HXL-CPLP / forum

Fórum do Grupo de Usuários do Padrão HXL da Comunidade dos Países de Língua Portuguesa, "HXL-CPLP"
https://github.com/HXL-CPLP/forum/issues
The Unlicense
2 stars 0 forks source link

[EN] Current state of tools with support for Frictionless Data (`datapackage.json`...) for HXL datasets #25

Open fititnt opened 3 years ago

fititnt commented 3 years ago

Quick links / see also


Note: this issue on @HXL-CPLP in special will be preferable written in English.

Like the [EN] Current state of tools with explicit support for the .hxl extension (in special over .csv) #24 issue, this one focus on tools that already support Frictionless Data (datapackage.json...), but our focus is with HXL datasets.

Some previous work was been made on @covid-taskforce-cplp. And now we're drafting the https://github.com/EticaAI/EticaAI-linguistic-datasets-pt (there mostly focused on actually work for create the datasets, not discuss the specification or implementations itself. So by creating this issue here, it make easier for cross reference later.

fititnt commented 3 years ago

Note: since at this moment the https://data.okfn.org/tools/view is offline, one way to allow testing of the https://github.com/EticaAI/EticaAI-linguistic-datasets-pt is just run locally the https://github.com/okfn/data.okfn.org-new.

So instead of https://data.okfn.org/tools/view?url=https%3A%2F%2Fgithub.com%2Fcovid-taskforce-cplp%2Fdados-v1, the http://localhost:5000/tools/view?url=https%3A%2F%2Fgithub.com%2Fcovid-taskforce-cplp%2Fdados-v1 can be used. (we still not have yet new datasets with HXL). Also the https://github.com/okfn/data.okfn.org-new needs a full URL. If using the / it will redirect for https://datahub.io/.

fititnt@bravo:/workspace/git/okfn/data.okfn.org-new$ npm install .
# (...)
fititnt@bravo:/workspace/git/okfn/data.okfn.org-new$ node app.js
404: Not Found Error: Invalid URI "/datapackage.json"
# (...)
Listening on 5000
# (...)
GET /tools/view?url=https%3A%2F%2Fgithub.com%2Fcovid-taskforce-cplp%2Fdados-v1 200 1787.039 ms - 51726

Captura de tela de 2020-12-03 16-53-25

fititnt commented 3 years ago

Also, this project may be pertinent https://github.com/mcarans/hdx-frictionless.

Update: after some feedback, we from HXL-CPLP may put 'on hold' the HXL on Frictionless Data. Not that this is not an nice feature, it is! But Frictionless Data requires that the datapackage.json must be manually edited and, at least if the dataset already is not strictly defined and will be shared with people that already use datapackage, datapackage.json may actually be more complex than just use plain HXL tools.

We from HXL-CPLP may get back to this topic on future. But at this point is not an priority in special because we're becoming more aware tools that HXL already have.