Pattern matching with particles

[tueda]

After generating diagrams, people usually want to manipulate them. It is unclear to me what the pattern-matching rules for particles are. Functions can't match them:

FORM 5.0.0-beta.1 (Jan 24 2025, v5.0.0-beta.1-115-g1d5e40b)  Run: Wed Jan 29 22:51:39 2025
    On names;

    Model PHI3;
      Particle phi,1;
      Vertex phi,phi,phi:g;
    EndModel;

    CF f1;
    V p1,q1,k1,k2;
    L F = diagrams_(PHI3,{phi},{phi},{p1,q1},{k1,k2},1);
    repeat id node_(?a,f1?(p1?),?b) = node_(?a,?b);
    P +ss;
    .end

 Symbols
   g
 Vectors
   p1 q1 k1 k2
 Commuting Functions
   phi(Particle) f1
 Sets
   PHI3: 0
   {}: phi
   {}: p1 q1
   {}: k1 k2
 Expressions
   F(local)
 Expressions to be printed
   F

Time =       0.00 sec    Generated terms =          1
               F         Terms in output =          1
                         Bytes used      =        504

   F =
       +
         1/2
         *topo_(1)
         *node_(1,1,phi(-p1))
         *node_(2,1,phi(-q1))
         *node_(3,g,phi(p1),phi(-k1),phi(-k2))
         *node_(4,g,phi(q1),phi(k1),phi(k2))
      ;

  0.00 sec out of 0.00 sec

Is this the intended behaviour?

A workaround is to declare f1 as a particle:

Model TEMP;
  Particle f1;
EndModel;

but it requires a unique model name unnecessarily, which is not nice.

A deeper question is whether we should distinguish between fermions and bosons or rather by other particle properties.

[jodavies]

One can still use phi itself for the pattern matching here, right? (Or another particle that might be declared in PHI3?)

This seems to match fine:

Model PHI3;
  Particle phi,1;
  Particle phi2,1;
  Vertex phi,phi,phi:g;
  Vertex phi2,phi2,phi2:g;
EndModel;

CF f1;
V p1,q1,k1,k2;
L F = diagrams_(PHI3,{phi},{phi},{p1,q1},{k1,k2},1);
L G = diagrams_(PHI3,{phi2},{phi2},{p1,q1},{k1,k2},1);

repeat id node_(?a,phi?(p1?),?b) = node_(?a,?b);

P +ss;
.end

[tueda]

You can use phi if you know you have the PHI3 model. Otherwise (when writing a model-independent general routine) you need a unique model:

Model TEMP;
  Particle phi;
EndModel;

[jodavies]

OK, I see your point. Though to some extent this works the same way with other types; a "context independent" procedure needs to make sure symbols etc used for pattern matching are defined, so defining a dummy model and fields for patterns is somehow the FORM style.