Catagolue Discussion Thread

[Ian07]

It looks like the new verification system is broken; I can't verify any hauls and none of the bold censuses have been updated since 02:13 UTC: https://catagolue.appspot.com/haul/b3s23/C1?committed=2

[calcyman]

It looks like the new verification system is broken; I can't verify any hauls and none of the bold censuses have been updated since 02:13 UTC: https://catagolue.appspot.com/haul/b3s23/C1?committed=2

Someone decided to search Day&Night... :/

[Ian07]

Someone decided to search Day&Night... :/

That's definitely not helping matters, but I don't think it's the main problem because all of the smaller censuses are working fine, meanwhile B3/S23/C1 and others which have peer review enabled are still frozen in time and gaining a growing backlog of uncommitted hauls.

[calcyman]

Someone decided to search Day&Night... :/

That's definitely not helping matters, but I don't think it's the main problem because all of the smaller censuses are working fine, meanwhile B3/S23/C1 and others which have peer review enabled are still frozen in time and gaining a growing backlog of uncommitted hauls.

It is the main problem: there have been consistent OutOfMemoryErrors since that time.

I've made some code modifications so that:

(a) these errors are less frequent;
(b) if they do occur, they'll cause the hauls to be dropped instead of freezing Catagolue indefinitely.

[Apple Bottom]

Someone decided to search Day&Night... :/

It may actually be worth including a warning message on the main Day & Night census page, urging people to not search that rule unless they know what they're doing so as to avoid overloading Catagolue.

[Ian07]

I just uploaded some hauls for b3aijnq7cs23-ckq4w/C1 which each contain several 20+ bit still lifes, but none of them appear on the main census page. The text census still works, and so do the subcensuses, but the only prefixes that appear on the census page are the ones that appeared in the initial haul. I remember having this same problem a couple of days ago while browsing a different census.

[calcyman]

I just uploaded some hauls for b3aijnq7cs23-ckq4w/C1 which each contain several 20+ bit still lifes, but none of them appear on the main census page. The text census still works, and so do the subcensuses, but the only prefixes that appear on the census page are the ones that appeared in the initial haul. I remember having this same problem a couple of days ago while browsing a different census.

The clue is the 'Last cache update: 2720 seconds ago' at the foot of the page.

I've now pushed a commit to reduce the refresh threshold from 3600 seconds to 600.

[muzik]

b3s23/C1 has now surpassed 300 trillion objects

[muzik]

Can Catagolue be updated to use the newest version of lifeviewer (either build 264 or build 265?)

https://catagolue.appspot.com/object/xp ... 5555555555

[calcyman]

Can Catagolue be updated to use the newest version of lifeviewer (either build 264 or build 265?)

https://catagolue.appspot.com/object/xp ... 5555555555

Yes, do you have a link to the lv-plugin.js for me to download?

[muzik]

I think it's in here

EDIT: btw, this links to the file now.

EDIT 2023: btw, this ACTUALLY links to the file now.

[calcyman]

Catagolue now has a /rules endpoint, yielding a more hierarchical organisation of information. Essentially, navigation through Catagolue forms a tree:

• Root: /rules
• Rule family: /rules/lifelike
• Individual rule: /census/b38s23
• Individual census: /census/b38s23/C1
• Individual tabulation: /census/b38s23/C1/xq16

Beyond the tabulations, there are another two levels -- object pages and sample soups -- but at this point it isn't strictly a tree, because different symmetries of the same rule link to the same object pages.

[muzik]

Can the rules be made to sort like a table (like the List of rules investigated on Catagolue wiki pages)? I can draw up a table concept on the wiki, complete with links to symmetries if you'd like that.

...and where are all the hexagonal rules?

[calcyman]

Can the rules be made to sort like a table (like the List of rules investigated on Catagolue wiki pages)? I can draw up a table concept on the wiki, complete with links to symmetries if you'd like that.

...and where are all the hexagonal rules?

Those are both good suggestions. I've replaced the LinkedHashMaps with TreeMaps to enforce lexicographical order:

https://catagolue.appspot.com/rules/lifelike

(I take it that's what you meant by 'sort like a table'?)

Also, the hexagonal rules are now separated, so we have things like this:

https://catagolue.appspot.com/rules/isogeny-hex-ot

[muzik]

I think it'd also be handy if for range-1 Moore and hexagonal outer-totalistic rules, the rule integer was also displayed next to the rule, and there existed an option to sort by this instead. Would probably make it easier to find rules.

Also, something weird: b026s1 is marked as only having one census when it clearly has two. - EDIT, fixed


Where exactly are rule names contributed?

[calcyman]

I think it'd also be handy if for range-1 Moore and hexagonal outer-totalistic rules, the rule integer was also displayed next to the rule, and there existed an option to sort by this instead.

Catagolue has no concept of rule integer.

Also, something weird: b026s1 is marked as only having one census when it clearly has two.

Thanks -- I've now amended this bug (and, by gosh, the number of censuses and rules have increased!): https://catagolue.appspot.com/rules

Where exactly are rule names contributed?

Common names for rules: https://gitlab.com/apgoucher/catagolue/ ... -rules.txt
Common names for rule families: https://gitlab.com/apgoucher/catagolue/ ... milies.txt
Rule family definitions: https://gitlab.com/apgoucher/lifelib/bl ... nuslist.py

[muzik]

The current list presentation still feels a bit wonky, which I personally don't like very much, especially since rules with names make it even more lumpier. I'd rather it was presented more like a table, so that everything still stays relatively straight, and ends up a bit easier to read. (Bonus points if it's sortable!)

Here's a really bad mockup of a table format that could be used, with information proudly and shamelessly ripped from the List of rules investigated on Catagolue page. (Note that the rule integer section would only be present on moore outer-totalistic rules and hexagonal outer-totalistic rules, assuming the rule integer will even be calculated at all for said rulespaces, which it probably won't be, and as such could probably be discounted.)

[calcyman]

(Bonus points if it's sortable!)

Yay! https://catagolue.appspot.com/rules/lifelike

(Click on the headers to sort and/or reverse direction of sort.)

I've also updated lv-plugin.js to match the one on these forums, which should hopefully be the build you needed.

[muzik]

The lack of lines feels a bit weird, but that aside, this is MUCH better than what it was. Thank you!

(from now on I'll try and keep discussion of this in the main suggestions thread)

(EDIT: which i just remembered does not exist)

(EDIT2: now it does, check it out!)

[muzik]

How often, exactly, do rule list pages update? Some rules I searched a few hours ago aren't showing up here: https://catagolue.appspot.com/rules/isotropic-hex-ot

EDIT: they are now

[calcyman]

How often, exactly, do rule list pages update?

The /rules pages are guaranteed to never be more than 6 hours old.

[muzik]

There are still some xs12s that use old names on Catagolue which should be updated, since their corresponding wiki pages have been moved to more appropriate titles (e.g. Boat-claw with tail --> Claw with boat with tail, Trans-aircraft carrier on table --> Trans-aircraft carrier down on table, Boat line tub --> Cis-boat line tub)

Also, quick reminder that I've put up a merge request that contains some more rule names, so don't forget to look over that.

[testitemqlstudop]

I can't access /chisquare

[77topaz]

To the anonymous user searching various rules with apgsearch "v0.54+0.32i-ab2-WHKD", could you please increase your haul size? You're submitting (10000-soup) hauls way too frequently, which is not good for the server. There's even a resultant backlog of hauls that aren't being committed.

EDIT: This made me think of something: in the last few months, there's been a lot of spam-searching, troll censuses, even Catagolue outages mostly caused by anonymous users. Maybe it would be a good idea to simply disable submitting hauls to Catagolue anonymously (i.e. without a payosha256 key)? If we did that, there'd be at least a little more accountability, and it'd be easier to identify and deal with those abusing the site.

[benetnasch85]

We have encountered server errors for more than 7 hours. Catagolue pages that do not need server data are displaying OK.