meta


meta command

alea      axis
box      brush
coe
displ
genetic
law
matrix       motif
name       neural network       normal       NP       number
put
rand       roll
sphere        spring
transf
validate       var
See also

meta

Command providing access to other commands.

meta alea

genetic        brush      network

meta alea genetic(id)

       Returns the reproducible probabilities of crossover and mutation of genetic id.

meta alea genetic(id)=croi,muta

     Changes these probabilities.
Options:
croi,muta,muta0,dmuta: forces the value muta0+randf(dmuta) to muta.

meta alea brush(id)

     Returns the bounds of random colors (reproducible) of the brush id.

meta alea brush=r1,g1,b1,r2,g2,b2

     Changes these bounds.
Note: the pixels of the brush id are random colors (reproducible) in the interval [r1,g1,b1] * [r2,g2,b2].

meta alea network(id)

     Returns the coefficients a1, a2, a, a0 of adaptive learning of the network id.

meta alea network(ida1,a2,a,a0

     Changes these coefficinets.
Note: When the error remains above 50% over a images (a1 to a2), the matrix of the network id is reset randomly, which can change a network trouble initialized.

meta axis

axis force
axis validate
axis vol force
axis vol vol

meta axis vol(id)

        Returns the meta axis property of volume id.

meta axis vol(id)=v1,si1,si2,r1, v2,s21,s22,r2,...

        Changes this property.
Note:
The points of volume id will be outside cylinders axis vertices si1,si2 of volume vi and radius ri.

meta axis force

meta axis force vol(id)

        Returns the meta axis force property of volume id.

meta axis force vol(id)=v1,si1,si2,r1,k1, v2,s21,s22,r2,k2,...

        Changes this property.
Note:
The points of volume id will be outside cylinders axis vertices si1,si2 of volume vi and radius ri by forces coefficient ki.

meta axis vol vol

meta axis vol vol(id)

        Returns the meta axis vol property of volume id.

meta axis vol vol(id)=d,v1,v2,...

        Changes this property.
Note:
The points of volume id will be outside cylinders axis CG volumes vi and radius d.

meta axis vol force

meta axis vol force vol(id)

        Returns the meta axis vol force property of volume id.

meta axis vol force vol(id)=d,k,v1,v2,...

        Changes this property.
Note:
The points of volume id will be outside cylinders axis CG volumes vi and radius d by forces coefficient k.

meta axis validate

meta axis validate vol(id)

        Returns the meta axis validate property of volume id.

meta axis vol force vol(id)=v

        Changes this property.
Note:
if v==0 inhibits properties meta axis vol meta axis vol force of volume id.

meta box vol

meta box vol(id)


     Returns the bos of volume num, which is tehe first mta box.
meta(n)box vol(num)
     Returns the last box of volume num.
See also:
generate meta box vol.

meta coe network

meta coe(1)network(id)

       Returns the parameters of the lower bound of the constant learning of the neural network id.

meta coe(2)network(id)

       Returns the parameters of the luppr bound of the constant learning of the neural network id.

meta coe(1)network(id)=c1_1,c1_2,c1,c1_0, dc1_1,dc1_2,dc1,dc1_0, k,k_min,k_max,k,cpt

       Changes these parameters of the of the lower bound c1 (varying between entre c1_1 and c1_2, with decrement dc1).

meta coe(2)network(id)=c2_1,c2_2,c2,c2_0, dc2_1,dc2_2,dc2,dc2_0

       Changes these parameters of the of theupper bound c2 (varying between c2_1 and c2_2, with decrement dc2).

Variation of ethe learnin constant

The learning constant of network id will vary between c1 (genening) to c2 (end).
If k and cpt_k present:
     k=k_min,k_max,k,cpt
     cpt_k=cpt_min,kcpt_max,cpt_k,cpt
c1 and c2 are multiplied by k every cpt_k when moy=moyenne() < 0.5
The command interaction displ network error displays 3 scales allowing to choose c1 between c1_min and c1_max, c2 between c2_min and c2_max, and k between k_min and k_max.

meta displ

meta displ

Returns the character allowing to hide the displays.

meta displ("c")

Change this character.
Notes:
1)in interaction mode, when typing the key c, the menus, the messages and the texts are not displayed (only volumes are displayed).
2) if c is . (dot) this feature is not active any more.

meta genetic

meta alea
meta genetic
meta rand

meta genetic(id)

Returns the average notes of genetic id if generate genetic adjust.

meta law

meta law brush(id)

Returns the transformed output (law) of the motif captured by the neural brush (of type luminance), allowing it to be displayed.

meta matrix network

meta matrix(m)

Returns the determinant of the square matrix m size 1, 4 or 9.

meta matrix network(id)

Returns the learning parameters of neural network id.

meta matrix network(num)=Max_noise,max_noise,d_noise

Changes these parameters.
Notes:
Max_noise: maximum duration resets.
max_noise: current duration resets.
d_noise: noisy coefficient.
During learning, if the error is greater than 0.25, the constants learning are multiplied by 0.75 each max_noise step, and if, moreover, the error remains greater than 0.25 and if its rate of change is less than 0001, the matrix and the learning constant are reseted.

meta motif

meta motif brush(id)

Returns the motif captured by a neural brush (of type luminance), allowing it to be displayed.

meta normal

meta normal vol(id)=v

With commands vol vertex vol(id)=v and allows illuminate the volume id by light vi according to the normal at the corresponding vertex of volume v.

meta NP obj

meta NP obj(id)

     Returns the blocks number of properties of the object identifier id.

meta number

meta number traj(id)

Returns the numbers n1,n2 of images of trajectory id type name.

meta number traj(id)n1,n2

Changes these numbers.

meta name traj(id)

Changes this name.

meta brush

For a neural brush id and behavioural this property stores the return law of the command motif network(idr) where motif network(idr) is the network associated with the brush id.

genetic
brush
network

meta rand genetic(id)

     Returns the non reproducible probabilities of crossover and mutation of genetic id.

meta rand genetic(id)=croi,muta

     Changes these non reproducible probabilities croi of crossover and muta of mutation of genetic id.
Options:
croi,muta,muta0,dmuta: forces value muta0+randf(dmuta) as muta.

meta rand brush(id)

     Returns the bounds of non reproductible random colors of the brush id.

meta rand brush=r1,g1,b1,r2,g2,b2

     Changes thes bounds.
Note: the pixels of the brush id are non reproducible random colors in the interval [r1,g1,b1] * [r2,g2,b2].

meta name

meta name traj(id)

Returns the name of the files read by the trajectory id type name.

meta name traj(id)

Changes this name.

meta network

alea
coe
error
validate
var

meta alea network(id)

       Returns the coefficients (a1,a2,a,a0) of random deformations of neural network id.

meta alea network(id)=a1,a2,a,a0

        Changes these coefficients (0,1000,200,200 default).

meta coe network(id)

       Returns the coefficients (d1,d2,d,d0) of variation of the learning constant of network id.

meta coe network(id)=d1,d2,d,d0

        Changes these coefficients (0,25,0,25 default).

meta coe(n) network(id)

       n=1: returns the variation coefficients (c1_1,c1_2,c1,c1_0) of the learning constant c1 of network id.        n=2: returns the variation coefficients (c2_1,c2_2,c2,c2_0) of the learning constant c2 of network id.

meta coe(n) network(id)=c1_1,c1_2,c1,c1_0

       Changes these coefficiejnts (c1,c2,c,c0=1,10,2,2 default).

meta error network(id)

       Returns the error curve dimensioned 32 by default, to change it to for example
     m=calloc(100,1);meta error network(1)=m;
The command displ network error display this curve.
It is necessary to execute it each training.

meta error network(id)=[1,n]

        Changes the dimension n of the error curve of network id.

meta validate network(id)

       Returns the validation coefficients (val_cpt,val_err,val_stat,val_nb) of network id.

meta validate network(id)=val_cpt,val_err,val_stat,val_nb

       Changes these coefficients.

meta var(n) network(id)

       n=1: returns the variation coefficients (dc1_1,dc1_2,dc1,dc1_0) of the learning constant c1 of network id.        n=2: returns the variation coefficients (dc2_1,dc2_2,dc2,dc2_0) of the learning constant c2 (dc1,dc2,dc,dc0=0,1,.01,.01 default).

meta var(n) network(id)=dc2,dc2,dc,dc2

       Changes these coefficients.

meta put

meta put vol(id1)

Returns property meta put of volume id1.

meta put vol(id)=id2,v1,v2

Changes this property.
Note:
vertices v1 of volume id1 are forced on vertices v2 of volume id2.

meta put

meta put vol(id1)

Returns the property meta put of volume id1.

meta put vol(id)=id2,v1,v2

Changes this property.
Note:
Vertices v1 of volume id1 are forced on sur vertices v2 of volume id2.

meta roll

meta roll light(id)

        Returns the meta roll property of light id.

meta roll light(id)=a,b

        Changes this property.
Note:
roll=a/(foc^b) is automatically calculated when foc is modified (a=3, b=3 are good values).

meta sphere

sphere force
sphere validate
sphere vol

meta sphere force

meta sphere force vol(id)

        Returns the meta sphere force property of volume id.

meta sphere force vol(id)=v1,d1,k1, v2,d2,k2,...

        Changes this property.
Note:
The points of volume id will be outside spheres center CG of volumes, vi, radius di by forces coefficients ki.

meta sphere validate

meta sphere validate vol(id)

        Returns the meta sphere validate property of volume id.

meta sphere validate vol(id)=v

        Changes this property.
Note:
if v==0 inhibits properties meta sphere vol meta sphere force of volume id.

meta sphere vol(id)

        Returns the meta sphere property of volume id.

meta sphere vol(id)=v1,r1, v2,r2,...

        Changes this property.
Note:
The points of volume id will be outside of the spheres centered at the CG of volumes vi and radius ri.

meta spring

meta spring poi vertex(s) vol(id)

        Returns the spring poi vertex property of vertices s of volume id.

meta spring poi vertex(s) vol(id)=raid,visc,x1,y1,z1,x2,y2,z2,...

        Changes this property.
Note:
In animation mode, if yes dynamic is active, vertices s of volume id are related to points (x1,y1,z1), (x2,y2,z2), ... by springs of stiffness raid and viscosity raid visc.

meta spring vol vertex(s)vol(id)

        Returns these coefficients.

meta spring vol vertex(s)vol(id)=r1,g1,f1,s1,r2,g2,f2,s2,...

       Changes these coefficients.
Notes:
1) The vertex s of volume id will experience a force due to the biasing spring through the vertices s1, s2, ... of volumes f1, f2, ... with stiffness r1, r2, .. and viscosities v1, v2, ...
2) if fi=0 then fi=id.

meta transf

meta T0 T vol(id0)

Returns the property meta T0 T of volume id0.

meta(0)T0 T vol(id0)=n,c

Adds such a property.
Note
T0 T are dilx dily dilz rota rotx roty rotz
       does: T0 matrix vol(id0)=c * (T matrix vol(n))
yes meta must be active.
Options:
meta(0)T0 T1 T2 vol(id0)=n1,c1:
       does: T0 matrix vol(id0)=c1 * (T1 matrix vol(n1)) + c1 * (T1 matrix vol(n1))
meta(0)T0 T follow vol(id0)=c:
does recursvly, for each follower of volume id0:
       Ti matrix vol(ni)=c * (T0 matrix vol(id2))
Examples:
meta(0)roty rota rota vol(1)=2,1,3,1;
       The roty matrix of volume 1 will be the sum of rota matrix of volumes 2 and 3.
meta(0)rota rota follow vol(1)=.75;
       The rota matrix of each follower of volume id0 will be .75 * (rota matrix) vol(id0) (usefull, for example, to automaticalt animate the fingers of an hand).

meta transf sin vol(id)

Returns the property meta transf sin of vol id.

meta transf sin vol(id)=c1,c2, w,t

Changes this property.
Note:
If volume id is type particle with property vol vol(id)=id2, the vertices of volume id2 will be transformed by transf(c) (c = sin[c1,c2]) t incremented every vertex.
Options:
c1,c2, w,t, t,dt: t incremented by dt every instance of volume id.

meta validate network

axis
force
network
rota
vol

meta validate axis

meta validate axis limit vol(id)
Returns the property n,x1,y1,z1,x2,y2,z2 of volume id.
meta validate axis limit vol(id)=n,x1,y1,z1,x2,y2,z2
Changes those parameters.
Notes:
meta validate axis limit vol(id)=0: starts the process.
the parameter n returned is the number of images when axis matrix vol(id) is in the gap [x1,y1,z1,x2,y2,z2].
usefull in some applications of genetic algorithmes as a way to notate an axis limited in an interval.
Options:
line: the parameter n returned is linerally interpolated such as:
     0 if ax < ax1 or ax > ax2
     1 if ax < ax1 or ax > ax2
by default thresholde: 0 if (ax < ax1 or ax > ax2) else 1. force

meta validate force

meta validate force axis vol(id)
Returns the property n,an1,an2,x1,y1,z1,x2,y2,z2 of volume id.
meta validate force axis vol(id)=n,an1,an2,x1,y1,z1,x2,y2,z2
Changes those parameters.
Notes:
meta validate force axis limit vol(id)=0: starts the process.
the parameter n returned is the number of images when axis matrix vol(id) is in the gap [x1,y1,z1,x2,y2,z2].
meta validate force rota vol(id)
Returns the property n,an1,an2 of volume id.
meta validate force axis vol(id)=n,an1,an2
Changes those parameters.
Notes:
meta validate force rota limit vol(id)=0: starts the process.
the parameter n returned is the number of images when rota matrix vol(id) is in the gap [an1,an2].

meta validate network(id)

Returns n,e,stat,nb of the learning network id:
     n=number of calls already made.
     e=error
     stat=1 the end of learning.
     nb=maximum calls, changing on scale np in displ network error.

meta validate rota

meta validate rota limit vol(id)
Returns the property cpt,a1,a2 of volume id.
meta validate rota limit vol(id)=n,a1,a2
Changes those parameters.
Notes:
meta validate rota limit vol(id)=0: starts the process.
the parameter n returned is the number of images when rota matrix vol(id) is in the gap [a1,a2].
usefull in some applications of genetic algorithmes as a way to notate an angle limited in an interval..
Options:
line: the parameter n returned is linerally interpolated such as:
     0 if a < a1 or a > a2
     1 if a < a1 or a > a2
by default thresholde: 0 if (a < a1 or a > a2) else 1.

meta validate vol(id)

Returns nb,cpt:
     n=number of images before the volume id will be invisible.
     cpt=meter

meta var

meta var network

meta var network(id)=min_noise,max_noise,cpt_noise,cpt

The matrix is reset when each cpt_noise when moy > 0.5.
The command displ network error displays a scale allowing to choose cpt_noise between min_noise and max_noise.

See also

no meta
yes meta