Datasets:

DigiGreen
/

agri-lexicon-hindi

Error code:   DatasetGenerationCastError
Exception:    DatasetGenerationCastError
Message:      An error occurred while generating the dataset

All the data files must have the same columns, but at some point there are 18 new columns ({'skeleton', 'word_norm', 'tier', 'frequency', 'word', 'variants', 'relevance_mode', 'n_variants_kept', 'n_dropped', 'category_mode', 'mispron_rate', 'meaning_change_rate', 'n_tokens', 'n_mispronounced', 'n_surface_forms', 'n_meaning_changed', 'phon', 'surface_forms'}) and 18 missing columns ({'phon_sim', 'norm_b', 'ortho_sim', 'freq_a', 'phon_a', 'same_category', 'is_variant_family', 'cat_a', 'phon_b', 'word_a', 'evidence', 'word_b', 'freq_b', 'phon_dist', 'norm_a', 'observed_confusions', 'risk', 'cat_b'}).

This happened while the csv dataset builder was generating data using

hf://datasets/DigiGreen/agri-lexicon-hindi/lexicon_clean.csv (at revision 7b72db30717d2abe149b6aa35e831b8ae1e96b62), ['hf://datasets/DigiGreen/agri-lexicon-hindi@7b72db30717d2abe149b6aa35e831b8ae1e96b62/confusion_pairs.csv', 'hf://datasets/DigiGreen/agri-lexicon-hindi@7b72db30717d2abe149b6aa35e831b8ae1e96b62/lexicon_clean.csv', 'hf://datasets/DigiGreen/agri-lexicon-hindi@7b72db30717d2abe149b6aa35e831b8ae1e96b62/pair_labels.csv', 'hf://datasets/DigiGreen/agri-lexicon-hindi@7b72db30717d2abe149b6aa35e831b8ae1e96b62/variant_families.csv']

Please either edit the data files to have matching columns, or separate them into different configurations (see docs at https://hf.co/docs/hub/datasets-manual-configuration#multiple-configurations)
Traceback:    Traceback (most recent call last):
                File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1848, in _prepare_split_single
                  writer.write_table(table)
                  ~~~~~~~~~~~~~~~~~~^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/arrow_writer.py", line 765, in write_table
                  self._write_table(pa_table, writer_batch_size=writer_batch_size)
                  ~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/arrow_writer.py", line 773, in _write_table
                  pa_table = table_cast(pa_table, self._schema)
                File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2378, in table_cast
                  return cast_table_to_schema(table, schema)
                File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2306, in cast_table_to_schema
                  raise CastError(
                  ...<3 lines>...
                  )
              datasets.table.CastError: Couldn't cast
              word: string
              word_norm: string
              phon: string
              skeleton: string
              n_tokens: int64
              frequency: int64
              relevance_mode: int64
              category_mode: string
              n_mispronounced: int64
              n_meaning_changed: int64
              n_dropped: int64
              mispron_rate: double
              meaning_change_rate: double
              n_surface_forms: int64
              surface_forms: string
              tier: string
              n_variants_kept: int64
              variants: string
              -- schema metadata --
              pandas: '{"index_columns": [{"kind": "range", "name": null, "start": 0, "' + 2451
              to
              {'word_a': Value('string'), 'word_b': Value('string'), 'norm_a': Value('string'), 'norm_b': Value('string'), 'phon_a': Value('string'), 'phon_b': Value('string'), 'phon_sim': Value('float64'), 'phon_dist': Value('float64'), 'observed_confusions': Value('int64'), 'evidence': Value('string'), 'cat_a': Value('string'), 'cat_b': Value('string'), 'same_category': Value('bool'), 'freq_a': Value('int64'), 'freq_b': Value('int64'), 'ortho_sim': Value('float64'), 'is_variant_family': Value('bool'), 'risk': Value('float64')}
              because column names don't match
              
              During handling of the above exception, another exception occurred:
              
              Traceback (most recent call last):
                File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 1369, 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 948, 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 1694, 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 1850, in _prepare_split_single
                  raise DatasetGenerationCastError.from_cast_error(
                  ...<4 lines>...
                  )
              datasets.exceptions.DatasetGenerationCastError: An error occurred while generating the dataset
              
              All the data files must have the same columns, but at some point there are 18 new columns ({'skeleton', 'word_norm', 'tier', 'frequency', 'word', 'variants', 'relevance_mode', 'n_variants_kept', 'n_dropped', 'category_mode', 'mispron_rate', 'meaning_change_rate', 'n_tokens', 'n_mispronounced', 'n_surface_forms', 'n_meaning_changed', 'phon', 'surface_forms'}) and 18 missing columns ({'phon_sim', 'norm_b', 'ortho_sim', 'freq_a', 'phon_a', 'same_category', 'is_variant_family', 'cat_a', 'phon_b', 'word_a', 'evidence', 'word_b', 'freq_b', 'phon_dist', 'norm_a', 'observed_confusions', 'risk', 'cat_b'}).
              
              This happened while the csv dataset builder was generating data using
              
              hf://datasets/DigiGreen/agri-lexicon-hindi/lexicon_clean.csv (at revision 7b72db30717d2abe149b6aa35e831b8ae1e96b62), ['hf://datasets/DigiGreen/agri-lexicon-hindi@7b72db30717d2abe149b6aa35e831b8ae1e96b62/confusion_pairs.csv', 'hf://datasets/DigiGreen/agri-lexicon-hindi@7b72db30717d2abe149b6aa35e831b8ae1e96b62/lexicon_clean.csv', 'hf://datasets/DigiGreen/agri-lexicon-hindi@7b72db30717d2abe149b6aa35e831b8ae1e96b62/pair_labels.csv', 'hf://datasets/DigiGreen/agri-lexicon-hindi@7b72db30717d2abe149b6aa35e831b8ae1e96b62/variant_families.csv']
              
              Please either edit the data files to have matching columns, or separate them into different configurations (see docs at https://hf.co/docs/hub/datasets-manual-configuration#multiple-configurations)

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.

word_a string	word_b string	norm_a string	norm_b string	phon_a string	phon_b string	phon_sim float64	phon_dist float64	observed_confusions int64	evidence string	cat_a string	cat_b string	same_category bool	freq_a int64	freq_b int64	ortho_sim float64	is_variant_family bool	risk float64
बैंगन	बैगन	बैंगन	बैगन	BenKaN	BeKaN	0.917	0.5	1,090	both	crop	crop	true	420	2,152	0.8	false	213.84
गेंहू	गेहूं	गेंहू	गेहूं	KenHu	KeHu	0.9	0.5	673	both	crop	crop	true	1,173	14,861	0.6	false	202.16
दवा	दवाई	दवा	दवाई	DaBa	DaBai	0.9	0.5	489	both	activity_method	activity_method	true	13,832	3,757	0.75	false	184.79
मकई	मक्का	मकई	मक्का	MaKai	MaKa	0.9	0.5	399	both	crop	crop	true	1,572	3,615	0.4	false	146.81
दवा	दावा	दवा	दावा	DaBa	DaBa	1	0	335	both	activity_method	activity_method	true	13,832	138	0.75	true	139.59
पौधा	पौधे	पौधा	पौधे	PoDa	PoDe	0.9	0.4	289	both	growth_stage	growth_stage	true	3,915	1,351	0.75	false	125.4
सरसो	सरसों	सरसो	सरसों	SaRaSo	SaRaSo	1	0	247	both	crop	crop	true	314	3,039	0.8	true	115.1
मकई	मकाई	मकई	मकाई	MaKai	MaKai	1	0	356	both	crop	crop	true	1,572	11	0.75	true	97.09
फुल	फूल	फुल	फूल	PuL	PuL	1	0	151	both	growth_stage	growth_stage	true	155	2,289	0.667	true	84.93
डाले	डालें	डाले	डालें	DaLe	DaLe	1	0	130	both	activity_method	activity_method	true	220	1,338	0.8	true	78.11
पता	पत्ता	पता	पत्ता	PaTa	PaTa	1	0	144	both	concern	concern	true	83	1,807	0.6	true	77.5
खाद	खाद्य	खाद	खाद्य	KaD	KaDY	0.75	1	194	both	fertilizer_chemical	fertilizer_chemical	true	4,772	206	0.6	false	77.26
कितना दिन	कितने दिन	कितना दिन	कितने दिन	KiTaNaDiN	KiTaNeDiN	0.956	0.4	123	both	quantity_unit	quantity_unit	true	465	931	0.889	true	74.99
बीच	बीज	बीच	बीज	BiC	BiJ	0.867	0.4	153	both	other	other	true	94	3,453	0.667	false	73.54
लगाए	लगाएं	लगाए	लगाएं	LaKae	LaKae	1	0	116	both	activity_method	activity_method	true	158	849	0.8	true	69.52
पौधा	पौधों	पौधा	पौधों	PoDa	PoDo	0.9	0.4	104	both	growth_stage	growth_stage	true	3,915	242	0.6	false	69.36
फूल गोभी	फूलगोभी	फूल गोभी	फूलगोभी	PuLaKoBi	PuLaKoBi	1	0	151	both	crop	crop	true	86	253	0.875	true	66.45
किट	कीट	किट	कीट	KiT	KiT	1	0	123	both	concern	concern	true	66	873	0.667	true	66.32
मंजर	मांजर	मंजर	मांजर	ManJaR	ManJaR	1	0	86	both	growth_stage	growth_stage	true	1,896	153	0.8	true	64.63
वर्मी कंपोस्ट	वर्मीकंपोस्ट	वर्मी कंपोस्ट	वर्मीकंपोस्ट	BaRMiKanPoST	BaRMiKanPoST	1	0	133	both	fertilizer_chemical	fertilizer_chemical	true	182	156	0.923	true	64.33
मुंग	मूंग	मुंग	मूंग	MunK	MunK	1	0	85	both	crop	crop	true	94	2,873	0.75	true	63.92
मशरुम	मशरूम	मशरुम	मशरूम	MaSaRuM	MaSaRuM	1	0	96	both	crop	crop	true	121	940	0.8	true	62.89
कोभी	गोभी	कोभी	गोभी	KoBi	KoBi	1	0	96	both	crop	crop	true	160	632	0.75	true	62.25
अमरुद	अमरूद	अमरुद	अमरूद	aMaRuD	aMaRuD	1	0	103	both	crop	crop	true	126	554	0.8	true	62.22
झड़	झर	झड	झर	JaD	JaR	0.667	1	234	empirical	concern	concern	true	330	115	0.5	false	57.33
कीड़ा	कीड़े	कीडा	कीडे	KiDa	KiDe	0.9	0.4	59	both	concern	concern	true	5,246	425	0.75	false	57.11
सुख	सूख	सुख	सूख	SuK	SuK	1	0	94	both	concern	concern	true	50	520	0.667	true	54.43
पत्ता	पत्ते	पत्ता	पत्ते	PaTa	PaTe	0.9	0.4	58	both	concern	concern	true	1,807	643	0.8	false	54.15
महीना	महीने	महीना	महीने	MaHiNa	MaHiNe	0.933	0.4	70	both	quantity_unit	quantity_unit	true	352	447	0.8	true	52.33
रोकने	रोपने	रोकने	रोपने	RoKaNe	RoPaNe	0.833	1	88	both	activity_method	activity_method	true	294	555	0.8	false	51.94
बुवाई	बोआई	बुवाई	बोआई	BuBai	Boai	0.72	1.4	108	both	activity_method	activity_method	true	1,836	128	0.4	false	50.74
सुख रहा	सूख रहा	सुख रहा	सूख रहा	SuKaRaHa	SuKaRaHa	1	0	75	both	concern	concern	true	59	450	0.857	true	49.26
सिंचाई	सिचाई	सिंचाई	सिचाई	SinCai	SiCai	0.917	0.5	65	both	activity_method	activity_method	true	1,751	75	0.833	false	48.97
पल्ला	पाला	पल्ला	पाला	PaLa	PaLa	1	0	47	both	concern	concern	true	95	1,863	0.6	true	47.48
दे	दें	दे	दें	De	De	1	0	67	both	activity_method	activity_method	true	194	121	0.667	true	46.28
पेड़	पैर	पेड	पैर	PeD	PeR	0.667	1	118	empirical	other	other	true	3,577	32	0.333	false	46.08
हेक्टर	हेक्टेयर	हेक्टर	हेक्टेयर	HeKTaR	HeKTeYaR	0.812	1.5	112	both	quantity_unit	quantity_unit	true	94	182	0.75	false	45.94
मकई	मक्के	मकई	मक्के	MaKai	MaKe	0.82	0.9	58	both	crop	crop	true	1,572	280	0.4	false	45.91
उड़द	उरद	उडद	उरद	uDaD	uRaD	0.75	1	142	both	crop	crop	true	245	40	0.667	false	44.61
चना	चने	चना	चने	CaNa	CaNe	0.9	0.4	50	both	crop	crop	true	1,816	113	0.667	false	44.44
फर	फल	फर	फल	PaR	PaL	0.867	0.4	49	both	concern	concern	true	103	2,809	0.5	false	43.61
रोपे	रोपें	रोपे	रोपें	RoPe	RoPe	1	0	61	both	activity_method	activity_method	true	184	103	0.8	true	43.45
उपयोग	प्रयोग	उपयोग	प्रयोग	uPaYoK	PRaYoK	0.75	1.5	46	both	activity_method	activity_method	true	1,468	1,914	0.5	false	43.34
पालक	पाला	पालक	पाला	PaLaK	PaLa	0.8	1	50	both	crop	concern	false	345	1,863	0.75	false	43.18
मिर्चा	मिर्ची	मिर्चा	मिर्ची	MiRCa	MiRCi	0.88	0.6	50	both	crop	crop	true	562	310	0.833	false	42.87
दबा	दवा	दबा	दवा	DaBa	DaBa	1	0	38	both	activity_method	activity_method	true	10	13,832	0.667	true	42.42
रबी	रवि	रबी	रवि	RaBi	RaBi	1	0	75	both	growth_stage	crop	false	71	85	0.333	false	42.17
सिम	सेम	सिम	सेम	SiM	SeM	0.867	0.4	73	both	crop	crop	true	261	95	0.667	false	41.9
किड़ा	कीड़ा	किडा	कीडा	KiDa	KiDa	1	0	34	both	concern	concern	true	40	5,246	0.75	true	41.87
बचाओ	बचाव	बचाओ	बचाव	BaCao	BaCaB	0.76	1.2	102	both	activity_method	activity_method	true	32	600	0.75	false	41.66
छींटे	छींटें	छींटे	छींटें	CinTe	CinTe	1	0	42	both	activity_method	activity_method	true	295	213	0.833	true	41.35
मसूर	मसूरी	मसूर	मसूरी	MaSuR	MaSuRi	0.917	0.5	31	both	crop	crop	true	1,509	596	0.8	false	41.27
आलु	आलू	आलु	आलू	aLu	aLu	1	0	31	both	crop	crop	true	36	7,081	0.667	true	40.89
सेब	सेव	सेब	सेव	SeB	SeB	1	0	58	both	crop	crop	true	121	76	0.667	true	39.41
मरने	मारने	मरने	मारने	MaRaNe	MaRaNe	1	0	55	both	concern	activity_method	false	30	374	0.8	true	39.32
कितनी बार	कितने बार	कितनी बार	कितने बार	KiTaNiBaR	KiTaNeBaR	0.956	0.4	42	both	quantity_unit	quantity_unit	true	220	261	0.889	true	39.2
कीड़ा	खीरा	कीडा	खीरा	KiDa	KiRa	0.75	1	37	both	concern	crop	false	5,246	479	0.5	false	39.15
खर पतवार	खरपतवार	खर पतवार	खरपतवार	KaRaPaTaBaR	KaRaPaTaBaR	1	0	64	both	concern	concern	true	12	482	0.875	true	39.1
कितना	कितने	कितना	कितने	KiTaNa	KiTaNe	0.933	0.4	26	both	quantity_unit	quantity_unit	true	1,668	335	0.8	true	37.67
मधुआ	मधुवा	मधुआ	मधुवा	MaDua	MaDuBa	0.833	1	105	both	concern	concern	true	117	24	0.6	false	37.38
लौका	लौकी	लौका	लौकी	LoKa	LoKi	0.85	0.6	55	both	crop	crop	true	95	341	0.75	false	37.19
लगा	लगाएं	लगा	लगाएं	LaKa	LaKae	0.9	0.5	46	both	activity_method	activity_method	true	46	849	0.6	false	37.06
पत्ती	पत्ते	पत्ती	पत्ते	PaTi	PaTe	0.9	0.4	35	both	concern	concern	true	197	643	0.8	false	36.58
गेम	गेहूं	गेम	गेहूं	KeM	KeHu	0.625	1.5	81	empirical	crop	crop	true	8	14,861	0.4	false	36.54
क्या करे	क्या करें	क्या करे	क्या करें	KYaKaRe	KYaKaRe	1	0	25	both	activity_method	activity_method	true	99	1,917	0.889	true	36.47
सिम	सीम	सिम	सीम	SiM	SiM	1	0	43	both	crop	crop	true	261	57	0.667	true	36.37
फल	फूल	फल	फूल	PaL	PuL	0.8	0.6	23	both	concern	growth_stage	false	2,809	2,289	0.667	false	36.35
दवा	दाना	दवा	दाना	DaBa	DaNa	0.75	1	27	both	activity_method	concern	false	13,832	351	0.5	false	35.77
कोबी	गोभी	कोबी	गोभी	KoBi	KoBi	1	0	58	both	crop	crop	true	6	632	0.5	false	35.64
मुंह	मूंग	मुंह	मूंग	Mu	MunK	0.625	1.5	74	empirical	concern	crop	false	32	2,873	0.5	false	34.31
मटर	मोटर	मटर	मोटर	MaTaR	MoTaR	0.92	0.4	40	both	crop	tool	false	1,029	23	0.75	true	33.98
तोड़ी	तोरी	तोडी	तोरी	ToDi	ToRi	0.75	1	60	both	crop	crop	true	83	363	0.75	false	33.86
टमाटर	मटर	टमाटर	मटर	TaMaTaR	MaTaR	0.786	1.5	24	both	crop	crop	true	2,004	1,029	0.6	false	33.7
जो	जौ	जो	जौ	Jo	Jo	1	0	62	both	crop	crop	true	14	131	0.5	false	33.55
अरहर	रहर	अरहर	रहर	aRaHaR	RaHaR	0.917	0.5	34	both	crop	crop	true	470	92	0.75	false	33.45
मक्का	मक्के	मक्का	मक्के	MaKa	MaKe	0.9	0.4	19	both	crop	crop	true	3,615	280	0.8	false	33.35
लगा है	लगाएं	लगा है	लगाएं	LaKaHe	LaKae	0.833	1	58	both	activity_method	activity_method	true	12	849	0.5	false	33.2
कितने दिन	कितने दिनों	कितने दिन	कितने दिनों	KiTaNeDiN	KiTaNeDiNo	0.95	0.5	22	both	quantity_unit	quantity_unit	true	931	227	0.818	true	33.15
बुआई	बुवाई	बुआई	बुवाई	Buai	BuBai	0.8	1	43	both	activity_method	activity_method	true	30	1,836	0.6	false	33.03
पिला	पीला	पिला	पीला	PiLa	PiLa	1	0	23	both	concern	concern	true	31	2,697	0.75	true	32.86
गेहू	गेहूं	गेहू	गेहूं	KeHu	KeHu	1	0	16	both	crop	crop	true	26	14,861	0.8	true	32.17
धन	धान	धन	धान	DaN	DaN	1	0	28	both	crop	crop	true	7	3,103	0.667	true	31.46
तिल	तेल	तिल	तेल	TiL	TeL	0.867	0.4	32	both	crop	fertilizer_chemical	false	288	168	0.667	false	31.14
खरपतवार नाशक	खरपतवार नाशी	खरपतवार नाशक	खरपतवार नाशी	KaRaPaTaBaRaNaSaK	KaRaPaTaBaRaNaSi	0.906	1.6	29	both	fertilizer_chemical	fertilizer_chemical	true	184	234	0.917	false	30.89
पहला	पाला	पहला	पाला	PaHaLa	PaLa	0.75	1.5	28	both	growth_stage	concern	false	255	1,863	0.75	false	30.85
पौधे	पौधों	पौधे	पौधों	PoDe	PoDo	0.85	0.6	22	both	growth_stage	growth_stage	true	1,351	242	0.6	false	30.72
कितनी	कितने	कितनी	कितने	KiTaNi	KiTaNe	0.933	0.4	24	both	quantity_unit	quantity_unit	true	202	335	0.8	true	30.64
कदुआ	कद्दू	कदुआ	कद्दू	KaDua	KaDu	0.9	0.5	29	both	crop	crop	true	51	827	0.4	false	30.63
बाव	भाव	बाव	भाव	BaB	BaB	1	0	33	both	activity_method	quantity_unit	false	56	151	0.667	true	30.57
मिर्च	मिर्चा	मिर्च	मिर्चा	MiRC	MiRCa	0.9	0.5	23	both	crop	crop	true	214	562	0.833	false	30.52
मछली	मसूरी	मछली	मसूरी	MaCaLi	MaSuRi	0.767	1.4	30	both	livestock	crop	false	359	596	0.4	false	30.49
करेला	करैला	करेला	करैला	KaReLa	KaReLa	1	0	27	both	crop	crop	true	675	26	0.8	true	30.32
चना	चाना	चना	चाना	CaNa	CaNa	1	0	22	both	crop	crop	true	1,816	20	0.75	true	29.9
मशरूम	मसूर	मशरूम	मसूर	MaSaRuM	MaSuR	0.7	2.1	25	both	crop	crop	true	940	1,509	0.4	false	29.75
आज	प्याज	आज	प्याज	aJ	PYaJ	0.5	2	57	empirical	other	crop	false	290	3,067	0.2	false	29.28
मंजर	मोजर	मंजर	मोजर	ManJaR	MoJaR	0.85	0.9	15	both	growth_stage	growth_stage	true	1,896	633	0.75	false	28.99
बिज	बीज	बिज	बीज	BiJ	BiJ	1	0	17	both	quantity_unit	other	false	22	3,453	0.667	true	28.81
बढ़ाए	बढ़ाएं	बढाए	बढाएं	BaDae	BaDae	1	0	32	both	activity_method	activity_method	true	37	145	0.8	true	28.67
छिड़काव	छिलका	छिडकाव	छिलका	CiDaKaB	CiLaKa	0.714	2	47	both	activity_method	other	false	3,023	8	0.667	false	28.35
श्री विधि	स्री विधि	श्री विधि	स्री विधि	SRiBiDi	SRiBiDi	1	0	30	both	activity_method	activity_method	true	188	32	0.889	true	28.27

End of preview.

Agricultural Domain Lexicon (Hindi) with Adjudicated Sound-Alike Confusion Pairs

A controlled agricultural vocabulary for Hindi, built for domain-aware ASR post-correction and for weighting agricultural terms in evaluation (AWWER). It is the lexical resource behind Digital Green's agricultural voice pipeline. Contains no farmer audio or transcripts — it is a mined, cleaned, and adjudicated word list.

Files

File	Rows	What it is
`lexicon_clean.csv`	18,646	The cleaned lexicon: term, normalized form, phoneme string, agronomic category, and a 4/3/2 criticality weight per term.
`pair_labels.csv`	13,904	Sound-alike term pairs adjudicated as SAME_WORD / SAME_MEANING / DIFF_MEANING / UNRELATED, with a disambiguating cue and phonetic similarity.
`confusion_pairs.csv`	23,857	The wider candidate confusion set (phonetically-similar neighbours) before adjudication, with observed-confusion counts and categories.
`variant_families.csv`	950	398 variant families grouped to a canonical head, in long format — one row per member word (950 member rows across the 398 families).
`agri_lexicon_master.xlsx`	—	All of the above as one workbook with an overview sheet.

How it was built (seven steps)

Mine agricultural terms from ~100,000 Hindi farmer queries → clean and normalize → tier and weight by agronomic criticality → convert to a script-unified phoneme representation → cluster phonetically-similar neighbours → adjudicate every confusable pair into same-meaning vs meaning-changing → group same-meaning variants into families with an over-merge guard. The process is repeatable per language.

Why the adjudication matters

Sound similarity alone does not tell you which confusions are dangerous. A sound-alike pair with a different meaning (e.g. दवा davā, medicine → दावा dāvā, claim) changes the farmer's question; a spelling variant of the same word does not. The pair_labels adjudication is what lets a corrector fold safe variants while leaving meaning-changing confusions for a contextual reader.

Intended use

Domain-aware ASR post-correction (recover garbled agricultural terms; defer meaning-changing homophones).
Domain-weighted evaluation of ASR (weight crop/pest/chemical/quantity terms more heavily).
A reusable resource for Hindi agricultural NLP.

The accompanying paper, Model-Agnostic and Language-Agnostic Voice Pipeline Improvement for the Agriculture Domain, is included in this repository (arXiv:2609.20504).

License

CC-BY-4.0. Please cite Digital Green's agricultural voice-pipeline paper.

Downloads last month: 33

Total file size:

13.2 MB

Paper for DigiGreen/agri-lexicon-hindi

Model-Agnostic and Language-Agnostic Voice Pipeline Improvement for the Agriculture Domain

Paper • 2609.20504 • Published 3 days ago