« first day (1048 days earlier)      last day (432 days later) » 

1:46 AM
IMO short/long should be by frequency not length
 
@hyper-neutrino it does
@emanresuA yep
the words closer to the top compress shorter
 
Yeah, what I mean is that more frequent words like because should be in short, while less frequent words like bluff should be in long (or something along those lines
 
iirc the "short dict" is in fact so named because the representation is shorter :P
 
Isn't it because of word length?
Holy hell have I gotten it wrong this whole time?
well if it's word frequency, any list can be used
sorted by google ngram frequency
 
1:59 AM
...
lmao
yeah the short dictionary takes less information to compress
recall that it takes 1 bit to determine which dictionary to use, then a word from the short dict takes log2(# of short-dict words) to compress and a word from the long dict takes log2(# of long-dict words)
so the dict length really refers to the dict length and not the words within them and the smaller the dictionary, the fewer bits it takes to encode because the less information the dict can hold to begin with
 
hey @hyper-neutrino
 
you could theoretically have more bits, like you could have a short, long, and super-long and use divmod 3 to determine which dict to use but then that adds extra overhead to all encoding so it's likely not really worth it
 
!!/status tingly
 
Bot status: Onwinye
Uptime: 12:28:12.632235
Wunnying since: 2023-02-28T13:32:32.820900
Ewwows since stawtup: 0
 
perish
 
2:00 AM
AHAHAHAAAAA
 
Could be good for word list sorting
 
Look on my works, ye Mighty, and despair!
 
How many words in the short dictionary?
 
!!/status
 
@user I am doing situps. 100 of them
ysthakur approved #18 (More statuses)
ysthakur merged pull request #18 (More statuses) in Vyxal/VyxalBot2
ysthakur deleted branch more-statuses in Vyxal/VyxalBot2
 
2:03 AM
more cowbell statuses
!!/status goofy
 
Bot status: Online🤓🤓
Uptime: 12:31:09.117872🤓🤓
Running since: 2023-02-28T13:32:32.820900🤓🤓🤓
Errors since startup: 0🤓🤓
 
goofy auh
 
ysthakur created branch existential-crisis in Vyxal/VyxalBot2
 
I hate it when that happens
 
ysthakur opened pull request #19 (10% probability of amilyxal being false for lyxal) in Vyxal/VyxalBot2
 
2:09 AM
I think I found a good word list already sorted by frequency and pretty large
Apparently it has 2747823 words
Which is a lot
Sorted by frequency on Wikipedia
!!/ping discussion how many words in the short dictionary and how many words in long dictionary?
 
@Ginger @lyxal @hyper-neutrino @Niko @The Thonnu @Steffan @mathcat @user ^
 
Ooh, and it's pretty up-to-date too
@lyxal We can do this experimentally
Get a humongous corpus, choose a size for the short and long dictionaries, use that to compress the corpus, see how small it gets, then try it again with different sizes
Does the wikipedia corpus include frequency? If so, that's all we need
 
@user words are sorted by frequency and include occurrences
 
@VyxalBot @mathcat approve plz
@lyxal poggers
 
Also, there's surely a limit on how many words are in the short dictionary
 
2:13 AM
Yes, but we can get a tighter bound on it experimentally
 
ysthakur deleted branch NOW'S-YOUR-CHANCE-TO-BE-A-BIG-SHOT in Vyxal/VyxalBot2
ysthakur closed issue #17 (More status variations) in Vyxal/VyxalBot2
 
!!/status nerd
 
Sorry Seggan, I'm afraid I can't do that.
 
aww
 
2:37 AM
!!/status tingly
 
Bot status: Onwinye
Uptime: 13:05:17.521001
Wunnying since: 2-2-2023-02-28T13:32:32.820900
Ewwows since stawtup: 0
 
your honour, death.
@user hey golfscript made it into the list :p
 
@VyxalBot well, guess I have no other choice
 
@lyxal Nice
 
@hyper-neutrino it's only fair I guess
 
2:42 AM
@hyper-neutrino lmao
Does it actually need to be a room owner?
 
yes
 
for pinning the releases, yes
 
^
 
Ah
 
also RO modifications are announced by feeds though I don't expect anyone to just happen to know that :P
 
2:43 AM
Ohhh
That's why I couldn't add it just now
 
(plus all of those links are example.com)
 
lmao
 
@hyper-neutrino didn't look at the pfp correctly lol
 
I did a bit of tomfoolery
 
because I did know that, I just forgot to look for the orange bit
I thought it looked a bit off, but just ignored it mentally lol
@user also, caird isn't able to exist with the Wikipedia dictionary :p
there's caird, but no coinhering or ahhing
which is fine with me, because it frees us from the shackels of jelly just that bit further :p
 
3:03 AM
@lyxal I've fixed the VyxalBot PR
You will have more existential crises now
 
Lyxal approved #19 (10% probability of amilyxal returning wrong result)
 
@lyxal i actually had no idea that was the actual sequence of tokens until i checked it by modifying sss
 
there was a discussion about it in TNB a while ago
 
didn't even know coinhering could be a word
and now that i've looked it up my mind is poisoned by the actual pronunciation :P
 
in The Nineteenth Byte, Sep 7, 2021 at 21:19, by caird coinheringaahing
The "words" of my username are caird, coinhering, aahing
in The Nineteenth Byte, Sep 7, 2021 at 21:23, by DLosc
@cairdcoinheringaahing Coinheringaah! So the username means "caird [who is] coinhering [and] aahing," but then "coinheringaahing" is a new verb derived from the username.
in The Nineteenth Byte, Jan 17, 2021 at 17:31, by caird coinheringaahing
Turns out my username consists of three words: caird, coinhering and aahing
in The Nineteenth Byte, Jul 14, 2022 at 18:06, by NoHaxJustRadvylf
It's caird + coinhering + aahing
 
3:07 AM
missing a leading space smh
https://tio.run/##dVNLboMwEF3DKaYrcIMQIWqkRmLbSyCKXDCfCjAytGnO0LN0101VdVP1IjkJHTskjtUUJIxn5r03H7vfjRXvVtNUCN7CI2uanV93IxO9YPiFuu25GOHatnNWwICPm/EWncPAcrKxLY07xuZ1Nta8o2JnWznT0RCB49iWZC@ROQIJTR/owNw44zlLe1oyv6i73M0qKggsYAkFFyB3UHegqRIPwpuA2Na2qhsGMyVmc2T3oEVG1MjrZ/xzT@YVkUHF7I4gkCCNyi6jbtcSZlaziOCUtcof812FCcaxZmCKtmhomQ5b2iPlHUWrNvY0U0LnjFeHBhn5hYpIJyjR/5VlCqpIpFyeeZTqvfYFRvFDJcf3lzxU5GPFUj1aDHP1zm94V3pnk/cVF4nVkhgqOF/2ckGlYZ1rahClu@VCHh3TFSuWZG7WqexDs2aEXOIg8aUjk6dMnihlXG5MpGyLAXXwPcReGrvjaS/ekifRGXNU9HYvsDJXXRjn@@vndf/xtob957tDyDT9Ag
 
it's been mentioned more times than I remember]
 
ysthakur merged pull request #19 (10% probability of amilyxal returning wrong result) in Vyxal/VyxalBot2
ysthakur deleted branch existential-crisis in Vyxal/VyxalBot2
 
!!/amilyxal
 
@lyxal You are lyxal.
 
3:11 AM
I think the bot needs to restart for it to take effect
Could one of y'all try running !!/die?
 
!!/die
 
Ah'll be bahk.
 
(sorry ginger in advance)
 
Well, here we are again.
 
!!/status exciting
 
3:11 AM
BOT STATUS: ONLINE!!!
UPTIME: 0:00:07.599592!!
RUNNING SINCE: 2023-03-01T03:11:51.960366!!!!
ERRORS SINCE STARTUP: 0!!!!
 
!!/amilyxal
 
@lyxal You are lyxal.
 
!!/amilyxal
 
@user You are not lyxal.
 
!!/status
 
3:13 AM
@user I am doing air traffic control. I'm a chatbot, not an AI
 
!!/amilyxal
 
@user You are not lyxal.
 
!!/amilyxal
 
@lyxal You are lyxal.
 
bruh
!!/amilyxal
 
3:13 AM
@user You are not lyxal.
 
!!/amilyxal
 
@lyxal You are lyxal.
 
Did I screw up the PR lol
 
probably needs an explicit restart
!!/pull
 
Sorry lyxal, I'm afraid I can't do that.
 
3:14 AM
I suppose that it doesn't auto-pull
 
!!/touch wsgi.py
 
Sorry user, I'm afraid I can't do that.
 
dumbo
 
!!/are you afraid you can't do that
 
Sorry lyxal, I'm afraid I can't do that.
 
3:29 AM
bruh guys the amount of hyperpings
!!/permissions revoke me discussion
 
@Steffan User Steffan is no longer a member of group discussion.
 
yeet
 
@Steffan ah yes, 5 hyperpings :p
 
3:47 AM
maybe the vyxal bot should have moods
 
 
2 hours later…
5:53 AM
is dictionary compress simply replace each common word with token?
maybe we can cut off after the part where it starts to take 5 token to compress a word(or something like that)
 
6:53 AM
@Niko it used to be, but now it's a more complicated algorithm I don't understand myself
It's the same one jelly uses
But ported over
 
!!/amilyxal
 
@mathcat You are not lyxal.
 
sad
 
 
4 hours later…
11:18 AM
Feb 23 at 11:03, by lyxal
227845 long and 20453 short
pasting that because I need quick reference to those numbers
thems the jelly stats
and so I'm going to use that as a starting point for balancing the word list lengths
 
11:31 AM
will v3 have modules?
 
that's the plan
and yes, it could be a good idea for flax to have them too
 
@lyxal nice
 
@user native doesn't like this
Holy hell
Vyxal's SSS is shorter than Jelly's SSS on a very important test case
String: "We're no strangers to love"
Jelly:  “©œ1Ṿ⁽TtṁO4ƬṅSḳXÄ» (18 bytes)
Vyxal: "'V«ᵡỌ}ṗ→\ᵜ;gṆᵘ” (16 bytes)
 
12:03 PM
short length:19000
long length:646609
that could be the sweet spot
import json

with open(
    r"file.txt", "r", encoding="utf-8"
) as f:
    lines = f.readlines()
    lines = map(lambda x: x.split(), lines)
    lines = list(filter(lambda x: int(x[1]) > 30, lines))
    short = list(map(lambda x: x[0], lines))[:19000]
    long = list(map(lambda x: x[0], lines))[19000:]
    print("short length:" + str(len(short)))
    print("long length:" + str(len(long)))

# import dictionary

with open("ShortDictionary.txt", "w", encoding="utf-8") as out:
    out.write("\n".join(short))
that's my sorting criteria
@Ginger does !!/die pull changes when restarting?
 
it does not
but I can do a pull pretty easily
 
ah well see we ran !!/die
thinking it did
sorry about the uptime reset
 
me when
alright, one second plz
 
Ah'll be bahk.
Well, here we are again.
 
It's always such a pleasure.
 
12:08 PM
!!/amilyxal
 
@Ginger An error occurred while handling message 63091502!
 
USERRRRRRRRRRRRRRRRRRRRRRRRRRRRRR!
🎵 uuuuser and his crazy explanations, the bot developer's gonna need their medication 🎵
 
!!/blame
 
@lyxal It was hyper-neutrino's fault!
 
not quite
 
12:11 PM
@user ...you used ^ instead of not
@lyxal you reviewed this, how tf did you not see that??? :p
 
Ah'll be bahk.
Well, here we are again.
 
!!/amilyxal
 
@Ginger You are not lyxal.
 
me when I am not lyxal
 
@Ginger because I thought the bitwise xor worked
also, not and
it needs to be xor
 
12:13 PM
ohhhhhh
yeah, that makes sense
 
so really, you're the bingus here
 
*laughs in cat*
...bruh just use !=
that's xor
 
Ah'll be bahk.
Well, here we are again.
 
!!/amilyxal
 
@Ginger You are not lyxal.
 
12:14 PM
ok, normality has been restored
 
@Ginger ^ should work though
because booleans get cast to int in python
 
but it doesn't, and they don't
 
well you or user clearly did a bingus somewhere, or otherwise the bot wouldn't'd've'd shid its pants
 
if a number is needed and you get a boolean, True becomes 1, and False becomes 0
what was the error
 
12:16 PM
lemme check
> TypeError: unsupported operand type(s) for ^: 'int' and 'float'
operator precedence stikes again
it xor'd the random.random() call with the "is this user lyxal" check, instead of the comparison
 
!!/amilyxal
 
@lyxal You are not lyxal.
3
 
oh
well
 
oh no
 
dang
 
12:19 PM
 
12:30 PM
@hyper-neutrino @UnrelatedString do either of y'all know if there's codepoints in the Jelly codepage that will never appear in an SSS compressed string
 
@VyxalBot AAAH
 
just pushed the entire lyrics of never gonna give you up as a test case
gaming
 
12:46 PM
@lyxal I feel like this should be a constant
 
12:59 PM
#PR 1646 is ready for review. Do note that the Native tests don't work for some reason
!!/amilyxal
 
@lyxal You are not lyxal.
 
still not?
how about now?
!!/amilyxal
 
@lyxal You are not lyxal.
 
huh
!!/amilyxal
 
@lyxal You are not lyxal.
 
1:03 PM
you sure?
!!/amilyxal
 
@lyxal You are not lyxal.
 
yeesh
I'll check one last time
maybe that'll change it
!!/amilyxal
 
@lyxal You are not lyxal.
 
 
1:18 PM
@VyxalBot finally
 
@VyxalBot lmao this doesn't seem right
 
I think I broke it
!!/amilyxal
 
@Ginger You are not lyxal.
 
uhh
 
this is fine.
 
1:22 PM
it seems that nobody is lyxal
eh whatever I'm sure it's fine
 
Ah'll be bahk.
Well, here we are again.
 
@lyxal I think I fixed it
 
!!/amilyxal
 
@mathcat You are not lyxal.
 
!!/amilyxal
 
1:24 PM
@Ginger You are not lyxal.
 
aw
 
!!/amilyxal
 
@lyxal You are lyxal.
 
Oh thank f*ck
Man it's good being me again
 
!!/amilyxal
 
1:28 PM
@Ginger You are not lyxal.
 
me when
 
!!/amilyxal
 
@mathcat You are not lyxal.
 
2:13 PM
@lyxal only the quotation marks (string opening and closing)
 
 
1 hour later…
3:36 PM
!!/amilyxal
 
@mathcat You are not lyxal.
 
I will never give up.
 
@mathcat Never gonna give you up!
 
3:55 PM
@lyxal Looks like Native doesn't work because of that io.Source.fromResource thing
gah
I'll try to fix later, got a class now
 
!!/amilyxal
 
@Ginger You are not lyxal.
 
!!/amilyxal
 
@Seggan You are lyxal.
6
 
YES
 
4:34 PM
!!/status
 
@user I am doing what's called a pro gamer move
 
!!/amilyxal
 
@user You are not lyxal.
 
oof
 
 
2 hours later…
6:25 PM
@Ginger not and ^ do entirely different things
Forgot about precedence, sorry about that
 
@user yeah, I had a square brain moment
 
 
1 hour later…
7:57 PM
I have not heard that phrase before but I will start using it now
 
@user glad to have expanded your lexicon :p
 
No, I just pushed a different word out of my brain to make room
 
glad to have right-shifted your lexicon, I guess
 
I can tell you what that word was because that word is in my brain anymore
 
ah yes, that makes sense
what is it then
 
8:01 PM
@Ginger Hmm, really
Because right-shifting implies there's an order
But that's the case
 
hm
 
It's really more of a heap, a list
 
y'know, I think I'm beginning to see a problem
 
Problem? I do see a problem
 
that makes two of us then
 
8:02 PM
No I meant there is no problem
 
eh, I think that "square brain" is a better word anyway
 
8:43 PM
@Ginger Did you figure out which word got yeeted out of my brain? I'm still trying to find it
Looks like you did
 
 
1 hour later…
10:04 PM
@user I sure did, but I decided not to replace it, because I like "square brain" more
 

« first day (1048 days earlier)      last day (432 days later) »