support for function incur in constraint functions about pyvsc

[hanfeng0114]

Like the following feature:
'
@vsc.constraint
def basic_c(self):
with vsc.foreach(self.GroupList, idx=True, it=False) as idx:
self.GroupList[idx]<= self.mThreadNum
with vsc.foreach(self.GroupList, idx=True, it=False) as idx:
self.GroupList[idx]<= self.mThreadNum
with vsc.foreach(self.GroupList, idx=True, it=False) as idx:
with vsc.if_then(self.GroupList[idx] > 1):
self.GroupListScore[idx] == 1
with vsc.else_then():
self.GroupListScore[idx] == 0
self.GroupListScore.sum < self.getSum(self.GroupListScore)
self.GroupList.sum == self.mThreadNum

        def getSum(self, aList):
            return sum(aList)

'

[mballance]

Can you provide a full example that includes the definition of data fields?
Currently, PyVSC invokes functions when the constraint is built. So, getSum would be invoked once to obtain the size of the GroupListScore. Obviously, this behavior might not be what is desired.

[hanfeng0114]

For example, The following code doesn't run well for line 94:

Line 94 doesn't have caculation functions.
My question is how to meet similar requirement?

[mballance]

Hi @hanfeng0114,
Thank you for the full example. I created a small testcase based on the example. The error I receive from running this testcase is shown below:

  File "/project/fun/pyvsc/pyvsc-20230821/src/vsc/rand_obj.py", line 208, in build_field_model
    fo.c(self)
  File "/project/fun/pyvsc/pyvsc-20230821/ve/unit/test_scalar_array.py", line 348, in basic_c
    self.GroupList.sum == self.mThreadNum * self.mSharePercent/100
  File "/project/fun/pyvsc/pyvsc-20230821/src/vsc/types.py", line 84, in __eq__
    return self.bin_expr(BinExprType.Eq, rhs)
  File "/project/fun/pyvsc/pyvsc-20230821/src/vsc/types.py", line 72, in bin_expr
    to_expr(rhs)
  File "/project/fun/pyvsc/pyvsc-20230821/src/vsc/types.py", line 338, in to_expr
    raise Exception("Element \"" + str(t) + "\" isn't recognized, and doesn't provide to_expr (type=%s)" % str(type(t)))
Exception: Element "6.4" isn't recognized, and doesn't provide to_expr (type=<class 'float'>)

The core issue is that the right-hand side of the equation produces a floating-point result and PyVSC doesn't currently handle this. I receive proper results if I explicitly convert the right-hand side to an integer:

    self.GroupList.sum == int(self.mThreadNum * self.mSharePercent/100)

I recommend using explicit conversion for now. In the future, I will either improve the error messaging or provide direct support for using floating-point expressions in integer constraint expressions.

[mballance]

The 0.8.4 release contains changes to automatically convert floating-point number used in constraints to integers. I believe this is the best way to address this issue.
Thank you for the issue report!