Add categories for tags

[alanfgh]

CK has suggested adding categories for tags:

http://a4esl.com/temporary/tatoeba/tags/

We also have "utility" tags (e.g., "@check", "@delete"). We use the convention that such tags start with "@". Other "prefix conventions": tags that list an author start with "by" and ones that list a source start with "from". While the ability to sort tags by name (mentioned in another issue ticket) will make it easier to group these tags, displaying only a single category at a time will be more difficult. For this reason, adding an enum to represent these categories would be a good idea. We undoubtedly want there to be fewer categories than 128, so a TINYINT would probably be sufficient.

We should make it possible to choose categories via the UI, but for a first pass, we could set some values via SQL statements (e.g., set all tags beginning with "by " to category "author"). It would be good to have an "unassigned" default category.

[ckjpn]

I re-uploaded the demo here.

http://tatoeba.ueuo.com/display_all_tags.html

I had to take a lot of the tatoeba support stuff offline, since it was slowing down my server.

[RyckRichards]

alexmarcelo and I are planning to reorganize the tags. Thanks so much for this data.

[alanfgh]

This issue should not be assigned to me anymore.

[ckjpn]

I would be willing to start categorizing tags.

I think I would need the following.

• a field added into the database called "category".
• a way for me to submit a category for each tag.
• for someone to set up a page similar to what I have at http://tatoeba.ueuo.com/display_all_tags.html.
  ** I think this page should be a table that is sortable.
• by default, all tags not yet categorized should be labeled "uncategorized", or perhaps "z-uncategorized" so that an alphabetical sort of the categories puts that tag at or near the bottom.

This is how the data in that page looks. (tag number, tag name, number of sentences, category)

"1218 less than 8 words 139537 length",
"5295 male name 58383 name",
"138 location 17821 function",
"5296 female name 12933 name",
"4070 present simple 11452 grammar",
"5305 names of 2 people 9614 pattern",

Javascript writes the table and a jQuery plugin is used for the sorting. (Show the source of the page to see how it's done.)

If the above data was written to a file named "tags.js" in the downloads folder, I think it could easily be imported into an HTML page like my demo page. That's not the ideal way, but while in the process of doing this, it would allow people to easily see the progress. The main problem with this is that the .js file would likely be cached on a visitor's computer, so the data might not be the most up-to-date. Perhaps, though, a setting on the server could change this, so the file wouldn't be cached very long.

Perhaps as a starter, TRANG or someone can create the field "category" and import all the categories I've already identified.

I can easily change that data to something like the following if it's not obvious how to do that.
Tag number + tab + category name (I would assume that internally categories would be identified by number.)

1218 length
5295 name
138 function
5296 name
4070 grammar
5305 pattern

[ckjpn]

I've noticed that this is tagged "unclear."

Here's a demo of what I mean.

I re-uploaded this page to the following URL on 2021-02-21, since the previous server is offline.

http://tatoeba.ueuo.com/stats-190323-tags.html

This would allow grouping "grammar" tags together, "function" tags together, etc.

[trang]

I think the "unclear" label is meant for things like:

• What categories do we want exactly?
• What would be the process for adding a new category? For deleting a category? For renaming a category?
• How do we decide which tag belongs to which category?

[ckjpn]

For starters, perhaps a few of the categories that I've suggested would work, and everything else could be put a "category not yet chosen" category, much like you haver "unknown" for languages that haven't yet been added to tatoeba.org.

[ckjpn]

I noticed that this is still tagged "unclear."
What other questions do you have?

[trang]

The questions I listed above still have no clear answer...

[ckjpn]

What categories do we want exactly?

I suggest we start with some of the ones I have on the demo page, and expand to include more as we need them.

What would be the process for adding a new category? For deleting a category? For renaming a category?

I think you could likely make a page accessible for admins to do this.
What would be needed is a way to search for the tag, and then have a field for that tag's category, hopefully with auto complete like tags are done now.

How do we decide which tag belongs to which category?

I've already suggested a few in the demo. Tags that can't yet easily be categorized could get a default tag, maybe "not in any category yet."

[jiru]

Note that the meaning of the "unclear" label is explained more in detail on the wiki.

In terms of UI, I had this idea from Hetzner’s console, they also have label feature that uses a key/value pattern: