is None not working

[Aimen]

Hi,

I'm facing a problem with the use of "is None"
I'm using forEachIncludingNullVars()
I have None values that i would like to penalize, but they dont seem to be detected.
When i try to reward the not None values it works.

Is it a bug or i'm missing something ?

Same when i try to avoir null pointer exceptions.
When i code, before making operations on some variables i set X is not None and Y is not None and ( X.a !=Y.a ). This avoids null pointer exceptions.
But with optapy, i still get null pointer exceptions.
Am I doing something wrong ?

Thanks :)

[Christopher Chianelli]

Can you please paste some code (that include your planning entities, problem facts, planning solution and constraints) that can be run to replicate the NullPointerException?

[Aimen]

Hi Christopher,

Here are the classes and more details about the error:

class Resp:
id: int
name: str
date_dispo: datetime.date
creneau_dispo: str
last_rdv: datetime.date
collab: str
param_strat: str
def init(self, id, name_resp, date_dispo, creneau_dispo, last_rdv,collab,param_strat):
self.id = id
self.name_resp = name_resp
self.date_dispo=date_dispo
self.creneau_dispo=creneau_dispo
self.last_rdv=last_rdv
self.collab=collab
self.param_strat=param_strat

@planning_id
def get_id(self):
    return self.id

def __str__(self):
    return f"id={self.id}, name_resp={self.name_resp}, date_dispo={self.date_dispo}, creneau_dispo={self.creneau_dispo}, last_rdv={self.last_rdv}, collab={self.collab}, param_strat={self.param_strat}\n"
def __repr__(self):
    return self.__str__()

@planning_entity
class Talent:
id: int
name: str
dateslot: Dateslot
resp: Resp
talent_creneau: Talent_creneau
def init(self, id, name, dateslot=None, resp=None, talent_creneau=None):
self.id = id
self.name = name
self.dateslot = dateslot
self.resp = resp
self.talent_creneau=talent_creneau

@planning_id
def get_id(self):
    return self.id

@planning_variable(Dateslot, ["dateslotRange"],nullable = False)
def get_dateslot(self):
    return self.dateslot

def set_dateslot(self, new_dateslot):
    self.dateslot = new_dateslot

@planning_variable(Resp, ["respRange"],nullable = True)
def get_resp(self):
    return self.resp

def set_resp(self, new_resp):
    self.resp = new_resp

contraint where the problem is located:

def pen_resp_notnone(constraint_factory: ConstraintFactory):
return constraint_factory.forEachIncludingNullVars(Talent)
.flatten_last(lambda talent: talent)
.filter(lambda talent:talent.resp is None )
.penalize("pen_resp_notnone" ,HardMediumSoftScore.ONE_MEDIUM)

Error :
java.lang.RuntimeException: java.lang.RuntimeException: org.optaplanner.jpyinterpreter.types.errors.AttributeError: object 'Talent(6, None, None, 32f2bd91-5482-46e4-a32e-7965a74191b5,None)
' does not have attribute 'resp'

[Aimen]

And when i write this, i don't get any penality, like if it didn't detect any thing :

def pen_resp_notnone(constraint_factory: ConstraintFactory):
return constraint_factory.forEachIncludingNullVars(Talent)
.filter(lambda talent:talent.resp is None )
.penalize("pen_resp_notnone" ,HardMediumSoftScore.ONE_MEDIUM)

[Christopher Chianelli]

Thanks for the reproducer. A tip on formatting code on GitHub: use

```language
code
```

when creating code blocks, which will cause the code to be properly formatted and indented. For instance,

```python

def my_fun():
    print('hello world')

```

would appear as

def my_fun():
    print('hello world')

The flatten_last 'trick' is interesting, since that code should error before the filter (since Talent is not iterable). Nonetheless, it is a bug, and can be observed without the flatten_last by using print (or in general, an operation that is not is).

[Aimen]

Thanks, i'm not familiar with github. If i understand correctly your answer, there is a bug in optapy? Or it's in my code ?

If it's in optapy, do you have any idea when the fix will be released?

Have a nice day!

[Christopher Chianelli]

Both (flatten_last should not be used to flatten something that not an iterable (user + optapy), and Java null is not being converted to Python None (optapy)). I can't give a firm time line for the fix, but I'll investigate it when I have time.

[Christopher Chianelli]

Fixed by #152