I encountered an error while replicating the dismod_MR tool

[yannie228]

I hope to get your help. The error is as follows:

[aflaxman]

I'm sorry to hear that you are having trouble using the dismod_mr tool. I can't tell what is going wrong from this screenshot alone. Can you produce a "minimal complete verifiable example" of the issue you are having? https://stackoverflow.com/help/minimal-reproducible-example

[yannie228]

Thanks for your prompt reply, maybe you can try running this ipynb file with your pd_sim_data folder？
DisMod-MR 2.1-Copy1.ipynb.zip

[aflaxman]

I had a little trouble getting the data to load (I had to change the directory to get dismod_mr.data.load('pd_sim_data') to succeed, but everything else in your notebook worked fine for me. Please try re-running everything in a fresh conda environment, following the directions on the README.md file, and see if there is perhaps something in your current python environment that is causing this issue.

[aflaxman]

I have not had a chance to look at this, but I thought I’d send a quick reply to say that I received it.

[fanwayne]

Hello, I think I have the same issue when running the example code recently.

The code was copied from the page "A motivating example: descriptive epidemiological meta-regression of Parkinson's Disease".

model.vars += dismod_mr.model.asr(model, 'p')

And below error occurs when running the asr function.

`
IndexError Traceback (most recent call last)
File c:\ProgramData\miniconda3\lib\site-packages\pymc\model.py:197, in ContextMeta.get_context(cls, error_if_none)
196 try:
--> 197 candidate: Optional[T] = cls.get_contexts()[-1]
198 except IndexError as e:
199 # Calling code expects to get a TypeError if the entity
200 # is unfound, and there's too much to fix.

IndexError: list index out of range

During handling of the above exception, another exception occurred:

TypeError Traceback (most recent call last)
File c:\ProgramData\miniconda3\lib\site-packages\pymc\distributions\distribution.py:276, in Distribution.new(cls, name, rng, dims, initval, observed, total_size, transform, *args, **kwargs)
274 from pymc.model import Model
--> 276 model = Model.get_context()
277 except TypeError:

File c:\ProgramData\miniconda3\lib\site-packages\pymc\model.py:202, in ContextMeta.get_context(cls, error_if_none)
201 if error_if_none:
--> 202 raise TypeError(f"No {cls} on context stack")
203 return None

TypeError: No <class 'pymc.model.Model'> on context stack
...
283 )
285 if "testval" in kwargs:
286 initval = kwargs.pop("testval")

TypeError: No model on context stack, which is needed to instantiate distributions. Add variable inside a 'with model:' block, or use the '.dist' syntax for a standalone distribution.
`

[aflaxman]

Sorry for the challenge of getting it to work… I believe that this issue you are running into is because dismod_mr is based on an older version of pymc. Please see if uninstalling your pymc and installing an older version makes the example work for you:

pip uninstall pymc
pip install pymc==2.3.8

[aflaxman]

Note also that the best python version for this is 3.6, as pymc==2.3.8 has some issues with the most current version of python (3.11). My apologies again for the challenge of getting this to work!

[fanwayne]

@aflaxman Many thanks for your support. It works fine in the local conda environment.

[xunluning]

Hello, I hope to get your help. The error is as follows:
pred=dismod_mr.model.predict_for(model,model.parameters['p'],'all',"female",2005,'GBR','female',2005,1.,model.vars['p'],0.,1.)
Traceback (most recent call last):
File "D:\Anaconda\envs\dismod_mr\lib\site-packages\networkx\classes\digraph.py", line 809, in successors
return iter(self._succ[n])
KeyError: 'GBR'
During handling of the above exception, another exception occurred:
Traceback (most recent call last):
File "D:\Anaconda\envs\dismod_mr\lib\code.py", line 91, in runcode
exec(code, self.locals)
File "", line 1, in
File "D:\Anaconda\envs\dismod_mr\lib\site-packages\dismod_mr\model\covariates.py", line 389, in predict_for
leaves = [n for n in nx.traversal.bfs_tree(area_hierarchy, area) if area_hierarchy.successors(n) == []]
File "D:\Anaconda\envs\dismod_mr\lib\site-packages\networkx\algorithms\traversal\breadth_first_search.py", line 196, in bfs_tree
T.add_edges_from(edges_gen)
File "D:\Anaconda\envs\dismod_mr\lib\site-packages\networkx\classes\digraph.py", line 681, in add_edges_from
for e in ebunch_to_add:
File "D:\Anaconda\envs\dismod_mr\lib\site-packages\networkx\algorithms\traversal\breadth_first_search.py", line 145, in bfs_edges
for e in generic_bfs_edges(G, source, successors, depth_limit):
File "D:\Anaconda\envs\dismod_mr\lib\site-packages\networkx\algorithms\traversal\breadth_first_search.py", line 76, in generic_bfs_edges
queue = deque([(source, depth_limit, neighbors(source))])
File "D:\Anaconda\envs\dismod_mr\lib\site-packages\networkx\classes\digraph.py", line 811, in successors
raise NetworkXError("The node %s is not in the digraph." % (n,))
networkx.exception.NetworkXError: The node GBR is not in the digraph.

[aflaxman]

I bet we are getting errors with modern versions of networkx now, too :( Can you check if your networkx version is 2.3?

[xunluning]

Thanks for your prompt reply, I have checked. Actually, the nextworkx version in my conda is 2.3.

[xunluning]

Thanks. It works successfully now. Mybe there was some error when l used 'model.keep()' command.

[aflaxman]

Yay! Mysterious, but I'll consider it a closed case. Please reopen if you have more trouble.