joblisttoday / data Goto Github PK

Research use cases for using DID and git based data structure:

• https://github.com/perma-id/w3id.org
• https://developer.tbd.website/
• perma-id/w3id.org#2709

Create a Github action/workflow, that will check if a company already exists in this repository's data.

If the check passes, the PR is allowed to be merged

If the check fails, a company exists, and the PR needs further review

Right now we use tags:

• on company.tags, to describe companies, it is a list of "free text field"
• as tags collection, but not currently used anywhere except the /tags/:tag page if it exist, to add data to a tag

But the tags are overall not so great, as a user experience.

• their semantic is unclear
• it is not really grouping companies together
• hard to figure what tag to add / remove, or what is available

We could improve tags:

• enforce companies.tags to pick from the tags collection, (create a relation between the two models, in the cms)
• clearly define the source of the tags collection list. Why does a tag exist, what value does it bring to the user

The tags need to evolve to be able to well describe a company.

Maybe it should be additional categories on the company model.

Examples:

• https://www.linkedin.com/pulse/complete-list-linkedins-service-categories-brynne-tillman
• https://en.wikipedia.org/wiki/Category:Types_of_business_entity

All in all this seems like a complex project, that needs to bring clear and accessible semantics, to add value for the end user when exploring companies and looking for jobs.

(related cms issue https://gitlab.com/joblist/cms/-/issues/4)

Because markdown is so annoying to work with:

• need a library to convert frontmatter and content to JSON (or HTML)
• all libraries are wacky, good, but you never know wich one to use
• writting markdown frontmatter config is hard, even more when JSON object are stringified (positions.[].map)

Therefore JSON will be used:

• hard to writte by hand, but easier than markdown
• all data in this repo should be consise, company.description does not need rich text, or font styles
• can/should use the cms for edition

Converted

• data (fix map from CMS test, can't read "non json map string")
• workers
• cms

Title

Example Company Test Github Issue Template

Slug

example-company

Description

This should not allow markdown!

Tags

hello,world,foo,bar

Project URL

https://example.org

Job Board URL

https://example.org/careers

Job Board Provider

greenhouse

Job Board Hostname

greenhouse

Currently the data in this repo, for companies, has issues, for examples with dates.

That leads to having the sqlite3 database tables to be generated with for example dates (created_at and updated_at) in different formats (cleanup here, in post processing https://gitlab.com/joblist/workers/-/merge_requests/2).

We should fix the data in two places:

• in the cms, to be sure that the data is created by the CMS (and user), in the correct format https://gitlab.com/joblist/cms
• in the data repo, to cleanup existing data in the correct format (sent also by the CMS)

• https://developer.personio.de/docs/retrieving-open-job-positions

used to be <hostname>-jobs.personio.de, now is <hostname>.jobs.personio.de, and here all hostnames are saved under hostname: <hostname>-jobs for companies with personio providers.

• https://github.com/perma-id/w3id.org
• https://w3id.org/

Company profile

https://profiles.joblist.today/companies/delivery-hero

Description

a test

Explore:

• https://github.com/duarteocarmo/esco-visualizations/
• https://esco.ec.europa.eu/select-language?destination=/node/1
• https://duarteocarmo.com/blog/every-job-world
• https://en.wikipedia.org/wiki/BERT_(language_model)
• https://projector.tensorflow.org/?config=https://raw.githubusercontent.com/duarteocarmo/esco-visualizations/master/projector_config.json