Fix bug with gauss command dumping

pykat/node_network.py

@@ -739,9 +739,9 @@ class Node(object):
                 raise pkex.BasePyKatException("Node {0} is not connected to {1}".format(self.name, self.__q_comp.name))
     
             if self.__q_x == self.__q_y:
-                rtn.append("gauss* {name} {comp} {node} {w0:.15g} {z:.15g}".format(name=name, node=self.name, comp=self.__q_comp.name, w0=self.__q_x.z, z=self.__q_x.zr))
+                rtn.append("gauss* {name} {comp} {node} {w0:.15g} {z:.15g}".format(name=name, node=self.name, comp=self.__q_comp.name, w0=self.__q_x.w0, z=self.__q_x.z))
             else:
-                rtn.append("gauss* {name} {comp} {node} {w0x:.15g} {zx:.15g} {w0y:.15g} {zy:.15g}".format(name=name, node=self.name, comp=self.__q_comp.name, w0x=self.__q_x.z, zx=self.__q_x.zr, w0y=self.__q_y.z, zy=self.__q_y.zr))
+                rtn.append("gauss* {name} {comp} {node} {w0x:.15g} {zx:.15g} {w0y:.15g} {zy:.15g}".format(name=name, node=self.name, comp=self.__q_comp.name, w0x=self.__q_x.w0, zx=self.__q_x.z, w0y=self.__q_y.w0, zy=self.__q_y.z))
             
         return rtn
         

[Daniel Brown]

I think it was correct before?

 gauss name component node w0 z [wy0 zy]          - set q parameter
 gauss* name component node q [qy] (q as 'z z_R') - set q parameter
 gauss** name component node w(z) Rc [wy(z) Rcy]  - set q parameter

[Sean Leavey]

Ah, so the gauss* command has different syntax. But it then looks like qy is not being set correctly. In the example above, the qx and qy parameters should be the same (gauss gauss1 laser n0 52.228253u -18.352737), but they're dumped as:

gauss* gauss1 laser n0 -18.352737 0.00805414127545816

[Sean Leavey]

At least it's not dumping complex numbers when it probably should, right?

[Sean Leavey]

And the string in "gauss* {name} {comp} {node} {w0:.15g} {z:.15g}".format(...) needs changed to reference q and qy instead of w0 and z.

[Daniel Brown]

Yep, the stupid thing in the original code is that the literal strings in the format are wrong:

if self.__q_x == self.__q_y:
    rtn.append("gauss* {name} {comp} {node} {z:.15g} {zr:.15g}".format(name=name, node=self.name, comp=self.__q_comp.name, z=self.__q_x.z, zr=self.__q_x.zr))
else:
    rtn.append("gauss* {name} {comp} {node} {zx:.15g} {zrx:.15g} {zy:.15g} {zry:.15g}".format(name=name, node=self.name, comp=self.__q_comp.name, zx=self.__q_x.z, zrx=self.__q_x.zr, zy=self.__q_y.z, zry=self.__q_y.zr))

but otherwise it should dump the correct values, q is a z zr pair no need for any j or i.

[Sean Leavey]

But the gauss* command wants:

gauss* name component node q [qy]

and it's being given:

gauss* {name} {comp} {node} {z:.15g} {zr:.15g}

i.e. q is being set to z and qy is being set to zr. If I understand correctly, when both q and qy are specified with gauss*, then it assumes q is qx. In that case, we're setting qx to z and qy to zr. That can't be right...?

[Sean Leavey]

Is q not parsed by Finesse as a single (complex) number? Does it allow two numbers with a space between?

[Daniel Brown]

It's the bit in the parenthesis at the end:

gauss* name component node q [qy] (q as 'z z_R')

is:

gauss* name component node qx_z qx_zr [qy_z qy_zr]

[Sean Leavey]

Ah I see, thanks for clarifying that. I'll close this PR.