Question

Here is the code:

import math
with open("test.stl") as file:
    vertices = [map(float, line.split()[1:4])
                for line in file
                if line.lstrip().startswith('vertex')]

    normals = [map(float, line.split()[2:5])
                for line in file
                if line.lstrip().startswith('facet')]

V=len(vertices)
ordering=[]
N=len(normals)

for i in range(0,N):
    p1=vertices[3*i]
    p2=vertices[3*i+1]
    p3=verticies[3*i+2]
    print p1

    x1=p1[0]
    y1=p1[1]
    z1=p1[2]

    x2=p2[0]
    y2=p2[1]
    z2=p2[2]

    x3=p3[0]
    y3=p3[1]
    z3=p3[2]

    a=[x2-x1,y2-y1,z2-z1]
    b=[x3-x1,y3-y1,z3-z1]

    a1=x2-x1
    a2=y2-y1
    a3=z2-z1
    b1=x3-x1
    b2=y3-y1
    b3=z3-z1

    normal=normals[i]

    cross_vector=[a2*b3-a3*b2,a3*b1-a1*b3,a1*b2-a2*b1]

    if cross_vector==normal:
        ordering.append([i,i+1,i+2])
    else:
        ordering.append([i,i+2,i+1])
print ordering
print cross_vector

If I try to add print p1 (or any of the other variables such as cross_vector) inside of the for loop, there aren’t any errors but no output and if I try to print them outside of the for loop it says NameError: name ‘(variable name)’ is not defined. So if none of these variables are being defined, obviously my ordering array prints as [] (blank). How can I change this. Do variables have to be declared before they are defined?

Edit: Here is the error output when the code above is run:

Traceback (most recent call last):
  File "convert.py", line 52, in <module>
    print cross_vector
NameError: name 'cross_vector' is not defined

As explained above this happens with any variable defined in the for loop, I am just using cross_vector as an example.

Answer 1

To solve the problem diagnosed by DSM, use:

import math
import itertools
with open("test.stl") as file:
    i1, i2 = itertools.tee(file)
    vertices = [map(float, line.split()[1:4])
                for line in i1
                if line.lstrip().startswith('vertex')]

    normals = [map(float, line.split()[2:5])
                for line in i2
                if line.lstrip().startswith('facet')]

You might also want to try and drop the list comprehension, and work with iterators throughout, to save on memory for large files.

---

Edit:

At present, you load the entire file into memory, and then create two more full size lists in memory. Instead, you can write it in a way that only reads from the file in memory as required. As an example, we can replace the list comprehensions with generator comprehensions:

import math
import itertools
with open("test.stl") as file:
    i1, i2 = itertools.tee(file)
    vertexIter = (map(float, line.split()[1:4])
                  for line in i1
                  if line.lstrip().startswith('vertex'))

    normalIter = (map(float, line.split()[2:5])
                  for line in i2
                  if line.lstrip().startswith('facet'))

Here, we’ve avoided using any memory at all.

For this to be useful, you need to be able to replace your loop, from:

for i in range(0,N):
    p1=vertices[3*i]
    p2=vertices[3*i+1]
    p3=verticies[3*i+2]
    normal = normals[i]

    # processing

To a single iterator:

for normal, p1, p2, p3 in myMagicIterator:
    # processing

One way I can think of doing this is:

myMagicIterator = itertools.izip(
    normalIter,
    itertools.islice(vertexIter, 0, 3),
    itertools.islice(vertexIter, 1, 3),
    itertools.islice(vertexIter, 2, 3)
)

Which is the iterator equivalent of:

myNormalList = zip(normals, vertices[0::3], vertices[1::3], vertices[2::3])