source: cpp/frams/genetics/f1/conv_f1.h @ 759

Last change on this file since 759 was 759, checked in by Maciej Komosinski, 6 years ago

f1 and f4 stick properties unified and moved to a separate structure and file (geneprops.cpp/h)

  • Property svn:eol-style set to native
File size: 2.4 KB
Line 
1// This file is a part of Framsticks SDK.  http://www.framsticks.com/
2// Copyright (C) 1999-2018  Maciej Komosinski and Szymon Ulatowski.
3// See LICENSE.txt for details.
4
5#ifndef _CONV_F1_H
6#define _CONV_F1_H
7
8#include <frams/genetics/genoconv.h>
9#include <frams/model/model.h>
10
11/**
12 Official Framsticks f1 -> f0 converter.
13 Source code is published as educational example, but currently not
14 well documented.
15 This converter features building creature using Model class
16 and some advanced tricks (changing model data during construction).
17 Final f0 genotype is generated by model class.
18
19 <b>SDK 1.0.5 NN extension: (for Framsticks 2.0)</B>
20
21 Until now, only the "standard" Framsticks v1 neuron classes could be used.
22 Starting with SDK 1.0.5 you can use all neuron types.
23
24 \b Syntax: [classname , input_or_parameter:weight , input_or_parameter:weight , ...]
25
26 The classname is optional, the default class will be used if it is omitted.
27 Note the comma after \b classname. It must be present if you provide the classname.
28 This is to distinguish between old muscle syntax and new classname specification:
29 - "[|,0:1]" is the neuron of class "|" (bend muscle) with one self connection.
30 \image html nn-ex1.gif
31 - "[|0:1]" describes the default neuron class (N) connected with the bend muscle (|).
32 \image html nn-ex2.gif
33
34
35 Neuron parameters can be specified along with inputs, eg: "[myclass,param1:34,param2:56]"
36 The names beginning with UPPER case character are assumed to be neuron classes.
37 "[myclass,g:1.23,G:2.34]" - the first 'g' is the parameter, the second one is the receptor
38 G ("gyroscope").
39
40 Another example:
41
42 "X[G][-1:2.3][-2:3.4]" - The first neuron is the standalone G receptor. Other neuron can use its output
43 signal by specifying it as regular input ("-1:2.3" and "-2:3.4"). This NN contains 3 neurons.
44 \image html nn-ex3.gif
45
46 "X[G:2.3][G:3.4]" - with the old syntax it is impossible to connect 2 neurons with the same gyroscope,
47 because everything inside brackets is private to the neuron, others can't refer to it.
48 Adding another neuron with "G" input will add another gyroscope object. This NN contains 4 neurons
49 (or 2 neurons if you try it in Framsticks v1).
50 \image html nn-ex4.gif
51
52 */
53class GenoConv_f1 : public GenoConverter
54{
55public:
56        GenoConv_f1()
57        {
58                name = "Recursive encoding";
59                in_format = '1';
60                mapsupport = 1;
61        }
62        SString convert(SString &i, MultiMap *map, bool using_checkpoints);
63        ~GenoConv_f1() {}
64};
65
66
67#endif
Note: See TracBrowser for help on using the repository browser.