JakeJing

" Pictures latest
filetype *.bmp,*.jpg,*.jpeg,*.png,*.xpm open -a Preview %f &
fileviewer *.bmp,*.jpg,*.jpeg,*.png,*.xpm
    \ kitten icat --silent --transfer-mode=file --place=%pwx%ph@%pxx%py %c >/dev/tty </dev/tty %N
    \ %pc
    \ kitten icat --clear --silent >/dev/tty </dev/tty %N

" for mac < 15.1.1
" fileviewer *.bmp,*.jpg,*.jpeg,*.png,*.xpm
"    \ kitty +kitten icat --transfer-mode=file --place=%pwx%ph@%pxx%py %c %N

JakeJing

" Pdf
filetype *.pdf open -a Preview %f &
fileviewer *.pdf
    \ tput cup %py %px > /dev/tty && sips -s format jpeg %c --out /tmp/tempfile.jpg > /dev/null&& kitty +kitten icat --transfer-mode=file --place=%pwx%ph@%pxx%py /tmp/tempfile.jpg %N
    \ %pc
    \ kitty +kitten icat --transfer-mode=file --place=%pwx%ph@%pxx%py --clear %N

JakeJing

" Video
filetype *.gif,*.avi,*.mp4,*.wmv,*.dat,*.3gp,*.ogv,*.mkv,*.mpg,*.mpeg,*.vob,
    \*.fl[icv],*.m2v,*.mov,*.webm,*.ts,*.mts,*.m4v,*.r[am],*.qt,*.divx,
    \*.as[fx]
    \ {Open in IINA}
    \ open -a IINA.app,
    \ {Open in QuickTime Player}
    \ open -a QuickTime\ Player.app,
    \ {Open in MPlayerX}
    \ open -a MPlayerX.app,

JakeJing

data {
  int N; // number of observations
  array[N] corr_matrix[2] y; // A n-element arrary of observed correlation matrix [N, 2, 2]
}
parameters {
  real<lower=0> eta;
}
model {
  for (i in 1 : N) {
    y[i,  : ,  : ] ~ lkj_corr(eta); // y[i,] and y[i, ,] should also be fine

JakeJing

my_lkjcorr_fun = "
functions {
  // generate lkj correlation matrix (R)
  matrix my_lkj_corr_rng(int K, real eta) {
    return lkj_corr_rng(K, eta);
  }
  
  // generate cholesky factor L_corr of a correlation matrix R
  matrix my_lkj_corr_chol_rng(int K, real eta){
  return lkj_corr_cholesky_rng(K, eta);

JakeJing

# randomly generate a correlation matrix (R)
my_lkj_corr_rng(K = 2, eta = 1.4)
#           [,1]       [,2]
#[1,]  1.0000000 -0.8743114
#[2,] -0.8743114  1.0000000

# randomly generate a Cholesky factor (L)
(L_chol = my_lkj_corr_chol_rng(K = 2, eta = 1.4))
#           [,1]     [,2]
#[1,] 1.00000000 0.000000

JakeJing

# y: a list of 100 matrices
y = map(1:100, ~ my_lkj_corr_rng(K = 2, eta = 1.4))

# data_lkj_corr: a list of list
data_lkj_corr = list(N = 100, y = y)

JakeJing

fit_lkj_corr = sampling(stan_lkj_corr, data=data_lkj_corr,
                        chains = 4, iter = 2000, cores = 4)
print(fit_lkj_corr)
#       mean se_mean   sd   2.5%    25%    50%    75%  97.5% n_eff Rhat
#eta    1.48    0.01 0.18   1.15   1.36   1.47   1.60   1.87  1263    1
#lp__ -65.07    0.02 0.73 -67.08 -65.21 -64.78 -64.61 -64.56  1664    1

JakeJing

data { 
  int J; // number of predictors or features
  int K; // outcome classes
  int N; 
  array[N] int y;
  matrix[N, J] X;
}
parameters {
  matrix[J, K-1] beta; // number of predictor * number of classes-1
}

JakeJing

N <- 1000
J <- 3 # number of predictors (intercept, slp1, slp2)
X <- cbind(1, rnorm(N, 0, 1), rnorm(N, 0, 1)) # N * J
K <- 3 # num of classes
beta <- rbind(c(1, -0.5, 0),
              c(-2, 4, 0),
              c(3, -1.5, 0)) 
# J * K (features * classes) with the last class as the reference
# [cls1_intercept, cls2_intercept, cls3_intercept]
# [cls1_x1,         cls2_x1,        cls3_x1      ]