
self,
|
539 |
+
first_party_names: pd.Series,
|
540 |
+
second_party_names: pd.Series,
|
541 |
+
text_column: pd.Series,
|
542 |
+
) -> pd.Series:
|
543 |
+
"""
|
544 |
+
Anonymize `text_column` by replacing `first_party_names` and
|
545 |
+
`second_party_names` wit "_PARTY_" tag.
|
546 |
+
|
547 |
+
Parameters:
|
548 |
+
------------
|
549 |
+
- first_party_names : pd.Series
|
550 |
+
Contains all first party names needed to be anonymized.
|
551 |
+
- second_party_names : pd.Series
|
552 |
+
Contains all second party names needed to be anonymized.
|
553 |
+
- text_column : pd.Series
|
554 |
+
Contains all texts needed to be anonymized.
|
555 |
+
|
556 |
+
Returns:
|
557 |
+
--------
|
558 |
+
- all_anonyimzed_facts : pd.Series
|
559 |
+
Contains anonymized version of `text_column`.
|
560 |
+
"""
|
561 |
+
|
562 |
+
all_anonymized_facts = []
|
563 |
+
|
564 |
+
for i in range(text_column.shape[0]):
|
565 |
+
facts = text_column.iloc[i]
|
566 |
+
first_party_name = first_party_names.iloc[i]
|
567 |
+
second_party_name = second_party_names.iloc[i]
|
568 |
+
anonymized_facts = self._anonymize_case_facts(
|
569 |
+
first_party_name, second_party_name, facts
|
570 |
+
)
|
571 |
+
all_anonymized_facts.append(anonymized_facts)
|
572 |
+
|
573 |
+
return pd.Series(all_anonymized_facts)
|
574 |
+
|
575 |
+
def preprocess_data(self, text_column: pd.Series) -> pd.Series:
|
576 |
+
"""
|
577 |
+
Preprocessing & cleaning all texts in `text_column`.
|
578 |
+
|
579 |
+
Parameters:
|
580 |
+
------------
|
581 |
+
- text_column : pd.Series
|
582 |
+
Contains all case facts.
|
583 |
+
|
584 |
+
Returns:
|
585 |
+
--------
|
586 |
+
- preprocessed_text : pd.Series
|
587 |
+
Contains all texts after being processed.
|
588 |
+
"""
|
589 |
+
|
590 |
+
preprocessed_text = text_column.apply(self._preprocess_text)
|
591 |
+
return preprocessed_text
|
src/style.css
ADDED
@@ -0,0 +1,94 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
@import url('https://fonts.googleapis.com/css2?family=Cairo:wght@300;400;500;600;700;800&display=swap');
|
2 |
+
|
3 |
+
* {
|
4 |
+
font-family: 'Cairo', sans-serif !important;
|
5 |
+
}
|
6 |
+
|
7 |
+
/* title */
|
8 |
+
.e16nr0p30 {
|
9 |
+
font-weight: 700;
|
10 |
+
font-size: 30px;
|
11 |
+
}
|
12 |
+
|
13 |
+
/* buttons */
|
14 |
+
.edgvbvh10,
|
15 |
+
.edgvbvh5 {
|
16 |
+
width: 100%;
|
17 |
+
height: 40px;
|
18 |
+
background-color: #4756ff;
|
19 |
+
color: #fff;
|
20 |
+
transition: 0.4s;
|
21 |
+
border: none;
|
22 |
+
}
|
23 |
+
|
24 |
+
.edgvbvh10:hover,
|
25 |
+
.edgvbvh5:hover {
|
26 |
+
background-color: #3747fd;
|
27 |
+
color: #fff;
|
28 |
+
border: none;
|
29 |
+
}
|
30 |
+
|
31 |
+
.edgvbvh10:focus,
|
32 |
+
.edgvbvh5:focus {
|
33 |
+
background-color: #3747fd;
|
34 |
+
color: #fff !important;
|
35 |
+
box-shadow: none;
|
36 |
+
border: none;
|
37 |
+
}
|
38 |
+
|
39 |
+
/* header */
|
40 |
+
.row_heading {
|
41 |
+
font-size: 14px;
|
42 |
+
}
|
43 |
+
|
44 |
+
/* spinner */
|
45 |
+
.css-1y04v0k.e17lx80j1,
|
46 |
+
.css-p6380s.e17lx80j1 {
|
47 |
+
margin: 0px;
|
48 |
+
border-color: #34e27f #b3b3b333 #cacaca33 !important;
|
49 |
+
-webkit-box-flex: 0;
|
50 |
+
flex-grow: 0;
|
51 |
+
flex-shrink: 0;
|
52 |
+
}
|
53 |
+
|
54 |
+
/* inputs styling */
|
55 |
+
.st-bf {
|
56 |
+
transition: 0.8s;
|
57 |
+
border: none !important;
|
58 |
+
}
|
59 |
+
|
60 |
+
.st-bf:hover {
|
61 |
+
box-shadow: 0 0 0 4px #dbdbdb !important;
|
62 |
+
}
|
63 |
+
|
64 |
+
/* text stylings */
|
65 |
+
.highlight-petitioner {
|
66 |
+
border-radius: 0.4rem;
|
67 |
+
background-color: rgba(253, 231, 142, 0.4);
|
68 |
+
color: #ffd061;
|
69 |
+
padding: 1px 5px;
|
70 |
+
margin-top: 10px;
|
71 |
+
margin-right: 5px;
|
72 |
+
}
|
73 |
+
|
74 |
+
.highlight-respondent {
|
75 |
+
border-radius: 0.4rem;
|
76 |
+
background-color: rgba(78, 170, 255, 0.2);
|
77 |
+
color: #6195ff;
|
78 |
+
padding: 1px 5px;
|
79 |
+
margin-top: 10px;
|
80 |
+
margin-right: 5px;
|
81 |
+
}
|
82 |
+
|
83 |
+
.bold-text {
|
84 |
+
font-weight: 700 !important;
|
85 |
+
}
|
86 |
+
|
87 |
+
.text-facts {
|
88 |
+
line-height: 40px;
|
89 |
+
}
|
90 |
+
|
91 |
+
/* footer */
|
92 |
+
footer {
|
93 |
+
display: none !important;
|
94 |
+
}
|
src/utils.py
ADDED
@@ -0,0 +1,389 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
from typing import Callable, List, Tuple
|
2 |
+
|
3 |
+
import numpy as np
|
4 |
+
import pandas as pd
|
5 |
+
|
6 |
+
from gensim.models.doc2vec import Doc2Vec, TaggedDocument
|
7 |
+
|
8 |
+
import tensorflow as tf
|
9 |
+
from tensorflow import keras
|
10 |
+
from keras.preprocessing.text import Tokenizer
|
11 |
+
|
12 |
+
|
13 |
+
def read_data(filepath="../csvs/"):
|
14 |
+
"""
|
15 |
+
Reading CSV files of the dataset.
|
16 |
+
|
17 |
+
Parameters:
|
18 |
+
----------
|
19 |
+
- filepath : str
|
20 |
+
Defines the path that contains the CSV files.
|
21 |
+
|
22 |
+
Returns:
|
23 |
+
--------
|
24 |
+
A tuple contains the following:
|
25 |
+
- X_train : pd.DataFrame
|
26 |
+
- y_train : pd.Series
|
27 |
+
- X_test : pd.DataFrame
|
28 |
+
- y_test : pd.Series
|
29 |
+
"""
|
30 |
+
|
31 |
+
X_train = pd.read_csv(filepath + "X_train.csv")
|
32 |
+
X_train = X_train.iloc[:, 1:]
|
33 |
+
|
34 |
+
X_test = pd.read_csv(filepath + "X_test.csv")
|
35 |
+
X_test = X_test.iloc[:, 1:]
|
36 |
+
|
37 |
+
y_train = pd.read_csv(filepath + "y_train.csv")
|
38 |
+
y_train = y_train.iloc[:, 1:]
|
39 |
+
|
40 |
+
y_test = pd.read_csv(filepath + "y_test.csv")
|
41 |
+
y_test = y_test.iloc[:, 1:]
|
42 |
+
|
43 |
+
return X_train, X_test, y_train, y_test
|
44 |
+
|
45 |
+
|
46 |
+
def train_model(
|
47 |
+
model_building_func: Callable[[], keras.models.Sequential],
|
48 |
+
X_train_vectors: pd.DataFrame | np.ndarray | tf.Tensor,
|
49 |
+
y_train: pd.Series,
|
50 |
+
k: int = 4,
|
51 |
+
num_epochs: int = 30,
|
52 |
+
batch_size: int = 64,
|
53 |
+
) -> Tuple[
|
54 |
+
List[keras.models.Sequential],
|
55 |
+
List[List[float]],
|
56 |
+
List[List[float]],
|
57 |
+
List[List[float]],
|
58 |
+
List[List[float]],
|
59 |
+
]:
|
60 |
+
"""
|
61 |
+
Trains a model on `X_train_vectors` and `y_train` using k-fold cross-validation.
|
62 |
+
|
63 |
+
Parameters:
|
64 |
+
-----------
|
65 |
+
- model_building_func : Callable[[], tf.keras.models.Sequential]
|
66 |
+
A function that builds and compiles a Keras Sequential model.
|
67 |
+
- X_train_vectors : pd.DataFrame
|
68 |
+
The training input data.
|
69 |
+
- y_train : pd.Series
|
70 |
+
The training target data.
|
71 |
+
- k : int, optional
|
72 |
+
The number of folds for cross-validation (default is 4).
|
73 |
+
- num_epochs : int, optional
|
74 |
+
The number of epochs to train for (default is 30).
|
75 |
+
- batch_size : int, optional
|
76 |
+
The batch size to use during training (default is 64).
|
77 |
+
|
78 |
+
Returns:
|
79 |
+
--------
|
80 |
+
A tuple containing the following items:
|
81 |
+
- all_models : List[keras.models.Sequential]
|
82 |
+
A list of `k` trained models.
|
83 |
+
- all_losses : List[List[float]]
|
84 |
+
A `k` by `num_epochs` list containing the training losses for each fold.
|
85 |
+
- all_val_losses : List[List[float]]
|
86 |
+
A `k` by `num_epochs` list containing the validation losses for each fold.
|
87 |
+
- all_acc : List[List[float]]
|
88 |
+
A `k` by `num_epochs` list containing the training accuracies for each fold.
|
89 |
+
- all_val_acc : List[List[float]]
|
90 |
+
A `k` by `num_epochs` list containing the validation accuracies for each fold.
|
91 |
+
"""
|
92 |
+
|
93 |
+
num_validation_samples = len(X_train_vectors) // k
|
94 |
+
|
95 |
+
all_models = []
|
96 |
+
all_losses = []
|
97 |
+
all_val_losses = []
|
98 |
+
all_accuracies = []
|
99 |
+
all_val_accuracies = []
|
100 |
+
|
101 |
+
for fold in range(k):
|
102 |
+
print(f"fold: {fold+1}")
|
103 |
+
validation_data = X_train_vectors[
|
104 |
+
num_validation_samples * fold : num_validation_samples * (fold + 1)
|
105 |
+
]
|
106 |
+
validation_targets = y_train[
|
107 |
+
num_validation_samples * fold : num_validation_samples * (fold + 1)
|
108 |
+
]
|
109 |
+
|
110 |
+
training_data = np.concatenate(
|
111 |
+
[
|
112 |
+
X_train_vectors[: num_validation_samples * fold],
|
113 |
+
X_train_vectors[num_validation_samples * (fold + 1) :],
|
114 |
+
]
|
115 |
+
)
|
116 |
+
training_targets = np.concatenate(
|
117 |
+
[
|
118 |
+
y_train[: num_validation_samples * fold],
|
119 |
+
y_train[num_validation_samples * (fold + 1) :],
|
120 |
+
]
|
121 |
+
)
|
122 |
+
|
123 |
+
model = model_building_func()
|
124 |
+
history = model.fit(
|
125 |
+
training_data,
|
126 |
+
training_targets,
|
127 |
+
validation_data=(validation_data, validation_targets),
|
128 |
+
epochs=num_epochs,
|
129 |
+
batch_size=batch_size,
|
130 |
+
)
|
131 |
+
|
132 |
+
all_models.append(model)
|
133 |
+
all_losses.append(history.history["loss"])
|
134 |
+
all_val_losses.append(history.history["val_loss"])
|
135 |
+
all_accuracies.append(history.history["accuracy"])
|
136 |
+
all_val_accuracies.append(history.history["val_accuracy"])
|
137 |
+
|
138 |
+
return (all_models, all_losses, all_val_losses, all_accuracies, all_val_accuracies)
|
139 |
+
|
140 |
+
|
141 |
+
def print_testing_loss_accuracy(
|
142 |
+
all_models: List[keras.models.Sequential],
|
143 |
+
X_test_vectors: pd.DataFrame | np.ndarray | tf.Tensor,
|
144 |
+
y_test: pd.Series,
|
145 |
+
) -> None:
|
146 |
+
"""
|
147 |
+
Displaying testing loss and testing accuracy of each model in `all_models`,
|
148 |
+
and displaying their average.
|
149 |
+
|
150 |
+
Parameters:
|
151 |
+
------------
|
152 |
+
- all_models : List[keras.models.Sequential]
|
153 |
+
A list of size `k` contains trained models.
|
154 |
+
- X_test_vectors : pd.DataFrame
|
155 |
+
Contains testing vectors.
|
156 |
+
- y_test : pd.Series
|
157 |
+
Contains testing labels.
|
158 |
+
"""
|
159 |
+
|
160 |
+
sum_testing_losses = 0.0
|
161 |
+
sum_testing_accuracies = 0.0
|
162 |
+
|
163 |
+
for i, model in enumerate(all_models):
|
164 |
+
print(f"model: {i+1}")
|
165 |
+
loss_accuracy = model.evaluate(X_test_vectors, y_test, verbose=1)
|
166 |
+
sum_testing_losses += loss_accuracy[0]
|
167 |
+
sum_testing_accuracies += loss_accuracy[1]
|
168 |
+
print("====" * 20)
|
169 |
+
|
170 |
+
num_models = len(all_models)
|
171 |
+
avg_testing_loss = sum_testing_losses / num_models
|
172 |
+
avg_testing_acc = sum_testing_accuracies / num_models
|
173 |
+
print(f"average testing loss: {avg_testing_loss:.3f}")
|
174 |
+
print(f"average testing accuracy: {avg_testing_acc:.3f}")
|
175 |
+
|
176 |
+
|
177 |
+
def calculate_average_measures(
|
178 |
+
all_losses: list[list[float]],
|
179 |
+
all_val_losses: list[list[float]],
|
180 |
+
all_accuracies: list[list[float]],
|
181 |
+
all_val_accuracies: list[list[float]],
|
182 |
+
) -> Tuple[
|
183 |
+
List[keras.models.Sequential],
|
184 |
+
List[List[float]],
|
185 |
+
List[List[float]],
|
186 |
+
List[List[float]],
|
187 |
+
List[List[float]],
|
188 |
+
]:
|
189 |
+
"""
|
190 |
+
Calculate the average measures of cross-validated results.
|
191 |
+
|
192 |
+
Parameters:
|
193 |
+
------------
|
194 |
+
- all_losses : List[List[float]]
|
195 |
+
A `k` by `num_epochs` list contains the values of training losses.
|
196 |
+
- all_val_losses : List[List[float]]
|
197 |
+
A `k` by `num_epochs` list contains the values of validation losses.
|
198 |
+
- all_accuracies : List[List[float]]
|
199 |
+
A `k` by `num_epochs` list contains the values of training accuracies.
|
200 |
+
- all_val_accuracies : List[List[float]]
|
201 |
+
A `k` by `num_epochs` list contains the values of validation accuracies.
|
202 |
+
|
203 |
+
Returns:
|
204 |
+
--------
|
205 |
+
A tuple containing the following items:
|
206 |
+
- avg_loss_hist : List[float]
|
207 |
+
A list of length `num_epochs` contains the average of training losses.
|
208 |
+
- avg_val_loss_hist : List[float]
|
209 |
+
A list of length `num_epochs` contains the average of validaton losses.
|
210 |
+
- avg_acc_hist : List[float]
|
211 |
+
A list of length `num_epochs` contains the average of training accuracies.
|
212 |
+
- avg_val_acc_hist : List[float]
|
213 |
+
A list of length `num_epochs` contains the average of validation accuracies.
|
214 |
+
"""
|
215 |
+
|
216 |
+
num_epochs = len(all_losses[0])
|
217 |
+
avg_loss_hist = [np.mean([x[i] for x in all_losses]) for i in range(num_epochs)]
|
218 |
+
avg_val_loss_hist = [
|
219 |
+
np.mean([x[i] for x in all_val_losses]) for i in range(num_epochs)
|
220 |
+
]
|
221 |
+
avg_acc_hist = [np.mean([x[i] for x in all_accuracies]) for i in range(num_epochs)]
|
222 |
+
avg_val_acc_hist = [
|
223 |
+
np.mean([x[i] for x in all_val_accuracies]) for i in range(num_epochs)
|
224 |
+
]
|
225 |
+
|
226 |
+
return (avg_loss_hist, avg_val_loss_hist, avg_acc_hist, avg_val_acc_hist)
|
227 |
+
|
228 |
+
|
229 |
+
class Doc2VecModel:
|
230 |
+
"""Responsible of creating, initializing, and training Doc2Vec embeddings model."""
|
231 |
+
|
232 |
+
def __init__(self, vector_size=50, min_count=2, epochs=100, dm=1, window=5) -> None:
|
233 |
+
"""
|
234 |
+
Initalize a Doc2Vec model.
|
235 |
+
|
236 |
+
Parameters:
|
237 |
+
------------
|
238 |
+
- vector_size : int, optional
|
239 |
+
Dimensionality of the feature vectors (Default is 50).
|
240 |
+
- min_count : int, optional
|
241 |
+
Ignores all words with total frequency lower than this (Default is 2).
|
242 |
+
- epochs : int, optional
|
243 |
+
Represents the number of training epochs (Default is 100).
|
244 |
+
- dm : int, optional
|
245 |
+
Defines the training algorithm. If `dm=1`, 'distributed memory' (PV-DM) is used.
|
246 |
+
Otherwise, `distributed bag of words` (PV-DBOW) is employed (Default is 1).
|
247 |
+
- window : int, optional
|
248 |
+
The maximum distance between the current and predicted word within a
|
249 |
+
sentence (Default is 5).
|
250 |
+
"""
|
251 |
+
|
252 |
+
self.doc2vec_model = Doc2Vec(
|
253 |
+
vector_size=vector_size,
|
254 |
+
min_count=min_count,
|
255 |
+
epochs=epochs,
|
256 |
+
dm=dm,
|
257 |
+
seed=865,
|
258 |
+
window=window,
|
259 |
+
)
|
260 |
+
|
261 |
+
def train_doc2vec_embeddings_model(
|
262 |
+
self, tagged_docs_train: List[TaggedDocument]
|
263 |
+
) -> Doc2Vec:
|
264 |
+
"""
|
265 |
+
Train Doc2Vec model on `tagged_docs_train`.
|
266 |
+
|
267 |
+
Parameters:
|
268 |
+
------------
|
269 |
+
- tagged_docs_train : list[TaggedDocument]
|
270 |
+
Contains the required format of training Doc2Vec model.
|
271 |
+
|
272 |
+
Returns:
|
273 |
+
--------
|
274 |
+
- doc2vec_model : Doc2Vec
|
275 |
+
The trained Doc2Vec model.
|
276 |
+
"""
|
277 |
+
|
278 |
+
self.doc2vec_model.build_vocab(tagged_docs_train)
|
279 |
+
self.doc2vec_model.train(
|
280 |
+
tagged_docs_train,
|
281 |
+
total_examples=self.doc2vec_model.corpus_count,
|
282 |
+
epochs=self.doc2vec_model.epochs,
|
283 |
+
)
|
284 |
+
|
285 |
+
return self.doc2vec_model
|
286 |
+
|
287 |
+
|
288 |
+
class GloveModel:
|
289 |
+
"""Responsible for creating and generating the glove embedding layer"""
|
290 |
+
|
291 |
+
def __init__(self) -> None:
|
292 |
+
pass
|
293 |
+
|
294 |
+
def _generate_glove_embedding_index(
|
295 |
+
self, glove_file_path: str = "GloVe/glove.6B.50d.txt"
|
296 |
+
) -> dict:
|
297 |
+
"""
|
298 |
+
Responsible for generating glove embedding index.
|
299 |
+
|
300 |
+
Parameters:
|
301 |
+
------------
|
302 |
+
- glove_file_path : str
|
303 |
+
Defines the path of the pretrained GloVe embeddings text file
|
304 |
+
(Default is "GloVe/glove.6B.50d.txt").
|
305 |
+
|
306 |
+
Returns:
|
307 |
+
--------
|
308 |
+
- embedding_index : dict
|
309 |
+
Contains each word as a key, and its co-effeicents as a value.
|
310 |
+
"""
|
311 |
+
|
312 |
+
embeddings_index = {}
|
313 |
+
with open(glove_file_path, encoding="utf8") as f:
|
314 |
+
for line in f:
|
315 |
+
values = line.split()
|
316 |
+
word = values[0]
|
317 |
+
coefs = np.asarray(values[1:], dtype="float32")
|
318 |
+
embeddings_index[word] = coefs
|
319 |
+
|
320 |
+
return embeddings_index
|
321 |
+
|
322 |
+
def _generate_glove_embedding_matrix(
|
323 |
+
self, word_index: dict, embedding_index: dict, max_length: int
|
324 |
+
) -> np.ndarray:
|
325 |
+
"""
|
326 |
+
Generating embedding matrix of each word in `word_index`.
|
327 |
+
|
328 |
+
Parameters:
|
329 |
+
-----------
|
330 |
+
- word_index : dict
|
331 |
+
Contains words as keys with there indicies as values.
|
332 |
+
- embedding_index : dict
|
333 |
+
Contains each word as a key, and its co-effeicents as a value.
|
334 |
+
- max_length : int
|
335 |
+
Defines the size of the embedding vector of each word in the
|
336 |
+
embedding matrix.
|
337 |
+
|
338 |
+
Returns:
|
339 |
+
--------
|
340 |
+
- embedding_matrix : np.ndarray
|
341 |
+
Contains all embedding vectors for each word in`word_index`.
|
342 |
+
"""
|
343 |
+
|
344 |
+
embedding_matrix = np.zeros((len(word_index) + 1, max_length))
|
345 |
+
|
346 |
+
for word, i in word_index.items():
|
347 |
+
embedding_vector = embedding_index.get(word)
|
348 |
+
if embedding_vector is not None:
|
349 |
+
embedding_matrix[i] = embedding_vector
|
350 |
+
|
351 |
+
return embedding_matrix
|
352 |
+
|
353 |
+
def generate_glove_embedding_layer(
|
354 |
+
self, glove_tokenizer: Tokenizer, max_length: int = 50
|
355 |
+
) -> keras.layers.Embedding:
|
356 |
+
"""
|
357 |
+
Create GloVe embedding layer for later usage in the neural network.
|
358 |
+
|
359 |
+
Paramters:
|
360 |
+
----------
|
361 |
+
- glove_tokenizer : Tokenizer
|
362 |
+
Trained tokenizer on training data to extract word index from it.
|
363 |
+
- max_length : int, optional
|
364 |
+
Defines the maximum length of the output embedding vector for
|
365 |
+
each word. (Default is 50).
|
366 |
+
|
367 |
+
Returns:
|
368 |
+
--------
|
369 |
+
- embedding_layer : keras.layers.Embedding
|
370 |
+
An embedding layer of size `word index + 1` by `max_length` with
|
371 |
+
trained weights that can be used a vectorizer of case facts.
|
372 |
+
"""
|
373 |
+
|
374 |
+
word_index = glove_tokenizer.word_index
|
375 |
+
|
376 |
+
embedding_index = self._generate_glove_embedding_index()
|
377 |
+
embedding_matrix = self._generate_glove_embedding_matrix(
|
378 |
+
word_index, embedding_index, max_length
|
379 |
+
)
|
380 |
+
|
381 |
+
embedding_layer = keras.layers.Embedding(
|
382 |
+
len(word_index) + 1,
|
383 |
+
max_length,
|
384 |
+
weights=[embedding_matrix],
|
385 |
+
input_length=max_length,
|
386 |
+
trainable=False,
|
387 |
+
)
|
388 |
+
|
389 |
+
return embedding_layer
|