problem in convolutional coding

i dont know what is the problem in this code:

% rate 1/2 convolutional encoder 
% define trellis 
trellis = poly2trellis(7,[171 133]); 
% define input and output 
x = randintrlv(1:100,3); 
y = convenc(x,trellis); 

Error using convenc
Expected MSG to be binary.

Error in convenc (line 88)
validateattributes(msg, {'numeric', 'logical'}, ...

% define generator polynomials 
g1 = [1 0 1 0 1 1 1 0]; 
g2 = [1 1 1 0 1 1 0 1]; 
% encode 
y1 = mod(conv(y,g1),2); 
y2 = mod(conv(y,g2),2); 
% interleave 
z = [y1;y2]; 
z = z(:)'; 
% modulate 
M = 4; 
k = log2(M); 
y_tx = bi2de(reshape(z,k,length(z)/k).','left-msb'); 
% plot constellation 
scatterplot(y_tx,1,0,'k*'); 
axis([-5 5 -5 5]); 
% add noise 
EbNo = 10; 
snr = EbNo + 10*log10(k) - 10*log10(log2(M)); 
y_rx = awgn(y_tx,snr,'measured'); 
% demodulate 
z_rx = de2bi(y_rx,k,'left-msb'); 
% de-interleave 
y1_hat = z_rx(1:2:end); 
y2_hat = z_rx(2:2:end); 
% decode 
tblen = 15; 
y_hat = vitdec(y1_hat,trellis,tblen,'trunc','hard'); 
% plot BER 
ber = berawgn(EbNo,'psk',M,'nondiff'); 
semilogy(EbNo,ber,'k*'); 
hold on; 
ber = berawgn(EbNo,'psk',M,'nondiff'); 
semilogy(EbNo,ber,'ko'); 
legend('Theoretical BER','Empirical BER'); 
xlabel('Eb/No (dB)'); 
ylabel('Bit Error Rate');

it shows this error:

Error using convenc

Expected MSG to be binary.

Error in convenc (line 88)

validateattributes(msg, {'numeric', 'logical'}, ...

Error in test2 (line 11)

y = convenc(x,trellis);

[Rik]

x = randintrlv(1:100,3)

x = 1×100

22 69 84 33 96 9 54 15 41 67 34 85 98 87 59 65 23 64 32 3 50 40 78 10 27 4 51 2 97 92

That isn't binary. Apparently the convenc function expects the input to be binary.

Why do you expect this to work?