Correct way to translate C++ code to Python

Question

I have a C++ code doing the following:

U[0]=rate[0];
map<int,double> Q[N+1]; 
map<int,double> r[N+1]; 
map<int,double> d[N+1]; 
r[0][0]=rate[0];
d[0][0]=1/(1+r[0][0]*dt);
Q[0][0]=1.0;
for(int i=1;i<=N;i++)
{
 for(int j=-i;j<=i;j=j+2)
 { 
 if(j==i)
 {Q[i][i]= 0.5*Q[i-1][i-1]*d[i-1][i-1];}
 if(j==-i)
 {Q[i][-i]=0.5*Q[i-1][-i+1]*d[i-1][-i+1];}
 Q[i][j]=0.5*(Q[i-1][j-1]*d[i-1][j-1]+Q[i-1][j+1]*d[i-1][j+1]);
 }
}

I need to write the equivalent Python version for this as I'm not too comfortable with C++. I have done the following:

Q, U, r, d = [], [], [], []
Q.append([])
Q[0].append(1)
U.append(rate[1])
r.append([])
r[0].append(rate[1])
d.append([])
d[0].append(1/(1+r[0][0]*dt))
for i in range(0, N):
 for j in range(-i, i, 2):
 if j == i:
 Q[i].append(0.5*Q[i-1][i-1]*d[i-1][i-1])
 elif j == -i:
 Q[i].insert(-i, (0.5*Q[i-1][-i+1]*d[i-1][-i+1]))
 Q[i].insert(j, 0.5*(Q[i-1][j-1]*d[i-1][j-1]+Q[i-1][j+1]*d[i-1][j+1]))

However, I'm getting an index out of range error in this line:

Q[i].insert(-i, (0.5*Q[i-1][-i+1]*d[i-1][-i+1]))

How do I correctly translate the C++ code to it's python equivalent?

• 463035818_is_not_an_ai: "I need to write the equivalent Python version for this as I'm not too comfortable with C++" huh? If you have the code already, why do you need to rewrite it? Rewriting c++ to some other language requires much more knowledge about c++ than compiling and running existing code (provided that code has no errors)
• user2201041: It looks like the C++ loops are including the final value, but the Python ones aren't. (range(n) gives [0, n))
• Brent Bradburn: @tobi303: Maybe they have a requirement that it run hundreds of times more slowly. :)
• Brent Bradburn: There are various unexplained differences. Why not make a more direct translation?

Answer 1

You get the index error because your arrays aren't prepopulated, as they would be in C++.

Try defining them like so:

from collections import defaultdict
...
#map<int,double> Q[N+1];
Q = [defaultdict(float) for _ in range(N+1)]
#map<int,double> r[N+1];
r = [defaultdict(float) for _ in range(N+1)]
#map<int,double> d[N+1];
d = [defaultdict(float) for _ in range(N+1)]

---

How do I correctly translate the C++ code to it's python equivalent?

If it were me, and I didn't understand the code I was starting with, I'd do a line-by-line, expression-by-expression translation with absolutely minimal changes.

I would also demand a test case that runs correctly with the C++ code and adapt that to confirm the Python code.

Here is my line-by-line, UNTESTED translation:

from collections import defaultdict
#NameError: name 'N' is not defined
N = 12
#NameError: name 'rate' is not defined
rate = [.03]
#NameError: name 'dt' is not defined
dt = 1.0/12
#U[0]=rate[0];
# unused
#map<int,double> Q[N+1];
Q = [defaultdict(float) for _ in range(N+1)]
#map<int,double> r[N+1];
r = [defaultdict(float) for _ in range(N+1)]
#map<int,double> d[N+1];
d = [defaultdict(float) for _ in range(N+1)]
#r[0][0]=rate[0];
r[0][0] = rate[0]
#d[0][0]=1/(1+r[0][0]*dt);
d[0][0] = 1 / (1 + r[0][0] * dt)
#Q[0][0]=1.0;
Q[0][0] = 1.0
#for(int i=1;i<=N;i++)
#{
for i in range(1, N+1, 1):
# for(int j=-i;j<=i;j=j+2)
# {
 for j in range(-i, i+1, 2):
# if(j==i)
# {Q[i][i]= 0.5*Q[i-1][i-1]*d[i-1][i-1];}
 if j == i:
 Q[i][i] = 0.5 * Q[i-1][i-1] * d[i-1][i-1]
# if(j==-i)
# {Q[i][-i]=0.5*Q[i-1][-i+1]*d[i-1][-i+1];}
 if j == -i:
 Q[i][-i] = 0.5 * Q[i-1][-i+1] * d[i-1][-i+1]
# Q[i][j]=0.5*(Q[i-1][j-1]*d[i-1][j-1]+Q[i-1][j+1]*d[i-1][j+1]);
 Q[i][j] = 0.5 * (Q[i-1][j-1] * d[i-1][j-1] + Q[i-1][j+1] * d[i-1][j+1])
# }
#}