`hxl2tab`: tab format, focused for compatibility with Orange Data Mining

[fititnt]

• Orange Data Miner specification about .tab
  • http://orange.biolab.si/docs/latest/reference/rst/Orange.data.formats.html
  • https://orange-data-mining-library.readthedocs.io/en/latest/reference/data.io.html
• File:
  • bin/hxl2tab
  • https://github.com/EticaAI/HXL-Data-Science-file-formats/blob/main/bin/hxl2tab

---

TODO: add more information.

[fititnt]

The EticaAI-Data_HXL-Data-Science-file-formats_Tab already have an draft of an table that could be used to make an Expert system without the need of full machine learning models.

But for this implementation, I think that we can simply implement both the more specific prefixes, like the +vt_orange_, and and maybe some special more generic attributes to be used with #3, like the one to mention the "class" (both Orange and Weka use class).

[fititnt]

Ok. Interesting. Here the Orange 'Simplified header' specification

---

While not ideal, the HXLated output without text headers actually are pretty similar to what orange would expect. The biggest difference is that everything after the # the orange consider as textual header, but before this is possible to add a few extra short variables.

[fititnt]

hxl2tab https://docs.google.com/spreadsheets/d/1Vqv6-EAdSHMSZvZtE426aXkDiwP8Mdrpft3tiGQ1RH0/edit#gid=0 temp/example-ebola-dataset-1_HXLated+tab.csv

fititnt@bravo:/workspace/git/EticaAI/HXL-Data-Science-file-formats$ head temp/example-ebola-dataset-1_HXLated+tab.csv
#status	#country	#adm1	#adm1+code	#loc	#loc	#org	#loc+type	#affected+dead	#affected+confirmed	#affected+suspected
Pending	Liberia	Margibi	LR09	Kakata 1	Kakata 2 AFL	AFL	ETC	0	0	0
Functional	Guinea	Nzerekore	GN008		Nzerekore	Ailema (?)	ETC	45	56	3
Pending	Liberia	River Gee	LR13	Fishtown	Fishtown ETC	American Red Cross	ETC	0	0	0
Functional	Sierra Leone	Western	SL04	Jui	Sierra Leone-China Friendship Hospital (Jui Hospital)	Chinese CDC	ETC	47	65	17
Pending	Guinea	Nzerekore	GN008			Croix-Rouge française	ETC	0	0	0
Pending	Sierra Leone	Western	SL04	Freetown	Goderich	EMERGENCY	ETC	0	0	0
Functional	Sierra Leone	Western	SL04	Lakka	Lakka Hospital ETU	EMERGENCY Italian NGO	ETC	3	17	11
Functional	Liberia	Margibi	LR09	Firestone	Firestone Medical Center	Firestone Company	ETC	14	29	19
Functional	Liberia	Montserrado	LR11	Monrovia	Monrovia, Congo Town - Old Ministry of Defence ETU 1	FMT	ETC	1	30	6

hxl2tab https://docs.google.com/spreadsheets/d/1Vqv6-EAdSHMSZvZtE426aXkDiwP8Mdrpft3tiGQ1RH0/edit#gid=0 temp/example-ebola-dataset-1_HXLated+tab_hxltabv15.tab

fititnt@bravo:/workspace/git/EticaAI/HXL-Data-Science-file-formats$ head temp/example-ebola-dataset-1_HXLated+tab_hxltabv15.tab
cD#status+vt_categorical+vt_class	D#country+vt_categorical	D#adm1+vt_categorical	D#adm1+code+vt_categorical	D#loc+vt_categorical	D#loc+vt_categorical	D#org+vt_categorical	#loc+type+vt_meta	C#affected+dead+number	C#affected+confirmed+number	C#affected+suspected+number
Pending	Liberia	Margibi	LR09	Kakata 1	Kakata 2 AFL	AFL	ETC	0	0	0
Functional	Guinea	Nzerekore	GN008		Nzerekore	Ailema (?)	ETC	45	56	3
Pending	Liberia	River Gee	LR13	Fishtown	Fishtown ETC	American Red Cross	ETC	0	0	0
Functional	Sierra Leone	Western	SL04	Jui	Sierra Leone-China Friendship Hospital (Jui Hospital)	Chinese CDC	ETC	47	65	17
Pending	Guinea	Nzerekore	GN008			Croix-Rouge française	ETC	0	0	0
Pending	Sierra Leone	Western	SL04	Freetown	Goderich	EMERGENCY	ETC	0	0	0
Functional	Sierra Leone	Western	SL04	Lakka	Lakka Hospital ETU	EMERGENCY Italian NGO	ETC	3	17	11
Functional	Liberia	Margibi	LR09	Firestone	Firestone Medical Center	Firestone Company	ETC	14	29	19
Functional	Liberia	Montserrado	LR11	Monrovia	Monrovia, Congo Town - Old Ministry of Defence ETU 1	FMT	ETC	1	30	6