Hacking apgsearch
Re: Hacking apgsearch
Just uploaded a haul of B3S23q, hope it worked!
Edit: woo hoo!
Edit: woo hoo!
\100\97\110\105
 A for awesome
 Posts: 2037
 Joined: September 13th, 2014, 5:36 pm
 Location: 0x1
 Contact:
Re: Hacking apgsearch
A small bugfix and update:
It fixes a bug that caused the auxiliary rules to be misgenerated sometimes and the rule to be randomly set to B3/S23 sometimes. Anyone currently using v0.20i is strongly encouraged to update to v0.21i because the bug severely limits v0.20i's capabilities. It also adds two small p4 ships to the common_names dict.
It fixes a bug that caused the auxiliary rules to be misgenerated sometimes and the rule to be randomly set to B3/S23 sometimes. Anyone currently using v0.20i is strongly encouraged to update to v0.21i because the bug severely limits v0.20i's capabilities. It also adds two small p4 ships to the common_names dict.
x₁=ηx
V ⃰_η=c²√(Λη)
K=(Λu²)/2
Pₐ=1−1/(∫^∞_t₀(p(t)ˡ⁽ᵗ⁾)dt)
$$x_1=\eta x$$
$$V^*_\eta=c^2\sqrt{\Lambda\eta}$$
$$K=\frac{\Lambda u^2}2$$
$$P_a=1\frac1{\int^\infty_{t_0}p(t)^{l(t)}dt}$$
http://conwaylife.com/wiki/A_for_all
Aidan F. Pierce
V ⃰_η=c²√(Λη)
K=(Λu²)/2
Pₐ=1−1/(∫^∞_t₀(p(t)ˡ⁽ᵗ⁾)dt)
$$x_1=\eta x$$
$$V^*_\eta=c^2\sqrt{\Lambda\eta}$$
$$K=\frac{\Lambda u^2}2$$
$$P_a=1\frac1{\int^\infty_{t_0}p(t)^{l(t)}dt}$$
http://conwaylife.com/wiki/A_for_all
Aidan F. Pierce
Re: Hacking apgsearch
I keep getting the error:
When I use a new rule with both 0.2i and 0.21i. What is happening? An example of a used rule:
Code: Select all

Golly warning:

Error reading APG_CoalesceObjects_B2aS12.rule on line 87: ba,bb,bc,2  too few entries

OK

Code: Select all
@RULE B2aS12
@TABLE
n_states:2
neighborhood:Moore
symmetries:rotate4reflect
0,1,0,1,0,0,0,0,0,1
0,1,0,0,1,0,0,0,0,1
0,1,0,0,0,1,0,0,0,1
0,0,1,0,1,0,0,0,0,1
0,0,1,0,0,1,0,0,0,1
0,0,1,0,0,0,1,0,0,1
1,0,0,0,0,0,0,0,0,0
1,1,1,1,0,0,0,0,0,0
1,1,1,0,1,0,0,0,0,0
1,1,1,0,0,1,0,0,0,0
1,1,1,0,0,0,1,0,0,0
1,1,1,0,0,0,0,1,0,0
1,1,1,0,0,0,0,0,1,0
1,1,0,1,0,1,0,0,0,0
1,1,0,1,0,0,1,0,0,0
1,1,0,0,1,0,1,0,0,0
1,0,1,0,1,0,1,0,0,0
1,1,1,1,1,0,0,0,0,0
1,1,1,1,0,1,0,0,0,0
1,1,1,1,0,0,1,0,0,0
1,1,1,0,1,1,0,0,0,0
1,1,1,0,1,0,1,0,0,0
1,1,1,0,1,0,0,1,0,0
1,1,1,0,1,0,0,0,1,0
1,1,1,0,0,1,1,0,0,0
1,1,1,0,0,1,0,1,0,0
1,1,1,0,0,1,0,0,1,0
1,1,1,0,0,0,1,1,0,0
1,1,0,1,0,1,0,1,0,0
1,0,1,0,1,0,1,0,1,0
1,0,0,0,1,1,1,1,1,0
1,0,0,1,0,1,1,1,1,0
1,0,0,1,1,0,1,1,1,0
1,0,0,1,1,1,0,1,1,0
1,0,0,1,1,1,1,0,1,0
1,0,0,1,1,1,1,1,0,0
1,0,1,0,1,0,1,1,1,0
1,0,1,0,1,1,0,1,1,0
1,0,1,1,0,1,0,1,1,0
1,1,0,1,0,1,0,1,1,0
1,0,0,1,1,1,1,1,1,0
1,0,1,0,1,1,1,1,1,0
1,0,1,1,0,1,1,1,1,0
1,0,1,1,1,0,1,1,1,0
1,1,0,1,0,1,1,1,1,0
1,1,0,1,1,1,0,1,1,0
1,0,1,1,1,1,1,1,1,0
1,1,0,1,1,1,1,1,1,0
1,1,1,1,1,1,1,1,1,0
@COLORS
0 0 0 0
1 255 255 255
\100\97\110\105
 A for awesome
 Posts: 2037
 Joined: September 13th, 2014, 5:36 pm
 Location: 0x1
 Contact:
Re: Hacking apgsearch
It's the @COLORS section; I will fix this in the next update, but for now, just put an amended version of the rule without the @COLORS section in Golly's rules folder (keeping the version including the colors in your rules folder). This takes advantage of the fact that Golly and apgsearch look up rules in the opposite order from each other, so Golly sees the colors and apgsearch doesn't.drc wrote:I keep getting the error:
When I use a new rule with both 0.2i and 0.21i. What is happening? An example of a used rule:Code: Select all
 Golly warning:  Error reading APG_CoalesceObjects_B2aS12.rule on line 87: ba,bb,bc,2  too few entries  OK 
Code: Select all
@RULE B2aS12 @TABLE n_states:2 neighborhood:Moore symmetries:rotate4reflect 0,1,0,1,0,0,0,0,0,1 0,1,0,0,1,0,0,0,0,1 0,1,0,0,0,1,0,0,0,1 0,0,1,0,1,0,0,0,0,1 0,0,1,0,0,1,0,0,0,1 0,0,1,0,0,0,1,0,0,1 1,0,0,0,0,0,0,0,0,0 1,1,1,1,0,0,0,0,0,0 1,1,1,0,1,0,0,0,0,0 1,1,1,0,0,1,0,0,0,0 1,1,1,0,0,0,1,0,0,0 1,1,1,0,0,0,0,1,0,0 1,1,1,0,0,0,0,0,1,0 1,1,0,1,0,1,0,0,0,0 1,1,0,1,0,0,1,0,0,0 1,1,0,0,1,0,1,0,0,0 1,0,1,0,1,0,1,0,0,0 1,1,1,1,1,0,0,0,0,0 1,1,1,1,0,1,0,0,0,0 1,1,1,1,0,0,1,0,0,0 1,1,1,0,1,1,0,0,0,0 1,1,1,0,1,0,1,0,0,0 1,1,1,0,1,0,0,1,0,0 1,1,1,0,1,0,0,0,1,0 1,1,1,0,0,1,1,0,0,0 1,1,1,0,0,1,0,1,0,0 1,1,1,0,0,1,0,0,1,0 1,1,1,0,0,0,1,1,0,0 1,1,0,1,0,1,0,1,0,0 1,0,1,0,1,0,1,0,1,0 1,0,0,0,1,1,1,1,1,0 1,0,0,1,0,1,1,1,1,0 1,0,0,1,1,0,1,1,1,0 1,0,0,1,1,1,0,1,1,0 1,0,0,1,1,1,1,0,1,0 1,0,0,1,1,1,1,1,0,0 1,0,1,0,1,0,1,1,1,0 1,0,1,0,1,1,0,1,1,0 1,0,1,1,0,1,0,1,1,0 1,1,0,1,0,1,0,1,1,0 1,0,0,1,1,1,1,1,1,0 1,0,1,0,1,1,1,1,1,0 1,0,1,1,0,1,1,1,1,0 1,0,1,1,1,0,1,1,1,0 1,1,0,1,0,1,1,1,1,0 1,1,0,1,1,1,0,1,1,0 1,0,1,1,1,1,1,1,1,0 1,1,0,1,1,1,1,1,1,0 1,1,1,1,1,1,1,1,1,0 @COLORS 0 0 0 0 1 255 255 255
x₁=ηx
V ⃰_η=c²√(Λη)
K=(Λu²)/2
Pₐ=1−1/(∫^∞_t₀(p(t)ˡ⁽ᵗ⁾)dt)
$$x_1=\eta x$$
$$V^*_\eta=c^2\sqrt{\Lambda\eta}$$
$$K=\frac{\Lambda u^2}2$$
$$P_a=1\frac1{\int^\infty_{t_0}p(t)^{l(t)}dt}$$
http://conwaylife.com/wiki/A_for_all
Aidan F. Pierce
V ⃰_η=c²√(Λη)
K=(Λu²)/2
Pₐ=1−1/(∫^∞_t₀(p(t)ˡ⁽ᵗ⁾)dt)
$$x_1=\eta x$$
$$V^*_\eta=c^2\sqrt{\Lambda\eta}$$
$$K=\frac{\Lambda u^2}2$$
$$P_a=1\frac1{\int^\infty_{t_0}p(t)^{l(t)}dt}$$
http://conwaylife.com/wiki/A_for_all
Aidan F. Pierce
 A for awesome
 Posts: 2037
 Joined: September 13th, 2014, 5:36 pm
 Location: 0x1
 Contact:
Re: Hacking apgsearch
I just noticed, in the Rule Request Thread, this:
I guess v0.2Ni has been outputting the rules in the wrong format all along. Here's the patched script: You are, again, strongly recommended to update. I also updated the Henseltest.py script in the main v0.2Ni post to reflect the change.dvgrn wrote:the correct "n" instead of the old Life32 "v".
x₁=ηx
V ⃰_η=c²√(Λη)
K=(Λu²)/2
Pₐ=1−1/(∫^∞_t₀(p(t)ˡ⁽ᵗ⁾)dt)
$$x_1=\eta x$$
$$V^*_\eta=c^2\sqrt{\Lambda\eta}$$
$$K=\frac{\Lambda u^2}2$$
$$P_a=1\frac1{\int^\infty_{t_0}p(t)^{l(t)}dt}$$
http://conwaylife.com/wiki/A_for_all
Aidan F. Pierce
V ⃰_η=c²√(Λη)
K=(Λu²)/2
Pₐ=1−1/(∫^∞_t₀(p(t)ˡ⁽ᵗ⁾)dt)
$$x_1=\eta x$$
$$V^*_\eta=c^2\sqrt{\Lambda\eta}$$
$$K=\frac{\Lambda u^2}2$$
$$P_a=1\frac1{\int^\infty_{t_0}p(t)^{l(t)}dt}$$
http://conwaylife.com/wiki/A_for_all
Aidan F. Pierce
Re: Hacking apgsearch
Possibly a bug: When searching in rule B2a/S0146, apgsearchisotropicv0.2 reports something like xq1_q1qz08, which is actually xp6_q1qz08. I have totally no idea about it.
 Attachments

 Screenshot of B2a/S0146 census, where the same spaceship has two different names.
 1.png (22.1 KiB) Viewed 17097 times
Still drifting.
 Apple Bottom
 Posts: 1033
 Joined: July 27th, 2015, 2:06 pm
 Contact:
Re: Hacking apgsearch
Not so much a bug as a problem: searching certain rules will cause 0.54+0.21i to rapidly consume all available memory and cause the machine it's running on to thrash.
Example rule: B2aei3a/S2i34i (rule files generated by isotropicrule.py):
I'd make a mostly uneducated guess that this is due to extremely highperiod oscillators. If so, I'd suggest capping the maximum number of generations the script'll do when investigating an object before giving up saying "ah ain't got no clue 'bout this here 'un".
Example rule: B2aei3a/S2i34i (rule files generated by isotropicrule.py):
Code: Select all
@RULE B2aei3a_S2i34i
*** File autogenerated by saverule. ***
This is a two state, isotropic, nontotalistic rule on the Moore neighbourhood.
The notation used to define the rule was originally proposed by Alan Hensel.
See http://www.ibiblio.org/lifepatterns/neighbors2.html for details
@TABLE
n_states:2
neighborhood:Moore
symmetries:rotate4reflect
var a={0,1}
var b={0,1}
var c={0,1}
var d={0,1}
var e={0,1}
var f={0,1}
var g={0,1}
var h={0,1}
# Birth
0,1,0,0,1,0,0,0,0,1
0,0,1,0,1,0,0,0,0,1
0,0,1,0,0,0,1,0,0,1
0,1,1,0,1,0,0,0,0,1
0,1,1,0,0,1,0,0,0,1
0,1,1,0,0,0,1,0,0,1
0,1,1,0,0,0,0,1,0,1
0,1,1,0,0,0,0,0,1,1
0,1,0,1,0,1,0,0,0,1
0,1,0,1,0,0,1,0,0,1
0,1,0,0,1,0,1,0,0,1
0,0,1,0,1,0,1,0,0,1
# Survival
1,1,1,0,0,0,0,0,0,1
1,1,0,1,0,0,0,0,0,1
1,1,0,0,1,0,0,0,0,1
1,0,1,0,1,0,0,0,0,1
1,0,1,0,0,0,1,0,0,1
1,1,1,1,0,0,0,0,0,1
1,1,1,0,1,0,0,0,0,1
1,1,1,0,0,1,0,0,0,1
1,1,1,0,0,0,1,0,0,1
1,1,1,0,0,0,0,1,0,1
1,1,1,0,0,0,0,0,1,1
1,1,0,1,0,1,0,0,0,1
1,1,0,1,0,0,1,0,0,1
1,1,0,0,1,0,1,0,0,1
1,0,1,0,1,0,1,0,0,1
1,1,1,1,1,0,0,0,0,1
1,1,1,1,0,1,0,0,0,1
1,1,1,1,0,0,1,0,0,1
1,1,1,0,1,0,1,0,0,1
1,1,1,0,1,0,0,1,0,1
1,1,1,0,1,0,0,0,1,1
1,1,1,0,0,1,1,0,0,1
1,1,1,0,0,1,0,1,0,1
1,1,1,0,0,1,0,0,1,1
1,1,1,0,0,0,1,1,0,1
1,1,0,1,0,1,0,1,0,1
1,0,1,0,1,0,1,0,1,1
# Death
1,a,b,c,d,e,f,g,h,0
@COLORS
If you speak, your speech must be better than your silence would have been. — Arabian proverb
Catagolue: Apple Bottom • Life Wiki: Apple Bottom • Twitter: @_AppleBottom_
Proud member of the Pattern Raiders!
Catagolue: Apple Bottom • Life Wiki: Apple Bottom • Twitter: @_AppleBottom_
Proud member of the Pattern Raiders!
 A for awesome
 Posts: 2037
 Joined: September 13th, 2014, 5:36 pm
 Location: 0x1
 Contact:
Re: Hacking apgsearch
Generally (and I've encountered this a lot), it has to do with the time it takes soups to come to stabilization. For this rule, that time tends to run over 100000 gens. The RuleLoader algorithm naturally runs these patterns slower and takes more memory to do so (thanks to the hashing, which is useless for running soups to stabilization). I've found that 0.54+0.21i chokes up once soups start regularly taking in excess of 20000 gens to stabilize. I see no way to solve this problem, other than to write 0.54+0.21i up into a standalone c++ implementation (which I don't really know how to do) and use a nonhashing algorithm or wait for Golly 3.0 to come out.Apple Bottom wrote:Not so much a bug as a problem: searching certain rules will cause 0.54+0.21i to rapidly consume all available memory and cause the machine it's running on to thrash.
Example rule: B2aei3a/S2i34i (rule files generated by isotropicrule.py):Code: Select all
image
I'd make a mostly uneducated guess that this is due to extremely highperiod oscillators.Code: Select all
rule
x₁=ηx
V ⃰_η=c²√(Λη)
K=(Λu²)/2
Pₐ=1−1/(∫^∞_t₀(p(t)ˡ⁽ᵗ⁾)dt)
$$x_1=\eta x$$
$$V^*_\eta=c^2\sqrt{\Lambda\eta}$$
$$K=\frac{\Lambda u^2}2$$
$$P_a=1\frac1{\int^\infty_{t_0}p(t)^{l(t)}dt}$$
http://conwaylife.com/wiki/A_for_all
Aidan F. Pierce
V ⃰_η=c²√(Λη)
K=(Λu²)/2
Pₐ=1−1/(∫^∞_t₀(p(t)ˡ⁽ᵗ⁾)dt)
$$x_1=\eta x$$
$$V^*_\eta=c^2\sqrt{\Lambda\eta}$$
$$K=\frac{\Lambda u^2}2$$
$$P_a=1\frac1{\int^\infty_{t_0}p(t)^{l(t)}dt}$$
http://conwaylife.com/wiki/A_for_all
Aidan F. Pierce
 Apple Bottom
 Posts: 1033
 Joined: July 27th, 2015, 2:06 pm
 Contact:
Re: Hacking apgsearch
OK, fair enough. The problem's easy to diagnose and fix, so it shouldn't be a problem in practice.A for awesome wrote:Generally (and I've encountered this a lot), it has to do with the time it takes soups to come to stabilization. For this rule, that time tends to run over 100000 gens. The RuleLoader algorithm naturally runs these patterns slower and takes more memory to do so (thanks to the hashing, which is useless for running soups to stabilization). I've found that 0.54+0.21i chokes up once soups start regularly taking in excess of 20000 gens to stabilize. I see no way to solve this problem, other than to write 0.54+0.21i up into a standalone c++ implementation (which I don't really know how to do) and use a nonhashing algorithm or wait for Golly 3.0 to come out.
Still, I'm wondering. Why run the pattern until stabilization no matter how generations it takes, instead of stopping after (say) 20,000? Rules like these aren't really suitable for soupsearching anyway, so perhaps the script could just pop up a message to this effect (possibly allowing the user to continue anyway, if they so choose).
Thanks for looking into this, BTW!
If you speak, your speech must be better than your silence would have been. — Arabian proverb
Catagolue: Apple Bottom • Life Wiki: Apple Bottom • Twitter: @_AppleBottom_
Proud member of the Pattern Raiders!
Catagolue: Apple Bottom • Life Wiki: Apple Bottom • Twitter: @_AppleBottom_
Proud member of the Pattern Raiders!

 Posts: 653
 Joined: June 14th, 2014, 5:03 pm
 Contact:
 A for awesome
 Posts: 2037
 Joined: September 13th, 2014, 5:36 pm
 Location: 0x1
 Contact:
Re: Hacking apgsearch
This seems like it might be a very good idea for nontotalistic rules. I will probably implement some variation of this feature if I ever get around to making 0.22i.Apple Bottom wrote:Still, I'm wondering. Why run the pattern until stabilization no matter how generations it takes, instead of stopping after (say) 20,000? Rules like these aren't really suitable for soupsearching anyway, so perhaps the script could just pop up a message to this effect (possibly allowing the user to continue anyway, if they so choose).
P.S. I should probably state as a disclaimer that the explanation given in my previous post was just an educated guess (read "wild speculation") about the source of the bug. My experience with 0.Ni and with Golly in general seems to suggest it, but it's not a 100% certainty.
I am mildly embarrassed to say that I don't actually know how to do that. It probably would still run slower than the equivalent outertotalistic search (due to inherent inefficiencies in the RuleLoader algorithm compared to QuickLife).fluffykitty wrote:What happens if you set the hash memory to 0?
x₁=ηx
V ⃰_η=c²√(Λη)
K=(Λu²)/2
Pₐ=1−1/(∫^∞_t₀(p(t)ˡ⁽ᵗ⁾)dt)
$$x_1=\eta x$$
$$V^*_\eta=c^2\sqrt{\Lambda\eta}$$
$$K=\frac{\Lambda u^2}2$$
$$P_a=1\frac1{\int^\infty_{t_0}p(t)^{l(t)}dt}$$
http://conwaylife.com/wiki/A_for_all
Aidan F. Pierce
V ⃰_η=c²√(Λη)
K=(Λu²)/2
Pₐ=1−1/(∫^∞_t₀(p(t)ˡ⁽ᵗ⁾)dt)
$$x_1=\eta x$$
$$V^*_\eta=c^2\sqrt{\Lambda\eta}$$
$$K=\frac{\Lambda u^2}2$$
$$P_a=1\frac1{\int^\infty_{t_0}p(t)^{l(t)}dt}$$
http://conwaylife.com/wiki/A_for_all
Aidan F. Pierce
 Apple Bottom
 Posts: 1033
 Joined: July 27th, 2015, 2:06 pm
 Contact:
Re: Hacking apgsearch
You mean the memory limit for HashLife in Golly's preferences? I'm not sure what that would accomplish, since setting it to 0 would actually remove the limit (set to 500 MB now).fluffykitty wrote:What happens if you set the hash memory to 0?
In any case I noticed something interesting: when you leave Golly and apgsearch 0.54+0.21i running undisturbed, memory usage only fluctuates by about 300 MB, barely a blip on the radar. However, if you interact with the view, say by zooming or panning, it sometimes suddenly starts consuming memory at a rate of 700 MB/second.
I haven't figured out exactly what's triggering this yet, but it appears that it's due to rapid UI interaction. If you don't do anything, patterns will happily evolve for tens of millions of generations.
I'm fairly sure right now that this IS a bug somewhere (in Golly, Golly's scripting subsystem, or apgsearch). Presumably it doesn't usually manifest because patterns stabilize too fast in "normal", soupsearchable rules for the user to be able to trigger it.
'course, given that, it's still just a curiosity.
EDIT: here's another bug. Soupsearching a related rule, B2ck3a/S2aei34i, produces the following error (sometimes quickly, sometimes after a while):
"period" is likely zero there.
If you speak, your speech must be better than your silence would have been. — Arabian proverb
Catagolue: Apple Bottom • Life Wiki: Apple Bottom • Twitter: @_AppleBottom_
Proud member of the Pattern Raiders!
Catagolue: Apple Bottom • Life Wiki: Apple Bottom • Twitter: @_AppleBottom_
Proud member of the Pattern Raiders!
 Apple Bottom
 Posts: 1033
 Joined: July 27th, 2015, 2:06 pm
 Contact:
Re: Hacking apgsearch
There's a small issue with chaoticgrowth patterns; these are reported as z_* instead of zz_*, e.g. z_QUADRATIC instead of zz_QUADRATIC, as seen here.
The culprit's powerlyse(), which returns the wrong descriptions.
The culprit's powerlyse(), which returns the wrong descriptions.
If you speak, your speech must be better than your silence would have been. — Arabian proverb
Catagolue: Apple Bottom • Life Wiki: Apple Bottom • Twitter: @_AppleBottom_
Proud member of the Pattern Raiders!
Catagolue: Apple Bottom • Life Wiki: Apple Bottom • Twitter: @_AppleBottom_
Proud member of the Pattern Raiders!
 A for awesome
 Posts: 2037
 Joined: September 13th, 2014, 5:36 pm
 Location: 0x1
 Contact:
Re: Hacking apgsearch
http://conwaylife.com/forums/viewtopic. ... &start=281Apple Bottom wrote:There's a small issue with chaoticgrowth patterns; these are reported as z_* instead of zz_*, e.g. z_QUADRATIC instead of zz_QUADRATIC, as seen here.
The culprit's powerlyse(), which returns the wrong descriptions.
x₁=ηx
V ⃰_η=c²√(Λη)
K=(Λu²)/2
Pₐ=1−1/(∫^∞_t₀(p(t)ˡ⁽ᵗ⁾)dt)
$$x_1=\eta x$$
$$V^*_\eta=c^2\sqrt{\Lambda\eta}$$
$$K=\frac{\Lambda u^2}2$$
$$P_a=1\frac1{\int^\infty_{t_0}p(t)^{l(t)}dt}$$
http://conwaylife.com/wiki/A_for_all
Aidan F. Pierce
V ⃰_η=c²√(Λη)
K=(Λu²)/2
Pₐ=1−1/(∫^∞_t₀(p(t)ˡ⁽ᵗ⁾)dt)
$$x_1=\eta x$$
$$V^*_\eta=c^2\sqrt{\Lambda\eta}$$
$$K=\frac{\Lambda u^2}2$$
$$P_a=1\frac1{\int^\infty_{t_0}p(t)^{l(t)}dt}$$
http://conwaylife.com/wiki/A_for_all
Aidan F. Pierce
 gameoflifeboy
 Posts: 474
 Joined: January 15th, 2015, 2:08 am
Re: Hacking apgsearch
Apple Bottom wrote:There's a small issue with chaoticgrowth patterns; these are reported as z_* instead of zz_*, e.g. z_QUADRATIC instead of zz_QUADRATIC, as seen here.
This should be fixed in v(0.54+0.22i).A for awesome wrote:http://conwaylife.com/forums/viewtopic. ... &start=281
Another thing that should be added is the ability to separate pairs of gliders that only run in nonLifelike rules. See here: http://catagolue.appspot.com/object/xq4 ... b3kq4js23
 A for awesome
 Posts: 2037
 Joined: September 13th, 2014, 5:36 pm
 Location: 0x1
 Contact:
Re: Hacking apgsearch
I'm actually working on a version of pseudobangbang that separates moving objects. I will probably include it in the next version, which would also get rid of things like this.gameoflifeboy wrote:Another thing that should be added is the ability to separate pairs of gliders that only run in nonLifelike rules. See here: http://catagolue.appspot.com/object/xq4 ... b3kq4js23
x₁=ηx
V ⃰_η=c²√(Λη)
K=(Λu²)/2
Pₐ=1−1/(∫^∞_t₀(p(t)ˡ⁽ᵗ⁾)dt)
$$x_1=\eta x$$
$$V^*_\eta=c^2\sqrt{\Lambda\eta}$$
$$K=\frac{\Lambda u^2}2$$
$$P_a=1\frac1{\int^\infty_{t_0}p(t)^{l(t)}dt}$$
http://conwaylife.com/wiki/A_for_all
Aidan F. Pierce
V ⃰_η=c²√(Λη)
K=(Λu²)/2
Pₐ=1−1/(∫^∞_t₀(p(t)ˡ⁽ᵗ⁾)dt)
$$x_1=\eta x$$
$$V^*_\eta=c^2\sqrt{\Lambda\eta}$$
$$K=\frac{\Lambda u^2}2$$
$$P_a=1\frac1{\int^\infty_{t_0}p(t)^{l(t)}dt}$$
http://conwaylife.com/wiki/A_for_all
Aidan F. Pierce
Re: Hacking apgsearch
Hi guys, I am currently testing my custom 3state lifelike rules with birthforcers implemented. Is there a way to adapt apgsearch for oscillators and spaceships in this rule (eelife3, I did not use state 4 or 5 or 6)?
The raw rule before undergoing table generation:
The rule table :
The raw rule before undergoing table generation:
Code: Select all
switch (c)
{
case 0:
if (state1 neighbors==8)
return 2;
else if ((state1 neighbors + state6 neighbors ==3 )(state2 neighbors + state4 neighbors >=1))
return 1;
else
return 0;
case 1 :
if (state1 neighbors==8)
return 2;
else if (
!(
( ( state1 neighbors + state6 neighbors )  1 )
==3)

(state3 neighbors + state4 neighbors>=1 )
)
return 0;
else
return 1;
case 2:
if ( ( ( state1 neighbors + state2 neighbors)  1 )==3 )
return 2;
else
return 0;
case 3: return 3;
case 4: return 4;
case 5: return 5;
case 6: return 6;
default:
return 0 ; // should throw an error here
}
Code: Select all
@RULE eelife3
# General Behavior:
#
# 0 is empty
# 1 is on
# 2 is birthforcer
# Specific Behavior
# 0 changes to 2 when state2 neighbors ==8
otherwise changes to 1 when (state1 neighbors==3)  (state2 neighbors>=1)
otherwise 0 stays 0
# 0 or 1 changes to 2 when state1 neighbors == 8
# 1 changes to 0 when (state1 neighbors) != 2, 3
# 2 changes to 0 when (state1 neighbors + state2 neighbors) !=2,3
# changes to 1 when state1 neighbors ==3
# otherwise 1,2 stays 1,2
#
@TABLE
# rules: 52
#
# Golly ruletable format.
# Each rule: C,N,NE,E,SE,S,SW,W,NW,C'
# N.B. Where the same variable appears multiple times in a transition,
# it takes the same value each time.
#
# Default for transitions not listed: no change
#
n_states:4
neighborhood:Moore
symmetries:rotate8
var a={0,1,2,3}
var b={0,1,2,3}
var c={0,1,2,3}
var d={0,1,2,3}
var e={0,1,2,3}
var f={0,1,2,3}
var g={0,1,2,3}
var h={0,3}
var i={0,3}
var j={0,3}
var k={0,3}
var l={0,3}
var m={0,1}
var n={0,2,3}
var o={0,2,3}
var p={0,2,3}
var q={0,2,3}
var r={0,2,3}
var s={0,2,3}
var t={0,2,3}
var u={1,2}
var v={0,3}
var w={0,3}
var x={1,2}
var y={1,2}
var z={1,2}
0,a,b,c,d,e,f,g,2,1
0,h,i,j,k,l,1,1,1,1
0,h,i,j,k,1,l,1,1,1
0,h,i,j,k,1,1,l,1,1
0,h,i,j,1,k,l,1,1,1
0,h,i,j,1,k,1,l,1,1
0,h,i,j,1,1,k,l,1,1
0,h,i,1,j,k,1,l,1,1
m,1,1,1,1,1,1,1,1,2
1,n,o,p,q,r,s,t,a,0
1,a,b,c,d,e,f,g,3,0
1,0,0,0,0,0,1,3,1,0
1,0,0,0,0,0,3,1,1,0
1,0,0,0,0,1,0,1,3,0
1,0,0,0,0,1,0,3,1,0
1,0,0,0,0,1,1,0,3,0
u,n,a,b,c,1,1,1,1,0
1,0,0,0,0,1,3,0,1,0
1,0,0,0,0,3,0,1,1,0
1,0,0,0,0,3,1,0,1,0
u,n,a,b,1,o,1,1,1,0
u,n,a,b,1,1,o,1,1,0
u,n,o,p,1,1,1,q,1,0
u,n,o,1,p,q,1,1,1,0
u,n,o,1,p,1,q,1,1,0
u,n,o,1,p,1,1,q,1,0
u,n,o,1,1,p,q,1,1,0
u,n,o,1,1,p,1,q,1,0
u,n,1,o,1,p,1,q,1,0
2,h,i,j,k,l,v,w,a,0
2,a,b,c,d,u,x,y,z,0
2,0,0,0,0,1,1,2,1,0
2,0,0,0,0,1,2,1,1,0
2,h,a,b,u,i,x,y,z,0
2,0,0,0,1,0,1,2,1,0
2,0,0,0,1,0,2,1,1,0
2,h,a,b,u,x,i,y,z,0
2,0,0,0,1,1,0,2,1,0
2,h,i,j,u,x,y,k,z,0
2,0,0,0,1,1,2,0,1,0
2,h,i,u,j,k,x,y,z,0
2,0,0,1,0,0,1,2,1,0
2,0,0,1,0,0,2,1,1,0
2,h,i,u,j,x,k,y,z,0
2,0,0,1,0,1,0,2,1,0
2,h,i,u,j,x,y,k,z,0
2,0,0,1,0,1,2,0,1,0
2,h,i,u,x,j,k,y,z,0
2,0,0,1,1,0,0,2,1,0
2,h,i,u,x,j,y,k,z,0
2,0,0,1,1,0,2,0,1,0
2,h,u,i,x,j,y,k,z,0
Re: Hacking apgsearch
This is getting really annoying:
Alright thanks!
Alright thanks!
Last edited by drc on April 14th, 2016, 6:07 pm, edited 1 time in total.
\100\97\110\105
 A for awesome
 Posts: 2037
 Joined: September 13th, 2014, 5:36 pm
 Location: 0x1
 Contact:
Re: Hacking apgsearch
Okay, I'm working on it! Give me time.drc wrote:This is getting really annoying:
Code: Select all
image
x₁=ηx
V ⃰_η=c²√(Λη)
K=(Λu²)/2
Pₐ=1−1/(∫^∞_t₀(p(t)ˡ⁽ᵗ⁾)dt)
$$x_1=\eta x$$
$$V^*_\eta=c^2\sqrt{\Lambda\eta}$$
$$K=\frac{\Lambda u^2}2$$
$$P_a=1\frac1{\int^\infty_{t_0}p(t)^{l(t)}dt}$$
http://conwaylife.com/wiki/A_for_all
Aidan F. Pierce
V ⃰_η=c²√(Λη)
K=(Λu²)/2
Pₐ=1−1/(∫^∞_t₀(p(t)ˡ⁽ᵗ⁾)dt)
$$x_1=\eta x$$
$$V^*_\eta=c^2\sqrt{\Lambda\eta}$$
$$K=\frac{\Lambda u^2}2$$
$$P_a=1\frac1{\int^\infty_{t_0}p(t)^{l(t)}dt}$$
http://conwaylife.com/wiki/A_for_all
Aidan F. Pierce
Re: Hacking apgsearch
Alright thanks!
Also, this is commonly clasiffied as pathological:
I'm guessing it separates the block and glider.
Also, this is commonly clasiffied as pathological:
Code: Select all
x = 6, y = 3, rule = B34k_S23
3b2o$o3b2o$2obo!
\100\97\110\105
 A for awesome
 Posts: 2037
 Joined: September 13th, 2014, 5:36 pm
 Location: 0x1
 Contact:
Re: Hacking apgsearch
Yes, that's what happens. That will be fixed in the next version as well.drc wrote:Also, this is commonly clasiffied as pathological:
I'm guessing it separates the block and glider.Code: Select all
rle
x₁=ηx
V ⃰_η=c²√(Λη)
K=(Λu²)/2
Pₐ=1−1/(∫^∞_t₀(p(t)ˡ⁽ᵗ⁾)dt)
$$x_1=\eta x$$
$$V^*_\eta=c^2\sqrt{\Lambda\eta}$$
$$K=\frac{\Lambda u^2}2$$
$$P_a=1\frac1{\int^\infty_{t_0}p(t)^{l(t)}dt}$$
http://conwaylife.com/wiki/A_for_all
Aidan F. Pierce
V ⃰_η=c²√(Λη)
K=(Λu²)/2
Pₐ=1−1/(∫^∞_t₀(p(t)ˡ⁽ᵗ⁾)dt)
$$x_1=\eta x$$
$$V^*_\eta=c^2\sqrt{\Lambda\eta}$$
$$K=\frac{\Lambda u^2}2$$
$$P_a=1\frac1{\int^\infty_{t_0}p(t)^{l(t)}dt}$$
http://conwaylife.com/wiki/A_for_all
Aidan F. Pierce
Re: Hacking apgsearch
Uploading some unneccesarily microscopic B2in3_S123a hauls. Is there a way to increase the searching speed for this (I'm searching the heck out of b3s12, my other favourite rule, at the same time)?
Bored of using the Moore neighbourhood for everything? Introducing the Range2 von Neumann isotropic nontotalistic rulespace!
 A for awesome
 Posts: 2037
 Joined: September 13th, 2014, 5:36 pm
 Location: 0x1
 Contact:
Re: Hacking apgsearch
As it is taking much longer than expected for me to develop a new version (which will probably NOT be numbered 0.54+0.22i due to significant differences from the previous version; I am also considering making a C++ version), I've decided to post the one dirty bug fix for 0.21i that I do know how to do (Warning: untested code ahead; read whole post before using, and use at your own risk):
Within the Soup.census() function, change the two lines reading to read and later in the same function change the lines reading to read .
This should cause much fewer pathological objects in rules with glider tagalongs and possible T tagalongs if any aside from the double T exist in any rule. However, this will also result in the misidentification of all close glider and T pairs whenever an errorcorrecting phase is triggered. This won't be a problem once there is an accurate way to separate close spaceship pairs, but for the moment I would recommend copying your version of 0.21i, then implementing this fix and only using the fixed version for rules that have a problem with glider or T tagalongs.
Within the Soup.census() function, change the two lines reading
Code: Select all
# Remove any gliders:
if (glidersexist):
Code: Select all
# Remove any gliders:
if (glidersexist and stepsize < 8):
Code: Select all
# Remove Ts, if they exist:
if self.rg.t:
Code: Select all
# Remove Ts, if they exist:
if self.rg.t and stepsize < 8:
This should cause much fewer pathological objects in rules with glider tagalongs and possible T tagalongs if any aside from the double T exist in any rule. However, this will also result in the misidentification of all close glider and T pairs whenever an errorcorrecting phase is triggered. This won't be a problem once there is an accurate way to separate close spaceship pairs, but for the moment I would recommend copying your version of 0.21i, then implementing this fix and only using the fixed version for rules that have a problem with glider or T tagalongs.
x₁=ηx
V ⃰_η=c²√(Λη)
K=(Λu²)/2
Pₐ=1−1/(∫^∞_t₀(p(t)ˡ⁽ᵗ⁾)dt)
$$x_1=\eta x$$
$$V^*_\eta=c^2\sqrt{\Lambda\eta}$$
$$K=\frac{\Lambda u^2}2$$
$$P_a=1\frac1{\int^\infty_{t_0}p(t)^{l(t)}dt}$$
http://conwaylife.com/wiki/A_for_all
Aidan F. Pierce
V ⃰_η=c²√(Λη)
K=(Λu²)/2
Pₐ=1−1/(∫^∞_t₀(p(t)ˡ⁽ᵗ⁾)dt)
$$x_1=\eta x$$
$$V^*_\eta=c^2\sqrt{\Lambda\eta}$$
$$K=\frac{\Lambda u^2}2$$
$$P_a=1\frac1{\int^\infty_{t_0}p(t)^{l(t)}dt}$$
http://conwaylife.com/wiki/A_for_all
Aidan F. Pierce
Re: Hacking apgsearch
With the new hack, this showed up:
http://catagolue.appspot.com/object/xq4_163xac4/b36ns23
http://catagolue.appspot.com/object/xq4_163xac4/b36ns23
\100\97\110\105
 A for awesome
 Posts: 2037
 Joined: September 13th, 2014, 5:36 pm
 Location: 0x1
 Contact:
Re: Hacking apgsearch
Yes. That's why I said to only use it in rules where glider tagalongs (or T tagalongs) cause too many pathological objects, not for rules with unidentifiable objects arising for any other reason (like replicators). Ultimately, this problem can be solved, but it will take quite a bit of time (and frankly, I'd almost prefer just to rewrite the whole thing).drc wrote:With the new hack, this showed up:
http://catagolue.appspot.com/object/xq4_163xac4/b36ns23
x₁=ηx
V ⃰_η=c²√(Λη)
K=(Λu²)/2
Pₐ=1−1/(∫^∞_t₀(p(t)ˡ⁽ᵗ⁾)dt)
$$x_1=\eta x$$
$$V^*_\eta=c^2\sqrt{\Lambda\eta}$$
$$K=\frac{\Lambda u^2}2$$
$$P_a=1\frac1{\int^\infty_{t_0}p(t)^{l(t)}dt}$$
http://conwaylife.com/wiki/A_for_all
Aidan F. Pierce
V ⃰_η=c²√(Λη)
K=(Λu²)/2
Pₐ=1−1/(∫^∞_t₀(p(t)ˡ⁽ᵗ⁾)dt)
$$x_1=\eta x$$
$$V^*_\eta=c^2\sqrt{\Lambda\eta}$$
$$K=\frac{\Lambda u^2}2$$
$$P_a=1\frac1{\int^\infty_{t_0}p(t)^{l(t)}dt}$$
http://conwaylife.com/wiki/A_for_all
Aidan F. Pierce