backpropagation algorithm in matlab - machine-learning

I'm writing a back propagation algorithm in matlab. But I can not get to write a good solution. I read a book Haykin and read some topics in Internet, how make it other people. I understand from door to door this algorithm in theory, but I have a much of error in practice. I have a NaN in my code.
You can see here.
I'm trying classification some points on plate. These are three ellipses, which are placed one inside the other.
I wrote this function. The second layer learn, but first layer dont learn.
function [E, W_1, W_2, B_1, B_2, X_3] = update(W_1, W_2, B_1, B_2, X_1, T, alpha)
V_1 = W_1 * X_1 + B_1;
X_2 = tansig(V_1);
V_2 = W_2 * X_2 + B_2;
X_3 = tansig(V_2);
E = 1 / 2 * sum((T - X_3) .^ 2);
dE = (T - X_3);
for j = 1 : size(X_2, 1)
delta_2_sum = 0;
for i = 1 : size(X_3, 1)
delta_2 = dE(i, 1) * dtansig(1, V_2(i, 1) );
W_2_tmp(i, j) = W_2(i, j) - alpha * delta_2 * X_2(j, 1);
B_2_tmp(i, 1) = B_2(i, 1) - alpha * delta_2;
end;
end;
for k = 1 : size(X_1, 1)
for j = 1 : size(X_2, 1)
delta_2_sum = 0;
for i = 1 : size(X_3, 1)
delta_2 = dE(i, 1) * dtansig(1, V_2(i, 1) );
delta_2_sum = delta_2_sum + W_2(i, j) * delta_2;
end;
delta_1 = delta_2_sum * dtansig(1, V_1(j, 1) );
W_1_tmp(j, k) = W_1(j, k) - alpha * delta_1 * X_1(k, 1);
B_1_tmp(j, 1) = B_1(j, 1) - alpha * delta_1;
end;
end;
if (min(W_1) < -10000 )
X = 1;
end;
B_1 = B_1_tmp;
B_2 = B_2_tmp;
W_1 = W_1_tmp
W_2 = W_2_tmp;
end
I wrote another variant code. And this code don't work. I calculated this code with 1-dimensional vector as input and as output. And I don't have truth result.
What can I do?
I use matlab nntool interface. But my backprop was written my hand.
How I can testing my code?
function [net] = backProp(net, epoch, alpha)
for u = 1 : epoch % Число эпох
for p = 1 : size(net.userdata{1, 1}, 2)
% Учим по всем элементам выборки
[~, ~, ~, De, Df, f] = frontProp(net, p, 1);
for l = size(net.LW, 1) : -1 : 1 % Обходим слои
if (size(net.LW, 1) == l )
delta{l} = De .* Df{l};
else
% size(delta{l + 1})
% size(net.LW{l + 1})
delta{l} = Df{l} .* (delta{l + 1}' * net.LW{l + 1} )';
end;
if (l == 1)
net.IW{l} + alpha * delta{l} * f{l}'
net.IW{l} = net.IW{l} + alpha * delta{l} * f{l}';
else
net.LW{l} + alpha * delta{l} * f{l}'
net.LW{l} = net.LW{l} + alpha * delta{l} * f{l}';
end;
end;
end;
end;
end

Related

Cost does not converge/converges very slowly in the soft-coded version?

I don't understand. When I hardcode my script, it converges excellent, but in the softcode version, given the same structure and learning rate, it converges very slowly and then simply stops converging from some point on.
Here is the softcode version:
def BCE_loss(Y_hat, Y):
m = Y_hat.shape[1]
cost = (-1 / m) * (np.dot(Y, np.log(Y_hat+1e-5).T) + np.dot(1-Y, np.log(1-Y_hat+1e-5).T))
cost = np.squeeze(cost)
return cost
def BCE_loss_backward(Y_hat, Y):
dA_prev = - (np.divide(Y, Y_hat) - np.divide(1-Y, 1-Y_hat))
return dA_prev
def gradient(dZ, A_prev):
dW = np.dot(dZ, A_prev.T) * (1 / A_prev.shape[1])
db = np.sum(dZ, axis=1, keepdims=True) * (1 / A_prev.shape[1])
return dW, db
def update(W, b, dW, db, learning_rate):
W -= np.dot(learning_rate, dW)
b -= np.dot(learning_rate, db)
return W, b
for i in range(epochs+1):
## Forward pass
for l in range(1, L):
if l==L-1:
if out_dim==1:
grads_GD['Z'+str(l)] = linear(params_GD['W'+str(l)], grads_GD['A'+str(l-1)], params_GD['b'+str(l)])
grads_GD['A'+str(l)] = sigmoid(grads_GD['Z'+str(l)])
else:
grads_GD['Z'+str(l)] = linear(params_GD['W'+str(l)], grads_GD['A'+str(l-1)], params_GD['b'+str(l)])
grads_GD['A'+str(l)] = softmax(grads_GD['Z'+str(l)])
else:
grads_GD['Z'+str(l)] = linear(params_GD['W'+str(l)], grads_GD['A'+str(l-1)], params_GD['b'+str(l)])
grads_GD['A'+str(l)] = relu(grads_GD['Z'+str(l)])
## Compute cost
if out_dim==1:
cost_GD = BCE_loss(grads_GD['A'+str(L-1)], Y)
cost_list_GD.append(cost_GD)
else:
cost_GD = CE_loss(grads_GD['A'+str(L-1)], Y)
cost_list_GD.append(cost_GD)
## Print cost
if i % print_num == 0:
print(f"Cost for gradient descent optimizer after epoch {i}: {cost_GD: .4f}")
elif cost_GD < cost_lim or i == epochs:
last_epoch_GD = i
print(f"Cost for gradient descent optimizer after epoch {i}: {cost_GD: .4f}")
break
else:
continue
## Backward pass
if out_dim==1:
grads_GD['dA'+str(L-1)] = BCE_loss_backward(grads_GD['A'+str(L-1)], Y)
grads_GD['dZ'+str(L-1)] = sigmoid_backward(grads_GD['dA'+str(L-1)], grads_GD['Z'+str(L-1)])
else:
grads_GD['dA'+str(L-1)] = CE_loss_backward(grads_GD['A'+str(L-1)], Y)
grads_GD['dZ'+str(L-1)] = softmax_backward(grads_GD['dA'+str(L-1)], grads_GD['Z'+str(L-1)])
grads_GD['dW'+str(L-1)], grads_GD['db'+str(L-1)] = gradient(grads_GD['dZ'+str(L-1)], grads_GD['A'+str(L-2)])
for l in reversed(range(1, L-1)):
grads_GD['dA'+str(l)] = linear_backward(params_GD['W'+str(l+1)], grads_GD['dZ'+str(l+1)])
grads_GD['dZ'+str(l)] = relu_backward(grads_GD['dA'+str(l)], grads_GD['Z'+str(l)])
grads_GD['dW'+str(l)], grads_GD['db'+str(l)] = gradient(grads_GD['dZ'+str(l)], grads_GD['A'+str(l-1)])
## Update parameters
for l in range(1, L):
params_GD['W'+str(l)], params_GD['b'+str(l)] = update(params_GD['W'+str(l)], params_GD['b'+str(l)], grads_GD['dW'+str(l)], grads_GD['db'+str(l)], learning_rate)
and here is the hardcode version:
def cost_function(Y, A4, N, epsilon):
cost = (-1 / N) * np.sum(np.multiply(Y, np.log(A4 + epsilon)) + np.multiply(1 - Y, np.log(1 - A4 + epsilon)))
return cost
for i in range(epochs):
Z1_GD = np.dot(W1_GD, X) + b1_GD
A1_GD = np.maximum(0, Z1_GD)
Z2_GD = np.dot(W2_GD, A1_GD) + b2_GD
A2_GD = np.maximum(0, Z2_GD)
Z3_GD = np.dot(W3_GD, A2_GD) + b3_GD
A3_GD = np.maximum(0, Z3_GD)
Z4_GD = np.dot(W4_GD, A3_GD) + b4_GD
A4_GD = class_layer(Z4_GD)
dZ4_GD = A4_GD - Y
dW4_GD = np.dot(dZ4_GD, A3_GD.T) * (1. / A3_GD.shape[1])
db4_GD = np.sum(dZ4_GD, axis=1, keepdims=True) * (1. / A3_GD.shape[1])
dA3_GD = np.dot(W4_GD.T, dZ4_GD)
dZ3_GD = np.array(dA3_GD, copy=True)
dZ3_GD[Z3_GD <= 0] = 0
dW3_GD = np.dot(dZ3_GD, A2_GD.T) * (1. / A2_GD.shape[1])
db3_GD = np.sum(dZ3_GD, axis=1, keepdims=True) * (1. / A2_GD.shape[1])
dA2_GD = np.dot(W3_GD.T, dZ3_GD)
dZ2_GD = np.array(dA2_GD, copy=True)
dZ2_GD[Z2_GD <= 0] = 0
dW2_GD = np.dot(dZ2_GD, A1_GD.T) * (1. / A1_GD.shape[1])
db2_GD = np.sum(dZ2_GD, axis=1, keepdims=True) * (1. / A1_GD.shape[1])
dA1_GD = np.dot(W2_GD.T, dZ2_GD)
dZ1_GD = np.array(dA1_GD, copy=True)
dZ1_GD[Z1_GD <= 0] = 0
dW1_GD = np.dot(dZ1_GD, X.T) * (1. / X.shape[1])
db1_GD = np.sum(dZ1_GD, axis=1, keepdims=True) * (1. / X.shape[1])
W1_GD = W1_GD - learning_rate * dW1_GD
b1_GD = b1_GD - learning_rate * db1_GD
W2_GD = W2_GD - learning_rate * dW2_GD
b2_GD = b2_GD - learning_rate * db2_GD
W3_GD = W3_GD - learning_rate * dW3_GD
b3_GD = b3_GD - learning_rate * db3_GD
W4_GD = W4_GD - learning_rate * dW4_GD
b4_GD = b4_GD - learning_rate * db4_GD
cost_GD = cost_function(Y, A4_GD, N, epsilon)
cost_GD = np.squeeze(cost_GD)
cost_list_GD.append(cost_GD)
I suppose something went wrong during softcoding.
I solved it myself. Apparently, the "else: continue" line in the print cost section caused the algorithm to do a backward pass only once. After that, it was just looping through the forward pass. Can anyone please explain the reason for such behavior?

How to solve logistic regression using gradient descent in octave?

I am learning Machine Learning course from coursera from Andrews Ng. I have written a code for logistic regression in octave. But, it is not working. Can someone help me?
I have taken the dataset from the following link:
Titanic survivors
Here is my code:
pkg load io;
[An, Tn, Ra, limits] = xlsread("~/ML/ML Practice/dataset/train_and_test2.csv", "Sheet2", "A2:H1000");
# As per CSV file we are reading columns from 1 to 7. 8-th column is Survived, which is what we are going to predict
X = [An(:, [1:7])];
Y = [An(:, 8)];
X = horzcat(ones(size(X,1), 1), X);
# Initializing theta values as zero for all
#theta = zeros(size(X,2),1);
theta = [-3;1;1;-3;1;1;1;1];
learningRate = -0.00021;
#learningRate = -0.00011;
# Step 1: Calculate Hypothesis
function g_z = estimateHypothesis(X, theta)
z = theta' * X';
z = z';
e_z = -1 * power(2.72, z);
denominator = 1.+e_z;
g_z = 1./denominator;
endfunction
# Step 2: Calculate Cost function
function cost = estimateCostFunction(hypothesis, Y)
log_1 = log(hypothesis);
log_2 = log(1.-hypothesis);
y1 = Y;
term_1 = y1.*log_1;
y2 = 1.-Y;
term_2 = y2.*log_2;
cost = term_1 + term_2;
cost = sum(cost);
# no.of.rows
m = size(Y, 1);
cost = -1 * (cost/m);
endfunction
# Step 3: Using gradient descent I am updating theta values
function updatedTheta = updateThetaValues(_X, _Y, _theta, _hypothesis, learningRate)
#s1 = _X * _theta;
#s2 = s1 - _Y;
#s3 = _X' * s2;
# no.of.rows
#m = size(_Y, 1);
#s4 = (learningRate * s3)/m;
#updatedTheta = _theta - s4;
s1 = _hypothesis - _Y;
s2 = s1 .* _X;
s3 = sum(s2);
# no.of.rows
m = size(_Y, 1);
s4 = (learningRate * s3)/m;
updatedTheta = _theta .- s4';
endfunction
costVector = [];
iterationVector = [];
for i = 1:1000
# Step 1
hypothesis = estimateHypothesis(X, theta);
#disp("hypothesis");
#disp(hypothesis);
# Step 2
cost = estimateCostFunction(hypothesis, Y);
costVector = vertcat(costVector, cost);
#disp("Cost");
#disp(cost);
# Step 3 - Updating theta values
theta = updateThetaValues(X, Y, theta, hypothesis, learningRate);
iterationVector = vertcat(iterationVector, i);
endfor
function plotGraph(iterationVector, costVector)
plot(iterationVector, costVector);
ylabel('Cost Function');
xlabel('Iteration');
endfunction
plotGraph(iterationVector, costVector);
This is the graph I am getting when I am plotting against no.of.iterations and cost function.
I am tired by adjusting theta values and learning rate. Can someone help me to solve this problem.
Thanks.
I have done a mathematical error. I should have used either power(2.72, -z) or exp(-z). Instead I have used as -1 * power(2.72, z). Now, I'm getting a proper curve.
Thanks.

Gradient Descent cost function explosion

I am writing this code for linear regression and trying Gradient Descent to minimize the RSS. The cost function seems to explode to infinity within 12 iterations. I know this is not supposed to happen. Maybe, I have used the wrong gradient function for RSS (can be seen in the function "grad()")?
NumberObservations=100
minVal=1
maxVal=20
X = np.random.uniform(minVal,maxVal,(NumberObservations,1))
e = np.random.normal(0, 1, (NumberObservations,1))
Y= 10 + 5*X + e
B = np.array([[0], [0]])
sum_y = sum(Y)
sum_x = sum(X)
sum_xy = sum(np.multiply(X, Y))
sum_x2 = sum(X*X)
alpha = 0.00001
iterations = 15
def cost_fun(X, Y, B):
b0 = B[0]
b1 = B[1]
s = (Y - (b0 + (b1*X)))**2
rss = sum(s)
return rss
def grad(X, Y, B):
print("B = " + str(B))
b0 = B[0]
b1 = B[1]
g0 = -2*(Y - b0 - (b1*X))
g1 = -2*((X*Y) - (b0*X) - (b1*X**2))
grad = np.concatenate((g0, g1), axis = 1)
return grad
def gradient_descent(X, Y, B, alpha, iterations):
cost_history = [0] * iterations
m = len(Y)
x0 = np.array(np.ones(m))
x0 = x0.reshape((100, 1))
X1 = np.concatenate((x0, X), axis = 1)
for iteration in range(iterations):
h = np.dot(X1, B)
h = h.reshape((100, 1))
loss = h - Y
g = grad(X, Y, B)
gradient = (np.dot(g.T, loss) / m)
B = B - alpha * gradient
cost = cost_fun(X, Y, B)
cost_history[iteration] = cost
print("Iteration %d | Cost: %f" % (iteration, cost))
print("-----------------------------------------------------------------------")
return B, cost_history
newB, cost_history = gradient_descent(X, Y, B, alpha, iterations)
# New Values of B
print(newB)
Please help.

SHA512 pure Lua 5.1 adaptation

I was searching for a pure Lua 5.1 adaptation for SHA512 and yielded no results anywhere I went. I found a similar question where someone tried to convert the SHA256 adaptation into SHA512 (except he was using Lua 5.3):
Adaptation of SHA2 512 gives incorrect results
Basically I couldn't use bitwise operators (not implemented in Lua 5.1) so I had to write my own implementations of them.
This is my code:
local MOD = 2^64;
local MODM = MOD-1;
local function memoize(f)
local mt = {}
local t = setmetatable({}, mt)
function mt:__index(k)
local v = f(k)
t[k] = v
return v
end
return t
end
local function make_bitop_uncached(t, m)
local function bitop(a, b)
local res,p = 0,1
while a ~= 0 and b ~= 0 do
local am, bm = a % m, b % m
res = res + t[am][bm] * p
a = (a - am) / m
b = (b - bm) / m
p = p*m
end
res = res + (a + b) * p
return res
end
return bitop
end
local function make_bitop(t)
local op1 = make_bitop_uncached(t,2^1)
local op2 = memoize(function(a) return memoize(function(b) return op1(a, b)
end) end)
return make_bitop_uncached(op2, 2 ^ (t.n or 1))
end
local bxor1 = make_bitop({[0] = {[0] = 0,[1] = 1}, [1] = {[0] = 1, [1] = 0}, n = 4})
local function bxor(a, b, c, ...)
local z = nil
if b then
a = a % MOD
b = b % MOD
z = bxor1(a, b)
if c then z = bxor(z, c, ...) end
return z
elseif a then return a % MOD
else return 0 end
end
local function band(a, b, c, ...)
local z
if b then
a = a % MOD
b = b % MOD
z = ((a + b) - bxor1(a,b)) / 2
if c then z = bit32_band(z, c, ...) end
return z
elseif a then return a % MOD
else return MODM end
end
local function bnot(x) return (-1 - x) % MOD end
local function rshift1(a, disp)
if disp < 0 then return lshift(a,-disp) end
return math.floor(a % 2 ^ 32 / 2 ^ disp)
end
local function rshift(x, disp)
if disp > 31 or disp < -31 then return 0 end
return rshift1(x % MOD, disp)
end
local function lshift(a, disp)
if disp < 0 then return rshift(a,-disp) end
return (a * 2 ^ disp) % 2 ^ 32
end
-- UTILITY FUNCTIONS
--
-- transform a string of bytes in a string of hexadecimal digits
local function str2hexa (s)
local h = string.gsub(s, ".", function(c)
return string.format("%02x", string.byte(c))
end)
return h
end
-- transforms number 'l' into a big-endian sequence of 'n' bytes
--(coded as a string)
local function num2string(l, n)
local s = ""
for i = 1, n do
--most significant byte of l
local remainder = l % 256
s = string.char(remainder) .. s
--remove from l the bits we have already transformed
l = (l-remainder) / 256;
end
return s
end
-- transform the big-endian sequence of eight bytes starting at
-- index 'i' in 's' into a number
local function s264num (s, i)
local n = 0
for i = i, i + 7 do
n = n*256 + string.byte(s, i)
end
return n
end
--
-- MAIN SECTION
--
-- FIRST STEP: INITIALIZE HASH VALUES
--(second 32 bits of the fractional parts of the square roots of the first
9th through 16th primes 23..53)
local HH = {}
local function initH512(H)
H = {0x6a09e667f3bcc908, 0xbb67ae8584caa73b, 0x3c6ef372fe94f82b, 0xa54ff53a5f1d36f1, 0x510e527fade682d1, 0x9b05688c2b3e6c1f, 0x1f83d9abfb41bd6b, 0x5be0cd19137e2179}
return H
end
-- SECOND STEP: INITIALIZE ROUND CONSTANTS
--(first 80 bits of the fractional parts of the cube roots of the first 80 primes 2..409)
local k = {
0x428a2f98d728ae22, 0x7137449123ef65cd, 0xb5c0fbcfec4d3b2f, 0xe9b5dba58189dbbc, 0x3956c25bf348b538,
0x59f111f1b605d019, 0x923f82a4af194f9b, 0xab1c5ed5da6d8118, 0xd807aa98a3030242, 0x12835b0145706fbe,
0x243185be4ee4b28c, 0x550c7dc3d5ffb4e2, 0x72be5d74f27b896f, 0x80deb1fe3b1696b1, 0x9bdc06a725c71235,
0xc19bf174cf692694, 0xe49b69c19ef14ad2, 0xefbe4786384f25e3, 0x0fc19dc68b8cd5b5, 0x240ca1cc77ac9c65,
0x2de92c6f592b0275, 0x4a7484aa6ea6e483, 0x5cb0a9dcbd41fbd4, 0x76f988da831153b5, 0x983e5152ee66dfab,
0xa831c66d2db43210, 0xb00327c898fb213f, 0xbf597fc7beef0ee4, 0xc6e00bf33da88fc2, 0xd5a79147930aa725,
0x06ca6351e003826f, 0x142929670a0e6e70, 0x27b70a8546d22ffc, 0x2e1b21385c26c926, 0x4d2c6dfc5ac42aed,
0x53380d139d95b3df, 0x650a73548baf63de, 0x766a0abb3c77b2a8, 0x81c2c92e47edaee6, 0x92722c851482353b,
0xa2bfe8a14cf10364, 0xa81a664bbc423001, 0xc24b8b70d0f89791, 0xc76c51a30654be30, 0xd192e819d6ef5218,
0xd69906245565a910, 0xf40e35855771202a, 0x106aa07032bbd1b8, 0x19a4c116b8d2d0c8, 0x1e376c085141ab53,
0x2748774cdf8eeb99, 0x34b0bcb5e19b48a8, 0x391c0cb3c5c95a63, 0x4ed8aa4ae3418acb, 0x5b9cca4f7763e373,
0x682e6ff3d6b2b8a3, 0x748f82ee5defb2fc, 0x78a5636f43172f60, 0x84c87814a1f0ab72, 0x8cc702081a6439ec,
0x90befffa23631e28, 0xa4506cebde82bde9, 0xbef9a3f7b2c67915, 0xc67178f2e372532b, 0xca273eceea26619c,
0xd186b8c721c0c207, 0xeada7dd6cde0eb1e, 0xf57d4f7fee6ed178, 0x06f067aa72176fba, 0x0a637dc5a2c898a6,
0x113f9804bef90dae, 0x1b710b35131c471b, 0x28db77f523047d84, 0x32caab7b40c72493, 0x3c9ebe0a15c9bebc,
0x431d67c49c100d4c, 0x4cc5d4becb3e42b6, 0x597f299cfc657e2a, 0x5fcb6fab3ad6faec, 0x6c44198c4a475817
}
-- THIRD STEP: PRE-PROCESSING (padding)
local function preprocess(toProcess, len)
--append a single '1' bit
--append K '0' bits, where K is the minimum number >= 0 such that L + 1 + K = 896mod1024
local extra = - (len + 17) % 128 + 8
len = num2string(8 * len, 8)
toProcess = toProcess .. "\128" .. string.rep("\0", extra) .. len
assert(#toProcess % 128 == 0)
return toProcess
end
local function rrotate(rot, n)
return rshift(rot, n) or (rshift(rot, 64 - n))
end
local function digestblock(msg, i, H)
local w = {}
for j = 1, 16 do w[j] = s264num(msg, i + (j - 1) * 8) end
for j = 17, 80 do
local v = w[j - 15]
local s0 = bxor(rrotate(v, 1), rrotate(v, 8), rshift(v, 7))
v = w[j - 2]
w[j] = w[j - 16] + s0 + w[j - 7] + bxor(rrotate(v, 19), rrotate(v, 61),
rshift(v, 6))
end
local a, b, c, d, e, f, g, h = H[1], H[2], H[3], H[4], H[5], H[6], H[7], H[8]
for i = 1, 80 do
a, b, c, d, e, f, g, h = a , b , c , d , e , f , g , h
local s0 = bxor(rrotate(a, 28), rrotate(a, 34), rrotate(a, 39))
local maj = bxor(band(a, b), band(a, c), band(b, c))
local t2 = s0 + maj
local s1 = bxor(rrotate(e, 14), rrotate(e, 18), rrotate(e, 41))
local ch = bxor (band(e, f), band(bnot(e), g))
local t1 = h + s1 + ch + k[i] + w[i]
h, g, f, e, d, c, b, a = g, f, e, d + t1, c, b, a, t1 + t2
end
H[1] = (H[1] + a)
H[2] = (H[2] + b)
H[3] = (H[3] + c)
H[4] = (H[4] + d)
H[5] = (H[5] + e)
H[6] = (H[6] + f)
H[7] = (H[7] + g)
H[8] = (H[8] + h)
end
local function finalresult512 (H)
-- Produce the final hash value:
return
str2hexa(num2string(H[1], 8)..num2string(H[2], 8)..num2string(H[3], 8)..num2string(H[4], 8)..
num2string(H[5], 8)..num2string(H[6], 8)..num2string(H[7], 8)..num2string(H[8], 8))
end
-- Returns the hash512 for the given string.
local function hash512 (msg)
msg = preprocess(msg, #msg)
local H = initH512(HH)
-- Process the message in successive 1024-bit (128 bytes) chunks:
for i = 1, #msg, 128 do
digestblock(msg, i, H)
end
return finalresult512(H)
end
print( hash512("a") )
At the end, when "a" is hashed, it turns into this:
8c14f3e36400000074d6c495c0000000fd2e4ad8b40000009a78880fb00000002c13f4fdc0000000bf50f67658000000cdf76c796c000000df8163cae8000000
Instead of the actual hash (which is this):
1F40FC92DA241694750979EE6CF582F2D5D7D28E18335DE05ABC54D0560E0F5302860C652BF08D560252AA5E74210546F369FBBBCE8C12CFC7957B2652FE9A75
So my question is, why is it wielding such different results. Is it a problem with the bitwise operator functions? I am stumped.
Here is a working implementation of SHA512 for Lua 5.1
File sha2for51.lua
-- This module contains functions to calculate SHA2 digest.
-- Supported hashes: SHA-224, SHA-256, SHA-384, SHA-512, SHA-512/224, SHA-512/256
-- This is a pure-Lua module, compatible with Lua 5.1
-- It works on Lua 5.1/5.2/5.3/5.4/LuaJIT, but it doesn't use benefits of Lua versions 5.2+
-- Input data may must be provided either as a whole string or as a sequence of substrings (chunk-by-chunk).
-- Result (SHA2 digest) is a string of lowercase hex digits.
--
-- Simplest usage example:
-- local your_hash = require("sha2for51").sha512("your string")
-- See file "sha2for51_test.lua" for more examples.
local unpack, table_concat, byte, char, string_rep, sub, string_format, floor, ceil, min, max =
table.unpack or unpack, table.concat, string.byte, string.char, string.rep, string.sub, string.format, math.floor, math.ceil, math.min, math.max
--------------------------------------------------------------------------------
-- BASIC BITWISE FUNCTIONS
--------------------------------------------------------------------------------
-- 32-bit bitwise functions
local AND, OR, XOR, SHL, SHR, ROL, ROR, HEX
-- Only low 32 bits of function arguments matter, high bits are ignored
-- The result of all functions (except HEX) is an integer (pair of integers) inside range 0..(2^32-1)
function SHL(x, n)
return (x * 2^n) % 4294967296
end
function SHR(x, n)
x = x % 4294967296 / 2^n
return x - x % 1
end
function ROL(x, n)
x = x % 4294967296 * 2^n
local r = x % 4294967296
return r + (x - r) / 4294967296
end
function ROR(x, n)
x = x % 4294967296 / 2^n
local r = x % 1
return r * 4294967296 + (x - r)
end
local AND_of_two_bytes = {} -- look-up table (256*256 entries)
for idx = 0, 65535 do
local x = idx % 256
local y = (idx - x) / 256
local res = 0
local w = 1
while x * y ~= 0 do
local rx = x % 2
local ry = y % 2
res = res + rx * ry * w
x = (x - rx) / 2
y = (y - ry) / 2
w = w * 2
end
AND_of_two_bytes[idx] = res
end
local function and_or_xor(x, y, operation)
-- operation: nil = AND, 1 = OR, 2 = XOR
local x0 = x % 4294967296
local y0 = y % 4294967296
local rx = x0 % 256
local ry = y0 % 256
local res = AND_of_two_bytes[rx + ry * 256]
x = x0 - rx
y = (y0 - ry) / 256
rx = x % 65536
ry = y % 256
res = res + AND_of_two_bytes[rx + ry] * 256
x = (x - rx) / 256
y = (y - ry) / 256
rx = x % 65536 + y % 256
res = res + AND_of_two_bytes[rx] * 65536
res = res + AND_of_two_bytes[(x + y - rx) / 256] * 16777216
if operation then
res = x0 + y0 - operation * res
end
return res
end
function AND(x, y)
return and_or_xor(x, y)
end
function OR(x, y)
return and_or_xor(x, y, 1)
end
function XOR(x, y, z) -- 2 or 3 arguments
if z then
y = and_or_xor(y, z, 2)
end
return and_or_xor(x, y, 2)
end
function HEX(x)
return string_format("%08x", x % 4294967296)
end
-- Arrays of SHA2 "magic numbers"
local sha2_K_lo, sha2_K_hi, sha2_H_lo, sha2_H_hi = {}, {}, {}, {}
local sha2_H_ext256 = {[224] = {}, [256] = sha2_H_hi}
local sha2_H_ext512_lo, sha2_H_ext512_hi = {[384] = {}, [512] = sha2_H_lo}, {[384] = {}, [512] = sha2_H_hi}
local common_W = {} -- a temporary table shared between all calculations
local function sha256_feed_64(H, K, str, W, offs, size)
-- offs >= 0, size >= 0, size is multiple of 64
for pos = offs, size + offs - 1, 64 do
for j = 1, 16 do
pos = pos + 4
local a, b, c, d = byte(str, pos - 3, pos)
W[j] = ((a * 256 + b) * 256 + c) * 256 + d
end
for j = 17, 64 do
local a, b = W[j-15], W[j-2]
W[j] = XOR(ROR(a, 7), ROL(a, 14), SHR(a, 3)) + XOR(ROL(b, 15), ROL(b, 13), SHR(b, 10)) + W[j-7] + W[j-16]
end
local a, b, c, d, e, f, g, h, z = H[1], H[2], H[3], H[4], H[5], H[6], H[7], H[8]
for j = 1, 64 do
z = XOR(ROR(e, 6), ROR(e, 11), ROL(e, 7)) + AND(e, f) + AND(-1-e, g) + h + K[j] + W[j]
h = g
g = f
f = e
e = z + d
d = c
c = b
b = a
a = z + AND(d, c) + AND(a, XOR(d, c)) + XOR(ROR(a, 2), ROR(a, 13), ROL(a, 10))
end
H[1], H[2], H[3], H[4] = (a + H[1]) % 4294967296, (b + H[2]) % 4294967296, (c + H[3]) % 4294967296, (d + H[4]) % 4294967296
H[5], H[6], H[7], H[8] = (e + H[5]) % 4294967296, (f + H[6]) % 4294967296, (g + H[7]) % 4294967296, (h + H[8]) % 4294967296
end
end
local function sha512_feed_128(H_lo, H_hi, K_lo, K_hi, str, W, offs, size)
-- offs >= 0, size >= 0, size is multiple of 128
-- W1_hi, W1_lo, W2_hi, W2_lo, ... Wk_hi = W[2*k-1], Wk_lo = W[2*k]
for pos = offs, size + offs - 1, 128 do
for j = 1, 32 do
pos = pos + 4
local a, b, c, d = byte(str, pos - 3, pos)
W[j] = ((a * 256 + b) * 256 + c) * 256 + d
end
local tmp1, tmp2
for jj = 17 * 2, 80 * 2, 2 do
local a_lo, a_hi, b_lo, b_hi = W[jj-30], W[jj-31], W[jj-4], W[jj-5]
tmp1 = XOR(SHR(a_lo, 1) + SHL(a_hi, 31), SHR(a_lo, 8) + SHL(a_hi, 24), SHR(a_lo, 7) + SHL(a_hi, 25)) + XOR(SHR(b_lo, 19) + SHL(b_hi, 13), SHL(b_lo, 3) + SHR(b_hi, 29), SHR(b_lo, 6) + SHL(b_hi, 26)) + W[jj-14] + W[jj-32]
tmp2 = tmp1 % 4294967296
W[jj-1] = XOR(SHR(a_hi, 1) + SHL(a_lo, 31), SHR(a_hi, 8) + SHL(a_lo, 24), SHR(a_hi, 7)) + XOR(SHR(b_hi, 19) + SHL(b_lo, 13), SHL(b_hi, 3) + SHR(b_lo, 29), SHR(b_hi, 6)) + W[jj-15] + W[jj-33] + (tmp1 - tmp2) / 4294967296
W[jj] = tmp2
end
local a_lo, b_lo, c_lo, d_lo, e_lo, f_lo, g_lo, h_lo, z_lo = H_lo[1], H_lo[2], H_lo[3], H_lo[4], H_lo[5], H_lo[6], H_lo[7], H_lo[8]
local a_hi, b_hi, c_hi, d_hi, e_hi, f_hi, g_hi, h_hi, z_hi = H_hi[1], H_hi[2], H_hi[3], H_hi[4], H_hi[5], H_hi[6], H_hi[7], H_hi[8]
for j = 1, 80 do
local jj = 2 * j
tmp1 = XOR(SHR(e_lo, 14) + SHL(e_hi, 18), SHR(e_lo, 18) + SHL(e_hi, 14), SHL(e_lo, 23) + SHR(e_hi, 9)) + AND(e_lo, f_lo) + AND(-1-e_lo, g_lo) + h_lo + K_lo[j] + W[jj]
z_lo = tmp1 % 4294967296
z_hi = XOR(SHR(e_hi, 14) + SHL(e_lo, 18), SHR(e_hi, 18) + SHL(e_lo, 14), SHL(e_hi, 23) + SHR(e_lo, 9)) + AND(e_hi, f_hi) + AND(-1-e_hi, g_hi) + h_hi + K_hi[j] + W[jj-1] + (tmp1 - z_lo) / 4294967296
h_lo = g_lo
h_hi = g_hi
g_lo = f_lo
g_hi = f_hi
f_lo = e_lo
f_hi = e_hi
tmp1 = z_lo + d_lo
e_lo = tmp1 % 4294967296
e_hi = z_hi + d_hi + (tmp1 - e_lo) / 4294967296
d_lo = c_lo
d_hi = c_hi
c_lo = b_lo
c_hi = b_hi
b_lo = a_lo
b_hi = a_hi
tmp1 = z_lo + AND(d_lo, c_lo) + AND(b_lo, XOR(d_lo, c_lo)) + XOR(SHR(b_lo, 28) + SHL(b_hi, 4), SHL(b_lo, 30) + SHR(b_hi, 2), SHL(b_lo, 25) + SHR(b_hi, 7))
a_lo = tmp1 % 4294967296
a_hi = z_hi + (AND(d_hi, c_hi) + AND(b_hi, XOR(d_hi, c_hi))) + XOR(SHR(b_hi, 28) + SHL(b_lo, 4), SHL(b_hi, 30) + SHR(b_lo, 2), SHL(b_hi, 25) + SHR(b_lo, 7)) + (tmp1 - a_lo) / 4294967296
end
tmp1 = H_lo[1] + a_lo
tmp2 = tmp1 % 4294967296
H_lo[1], H_hi[1] = tmp2, (H_hi[1] + a_hi + (tmp1 - tmp2) / 4294967296) % 4294967296
tmp1 = H_lo[2] + b_lo
tmp2 = tmp1 % 4294967296
H_lo[2], H_hi[2] = tmp2, (H_hi[2] + b_hi + (tmp1 - tmp2) / 4294967296) % 4294967296
tmp1 = H_lo[3] + c_lo
tmp2 = tmp1 % 4294967296
H_lo[3], H_hi[3] = tmp2, (H_hi[3] + c_hi + (tmp1 - tmp2) / 4294967296) % 4294967296
tmp1 = H_lo[4] + d_lo
tmp2 = tmp1 % 4294967296
H_lo[4], H_hi[4] = tmp2, (H_hi[4] + d_hi + (tmp1 - tmp2) / 4294967296) % 4294967296
tmp1 = H_lo[5] + e_lo
tmp2 = tmp1 % 4294967296
H_lo[5], H_hi[5] = tmp2, (H_hi[5] + e_hi + (tmp1 - tmp2) / 4294967296) % 4294967296
tmp1 = H_lo[6] + f_lo
tmp2 = tmp1 % 4294967296
H_lo[6], H_hi[6] = tmp2, (H_hi[6] + f_hi + (tmp1 - tmp2) / 4294967296) % 4294967296
tmp1 = H_lo[7] + g_lo
tmp2 = tmp1 % 4294967296
H_lo[7], H_hi[7] = tmp2, (H_hi[7] + g_hi + (tmp1 - tmp2) / 4294967296) % 4294967296
tmp1 = H_lo[8] + h_lo
tmp2 = tmp1 % 4294967296
H_lo[8], H_hi[8] = tmp2, (H_hi[8] + h_hi + (tmp1 - tmp2) / 4294967296) % 4294967296
end
end
--------------------------------------------------------------------------------
-- CALCULATING THE MAGIC NUMBERS (roots of primes)
--------------------------------------------------------------------------------
do
local function mul(src1, src2, factor, result_length)
-- Long arithmetic multiplication: src1 * src2 * factor
-- src1, src2 - long integers (arrays of digits in base 2^24)
-- factor - short integer
local result = {}
local carry = 0
local value = 0.0
local weight = 1.0
for j = 1, result_length do
local prod = 0
for k = max(1, j + 1 - #src2), min(j, #src1) do
prod = prod + src1[k] * src2[j + 1 - k]
end
carry = carry + prod * factor
local digit = carry % 16777216
result[j] = digit
carry = floor(carry / 16777216)
value = value + digit * weight
weight = weight * 2^24
end
return
result, -- long integer
value -- and its floating point approximation
end
local idx, step, p, one = 0, {4, 1, 2, -2, 2}, 4, {1}
local sqrt_hi, sqrt_lo, idx_disp = sha2_H_hi, sha2_H_lo, 0
repeat
p = p + step[p % 6]
local d = 1
repeat
d = d + step[d % 6]
if d * d > p then
idx = idx + 1
local root = p^(1/3)
local R = mul({floor(root * 2^40)}, one, 1, 2)
local _, delta = mul(R, mul(R, R, 1, 4), -1, 4)
local hi = R[2] % 65536 * 65536 + floor(R[1] / 256)
local lo = R[1] % 256 * 16777216 + floor(delta * (2^-56 / 3) * root / p)
sha2_K_hi[idx], sha2_K_lo[idx] = hi, lo
if idx < 17 then
root = p^(1/2)
R = mul({floor(root * 2^40)}, one, 1, 2)
_, delta = mul(R, R, -1, 2)
hi = R[2] % 65536 * 65536 + floor(R[1] / 256)
lo = R[1] % 256 * 16777216 + floor(delta * 2^-17 / root)
sha2_H_ext256[224][idx + idx_disp] = lo
sqrt_hi[idx + idx_disp], sqrt_lo[idx + idx_disp] = hi, lo
if idx == 8 then
sqrt_hi, sqrt_lo, idx_disp = sha2_H_ext512_hi[384], sha2_H_ext512_lo[384], -8
end
end
break
end
until p % d == 0
until idx > 79
end
-- Calculating IV for SHA512/224 and SHA512/256
for width = 224, 256, 32 do
local H_lo, H_hi = {}, {}
for j = 1, 8 do
H_lo[j] = XOR(sha2_H_lo[j], 0xa5a5a5a5)
H_hi[j] = XOR(sha2_H_hi[j], 0xa5a5a5a5)
end
sha512_feed_128(H_lo, H_hi, sha2_K_lo, sha2_K_hi, "SHA-512/"..tonumber(width).."\128"..string_rep("\0", 115).."\88", common_W, 0, 128)
sha2_H_ext512_lo[width] = H_lo
sha2_H_ext512_hi[width] = H_hi
end
--------------------------------------------------------------------------------
-- FINAL FUNCTIONS
--------------------------------------------------------------------------------
local function sha256ext(width, text)
-- Create an instance (private objects for current calculation)
local H, length, tail = {unpack(sha2_H_ext256[width])}, 0, ""
local function partial(text_part)
if text_part then
if tail then
length = length + #text_part
local offs = 0
if tail ~= "" and #tail + #text_part >= 64 then
offs = 64 - #tail
sha256_feed_64(H, sha2_K_hi, tail..sub(text_part, 1, offs), common_W, 0, 64)
tail = ""
end
local size = #text_part - offs
local size_tail = size % 64
sha256_feed_64(H, sha2_K_hi, text_part, common_W, offs, size - size_tail)
tail = tail..sub(text_part, #text_part + 1 - size_tail)
return partial
else
error("Adding more chunks is not allowed after asking for final result", 2)
end
else
if tail then
local final_blocks = {tail, "\128", string_rep("\0", (-9 - length) % 64 + 1)}
tail = nil
-- Assuming user data length is shorter than 2^53 bytes
-- Anyway, it looks very unrealistic that one would spend enough time to process a 2^53 bytes of data by using this Lua script :-)
-- 2^53 bytes = 2^56 bits, so "bit-counter" fits in 7 bytes
length = length * (8 / 256^7) -- convert "byte-counter" to "bit-counter" and move floating point to the left
for j = 4, 10 do
length = length % 1 * 256
final_blocks[j] = char(floor(length))
end
final_blocks = table_concat(final_blocks)
sha256_feed_64(H, sha2_K_hi, final_blocks, common_W, 0, #final_blocks)
local max_reg = width / 32
for j = 1, max_reg do
H[j] = HEX(H[j])
end
H = table_concat(H, "", 1, max_reg)
end
return H
end
end
if text then
-- Actually perform calculations and return the SHA256 digest of a message
return partial(text)()
else
-- Return function for partial chunk loading
-- User should feed every chunks of input data as single argument to this function and receive SHA256 digest by invoking this function without an argument
return partial
end
end
local function sha512ext(width, text)
-- Create an instance (private objects for current calculation)
local length, tail, H_lo, H_hi = 0, "", {unpack(sha2_H_ext512_lo[width])}, {unpack(sha2_H_ext512_hi[width])}
local function partial(text_part)
if text_part then
if tail then
length = length + #text_part
local offs = 0
if tail ~= "" and #tail + #text_part >= 128 then
offs = 128 - #tail
sha512_feed_128(H_lo, H_hi, sha2_K_lo, sha2_K_hi, tail..sub(text_part, 1, offs), common_W, 0, 128)
tail = ""
end
local size = #text_part - offs
local size_tail = size % 128
sha512_feed_128(H_lo, H_hi, sha2_K_lo, sha2_K_hi, text_part, common_W, offs, size - size_tail)
tail = tail..sub(text_part, #text_part + 1 - size_tail)
return partial
else
error("Adding more chunks is not allowed after asking for final result", 2)
end
else
if tail then
local final_blocks = {tail, "\128", string_rep("\0", (-17-length) % 128 + 9)}
tail = nil
-- Assuming user data length is shorter than 2^53 bytes
-- 2^53 bytes = 2^56 bits, so "bit-counter" fits in 7 bytes
length = length * (8 / 256^7) -- convert "byte-counter" to "bit-counter" and move floating point to the left
for j = 4, 10 do
length = length % 1 * 256
final_blocks[j] = char(floor(length))
end
final_blocks = table_concat(final_blocks)
sha512_feed_128(H_lo, H_hi, sha2_K_lo, sha2_K_hi, final_blocks, common_W, 0, #final_blocks)
local max_reg = ceil(width / 64)
for j = 1, max_reg do
H_lo[j] = HEX(H_hi[j])..HEX(H_lo[j])
end
H_hi = nil
H_lo = table_concat(H_lo, "", 1, max_reg):sub(1, width / 4)
end
return H_lo
end
end
if text then
-- Actually perform calculations and return the SHA256 digest of a message
return partial(text)()
else
-- Return function for partial chunk loading
-- User should feed every chunks of input data as single argument to this function and receive SHA256 digest by invoking this function without an argument
return partial
end
end
local sha2for51 = {
sha224 = function (text) return sha256ext(224, text) end, -- SHA-224
sha256 = function (text) return sha256ext(256, text) end, -- SHA-256
sha384 = function (text) return sha512ext(384, text) end, -- SHA-384
sha512 = function (text) return sha512ext(512, text) end, -- SHA-512
sha512_224 = function (text) return sha512ext(224, text) end, -- SHA-512/224
sha512_256 = function (text) return sha512ext(256, text) end, -- SHA-512/256
}
return sha2for51
File sha2for51_test.lua
--------------------------------------------------------------------------------
-- TESTS
--------------------------------------------------------------------------------
local sha2 = require"sha2for51"
local function test_sha256()
local sha256 = sha2.sha256
-- some test strings
assert(sha256("The quick brown fox jumps over the lazy dog") == "d7a8fbb307d7809469ca9abcb0082e4f8d5651e46d3cdb762d02d0bf37c9e592")
assert(sha256("The quick brown fox jumps over the lazy cog") == "e4c4d8f3bf76b692de791a173e05321150f7a345b46484fe427f6acc7ecc81be")
assert(sha256("abc") == "ba7816bf8f01cfea414140de5dae2223b00361a396177a9cb410ff61f20015ad")
assert(sha256("123456") == "8d969eef6ecad3c29a3a629280e686cf0c3f5d5a86aff3ca12020c923adc6c92")
assert(sha256("abcdbcdecdefdefgefghfghighijhijkijkljklmklmnlmnomnopnopq") == "248d6a61d20638b8e5c026930c3e6039a33ce45964ff2167f6ecedd419db06c1")
assert(sha256("abcdefghbcdefghicdefghijdefghijkefghijklfghijklmghijklmnhijklmnoijklmnopjklmnopqklmnopqrlmnopqrsmnopqrstnopqrstu") == "cf5b16a778af8380036ce59e7b0492370b249b11e8f07a51afac45037afee9d1")
-- chunk-by-chunk loading: sha256("string") == sha256()("st")("ri")("ng")()
local append_next_chunk = sha256() -- create a private closure for calculating digest of single string
append_next_chunk("The quick brown fox")
append_next_chunk(" jumps ")
append_next_chunk("") -- chunk may be empty string
append_next_chunk("over the lazy dog")
assert(append_next_chunk() == "d7a8fbb307d7809469ca9abcb0082e4f8d5651e46d3cdb762d02d0bf37c9e592") -- asking for final result (invocation without an argument)
assert(append_next_chunk() == "d7a8fbb307d7809469ca9abcb0082e4f8d5651e46d3cdb762d02d0bf37c9e592") -- you can ask the same result multiple times if needed
-- append_next_chunk("more text") will fail here: no more chunks are allowed after receiving the result, the closure is useless now, let it be GC-ed
assert(not pcall(append_next_chunk, "more text"))
-- one-liner is possible due to "append_next_chunk(chunk)" returns the function "append_next_chunk"
assert(sha256()("The quick brown fox")(" jumps ")("")("over the lazy dog")() == "d7a8fbb307d7809469ca9abcb0082e4f8d5651e46d3cdb762d02d0bf37c9e592")
-- empty string
assert(sha256("") == "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855")
assert(sha256()() == "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855")
-- computations of different strings don't interfere with each other
local chunk_for_digits = sha256()
chunk_for_digits("123")
local chunk_for_fox = sha256()
chunk_for_fox("The quick brown fox jumps ")
chunk_for_digits("45")
chunk_for_fox("over the lazy dog")
chunk_for_digits("6")
assert(chunk_for_digits() == "8d969eef6ecad3c29a3a629280e686cf0c3f5d5a86aff3ca12020c923adc6c92")
assert(chunk_for_fox() == "d7a8fbb307d7809469ca9abcb0082e4f8d5651e46d3cdb762d02d0bf37c9e592")
-- "00...0\n"
for i, dgst in pairs{ -- from 50 to 70 zeroes
[50] = "9660acb8046abf46cf27280e61abd174ebac98ad6855e093772b78df85523129",
[51] = "31e1c552b357ace9bcb924691799a3c0d3aa10d8b428d9de28a278e3c79ecb7b",
[52] = "0be5c4bcb6f47e30c13515594dbef4faa3a6485af67c177179fee8b33cd4f2a0",
[53] = "d368c7f6038c1743bdbfe6a9c3a72d4e6916aa219ed8d559766c9e8f9845f3b8",
[54] = "7080a4aa6ff030ae152fe610a62ee29464f92afeb176474551a69d35aab154a0",
[55] = "149c1cda81fa9359c0c2a5e405ca972986f1d53e05f6282871dd1581046b3f44",
[56] = "eb2d4d41948ce546c8adff07ee97342070c5b89789f616a33efe52c7d3ec73d4",
[57] = "c831db596ccbbf248023461b1c05d3ae084bcc79bcb2626c5ec179fb34371f2a",
[58] = "1345b8a930737b1069bbf9b891ce095850f6cdba6e25874ea526a2ccb611fe46",
[59] = "380ad21e466885fae080ceeada75ac04944687e626e161c0b24e91af3eec2def",
[60] = "b9ab06fa30ef8531c5eee11651aa86f8279a245e0a3c29bf6228c59475cc610a",
[61] = "bcc187de6605d9e11a0cc6edf02b67fb651fe1779ec59438788093d8e376c07c",
[62] = "ae0b3681157b83b34de8591d2453915e40c3105ae79434e241d82d4035218e01",
[63] = "68a27b4735f6806fb5983c1805a23797aa93ea06e0ebcb6daada2ea1ab5a05af",
[64] = "827d096d92f3deeaa0e8070d79f45beb176768e57a958a1cd325f5f4b754b048",
[65] = "6c7bd8ec0fe9b4e05a2d27dd5e41a8687a9716a2e8926bdfa141266b12942ec1",
[66] = "2f4b4c41017a2ddd1cc8cd75478a82e9452e445d4242f09782535376d6f4ba50",
[67] = "b777b86e005807a446ead00986fcbf3bdd6c022524deabf017eeb3f0c30b6eed",
[68] = "777da331f60c793f582e4ca33223778218ddfd241981f15be5886171fb8301b5",
[69] = "06ed0c4cbf7d2b38de5f01eab2d2cd552d9cb87f97b714b96bb7a9d1b6117c6d",
[70] = "e82223344d5f3c024514cfbe6d478b5df98bb878f34d7a07e7b064fa7fa91946"
} do
assert(sha256(("0"):rep(i).."\n") == dgst)
end
-- "aa...a"
assert(sha256(("a"):rep(55)) == "9f4390f8d30c2dd92ec9f095b65e2b9ae9b0a925a5258e241c9f1e910f734318")
assert(sha256(("a"):rep(56)) == "b35439a4ac6f0948b6d6f9e3c6af0f5f590ce20f1bde7090ef7970686ec6738a")
-- "aa...a\n" in chunk-by-chunk mode
local next_chunk = sha256()
for i = 1, 65 do
next_chunk("a")
end
next_chunk("\n")
assert(next_chunk() == "574883a9977284a46845620eaa55c3fa8209eaa3ebffe44774b6eb2dba2cb325")
local function split_and_calculate_sha256(s, len) -- split string s in chunks of length len
local next_chunk = sha256()
for idx = 1, #s, len do
next_chunk(s:sub(idx, idx + len - 1))
end
return next_chunk()
end
-- "00...0\n00...0\n...00...0\n" (80 lines of 80 zeroes each) in chunk-by-chunk mode with different chunk lengths
local s = (("0"):rep(80).."\n"):rep(80)
assert(split_and_calculate_sha256(s, 1) == "736c7a8b17e2cfd44a3267a844db1a8a3e8988d739e3e95b8dd32678fb599139")
assert(split_and_calculate_sha256(s, 2) == "736c7a8b17e2cfd44a3267a844db1a8a3e8988d739e3e95b8dd32678fb599139")
assert(split_and_calculate_sha256(s, 7) == "736c7a8b17e2cfd44a3267a844db1a8a3e8988d739e3e95b8dd32678fb599139")
assert(split_and_calculate_sha256(s, 70) == "736c7a8b17e2cfd44a3267a844db1a8a3e8988d739e3e95b8dd32678fb599139")
end
local function test_sha512()
local sha512 = sha2.sha512
assert(sha512("abc") == "ddaf35a193617abacc417349ae20413112e6fa4e89a97ea20a9eeee64b55d39a2192992a274fc1a836ba3c23a3feebbd454d4423643ce80e2a9ac94fa54ca49f")
assert(sha512("abcdefghbcdefghicdefghijdefghijkefghijklfghijklmghijklmnhijklmnoijklmnopjklmnopqklmnopqrlmnopqrsmnopqrstnopqrstu") ==
"8e959b75dae313da8cf4f72814fc143f8f7779c6eb9f7fa17299aeadb6889018501d289e4900f7e4331b99dec4b5433ac7d329eeb6dd26545e96e55b874be909")
-- "aa...a"
for i, dgst in pairs{ -- from 109 to 116 letters "a"
[109] = "0cda6b04d9466bb7f3995c16732e1347f29c23a64fe0b085fadba0995644cc5aa71587423c274c10e09518310c5f866cfaceb229fabb574219f12182eb114182",
[110] = "c825949632e509824543f7eaf159fb6041722fce3c1cdcbb613b3d37ff107c519417baac32f8e74fe29d7f4823bf6886956603dca5354a6ed6e4a542e06b7d28",
[111] = "fa9121c7b32b9e01733d034cfc78cbf67f926c7ed83e82200ef86818196921760b4beff48404df811b953828274461673c68d04e297b0eb7b2b4d60fc6b566a2",
[112] = "c01d080efd492776a1c43bd23dd99d0a2e626d481e16782e75d54c2503b5dc32bd05f0f1ba33e568b88fd2d970929b719ecbb152f58f130a407c8830604b70ca",
[113] = "55ddd8ac210a6e18ba1ee055af84c966e0dbff091c43580ae1be703bdb85da31acf6948cf5bd90c55a20e5450f22fb89bd8d0085e39f85a86cc46abbca75e24d",
[114] = "5e9eb0e4b270d086e77eeaf3ce8b1cfc615031b8c463dc34f5c139786f274f22accb4d89e8f40d1a0c2acc84c4dc0f2bab390a9d9495493bd617ed004271bb64",
[115] = "eaa30f93760743ac7d0a6cb8ed5ef3b30c59097bc44d0ec337344301deba9fb92b20c488d55de415f6aaed0df4925b42894b81d2e1cde89d91ec7f6cc67262b4",
[116] = "a8bff469314a1ce0c990bb3fd539d92accb6249cc674b559bc9d3898b7a126fee597197fa42c971443470053c7d7f54b09371a59b0f7af87b1917c5347e8f8e0",
} do
assert(sha512(("a"):rep(i)) == dgst)
end
end
local function all_tests_sha2()
test_sha256()
assert(sha2.sha224"abc" == "23097d223405d8228642a477bda255b32aadbce4bda0b3f7e36c9da7")
assert(sha2.sha224"abcdbcdecdefdefgefghfghighijhijkijkljklmklmnlmnomnopnopq" == "75388b16512776cc5dba5da1fd890150b0c6455cb4f58b1952522525")
test_sha512()
assert(sha2.sha384"abc" == "cb00753f45a35e8bb5a03d699ac65007272c32ab0eded1631a8b605a43ff5bed8086072ba1e7cc2358baeca134c825a7")
assert(sha2.sha384"abcdefghbcdefghicdefghijdefghijkefghijklfghijklmghijklmnhijklmnoijklmnopjklmnopqklmnopqrlmnopqrsmnopqrstnopqrstu" == "09330c33f71147e83d192fc782cd1b4753111b173b3b05d22fa08086e3b0f712fcc7c71a557e2db966c3e9fa91746039")
assert(sha2.sha512_224"abc" == "4634270f707b6a54daae7530460842e20e37ed265ceee9a43e8924aa")
assert(sha2.sha512_224"abcdefghbcdefghicdefghijdefghijkefghijklfghijklmghijklmnhijklmnoijklmnopjklmnopqklmnopqrlmnopqrsmnopqrstnopqrstu" == "23fec5bb94d60b23308192640b0c453335d664734fe40e7268674af9")
assert(sha2.sha512_256"abc" == "53048e2681941ef99b2e29b76b4c7dabe4c2d0c634fc6d46e0e2f13107e7af23")
assert(sha2.sha512_256"abcdefghbcdefghicdefghijdefghijkefghijklfghijklmghijklmnhijklmnoijklmnopjklmnopqklmnopqrlmnopqrsmnopqrstnopqrstu" == "3928e184fb8690f840da3988121d31be65cb9d3ef83ee6146feac861e19b563a")
print"All tests passed"
end
all_tests_sha2()
local function benchmark()
print("Benchmarking (calculating SHA512 of 1MByte string of letters 'a')...")
local time_intervals = {}
local length = 2^20
local part = ("a"):rep(2^12)
local N = length/#part
local result
local k = 2
for j = 1, 2*k-1 do
local clk0 = os.clock()
local x = sha2.sha512()
for j = 1, N do
x(part)
end
result = x()
time_intervals[j] = os.clock() - clk0
end
--print("Result = "..result)
-- get median time
table.sort(time_intervals)
print('CPU seconds:', time_intervals[k])
end
benchmark() -- about 15 seconds per megabyte

Spline with multiple durations

I'm trying to make a spline that has multiple durations. Basically, I want to create a spline generated through user supplied key frames. There may be multiple key frames with different durations. So I ask, how would I make a that travels at different speeds through different key frames. Let's say I want there to be 1 second between key frame A and B, but 5 seconds between B and C. Similar to the way animation editors like Source Filmmaker, or Autodesk Maya do it. Every time I've seen someone make a spline, it always has one, constant speed. Always X seconds between key frames, but this is not how animation editors work, as they have different speeds, that's what I want.
Just to note, I have tried. I made the duration change once it gets to a different key frame, but that just instantly slows it down, like some slow motion movie effect, which is not what I'm looking for. Do I just gradually transition the speed to fit that of the next key frame? Is there just some equation?
function smooth( points, steps ) --points is an array, steps is how many frames inbetween spline points
if #points < 3 then
return points
end
local steps = steps or 5
local spline = {}
local count = #points - 1
local p0, p1, p2, p3, x, y, z
for i = 1, count do
if i == 1 then
p0, p1, p2, p3 = points[i], points[i], points[i + 1], points[i + 2]
elseif i == count then
p0, p1, p2, p3 = points[#points - 2], points[#points - 1], points[#points], points[#points]
else
p0, p1, p2, p3 = points[i - 1], points[i], points[i + 1], points[i + 2]
end
for t = 0, 1, 1 / steps do
-- Main spline equation
x = 1 * ( ( 2 * p1.x ) + ( p2.x - p0.x ) * t + ( 2 * p0.x - 5 * p1.x + 4 * p2.x - p3.x ) * t * t + ( 3 * p1.x - p0.x - 3 * p2.x + p3.x ) * t * t * t )
y = 1 * ( ( 2 * p1.y ) + ( p2.y - p0.y ) * t + ( 2 * p0.y - 5 * p1.y + 4 * p2.y - p3.y ) * t * t + ( 3 * p1.y - p0.y - 3 * p2.y + p3.y ) * t * t * t )
z = 1 * ( ( 2 * p1.z ) + ( p2.z - p0.z ) * t + ( 2 * p0.z - 5 * p1.z + 4 * p2.z - p3.z ) * t * t + ( 3 * p1.z - p0.z - 3 * p2.z + p3.z ) * t * t * t )
if not(#spline > 0 and spline[#spline].x == x and spline[#spline].y == y and spline[#spline].z == z) then
table.insert( spline , { x = x , y = y, z = z } )
end
end
end
return spline
end
Straightforward approach was used:
local zero_vector = {0, 0, 0}
local function get_slope(is_edge, left, right)
if is_edge then
return zero_vector
else
local t = right.time - left.time
assert(t > 0, "Non-increasing time sequence")
return {(right[1] - left[1])/t,
(right[2] - left[2])/t,
(right[3] - left[3])/t}
end
end
function smooth(checkpoints, frames_per_second)
frames_per_second = frames_per_second or 5
if #checkpoints < 2 then
return checkpoints
end
-- Prepare formulas for each segment of spline
local formulas = {}
for segment = 1, #checkpoints - 1 do
local left = checkpoints[segment]
local right = checkpoints[segment + 1]
local t = right.time - left.time
assert(t > 0, "Non-increasing time sequence")
local left_deriv = get_slope(segment == 1,
checkpoints[segment - 1], right)
local right_deriv = get_slope(segment == #checkpoints - 1,
left, checkpoints[segment + 2])
formulas[segment] = {}
for j = 1, 3 do
local d = left[j]
local c = left_deriv[j]
local a = (right[j] - d - c*t) / (t*t)
local b = 3*a + (c - right_deriv[j])/t
formulas[segment][j] = {(a - b)/t, b, c, d}
end
end
-- Calculate spline points
local total_seconds = checkpoints[#checkpoints].time - checkpoints[1].time
local segment = 1
local spline = {}
for frame_no = 0, total_seconds * frames_per_second do
local t = checkpoints[1].time + frame_no / frames_per_second
local point = {time = t}
while segment < #formulas and t > checkpoints[segment + 1].time do
segment = segment + 1
end
t = t - checkpoints[segment].time
for j = 1, 3 do
local c = formulas[segment][j]
point[j] = ((c[1]*t + c[2])*t + c[3])*t + c[4]
end
table.insert(spline, point)
end
return spline
end
Usage example:
-- x y z "timestamp in seconds"
local checkpoint_A = {11, 12, 13, time = 0}
local checkpoint_B = {21, 15, 18, time = 1} -- 1 second between A and B
local checkpoint_C = {13, 24, 20, time = 6} -- 5 seconds between B and C
local checkpoints = {checkpoint_A, checkpoint_B, checkpoint_C}
-- total duration is 6 seconds, 10 frames per second, 61 points returned
local array_of_61_points = smooth(checkpoints, 10)
for _, point in ipairs(array_of_61_points) do
print(string.format("time = %.1f, x = %.3f, y = %.3f, z = %.3f",
point.time, point[1], point[2], point[3]))
end

Resources