Question about is_modellist_updated

[bakerwy]

Hi,

Before PR, I would like to ask if it is intended code.

mx_get_modellist(kernel) makes list [current model(or None), model1, model2, ...]
and docstring said

"""Returns a list of model info.

Returns a list of dicts of basic model attributes.
The first element of the list is the current model info.
None if not current model is set.
"""

In is_modellist_updated, when comparing the old list and the new one, the None case is handled for old element (cur) but for the new element (oth) is not handled

for cur, oth in zip(self.modellist, modellist):
    if cur is None:              # old element guarded
        if oth is None:
            continue
        else:
            return True
    if cur["name"] == oth["name"] and cur["id"] == oth["id"]:   # line 780
        ...

so when the current model transitions from set -> None (old element is a dict and new element is None),
it evaluates oth["name"] as None["name"] and raises error. Since both lists come from the same mx_get_modellist,
my understanding is that if cur can be None, then oth can equally be None.

the following code supports your intend? ( I am afraid of misreading the intend)

for cur, oth in zip(self.modellist, modellist):
    if cur is None or oth is None:
        if cur is None and oth is None:
            continue
        return True
    if cur["name"] == oth["name"] and cur["id"] == oth["id"]:
        continue
    return True
return False

[fumitoh]

Thank you for the post. It could be a bug, but when I tested below in MxConsole, I didn't get an error. Am I missing something?
Can you provide sample code that raises the error?

>>> import modelx as mx

>>> m1 = mx.new_model()  # Current Model set to Model1

>>> m1.close() # Current Model set to None

[bakerwy]

Here is a reproduction of the error

Run these in the IPython console as separate executions:

In [1]: import modelx as mx

In [2]: model = mx.new_model("Pricing")
   ...: space = model.new_space("Data")
   ...: space.new_cells(name="premium", formula=lambda t: 100 + t)

In [3]: model.write("pricing_v1")

In [4]: backup = mx.read_model("pricing_v1", name="Backup")
   ...: print(backup.Data.premium(0))   # check a value in the saved copy
   ...: backup.close()                  # done with it, free memory

The refresh after In [4] raises in MxModelList.is_modellist_updated:

TypeError: 'NoneType' object is not subscriptable

Note : In [4] must be a single execution

[fumitoh]

@bakerwy Thank you. I reproduced the error. This is a bug. Creating a PR for this would be very much appreciated.

[bakerwy]

I just created the PR. Thanks.

[fumitoh]

Thank you!