Darwinbots Forum
Code center => Suggestions => Topic started by: Old Henk on March 30, 2006, 07:28:02 AM
-
Hi,
Currently snapshots are a great way to get a ... snapshot of the whole simulation. ;) But the files they are exported to are rather large chunks of data and getting (all) the DNA files out as .txt's is difficult, even with the Excel utility (which I find difficult to handle).
Would it be possible to create a small utility to export all dna's conceiled in a snapshot to seperate .txt files?
I imagine a small program-window in which I'd - 1)be able to select my snapshot file, after which I'd click "Export" (or 'Bunnies' , but export seems more appropriate).
- 2)get a popup screen in which I'd be able to select my output directory.
- 3)Then all DNA in the snapshot would be converted into .txt files (one txt file for one bot)
IMHO This's be great and would greatly simplify evolutionary analysis in DB.
What do you guys think?
-
yes ... this would be wonderful.
along the same lines ...
I was wondering if one might be able to extract to a file ...
a list ... say of the 10 oldest bots, or those with most offspring
or greatest number of mutations or some combination ...
along the lines of the King Bot selection ...
but with user controlled selection criteria ...
and not for just one King Bot ...
but a ranking of them, be it 5 or 10 or whatever ...
also user selectable.
just something I've wondered about ...
as henk says ...
to greatly simplify evolutionary analysis.
-
I am still only about half way through my updated excel Snapshot decoder.
Unfortunately I keep getting pulled away from it to do other stuff.
It is actually working to a point but it doesn't yet have all the tools needed to fully analyse a DNA file.
What it can do already is to convert the messy snapshot output into DNA files that are usable directly in DB.
You simply choose the robot number from the output page and compile the DNA from it.
Most of the controls are now handled through a Visual Basic interface so it is a lot easier to work with too.
-
sounds good PY ...
gives you something to do with all your 'spare time'. ;)
-
Proper analysis tools for the simulation are essential. They just take an inordinate amount of time to program, and so tend to take a back burner to issues which might be more favorable or urgent.
-
I see. well it's your time so take all you need, to do the stuff you want :)
Kudos to the programmers for putting their spare time in DB! :clap:
-
Woah, I suddenly feel like some sort of thought police, emotionally manipulative...
Feel free to criticize my time/coding priorities! If enough people would really like some better analysis tools, I'll add it to my list of things to work on.
-
Yes, I would like better graphs/analysis thingies. I also want something done to increase the environment complexity.
-
Proper analysis tools for the simulation are essential. They just take an inordinate amount of time to program, and so tend to take a back burner to issues which might be more favorable or urgent.
to whom?
-
To my understanding of what the community wants, and what I feel like I have to energy to program.
-
I am still only about half way through my updated excel Snapshot decoder.
Unfortunately I keep getting pulled away from it to do other stuff.
It is actually working to a point but it doesn't yet have all the tools needed to fully analyse a DNA file.
What it can do already is to convert the messy snapshot output into DNA files that are usable directly in DB.
You simply choose the robot number from the output page and compile the DNA from it.
Most of the controls are now handled through a Visual Basic interface so it is a lot easier to work with too.
Can I get a copy of this? All I need is to be able to get the DNA of the best veg in my evo sim, and as there are currently 500 of them it's a bit of a pain trying to find it. Cheers!
-
I'm not sure if it ever ended up anywhere. I looked on the FTP but didn't see anything like that.
-
well Findbest is used to find the best robot in the program (it is the crown icon on the toolbar), and once we implement chloroplasts you should not be able to distinguish between vegs and other robots.
However, we can do something like add how good a robot is to the snapshot and then make a program that extracts all robots based on a certain criteria. Such as: Find all robots where FounderName = test.txt and performance = max. It can work like a Database query.
Maybe we can do
max
min
/
to
and then we can do Find all robots where FounderName = test.txt and performance = min to max / 2
this will search all robots whose performance is between the minimum value and half to the maximum value.
stuff like that, what do you think?