Fix custom configs from script

[lhoestq]

We should not use the parquet export when the user is passing config_kwargs

I also fixed a regression that would disallow creating a custom config when a dataset has multiple predefined configs

fix #6533

[HuggingFaceDocBuilderDev]

The docs for this PR live here. All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.

[github-actions]

Show benchmarks

PyArrow==8.0.0

Show updated benchmarks!

Benchmark: benchmark_array_xd.json

metric	read_batch_formatted_as_numpy after write_array2d	read_batch_formatted_as_numpy after write_flattened_sequence	read_batch_formatted_as_numpy after write_nested_sequence	read_batch_unformated after write_array2d	read_batch_unformated after write_flattened_sequence	read_batch_unformated after write_nested_sequence	read_col_formatted_as_numpy after write_array2d	read_col_formatted_as_numpy after write_flattened_sequence	read_col_formatted_as_numpy after write_nested_sequence	read_col_unformated after write_array2d	read_col_unformated after write_flattened_sequence	read_col_unformated after write_nested_sequence	read_formatted_as_numpy after write_array2d	read_formatted_as_numpy after write_flattened_sequence	read_formatted_as_numpy after write_nested_sequence	read_unformated after write_array2d	read_unformated after write_flattened_sequence	read_unformated after write_nested_sequence	write_array2d	write_flattened_sequence	write_nested_sequence
new / old (diff)	0.005462 / 0.011353 (-0.005891)	0.003918 / 0.011008 (-0.007090)	0.065021 / 0.038508 (0.026513)	0.032620 / 0.023109 (0.009511)	0.249794 / 0.275898 (-0.026104)	0.277330 / 0.323480 (-0.046150)	0.002962 / 0.007986 (-0.005023)	0.003435 / 0.004328 (-0.000894)	0.048992 / 0.004250 (0.044742)	0.046841 / 0.037052 (0.009788)	0.252459 / 0.258489 (-0.006030)	0.287889 / 0.293841 (-0.005952)	0.028322 / 0.128546 (-0.100224)	0.011214 / 0.075646 (-0.064432)	0.208555 / 0.419271 (-0.210717)	0.037004 / 0.043533 (-0.006529)	0.262537 / 0.255139 (0.007398)	0.307418 / 0.283200 (0.024218)	0.021552 / 0.141683 (-0.120131)	1.144252 / 1.452155 (-0.307903)	1.195687 / 1.492716 (-0.297029)

Benchmark: benchmark_getitem_100B.json

metric	get_batch_of_1024_random_rows	get_batch_of_1024_rows	get_first_row	get_last_row
new / old (diff)	0.004766 / 0.018006 (-0.013240)	0.301926 / 0.000490 (0.301436)	0.000218 / 0.000200 (0.000018)	0.000042 / 0.000054 (-0.000012)

Benchmark: benchmark_indices_mapping.json

metric	select	shard	shuffle	sort	train_test_split
new / old (diff)	0.017891 / 0.037411 (-0.019521)	0.066848 / 0.014526 (0.052322)	0.075522 / 0.176557 (-0.101035)	0.120762 / 0.737135 (-0.616374)	0.075980 / 0.296338 (-0.220359)

Benchmark: benchmark_iterating.json

metric	read 5000	read 50000	read_batch 50000 10	read_batch 50000 100	read_batch 50000 1000	read_formatted numpy 5000	read_formatted pandas 5000	read_formatted tensorflow 5000	read_formatted torch 5000	read_formatted_batch numpy 5000 10	read_formatted_batch numpy 5000 1000	shuffled read 5000	shuffled read 50000	shuffled read_batch 50000 10	shuffled read_batch 50000 100	shuffled read_batch 50000 1000	shuffled read_formatted numpy 5000	shuffled read_formatted_batch numpy 5000 10	shuffled read_formatted_batch numpy 5000 1000
new / old (diff)	0.284843 / 0.215209 (0.069634)	2.816260 / 2.077655 (0.738605)	1.484370 / 1.504120 (-0.019750)	1.362090 / 1.541195 (-0.179104)	1.421729 / 1.468490 (-0.046762)	0.561673 / 4.584777 (-4.023104)	2.370793 / 3.745712 (-1.374919)	2.982639 / 5.269862 (-2.287223)	1.834614 / 4.565676 (-2.731063)	0.063158 / 0.424275 (-0.361117)	0.005044 / 0.007607 (-0.002563)	0.339834 / 0.226044 (0.113790)	3.369051 / 2.268929 (1.100122)	1.821040 / 55.444624 (-53.623584)	1.544009 / 6.876477 (-5.332468)	1.603902 / 2.142072 (-0.538171)	0.638151 / 4.805227 (-4.167076)	0.117012 / 6.500664 (-6.383652)	0.042999 / 0.075469 (-0.032470)

Benchmark: benchmark_map_filter.json

metric	filter	map fast-tokenizer batched	map identity	map identity batched	map no-op batched	map no-op batched numpy	map no-op batched pandas	map no-op batched pytorch	map no-op batched tensorflow
new / old (diff)	0.941809 / 1.841788 (-0.899978)	12.279635 / 8.074308 (4.205326)	10.212876 / 10.191392 (0.021484)	0.129904 / 0.680424 (-0.550519)	0.014210 / 0.534201 (-0.519991)	0.286140 / 0.579283 (-0.293143)	0.267453 / 0.434364 (-0.166911)	0.324417 / 0.540337 (-0.215921)	0.428262 / 1.386936 (-0.958674)

PyArrow==latest

Show updated benchmarks!

Benchmark: benchmark_array_xd.json

metric	read_batch_formatted_as_numpy after write_array2d	read_batch_formatted_as_numpy after write_flattened_sequence	read_batch_formatted_as_numpy after write_nested_sequence	read_batch_unformated after write_array2d	read_batch_unformated after write_flattened_sequence	read_batch_unformated after write_nested_sequence	read_col_formatted_as_numpy after write_array2d	read_col_formatted_as_numpy after write_flattened_sequence	read_col_formatted_as_numpy after write_nested_sequence	read_col_unformated after write_array2d	read_col_unformated after write_flattened_sequence	read_col_unformated after write_nested_sequence	read_formatted_as_numpy after write_array2d	read_formatted_as_numpy after write_flattened_sequence	read_formatted_as_numpy after write_nested_sequence	read_unformated after write_array2d	read_unformated after write_flattened_sequence	read_unformated after write_nested_sequence	write_array2d	write_flattened_sequence	write_nested_sequence
new / old (diff)	0.005351 / 0.011353 (-0.006002)	0.003591 / 0.011008 (-0.007417)	0.048755 / 0.038508 (0.010247)	0.030857 / 0.023109 (0.007748)	0.270301 / 0.275898 (-0.005597)	0.294459 / 0.323480 (-0.029021)	0.004265 / 0.007986 (-0.003720)	0.002712 / 0.004328 (-0.001616)	0.047725 / 0.004250 (0.043475)	0.048392 / 0.037052 (0.011339)	0.274226 / 0.258489 (0.015737)	0.304010 / 0.293841 (0.010169)	0.029283 / 0.128546 (-0.099263)	0.011196 / 0.075646 (-0.064450)	0.057213 / 0.419271 (-0.362058)	0.057504 / 0.043533 (0.013971)	0.266091 / 0.255139 (0.010952)	0.285991 / 0.283200 (0.002791)	0.020030 / 0.141683 (-0.121653)	1.121514 / 1.452155 (-0.330641)	1.192608 / 1.492716 (-0.300108)

Benchmark: benchmark_getitem_100B.json

metric	get_batch_of_1024_random_rows	get_batch_of_1024_rows	get_first_row	get_last_row
new / old (diff)	0.095041 / 0.018006 (0.077035)	0.301255 / 0.000490 (0.300765)	0.000218 / 0.000200 (0.000018)	0.000044 / 0.000054 (-0.000010)

Benchmark: benchmark_indices_mapping.json

metric	select	shard	shuffle	sort	train_test_split
new / old (diff)	0.022265 / 0.037411 (-0.015146)	0.078416 / 0.014526 (0.063890)	0.091097 / 0.176557 (-0.085460)	0.129864 / 0.737135 (-0.607272)	0.091683 / 0.296338 (-0.204655)

Benchmark: benchmark_iterating.json

metric	read 5000	read 50000	read_batch 50000 10	read_batch 50000 100	read_batch 50000 1000	read_formatted numpy 5000	read_formatted pandas 5000	read_formatted tensorflow 5000	read_formatted torch 5000	read_formatted_batch numpy 5000 10	read_formatted_batch numpy 5000 1000	shuffled read 5000	shuffled read 50000	shuffled read_batch 50000 10	shuffled read_batch 50000 100	shuffled read_batch 50000 1000	shuffled read_formatted numpy 5000	shuffled read_formatted_batch numpy 5000 10	shuffled read_formatted_batch numpy 5000 1000
new / old (diff)	0.294104 / 0.215209 (0.078895)	2.886809 / 2.077655 (0.809154)	1.601931 / 1.504120 (0.097811)	1.469353 / 1.541195 (-0.071842)	1.525132 / 1.468490 (0.056642)	0.565164 / 4.584777 (-4.019613)	2.432873 / 3.745712 (-1.312839)	2.885849 / 5.269862 (-2.384013)	1.780474 / 4.565676 (-2.785203)	0.064358 / 0.424275 (-0.359917)	0.005186 / 0.007607 (-0.002421)	0.349374 / 0.226044 (0.123329)	3.424751 / 2.268929 (1.155823)	1.956874 / 55.444624 (-53.487750)	1.679002 / 6.876477 (-5.197475)	1.718821 / 2.142072 (-0.423252)	0.656974 / 4.805227 (-4.148254)	0.120645 / 6.500664 (-6.380019)	0.042355 / 0.075469 (-0.033114)

Benchmark: benchmark_map_filter.json

metric	filter	map fast-tokenizer batched	map identity	map identity batched	map no-op batched	map no-op batched numpy	map no-op batched pandas	map no-op batched pytorch	map no-op batched tensorflow
new / old (diff)	1.001923 / 1.841788 (-0.839864)	13.208127 / 8.074308 (5.133819)	11.164863 / 10.191392 (0.973471)	0.131964 / 0.680424 (-0.548460)	0.015344 / 0.534201 (-0.518857)	0.287961 / 0.579283 (-0.291322)	0.273986 / 0.434364 (-0.160378)	0.327280 / 0.540337 (-0.213058)	0.426761 / 1.386936 (-0.960175)

[albertvillanova]

Thanks for the fix and the patch release. This confirms that, as I suggested in the Summer, maybe we should avoid making a release right before leaving on holidays.