Material Specifications: cannot unpack non-iterable ValueNode object

[SterlingButters]

@MicahGale How are material cards supposed to be formatted for MontePy to parse them? I keep getting the error in the subject line. When I inspect the isotope_fractions variable in \data_inputs\material.py, it is a ListNode object containing ValueNode objects. For any ValueNode object, I see no attribute that corresponds to the isotopic fraction (at least not one intuitively named). Additionally this object is non-iterable so the usage isn't correct either. Am I missing something?

[MicahGale]

This seems like a bug most likely. MontePy should be able to handle a material definition as long as the parameters come at the end.

So this should work

m1 1001.80c 0.33
      8016.80c 0.666667
      plib=80p

This syntax breaks it:

m1 1001.80c 0.33
      plib=80p
      8016.80c 0.666667

These object types that you are seeing are data types from the syntax tree. A bad syntax should be handled with syntax errors. So this is definitely odd.

Could you provide an example made up material that causes this bug?

[SterlingButters]

Is the plib keyword required? I specify my materials like (without plib):

m1 1001.80c 0.33
      8016.80c 0.666667

[MicahGale]

No plib is not at all required. It was just an example of key-value parameter that you can use.

That looks like it should work. I'll have to look into this further.

[SterlingButters]

I've never used SLY so I'm just throwing ideas at the wall but is the issue possibly linked to _flush_phrase never being called for the isotope_fractions (since _flush_phrase is what generates the ValueNode - although this doesn't make a ValueNode iterable but maybe you could zip the ZAID and fractions ValueNode's)?

[MicahGale]

That example code I just dropped in MontePy and it didn't have any issues as I suspected. Could you get us a stack trace and MWE?

As for:

I've never used SLY so I'm just throwing ideas at the wall but is the issue possibly linked to _flush_phrase never being called for the isotope_fractions (since _flush_phrase is what generates the ValueNode - although this doesn't make a ValueNode iterable but maybe you could zip the ZAID and fractions ValueNode's)?

So _flush_phrase isn't a SLY thing, that's just a utility function I made. First this guide should give a quick intro to it. Basically _flush_phrase is just used for generating ValueNode as leaves of the tree quickly. isotopes_factions which is of type IsotopeNode is a list of pairs of ValueNodes that represent the ZAID and then the atom/weight fraction. So the issue is that somehow a ValueNode is getting put too high in the syntax tree that the Material object as it tries to walk the Syntax Tree is finding it instead of a IsotopesNode (which IIRC is a subclass of ListNode) for some reason.

Dynamic typing strikes again.

Test I ran:

In [1]: import montepy
 
In [2]: input = montepy.input_parser.mcnp_input.Input("""m1 1001.80c 0.33
   ...:       8016.80c 0.666667""".split("\n"), montepy.input_parser.block_type.BlockType.DATA)

In [3]: material = montepy.data_inputs.material.Material(input)

In [4]:

In [4]: material
Out[4]:
MATERIAL: 1 fractions: atom
 H-1   (80c) 0.33
 O-16  (80c) 0.666667

[SterlingButters]

@MicahGale Ahah! Found the issue - I lied to you. I did not, in fact, supply the .abx suffix - which, as you probably know:

If .abx is omitted the first entry from the xsdir file corresponding to the proper library class will
be used. This same approach will be used for all libraries not specified.

[SterlingButters]

@MicahGale By the way, are we adding of these to your unit tests (or maybe these aren't worthy)?

[MicahGale]

So to confirm the problem is:

m1 1001 0.333
    8016 0.6667

Yes I know this absolutely valid and can't believe I missed this as a test.

So yes we will absolutely add this to our test suite. It is the expectation for any bug fixes. What I will usually do first is write/modify a test to replicate the bug. Ensure that it fails, then push that test to the PR so that way a record of the test failing is preserved. This just makes it nicer for the the reviewer ok really @tjlaboss to verify that the test is effective. I think 100% of the tests in tests/test_edge_cases.py comes from bug reports.

[MicahGale]

Error confirmed. Full stack trace:

---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
Cell In[3], line 1
----> 1 material = montepy.data_inputs.material.Material(input)

File ~/dev/montepy/montepy/data_inputs/material.py:40, in Material.__init__(self, input)
     38 set_atom_frac = False
     39 isotope_fractions = self._tree["data"]
---> 40 for isotope_node, fraction in isotope_fractions:
     41     isotope = Isotope(node=isotope_node)
     42     fraction.is_negatable_float = True

TypeError: cannot unpack non-iterable ValueNode object