ln 323 in streamlit_post_processing.py was recently changed "x_value": y_scaling_x → "x_value": str(y_scaling_x) to fix a config export issue, however this broke scaling functionality in the web app.
We could try to fix this by typing the scaling x-value before performing comparisons with it, such as:
under ln 251 in streamlit_post_processing.py adding and using typed_x_val = state.post.val_as_col_dtype(axis["scaling"]["column"]["x_value"], state.x_axis_column).iloc[0]
However, these fixes alone are not enough because we must find a way to get around the error thrown if the app tries to type a "None" string as numeric (specifically, I found this was an issue when typing to Int64).
ln 323 in streamlit_post_processing.py was recently changed
"x_value": y_scaling_x
→"x_value": str(y_scaling_x)
to fix a config export issue, however this broke scaling functionality in the web app.We could try to fix this by typing the scaling x-value before performing comparisons with it, such as:
scaling_x_value_mask = (self.df[x_column] == scaling_column.get("x_value"))
→scaling_x_value_mask = (self.df[x_column] == self.val_as_col_dtype(scaling_column["x_value"], x_column).iloc[0])
typed_x_val = state.post.val_as_col_dtype(axis["scaling"]["column"]["x_value"], state.x_axis_column).iloc[0]
However, these fixes alone are not enough because we must find a way to get around the error thrown if the app tries to type a "None" string as numeric (specifically, I found this was an issue when typing to Int64).