Dataset Preview
Duplicate
The full dataset viewer is not available (click to read why). Only showing a preview of the rows.
The dataset generation failed
Error code:   DatasetGenerationError
Exception:    CastError
Message:      Couldn't cast
(0.5, 1.0, 0.5, 0.0): list<item: int64>
  child 0, item: int64
focus: list<item: int64>
  child 0, item: int64
gold_name_in_issue: bool
hybrid_tok: int64
anchor_dist: int64
skip: string
n_gold_files: int64
n_anchors: int64
repo: string
n_gold: int64
graph_tok: int64
iid: string
gold_file_in_issue: bool
q_tokens: int64
graph_frank: int64
anchors_frank: int64
hybrid_rank: int64
repo_tokens: int64
new_only: bool
bm25_rank: int64
graph_rank: int64
hybrid_frank: int64
bm25_frank: int64
anchors_tok: int64
bm25_tok: int64
split: string
n_nodes: int64
anchors_rank: int64
to
{'iid': Value('string'), 'repo': Value('string'), 'split': Value('string'), 'n_nodes': Value('int64'), 'repo_tokens': Value('int64'), 'n_gold': Value('int64'), 'n_gold_files': Value('int64'), 'q_tokens': Value('int64'), 'gold_name_in_issue': Value('bool'), 'gold_file_in_issue': Value('bool'), 'anchor_dist': Value('int64'), 'n_anchors': Value('int64'), 'bm25_rank': Value('int64'), 'bm25_tok': Value('int64'), 'bm25_frank': Value('int64'), 'anchors_rank': Value('int64'), 'anchors_tok': Value('int64'), 'anchors_frank': Value('int64'), 'graph_rank': Value('int64'), 'graph_tok': Value('int64'), 'graph_frank': Value('int64'), 'hybrid_rank': Value('int64'), 'hybrid_tok': Value('int64'), 'hybrid_frank': Value('int64'), 'skip': Value('string'), 'new_only': Value('bool')}
because column names don't match
Traceback:    Traceback (most recent call last):
                File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1835, in _prepare_split_single
                  for key, table in generator:
                                    ^^^^^^^^^
                File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 621, in wrapped
                  for item in generator(*args, **kwargs):
                              ~~~~~~~~~^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 361, in _generate_tables
                  self._cast_table(pa_table, json_field_paths=json_field_paths),
                  ~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 132, in _cast_table
                  pa_table = table_cast(pa_table, self.info.features.arrow_schema)
                File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2381, in table_cast
                  return cast_table_to_schema(table, schema)
                File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2309, in cast_table_to_schema
                  raise CastError(
                  ...<3 lines>...
                  )
              datasets.table.CastError: Couldn't cast
              (0.5, 1.0, 0.5, 0.0): list<item: int64>
                child 0, item: int64
              focus: list<item: int64>
                child 0, item: int64
              gold_name_in_issue: bool
              hybrid_tok: int64
              anchor_dist: int64
              skip: string
              n_gold_files: int64
              n_anchors: int64
              repo: string
              n_gold: int64
              graph_tok: int64
              iid: string
              gold_file_in_issue: bool
              q_tokens: int64
              graph_frank: int64
              anchors_frank: int64
              hybrid_rank: int64
              repo_tokens: int64
              new_only: bool
              bm25_rank: int64
              graph_rank: int64
              hybrid_frank: int64
              bm25_frank: int64
              anchors_tok: int64
              bm25_tok: int64
              split: string
              n_nodes: int64
              anchors_rank: int64
              to
              {'iid': Value('string'), 'repo': Value('string'), 'split': Value('string'), 'n_nodes': Value('int64'), 'repo_tokens': Value('int64'), 'n_gold': Value('int64'), 'n_gold_files': Value('int64'), 'q_tokens': Value('int64'), 'gold_name_in_issue': Value('bool'), 'gold_file_in_issue': Value('bool'), 'anchor_dist': Value('int64'), 'n_anchors': Value('int64'), 'bm25_rank': Value('int64'), 'bm25_tok': Value('int64'), 'bm25_frank': Value('int64'), 'anchors_rank': Value('int64'), 'anchors_tok': Value('int64'), 'anchors_frank': Value('int64'), 'graph_rank': Value('int64'), 'graph_tok': Value('int64'), 'graph_frank': Value('int64'), 'hybrid_rank': Value('int64'), 'hybrid_tok': Value('int64'), 'hybrid_frank': Value('int64'), 'skip': Value('string'), 'new_only': Value('bool')}
              because column names don't match
              
              The above exception was the direct cause of the following exception:
              
              Traceback (most recent call last):
                File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 1379, in compute_config_parquet_and_info_response
                  parquet_operations, partial, estimated_dataset_info = stream_convert_to_parquet(
                                                                        ~~~~~~~~~~~~~~~~~~~~~~~~~^
                      builder, max_dataset_size_bytes=max_dataset_size_bytes
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                  )
                  ^
                File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 956, in stream_convert_to_parquet
                  builder._prepare_split(split_generator=splits_generators[split], file_format="parquet")
                  ~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1702, in _prepare_split
                  for job_id, done, content in self._prepare_split_single(
                                               ~~~~~~~~~~~~~~~~~~~~~~~~~~^
                      gen_kwargs=gen_kwargs, job_id=job_id, **_prepare_split_args
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                  ):
                  ^
                File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1887, in _prepare_split_single
                  raise DatasetGenerationError("An error occurred while generating the dataset") from e
              datasets.exceptions.DatasetGenerationError: An error occurred while generating the dataset

Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.

iid
string
repo
string
split
string
n_nodes
int64
repo_tokens
int64
n_gold
int64
n_gold_files
int64
q_tokens
int64
gold_name_in_issue
bool
gold_file_in_issue
bool
anchor_dist
int64
n_anchors
int64
bm25_rank
int64
bm25_tok
int64
bm25_frank
int64
anchors_rank
int64
anchors_tok
int64
anchors_frank
int64
graph_rank
int64
graph_tok
int64
graph_frank
int64
hybrid_rank
int64
hybrid_tok
int64
hybrid_frank
int64
skip
string
new_only
bool
fastapi_15661
fastapi/fastapi
public4
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
no_indexable_gold
false
fastapi_15588
fastapi/fastapi
public4
656
394,445
3
1
472
false
true
3
1
3
3,937
3
null
null
null
148
253,273
null
3
47,835
3
null
null
fastapi_15589
fastapi/fastapi
public4
658
394,670
1
1
311
false
false
null
0
9
8,417
7
null
null
null
null
null
null
9
8,417
7
null
null
fastapi_15030
fastapi/fastapi
public4
645
387,571
7
9
37
false
false
2
1
105
231,331
1
null
null
null
47
155,640
2
75
211,136
2
null
null
fastapi_14962
fastapi/fastapi
public4
636
380,287
4
3
52
false
false
null
0
2
3,632
0
null
null
null
null
null
null
3
7,366
0
null
null
fastapi_15280
fastapi/fastapi
public4
652
395,528
3
2
16
false
false
null
0
41
57,423
1
null
null
null
null
null
null
10
46,839
3
null
null
fastapi_14964
fastapi/fastapi
public4
637
381,192
3
1
92
true
true
0
3
0
440
0
0
136
0
0
122
0
0
122
0
null
null
fastapi_15023
fastapi/fastapi
public4
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
no_indexable_gold
false
fastapi_14986
fastapi/fastapi
public4
642
384,695
3
2
250
false
false
3
1
0
600
0
null
null
null
19
64,362
4
1
722
1
null
null
fastapi_14978
fastapi/fastapi
public4
642
383,090
9
3
28
false
false
null
0
2
3,984
2
null
null
null
null
null
null
3
4,176
2
null
null
requests_7505
psf/requests
public4
326
108,786
6
2
129
false
false
1
2
0
3,067
0
null
null
0
2
1,499
0
2
1,499
0
null
null
requests_7502
psf/requests
public4
326
108,708
1
1
16
true
false
0
2
0
659
0
0
659
0
0
659
0
0
659
0
null
null
requests_7433
psf/requests
public4
326
109,186
1
1
124
true
false
0
5
0
757
0
1
848
1
2
1,123
1
0
757
0
null
null
requests_7427
psf/requests
public4
326
109,019
1
1
98
false
false
null
0
0
535
0
null
null
null
null
null
null
0
535
0
null
null
requests_7315
psf/requests
public4
296
93,700
1
1
173
true
false
0
6
1
585
1
0
306
0
0
306
0
0
306
0
null
null
requests_7328
psf/requests
public4
296
93,700
1
1
145
false
false
1
1
9
13,031
0
null
null
0
5
7,540
0
2
5,514
0
null
null
requests_7309
psf/requests
public4
296
93,730
1
1
101
true
false
0
1
0
193
0
0
193
0
0
193
0
0
193
0
null
null
requests_7205
psf/requests
public4
296
93,748
1
1
152
false
false
3
1
0
320
0
null
null
null
87
65,613
6
3
6,434
1
null
null
rich_4070
Textualize/rich
public4
1,169
474,384
26
9
774
true
true
0
20
0
3,114
0
0
2,958
0
0
2,958
0
0
2,958
0
null
null
rich_4077
Textualize/rich
public4
1,169
475,198
2
1
16
false
false
null
0
27
53,411
null
null
null
null
null
null
null
39
67,539
null
null
null
rich_4079
Textualize/rich
public4
1,170
475,247
2
1
17
false
false
null
0
59
81,574
null
null
null
null
null
null
null
30
60,403
null
null
null
rich_4076
Textualize/rich
public4
1,169
475,179
2
1
17
false
false
null
0
631
140,810
null
null
null
null
null
null
null
465
336,180
null
null
null
rich_4075
Textualize/rich
public4
1,169
475,114
3
1
25
false
false
null
0
137
126,032
null
null
null
null
null
null
null
89
115,384
null
null
null
rich_4006
Textualize/rich
public4
1,169
473,642
1
1
24
true
false
0
1
0
477
0
0
477
0
0
477
0
0
477
0
null
null
fastapi_14258
fastapi/fastapi
public4
630
380,074
1
1
104
false
false
null
0
9
8,731
7
null
null
null
null
null
null
9
8,731
6
null
null
fastapi_14851
fastapi/fastapi
public4
628
380,133
5
1
116
true
false
0
3
1
466
1
0
248
0
1
45,023
1
0
248
0
null
null
fastapi_14873
fastapi/fastapi
public4
630
380,074
1
1
168
true
true
0
5
4
3,084
0
0
2,535
0
0
2,535
0
0
2,535
0
null
null
fastapi_13920
fastapi/fastapi
public4
627
380,038
2
1
56
false
false
null
0
77
151,082
null
null
null
null
null
null
null
19
35,119
9
null
null
fastapi_14794
fastapi/fastapi
public4
627
380,009
1
1
204
true
false
0
4
0
1,704
0
3
2,637
3
6
12,560
3
1
2,534
1
null
null
fastapi_14786
fastapi/fastapi
public4
626
375,131
2
1
575
true
true
0
8
70
209,621
9
1
283
1
1
283
1
2
519
2
null
null
fastapi_14492
fastapi/fastapi
public4
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
no_indexable_gold
false
fastapi_14479
fastapi/fastapi
public4
659
391,921
1
1
613
true
true
0
10
1
2,765
0
1
5,481
0
1
5,481
0
1
5,481
0
null
null
fastapi_14463
fastapi/fastapi
public4
592
365,018
2
1
2,523
true
true
0
9
1
7,874
1
2
49,316
2
2
49,316
2
0
4,155
0
null
null
fastapi_12942
fastapi/fastapi
public4
530
349,897
1
1
348
false
false
3
7
246
315,629
null
null
null
null
209
303,670
null
201
310,325
9
null
null
fastapi_14099
fastapi/fastapi
public4
519
346,898
8
5
161
false
false
3
1
0
1,516
0
null
null
null
34
17,959
3
3
4,261
2
null
null
fastapi_14609
fastapi/fastapi
public4
660
392,125
133
20
36
true
true
0
2
0
2,229
0
0
1,705
0
0
1,705
0
0
1,705
0
null
null
fastapi_14372
fastapi/fastapi
public4
640
385,951
4
2
290
true
false
0
5
71
209,701
null
0
41
0
4
7,686
1
3
46,058
2
null
null
fastapi_14349
fastapi/fastapi
public4
640
385,931
1
1
42
false
false
null
0
9
9,556
0
null
null
null
null
null
null
7
9,439
1
null
null
fastapi_14512
fastapi/fastapi
public4
666
393,769
2
1
72
false
false
3
3
198
308,417
null
null
null
null
38
50,124
7
90
240,067
null
null
null
fastapi_14487
fastapi/fastapi
public4
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
no_indexable_gold
false
fastapi_11355
fastapi/fastapi
public4
658
391,076
3
1
74
false
false
null
0
230
293,524
6
null
null
null
null
null
null
117
237,507
8
null
null
fastapi_14485
fastapi/fastapi
public4
665
393,709
2
1
79
false
false
null
0
471
352,245
5
null
null
null
null
null
null
410
345,768
6
null
null
fastapi_14482
fastapi/fastapi
public4
665
393,551
4
1
66
false
false
null
0
198
323,782
7
null
null
null
null
null
null
144
247,925
9
null
null
fastapi_14306
fastapi/fastapi
public4
661
391,836
13
2
697
true
true
0
3
0
1,289
0
0
92
0
0
92
0
0
17
0
null
null
fastapi_14458
fastapi/fastapi
public4
661
391,709
3
1
147
true
false
0
17
0
259
0
0
304
0
2
1,882
0
0
259
0
null
null
fastapi_14459
fastapi/fastapi
public4
658
391,190
9
3
203
false
false
3
1
27
110,278
4
null
null
null
66
81,923
1
34
146,571
5
null
null
fastapi_14448
fastapi/fastapi
public4
657
389,377
8
2
306
false
false
null
0
44
140,693
null
null
null
null
null
null
null
91
237,811
4
null
null
fastapi_14455
fastapi/fastapi
public4
658
391,092
1
1
70
false
false
3
1
35
30,464
6
null
null
null
66
81,825
1
56
107,704
4
null
null
fastapi_9753
fastapi/fastapi
public4
657
389,291
3
1
172
false
false
null
0
21
68,989
9
null
null
null
null
null
null
68
184,177
null
null
null
fastapi_14430
fastapi/fastapi
public4
657
389,265
2
1
83
false
false
null
0
86
253,864
null
null
null
null
null
null
null
53
206,311
null
null
null
fastapi_5077
fastapi/fastapi
public4
657
389,282
2
1
669
true
true
0
7
0
130
0
0
130
0
1
3,710
0
0
130
0
null
null
fastapi_9555
fastapi/fastapi
public4
656
389,030
4
1
733
false
false
2
4
46
96,558
7
null
null
null
125
217,635
null
76
206,927
9
null
null
fastapi_14361
fastapi/fastapi
public4
647
386,802
1
1
24
false
true
null
0
2
932
0
null
null
null
null
null
null
0
389
0
null
null
fastapi_13207
fastapi/fastapi
public4
656
388,938
2
1
88
false
false
2
2
1
573
1
null
null
null
32
106,936
5
4
48,382
3
null
null
fastapi_14360
fastapi/fastapi
public4
656
388,821
1
1
188
false
false
4
2
0
845
0
null
null
null
65
48,228
4
5
7,051
3
null
null
fastapi_14356
fastapi/fastapi
public4
656
388,566
1
1
370
false
false
2
4
65
229,405
3
null
null
null
114
222,718
5
85
240,111
6
null
null
fastapi_5624
fastapi/fastapi
public4
654
387,830
1
1
15
false
false
null
0
0
806
0
null
null
null
null
null
null
0
806
0
null
null
fastapi_14419
fastapi/fastapi
public4
654
387,831
7
2
172
false
false
1
1
3
3,144
2
null
null
null
3
5,515
1
1
1,787
1
null
null
fastapi_14301
fastapi/fastapi
public4
640
385,944
1
1
184
true
false
0
5
0
118
0
1
924
0
1
924
0
0
118
0
null
null
fastapi_14262
fastapi/fastapi
public4
638
384,150
24
10
609
true
false
0
4
4
7,374
0
0
424
0
0
2,741
0
0
424
0
null
null
fastapi_14266
fastapi/fastapi
public4
638
384,217
1
1
61
false
false
4
1
146
255,166
null
null
null
null
110
105,968
8
85
226,146
null
null
null
fastapi_14246
fastapi/fastapi
public4
643
384,172
1
1
68
false
false
null
0
20
18,369
0
null
null
null
null
null
null
22
111,970
3
null
null
fastapi_14186
fastapi/fastapi
public4
642
382,775
39
10
252
true
true
0
1
1
2,714
1
0
2,753
0
0
2,753
0
0
2,753
0
null
null
fastapi_9425
fastapi/fastapi
public4
519
346,855
1
1
58
false
false
null
0
31
79,572
4
null
null
null
null
null
null
45
177,840
5
null
null
fastapi_13713
fastapi/fastapi
public4
518
346,572
3
2
189
true
false
1
2
0
8,365
0
null
null
0
3
48,300
0
3
54,551
0
null
null
fastapi_14077
fastapi/fastapi
public4
518
346,594
4
3
169
false
false
2
1
0
333
0
null
null
null
54
177,155
4
3
47,859
3
null
null
rich_3718
Textualize/rich
public4
1,141
418,535
1
1
187
false
true
1
1
43
55,493
null
null
null
0
3
5,797
0
18
25,793
0
null
null
rich_3934
Textualize/rich
public4
1,166
470,902
4
2
16
false
true
null
0
41
55,990
null
null
null
null
null
null
null
58
75,339
null
null
null
rich_3953
Textualize/rich
public4
1,169
473,611
2
1
32
false
false
null
0
957
227,497
null
null
null
null
null
null
null
null
null
null
null
null
rich_3944
Textualize/rich
public4
1,169
473,611
1
1
15
false
false
null
0
948
218,286
null
null
null
null
null
null
null
null
null
null
null
null
rich_3942
Textualize/rich
public4
1,168
473,552
7
2
13
true
true
null
0
15
10,018
2
null
null
null
null
null
null
10
5,950
0
null
null
rich_3938
Textualize/rich
public4
1,167
472,872
3
2
22
false
false
null
0
19
46,407
null
null
null
null
null
null
null
26
76,669
null
null
null
rich_3882
Textualize/rich
public4
1,141
418,463
2
1
163
true
true
0
2
7
2,252
0
0
78
0
0
78
0
0
78
0
null
null
rich_3894
Textualize/rich
public4
1,141
418,463
1
1
493
false
false
2
2
8
12,492
1
null
null
null
42
61,792
2
9
16,297
2
null
null
rich_3935
Textualize/rich
public4
1,167
472,387
5
1
18
false
false
null
0
29
61,181
null
null
null
null
null
null
null
51
84,681
null
null
null
rich_3905
Textualize/rich
public4
1,167
472,335
2
1
418
true
true
0
9
1
5,872
1
0
5,501
0
0
5,501
0
0
5,501
0
null
null
rich_3930
Textualize/rich
public4
1,141
418,463
8
26
306
false
false
null
0
1
1,621
1
null
null
null
null
null
null
0
1,250
0
null
null
httpx_3672
encode/httpx
public4
728
140,768
24
7
60
true
false
0
2
0
193
0
0
193
0
0
196
0
0
193
0
null
null
fastapi_14791
fastapi/fastapi
public4
626
375,168
1
1
147
true
true
0
5
0
4,155
0
3
4,771
3
1
4,425
1
0
4,155
0
null
null
fastapi_15800
fastapi/fastapi
public4
713
408,498
13
8
309
false
false
null
0
15
61,418
7
null
null
null
null
null
null
17
72,838
7
null
null
fastapi_15785
fastapi/fastapi
public4
702
407,830
6
2
315
false
false
null
0
23
122,544
6
null
null
null
null
null
null
22
75,400
7
null
null
fastapi_14371
fastapi/fastapi
public4
667
393,999
13
4
99
false
false
null
0
74
224,154
5
null
null
null
null
null
null
45
160,342
3
null
null
fastapi_14605
fastapi/fastapi
public4
659
391,921
16
7
60
false
false
null
0
1
12,278
1
null
null
null
null
null
null
0
3,611
0
null
null
fastapi_15763
fastapi/fastapi
public4
702
407,203
1
1
266
false
false
null
0
28
73,799
7
null
null
null
null
null
null
46
156,888
7
null
null
fastapi_14583
fastapi/fastapi
public4
663
391,573
4
2
26
false
false
3
1
47
48,685
8
null
null
null
38
24,685
1
67
56,169
3
null
null
fastapi_15745
fastapi/fastapi
public4
668
397,915
19
4
1,076
true
false
0
7
0
2,812
0
0
2,352
0
0
2,352
0
0
38,850
0
null
null
fastapi_14297
fastapi/fastapi
public4
656
388,721
1
1
304
false
true
1
7
3
4,351
2
null
null
2
138
236,292
2
13
69,028
2
null
null
fastapi_14303
fastapi/fastapi
public4
656
388,526
1
1
661
false
true
1
8
17
37,111
3
null
null
2
77
198,779
2
27
95,479
3
null
null
fastapi_13537
fastapi/fastapi
public4
656
388,509
1
1
412
true
false
0
3
2
1,583
0
1
549
0
1
549
0
1
1,128
0
null
null
fastapi_14953
fastapi/fastapi
public4
634
380,202
3
4
443
true
true
3
1
5
10,473
2
null
null
null
8
17,096
3
4
8,539
2
null
null
fastapi_13786
fastapi/fastapi
public4
647
386,802
25
11
1,332
true
true
0
10
1
3,380
1
0
803
0
0
803
0
0
803
0
null
null
fastapi_14616
fastapi/fastapi
public4
627
380,074
2
1
294
false
false
1
8
35
146,947
2
null
null
null
74
194,864
null
29
144,104
7
null
null
fastapi_11194
fastapi/fastapi
public4
519
346,911
1
1
154
false
false
3
5
57
173,326
5
null
null
null
149
230,510
2
83
204,614
3
null
null
requests_6757
psf/requests
public4
300
95,358
3
2
67
false
false
null
0
0
3,203
0
null
null
null
null
null
null
0
3,203
0
null
null
requests_6629
psf/requests
public4
295
91,388
1
1
209
true
true
1
1
3
1,602
0
null
null
0
1
1,089
0
2
1,198
0
null
null
requests_6644
psf/requests
public4
296
91,724
1
1
73
false
false
null
0
24
32,407
0
null
null
null
null
null
null
16
24,026
0
null
null
requests_6589
psf/requests
public4
295
91,388
1
1
59
true
false
0
1
0
621
0
0
621
0
0
621
0
0
621
0
null
null
requests_6592
psf/requests
public4
295
91,396
1
1
59
false
false
null
0
0
1,057
0
null
null
null
null
null
null
1
4,623
1
null
null
rich_3180
Textualize/rich
public4
1,134
409,759
5
2
297
false
false
null
0
4
5,713
2
null
null
null
null
null
null
0
1,095
0
null
null
rich_3486
Textualize/rich
public4
1,137
413,822
10
3
24
true
true
0
1
3
628
1
0
3,986
0
0
3,986
0
0
3,986
0
null
null
End of preview.

Prose Is Noise: Where a 32K-Context Gemma 4 Agent Should Look First

Code and per-task results for our Kaggle Gemma 4 paper-track submission. Everything is pure Python (stdlib only) and runs on a CPU.

Headline (held-out, 1,040 tasks, 31 repos; tuned on dev only)

Strategy sym@1 gold within 4K tokens
BM25 (full issue) 31.4% 51.9%
Graph (PPR from issue anchors) 23.8% 34.1%
Fusion (focused BM25 + 0.5 full BM25 + 0.5 anchors) 42.4% 64.3%

Files

  • localize.py – the localizer (BM25, anchors, personalized PageRank, focused query, fusion). shortlist(rec, issue_text, budget_tokens=4000) returns ranked symbols.
  • extract.py – builds per-task symbol graphs + gold labels at each task's base_commit.
  • evaluate.py, tune.py, report.py – evaluation, dev-only weight grid, summary tables.
  • results_full.jsonl, results_title.jsonl – per-task ranks / tokens-to-gold for every baseline (full issue and title-only).
  • tune_public4.json (dev grid, 72 settings) and tune_heldout.json (single held-out run of the chosen weights (0.5, 1.0, 0.5, 0.0)).
  • writeup.md – the paper.

Reproduce

  1. Put the competition's public task files in this folder as tasks.jsonl (dev, 129 tasks) and heldout1k.jsonl.gz (1,046 tasks).
  2. bash clone_full.sh (bare clones of the 35 repos in repos.txt into repos/)
  3. python3 extract.py (~5 min, 2 CPUs) β†’ idx/
  4. python3 evaluate.py full && python3 evaluate.py title && python3 report.py full title
  5. python3 tune.py public4 then python3 tune.py heldout 0.5,1.0,0.5,0

Bootstrap seeds: 0 (CIs) and 1 (paired differences), 1,000 / 2,000 resamples.

Downloads last month
55