Skip to content

test data: same data, different formats#990

Open
artoonie wants to merge 1 commit intodevelopfrom
feature/same-data-different-format-tests
Open

test data: same data, different formats#990
artoonie wants to merge 1 commit intodevelopfrom
feature/same-data-different-format-tests

Conversation

@artoonie
Copy link
Copy Markdown
Collaborator

@artoonie artoonie commented Mar 24, 2026

These tests are designed to aid with testing autoload (PR #987)

This PR takes an existing CDF CVR, duplicates it, and copies the data into the ES&S format. Still TODO: handle other formats. I think this PR can be merged without having all providers in here.

How this works:

  • A base config file has an empty CVR sources section, base_config
  • There are stem_config.json files which populate the CVR sources to generate a runnable config
  • We generate the config for each stem and tabulate
  • We compare all the output files, JSON and CSV, to ensure equivalence between providers

@artoonie artoonie force-pushed the feature/same-data-different-format-tests branch from f362e75 to f0d3584 Compare March 24, 2026 19:53
@artoonie artoonie force-pushed the feature/same-data-different-format-tests branch from f0d3584 to 50c2a66 Compare March 24, 2026 20:01
@yezr
Copy link
Copy Markdown
Collaborator

yezr commented Mar 27, 2026

This is great to enable the minor config changes needed for these new tests!

The sameDataDifferentFormats test that was added here, does that test any tabulator functionality specifically? Or is that mainly just a basic example of how to use this new harness to change the configs? I guess for now it is a test that whatever we did to copy the CVR data from one provider to another works? And if we can count on that to be the same, then we can use it for the GuiTests that are coming with #991?

That CVR copying code could also be interesting to explicitly check the assumptions we make for each provider? Though that could bloat this repo.

@artoonie
Copy link
Copy Markdown
Collaborator Author

This essentially does not test anything new, other than the confirmation that there is no provider-specific information in the outputs.

The goal is just to create a basic example of this new test harness, which should eventually include all providers.

Creating the ES&S data was easy, as I just used the RCTab CVR format and modified it a touch. I started with an existing CDF file because that data is the hardest to manually create.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants