Why does is_adult_for_medicaid return different results for 2024 when printed locally vs. running in a Python notebook?

[myang1010]

Hi all, I'm experimenting with the Python notebook and ran into a question. When I check is_adult_for_medicaid for the 2024 period locally, I get False, but when I run the same test in the Python notebook, I get True. Here's the PE household link. Am I missing something here? Thanks!

• Python notebook output

• My local print

To print locally, I added below print lines inside policyengine_us/variables/gov/hhs/medicaid/eligibility/categories/medicaid_category.py,

        for variable in variable_to_category.keys():
            print("variable:", variable, "period: ", period)
            result = person(variable, period)
            print("Result of person:", result)

[PavelMakarchuk]

@myang1010 thank you for filing - a couple of question that could help clarify.

1. What operating system are you using?
2. Could you elaborate on what you mean by "running locally". Are you defining the household situation in a local .ipynb file or are you constructing a .yaml style integration test with the same household inputs as the linked example?

I also noticed that in the shared notebook output you print is_adult_for_medicaid_fc which only checks the adult status based on the financial criteria - to get the actual adult status you can print is_adult_for_medicaid instead

[PavelMakarchuk]

Could you share the .yaml test code, I will look into this?

[myang1010]

Hi @PavelMakarchuk , sorry about my earlier inaccurate reply. I don’t have any specific .yaml test code, but I was testing using our screener. I've attached the input data sent to the household API in case it helps! If not, I'm open to have a call and show you what I did to reproduce this. Thanks.

person_used_for_testing.txt

[MaxGhenis]

Please determine the PolicyEngine request from the screener call -- we cannot debug the MyFriendBen tool.

[myang1010]

Here's the cleaned up PolicyEngine request,
pe_input_data.txt

[myang1010]

I just realized that the "employment_income" for the first person in my notebook was set to monthly. After updating it to yearly, the results are now as expected.