AppTest bug that returns AttributeError: _widget_state

[Breck_Baldwin]

The following test code fails as indicated depending on the params to a widget that I am not even testing:

test_col_val_filter.py

from streamlit.testing.v1 import AppTest
import pandas as pd


def test_col_val_filter():
    at = AppTest.from_file("col_val_filter.py")

    SS = at.session_state
    SS.data = pd.DataFrame(
        {
            "col1": ["v1", "v1", "v3"],
            "col3": [1, 2, 2],
            "col2": ["a", "b", "a"],
            "id": ["i1", "i2", "i3"],
        }
    )
    SS.testing = True
    SS.config = [
        {
            "fn": "column_filter",
            "args": [{"var": "UI"}, 1],
            "import": "col_val_filter",
            "setup": [{"col": "col1", "vals": ["v1"]}],
        }
    ]
    at.run()

    # canary
    test_sb = at.selectbox(key="test_sb")
    assert test_sb.value == ""

    test_sb.set_value(
        "foo"
    ).run()  # fails here with FAILED test/test_col_val_filter.py::test_col_val_filter - AttributeError: _widget_state

    assert test_sb.value == "foo"

This is the code being tested: col_val_filter.py
Commenting any of the spots as indicated breaks the test, otherwise it works.

import streamlit as UI

SS = UI.session_state


def count_cols(SS):
    if "cols_counts" in SS and not SS.get("reset_data", True):
        return
    SS.col_counts = {}
    SS.cols = []
    for col in sorted(SS.data.columns):
        SS.col_counts[col] = len(SS.data[col].unique())
        SS.cols.append(col)
    SS.reset_data = False


def count_chats(col, val):
    non_none_df = SS.data[SS.data[col] == val]
    return len(non_none_df["id"].unique())


def add_count(option):
    if option == "":
        return ""
    return f"{option}:{SS.col_counts[option]}"


def column_filter(UI, i_cell, Context):
    SS = UI.session_state
    count_cols(SS)
    Context.selectbox("Test", options=["", "foo", "bar"], key="test_sb")
    for i, setting in enumerate(SS.config[i_cell]["setup"]):
        option_list = [""] + SS.cols  # comment out SS.cols, works
        Context.selectbox(
            f"Select",
            options=option_list,
            format_func=add_count,  # comment param out and it works
            index=option_list.index(setting["col"]),  # or comment out and it works
            key=f"coll_{i}_sb",
        )


if SS.get("testing", False):  # stands up UI for testing with simple example
    column_filter(UI, 0, UI)

pytest test/test_col_val_filter.py
=========================== test session starts ============================
platform darwin -- Python 3.12.2, pytest-8.1.1, pluggy-1.5.0
rootdir: /Users/fbaldw522/ghe/SpeakPeek/streamlit/cells
collected 1 item                                                           

test/test_col_val_filter.py F                                        [100%]

================================= FAILURES =================================
___________________________ test_col_val_filter ____________________________

    def test_col_val_filter():
        at = AppTest.from_file("col_val_filter.py")
    
        SS = at.session_state
        SS.data = pd.DataFrame(
            {
                "col1": ["v1", "v1", "v3"],
                "col3": [1, 2, 2],
                "col2": ["a", "b", "a"],
                "id": ["i1", "i2", "i3"],
            }
        )
        SS.testing = True
        SS.config = [
            {
                "fn": "column_filter",
                "args": [{"var": "UI"}, 1],
                "import": "col_val_filter",
                "setup": [{"col": "col1", "vals": ["v1"]}],
            }
        ]
        at.run()
    
        # canary
        test_sb = at.selectbox(key="test_sb")
        assert test_sb.value == ""
    
        test_sb.set_value(
            "foo"
>       ).run()  # fails here with FAILED test/test_col_val_filter.py::test_col_val_filter - AttributeError: _widget_state

test/test_col_val_filter.py:34: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
../../../../miniforge3/envs/sppk/lib/python3.12/site-packages/streamlit/testing/v1/element_tree.py:153: in run
    return self.root.run(timeout=timeout)
../../../../miniforge3/envs/sppk/lib/python3.12/site-packages/streamlit/testing/v1/element_tree.py:1839: in run
    widget_states = self.get_widget_states()
../../../../miniforge3/envs/sppk/lib/python3.12/site-packages/streamlit/testing/v1/element_tree.py:1822: in get_widget_states
    w = get_widget_state(node)
../../../../miniforge3/envs/sppk/lib/python3.12/site-packages/streamlit/testing/v1/element_tree.py:1765: in get_widget_state
    return node._widget_state
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

self = Selectbox(key='coll_0_sb', _value=InitialValue(), label='Select', options=['', 'col1:2', 'col2:2', 'col3:2', 'id:3'])
name = '_widget_state'

    def __getattr__(self, name: str) -> Any:
        """Fallback attempt to get an attribute from the proto"""
>       return getattr(self.proto, name)
E       AttributeError: _widget_state

../../../../miniforge3/envs/sppk/lib/python3.12/site-packages/streamlit/testing/v1/element_tree.py:142: AttributeError
--------------------------- Captured stderr call ---------------------------
2024-04-28 13:36:50.469 Session state does not function when running a script without `streamlit run`
============================= warnings summary =============================
<frozen importlib._bootstrap>:488
  <frozen importlib._bootstrap>:488: DeprecationWarning: Type google._upb._message.MessageMapContainer uses PyType_Spec with a metaclass that has custom tp_new. This is deprecated and will no longer be allowed in Python 3.14.

<frozen importlib._bootstrap>:488
  <frozen importlib._bootstrap>:488: DeprecationWarning: Type google._upb._message.ScalarMapContainer uses PyType_Spec with a metaclass that has custom tp_new. This is deprecated and will no longer be allowed in Python 3.14.

-- Docs: https://docs.pytest.org/en/stable/how-to/capture-warnings.html
========================= short test summary info ==========================
FAILED test/test_col_val_filter.py::test_col_val_filter - AttributeError: _widget_state
====================== 1 failed, 2 warnings in 0.69s =======================

OSX Sonomo 14.4
Python 3.12.2
Streamlit 1.33.0

Any help appreciated–awesome product btw.

[Lukas321]

I recently had to solve an issue just like this one. It was with a radio button instead of a selectbox, but the error was exactly the same.

In my case, I was erroneously providing a string as the radio button option, when the widget options actually came from an Enum.

i.e., I was doing this:

# failed here with exactly the same error that you got
app_test.radio[0].set_value("option1").run()

But in the radio button that I was trying to access, the options were actually part of an enum, like this:

import streamlit as st
from enum import Enum

class RadioOptions(Enum):
    option1 = "option1"
    option2 = "option2"
    option3 = "option3"

st.radio(
    "My Radio Button",
    options=[option for option in RadioOptions],
    format_func=lambda option: option.value,
    index=0,
)

So the solution was to do this instead:

app_test.radio[0].set_value(RadioOptions.option1).run()

Maybe your situation could be fixed by using an enum to enumerate the options for your selectbox, if possible? Otherwise, I’m not sure if the error comes from using a string in the set_value method, or if it happens if the value passed in the set_value method is not actually a member of the widget options.