Working to add checks for duplicated comparison features, feature selection, and column mappings #113
Conversation
…s, and column_mappings
|
For some reason, this isn't letting me make it a draft right now...but I am adding additional tests. Also, no clue why I made the name of the branch same_output. |
|
Ok, now it's ready |
|
This is for issue #50. |
There was a problem hiding this comment.
Looking mostly good! I left a few comments.
I think that pipeline_features could use a similar check. That wasn't mentioned in the issue, sorry about that.
In the error messages I would add some more information. I would include
- the name of the section, like
[[comparison_features]] - short instructions on the problem and how to fix it, maybe like "duplicate aliases are not allowed. Please edit your config file to use unique aliases for each comparison feature."
- a comma-separated list of the offending aliases or column names, not in Python format
|
I'm realizing I didn't actually address the big top comment and only the little ones |
…read multiple times with different aliases
|
Doing some more testing of this this morning. I think we may need a few more changes. Not ready to merge yet. |
There was a problem hiding this comment.
Ah, I think I was wrong in one of my comments, sorry.
comparison_features are allowed to overwrite column_mappings. We do this often with the fetch_a comparison feature. So the comparison_features check should just be for duplicate aliases.
The column_mappings code still looks good. We can keep the stricter checking for feature_selections too, I think. I will double-check that.
|
Just updated! Is that back to what you're expecting? |
This is meant to address one of the GitHub issues. This adds checks for duplicates in those three sections to hopefully prevent against errors later on.
Do these checks look like enough?