The Online Life-Like CA Soup Search

[Nathaniel]

And I might have already asked this but is there any way I or other forum members could help fix the data for objects not shown on the chart?

I just started coding some stuff for the CA soup search again this morning, so I'll try to add features like this. The only reason it hasn't been there from the get-go is that it's painfully difficult to track down and correct errors if they are made (eg. if a pattern containing two blinkers and a block is accidentally said to have three blinkers and a block, or some such thing). Not that I expect people to be making errors all the time, but it's difficult enough to manage just with one person doing it. For the next version of the script, I'm experiementing with a feature that downloads the (some number) most common patterns found so far for that rule, and automatically checks whether or not the pattern under consideration can be split into those subpatterns. If this works out like I hope it will, then there will be very very little post-processing that will have to be done by me after the fact.

[Nathaniel]

OK, version 1.03 of the soup search script is now online. Notable changes include:

• The user stats on the soup webpage have been extended quite a bit. There are now all-time and this-week-only stats, as well as stats per rule and user pages that show detailed stats for individual users.
• When running the script, you now see some stats detailing what has been found on that particular run (see screenshot below).
• Census data cleaning will now be quite a bit easier for me. Expect to see the census lists expanded a fair bit tomorrow.
• The script now remembers your name and e-mail address

Script screenshot

soup.png (109.9 KiB) Viewed 16991 times

[iconmaster]

I am having a problem installing your program. I did install Python 2.6.2, But I cannot figure out which DLL I should tell Golly to use to run the script. Please help me!

[Nathaniel]

I am having a problem installing your program. I did install Python 2.6.2, But I cannot figure out which DLL I should tell Golly to use to run the script. Please help me!

Unfortunately it depends a *lot* on your computer and there's no blanket answer. If you're using a Windows machine, then some places to check are:

C:\Python26\python26.dll
C:\Windows\SysWOW64\python26.dll

And slight variations of those -- you want to find python26.dll
Maybe try doing a search on your computer if it's not in one of those locations (be sure to search non-indexed and hidden files).

Edit: I've attached a zipped version of the Python 2.6.2 DLL to this post. I think it should work if you just extract that DLL wherever you want and point Golly to it.

[iconmaster]

Thanks. I will try right away!

[iconmaster]

I have sucsessfully activated the Soup Search, but I have another question:

How do you change the rule to work on? Myself, I want to start anlyzing other soups than Life soups.

[Nathaniel]

How do you change the rule to work on? Myself, I want to start anlyzing other soups than Life soups.

Just change the active rule in Golly before starting the script. Go to Control -> Set Rule... and then enter the rule that you want the script to work on. The only rules currently supported by the soup search script are the ones listed at the left under "Census Results" at the soup search homepage.

[iconmaster]

It would be a good idea to awnser some of these question in the FAQ.
A quick question: When i look up myself in the Stats page, it says I do not exist. Do I have to wait a while before it records me?

[Nathaniel]

A quick question: When i look up myself in the Stats page, it says I do not exist. Do I have to wait a while before it records me?

Yep, the script only uploads census data to the server once every 15 minutes. You will "exist" after the script does its first upload.

And yeah, I'll update the FAQ tonight.

[iconmaster]

Thanks for awnsering my questions. Now I can finally do the Soup Search!

EDIT: I have a few suggestions:

1. Include Diomeba,Balanced Ameoba, and B3468/S03.

2. In the Stats page, include a "Least Common Object Found" section. Not all that useful, but still fun.

[iconmaster]

It seems I suddenly get an "ImportError: no module named soup_search_exe" error when trying to run the script.
I have ran it sucsessfully before.

[ddt]

Hi,

I am trying to get 1.03 to run here. I am on amd64 so I first modified wxpython.cpp to use Py_InitModule4_64 in place of Py_InitModule4, rebuilt golly and tested with the python scripts supplied with golly. So far so good. Then I unzipped the soup stuff in the Python dir and tried running the soup_search.py script and got:

Traceback (most recent call last):
File "<string>", line 1, in <module>
File "/mnt/sdc4/home/ed/golly-2.0-src/Scripts/Python/soup_search.py", line 13, in <module>
import golly,soup_search_exe;reload(soup_search_exe)
ImportError: Bad magic number in /mnt/sdc4/home/ed/golly-2.0-src/Scripts/Python/soup_search_exe.pyc

Where is the source to soup_search_exe.pyc? Suspect that 64 bit python needs this recompiled.

$ python --version
Python 2.6.2

Thanks

[Nathaniel]

Traceback (most recent call last):
File "<string>", line 1, in <module>
File "/mnt/sdc4/home/ed/golly-2.0-src/Scripts/Python/soup_search.py", line 13, in <module>
import golly,soup_search_exe;reload(soup_search_exe)
ImportError: Bad magic number in /mnt/sdc4/home/ed/golly-2.0-src/Scripts/Python/soup_search_exe.pyc

Where is the source to soup_search_exe.pyc? Suspect that 64 bit python needs this recompiled.

Thanks

The problem is likely that you are using an incompatible version of Python -- unfortunately the soup search only works with Python 2.6.0 and 2.6.2 (and *maybe* with a couple other versions, if you're lucky). What version of Python are you using?

Unfortunately, the source code to soup_search_exe.pyc is not made public for security reasons -- I don't want people altering the source code and using it to upload erroneous data. However, I compiled the source code on Vista 64-bit so I don't expect that 32-bit vs. 64-bit would be the issue.

[ddt]

I just updated the post - its 2.6.2 here. If you want to send me the source I will just recompile it and see if that is the problem - I will not redistribute/modify it.

[Nathaniel]

If you want to send me the source I will just recompile it and see if that is the problem - I will not redistribute/modify it.

Sorry, but I don't think I can do that :/ I'll experiment with py2exe to see if I can make a more platform-independent version of the script.

Also, congrats to ErikDeNeve for topping the methuselah leaderboard with a lifespan of 26726!

[iconmaster]

I have another suggestion:

Add Pseudo Life to the rules.

[Lewis]

Could we add B3/S2456 to the rules? It turns up literally hundreds of still lifes (making it difficult to get statistics on by hand) and also many oscillators of periods 2, 3, 4, 5, 6, and a 3c/16 ship.

[Sokwe]

Why does the results for each rule only go down to patterns that have occured 10 or more times?

Are these rare patterns still making it to the database (I assume so)? On this note, the results for B35/S23 include objects with counts that are below this limit (long long long barge has been seen only three times after censusing almost 200,000,000 soups).

[Nathaniel]

Are these rare patterns still making it to the database (I assume so)? On this note, the results for B35/S23 include objects with counts that are below this limit (long long long barge has been seen only three times after censusing almost 200,000,000 soups).

Yep, everything's still in the database -- nothing should ever get lost, even if I'm behind on updating what is displayed. I went a bit overboard when I "cleaned" the B35/S23 table recently (it was easy since there's so few different objects in that rule). Now I just need to get off my fanny and clean the rest of the tables further down.

[ddt]

If you want to send me the source I will just recompile it and see if that is the problem - I will not redistribute/modify it.

Sorry, but I don't think I can do that :/ I'll experiment with py2exe to see if I can make a more platform-independent version of the script.

Also, congrats to ErikDeNeve for topping the methuselah leaderboard with a lifespan of 26726!

If you get stuck and want access to my box for some testing. I can setup an ssh account for you. Just generate a key pair using a pass phase; this protects me if the private key becomes public. Then send me the public key. If you are stuck on windows, putty makes a good ssh client.

[iconmaster]

There are some rules that form regular shapes, but still expand forever when given a large room, Like certain Generations rules. You could make the script use a torus universe with those rules, wait until stabilization, and then count them as normal. This would allow more rules to be censused.

[Lewis]

When will the tables next be updated? (some of the pictures for oscillators also seem to be missing)

[Nathaniel]

Here's a slightly cleaned up 23-cell version of the methuselah that just topped the leaderboard. It has a lifespan of 28,138. Slowly but surely we're closing in on Lidka

Code: Select all

x = 25, y = 13, rule = B3/S23
2$11b2o$9b2o2bo$10bo$12bo$10b2o$8bo2b2o$4bo$3b2ob2o14b2o$4bobo15bo$5bo!

[ddt]

If you want to send me the source I will just recompile it and see if that is the problem - I will not redistribute/modify it.

Sorry, but I don't think I can do that :/ I'll experiment with py2exe to see if I can make a more platform-independent version of the script.

Also, congrats to ErikDeNeve for topping the methuselah leaderboard with a lifespan of 26726!

Soup 1.03 works with 64 bit linux using golly 2.1 (after fixing golly to understand 64 bit python).

[knightlife]

Here's a slightly cleaned up 23-cell version of the methuselah

I found this 21-cell version: Is it worth it to trade off lifespan (11 tics shorter) for minimum size?

Code: Select all

x = 24, y = 16, rule = B3/S23
10b2o$10bob2o2$12bo4$b2o$obo19b2o$o2bo18bo$bo$2bo2$4bobo$5bo$5bo!

I suppose each initial cell count should have its own associated maximum known lifespan.

Edit:
Has anyone plotted initial cell count versus maximum lifespan for methuselahs?