Convert tags to smaller set of blog post categories
In the process of migrating the Drupal site to Lektor, @duncan and I agreed it might make sense to replace the blog post tags with a predefined set of categories. This would simplify the Lektor site structure since we'd be able to forgo the use of the lektor-tags module and allow us to provide RSS/Atom feeds of posts in each category.
The list of tags currently used on the blog is available here: https://pad.riseup.net/p/QqaP7feFvQlPJrDHmSyX
Designs
- #40004Launch support's Forum and Blog migration
Activity
added Backlog label
assigned to @lavamind
changed milestone to %Launch support's Forum and Blog migration
assigned to @duncan and unassigned @lavamind
@emmapeel, you mentioned that you'd like to keep tags on the new blog. Could you explain why you would like to keep them?
mentioned in issue #40004 (closed)
@lavamind: I think there is a misunderstanding: if I remember correctly, I was more in favor to have more tags in the new forum, as opposed to some of the categories we were using. My main reason was that categories are mutually exclusive, and some forum topics may pertain to more than one category.
One thing I said about blog tags was that maybe we should do some tag gardening once in a while, merging similar tags, or having a set and not allowing everybody to add tags.
Tags are good for SEO and for allowing a transversal way of reading the posts. But of course that should be also provided with the interface (like showing similarly tagged posts).
-
Alright, thanks for clarifying. I think we're in agreement then! Of course the basic idea is to work these categories out of the existing tags, so in that sense it would be tidying up our "garden" of tags which has more or less been run over by weeds :)
Those categories would work the same way as tags in that they'd be shown under each post. But since we'd have a more manageable number (around a dozen maybe?), we could show the list of categories directly on the sidebar, which would link to the list of posts in each category, adding to the discoverability of the diversity of topics represented on our blog (eg. not just release announcements).
Edited by Jérôme Charaoui
In 2020, an Outreachy intern Annalee cleaned and organized the tags.
Here's her final report and tag status:
assigned to @gaba
removed Next label
added Doing label
-
A few ideas:
- There are still many many tags left, nearly 200
- Consistency of capitalization could be improved, eg. which do we capitalize of continents, countries, organizations, or none?
- Do we keep tags for countries/continents? Maybe we could group them all (eg. using a category named `regions?)
- Maybe also consider grouping more things like operating systems, instead of having
windows,os x,debianfor example
I made
Sheet3on the spreadsheet to combine/sort the new tag set. 
I also did a small review using your new tab
Sheet3, @lavamind.What I did was:
- removing language (just one listed)
- comparing any tag removed with the pad to see how many times that tag was mentioned (tags should be words that can gather many posts together)
- removed duplicated tags, like (UK,uk; release,releases, etc)
- research centers under "academia"
I think we should also remove organizations names, but i was not sure, so i left them there.
Edited by Nah
Yes. While I was going through the tags I thought that maybe we should remove countries/continents. Although sometimes we want to track situation on a specific places. Maybe all in one bag as censored may makes sense.
About operating systems, yes. We can have windows, mac, linux... instead.
@nah which language you removed? Maybe it would be a good idea to have a tag for a language but we can add a tag later if we need posts in Spanish for example.
I reviewed the
sheet3tab,review 05/octobercolumn and made these suggestions:- anonymous donor -> remove & merge with fundraising
- bitcoin -> either remove this entirely or replace with cryptocurrency. i'd remove it.
- crowdfunding -> remove & merge with fundraising
- donations -> remove & merge with fundraising
(not sure how I didn't get this notification and request for review, sorry for the delay).
Edited by al smith
marked this issue as related to #40004 (closed)
-
@lavamind do you need anything else in this ticket?
added Next label
removed Doing label
marked this issue as related to #40009
-
At our meeting on October 20 we settled on the following list of categories for the new blog, grouped by theme:
areas of work topics operations circumvention advocacy jobs network releases fundraising applications relays announcements community human rights devops usability research reports metrics onion services localization global south partners We can still adjust this list, are there any changes we'd like to make?
I may be a little biased here but we at TPA sometimes do cool sysadmin stuff that's interesting enough to share with the community, so maybe we could add a
sysadmincategory? /cc @anarcat -
Today I wrote a Python script that will apply the mapping to all the imported blog posts. It works by reading from a CSV file
We're still missing a few mappings for some of the most-used tags at the top of the list, for example those related to Tails or GSoC. If anyone would like to propose changes to the mapping, you can fork this repository, edit the CSV file right in GitLab (really convenient!), commit and create a Merge Request. Or simply commit directly in this repository, if you have access.
When running the script against the latest 150 or so blog post which we've imported so far, these issues are raised:
WARN: <announcing-arti> has zero categories WARN: <new-releases-with-security-fixes-tor-03514-0448-and-0457> has zero categories WARN: <entering-the-matrix> has zero tags WARN: <privacy-lgbtqia-human-rights> has zero categories WARN: <dreaming-at-dusk> has zero tags WARN: <transparency-openness-and-our-2018-and-2019-financials> has zero categories WARN: <from-trac-into-gitlab-for-tor> has zero categories WARN: <new-stable-releases-tor-03516-04510-and-0467> has zero categories WARN: <welcome-new-tor-board-members> has zero categories WARN: <tor-media-2020> has zero categories WARN: <transparency-openness-and-our-2019-and-2020-financials> has zero categories WARN: <new-release-tor-browser-1055> has zero tags WARN: <tors-bug-smash-fund-2021> has zero tags WARN: <new-release-candidate-tor-0455-rc> has zero tags WARN: <in-memoriam-of-karsten-loesing> has zero tags WARN: <sign-to-reclaim-your-face> has zero tags WARN: <anonymous-gitlab> has zero categories WARN: <improving-ux-connecting-to-tor-105> has zero tags WARN: <announcing-board-alissa-desigan-kendra> has zero categories WARN: <first-global-encryption-day> has zero tags WARN: <snowflake-in-tor-browser-stable> has zero tags WARN: <docshackathon-2021> has zero categories WARN: <check-status-of-tor-services> has zero categories-
zero tagsmeans the Drupal post has no tags at all, so the mapping is useless here -
zero categoriesmeans the mapping produced no categories from the tags of this post
-
-
Thanks for this list! I understand that some posts do not have tags now, but would it be possible to have on the error entry the current tags the posts with tags have?
This way it will be easier to map them to our restricted set
Edited by emmapeel -
Here (again, only for a subset of the blog posts).
WARN: <announcing-arti> tags mapping did not yield any categories (got tags [arti]) WARN: <new-releases-with-security-fixes-tor-03514-0448-and-0457> tags mapping did not yield any categories (got tags [security]) WARN: <entering-the-matrix> has no tags and no categories WARN: <dreaming-at-dusk> has no tags and no categories WARN: <transparency-openness-and-our-2018-and-2019-financials> tags mapping did not yield any categories (got tags [Form 990,Form PC,financial statements]) WARN: <from-trac-into-gitlab-for-tor> tags mapping did not yield any categories (got tags [gitlab,development roadmap]) WARN: <new-stable-releases-tor-03516-04510-and-0467> tags mapping did not yield any categories (got tags [stable, release, tor, security]) WARN: <welcome-new-tor-board-members> tags mapping did not yield any categories (got tags [board of directors,press release]) WARN: <tor-media-2020> tags mapping did not yield any categories (got tags [media coverage]) WARN: <transparency-openness-and-our-2019-and-2020-financials> tags mapping did not yield any categories (got tags [financial statements,Form 990]) WARN: <new-release-tor-browser-1055> has no tags and no categories WARN: <tors-bug-smash-fund-2021> has no tags and no categories WARN: <new-release-candidate-tor-0455-rc> has no tags and no categories WARN: <in-memoriam-of-karsten-loesing> has no tags and no categories WARN: <sign-to-reclaim-your-face> has no tags and no categories WARN: <anonymous-gitlab> tags mapping did not yield any categories (got tags [privacy,gitlab,anonymity,Outreachy]) WARN: <improving-ux-connecting-to-tor-105> has no tags and no categories WARN: <announcing-board-alissa-desigan-kendra> tags mapping did not yield any categories (got tags [board of directors]) WARN: <first-global-encryption-day> has no tags and no categories WARN: <snowflake-in-tor-browser-stable> has no tags and no categories WARN: <docshackathon-2021> tags mapping did not yield any categories (got tags [documentation]) WARN: <check-status-of-tor-services> tags mapping did not yield any categories (got tags [news,sysadmin])
removed Next label
added Doing label
I may be a little biased here but we at TPA sometimes do cool sysadmin stuff that's interesting enough to share with the community, so maybe we could add a
sysadmincategory? /cc @anarcatagreed, i posted at least one blog post so far. i would post more if it was done through MRs in git too. :p
-
Done here. The mapping can still be improved but we'll track this in #40009
removed Doing label
mentioned in issue tpo/applications/tor-browser#40705 (closed)
mentioned in merge request !1 (closed)