Fix vignette correctness comparison ignoring metadata #12
Conversation
There was a problem hiding this comment.
Codex Review: Here are some suggestions.
Reply with @codex fix comments to fix any unresolved comments.
About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you open a pull request for review, mark a draft as ready, or comment "@codex review". If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex fix this CI failure" or "@codex address that feedback".
| # Drop the bloomjoin-specific metadata/class so we compare the raw data only | ||
| arranged_bloom <- as.data.frame(arranged_bloom) | ||
| arranged_baseline <- as.data.frame(arranged_baseline) | ||
| stopifnot(identical(arranged_bloom, arranged_baseline)) |
There was a problem hiding this comment.
[P1] Equality check still fails to remove bloom metadata
The new as.data.frame() calls only strip the custom bloomjoin class but they do not remove the bloom_metadata attribute added in bloom_join() (R/bloomjoin.R sets attr(result, "bloom_metadata") <- metadata). identical() compares attributes as well as classes, so this chunk will still raise because arranged_bloom retains bloom_metadata while arranged_baseline does not. To actually compare only the row data you also need to clear that attribute (e.g., attr(arranged_bloom, "bloom_metadata") <- NULL) or use an attribute-insensitive comparison.
Useful? React with 👍 / 👎.
Summary
Testing
Rscript -e 'rmarkdown::render("vignettes/benchmarking-bloomjoin.Rmd", output_dir = tempdir())'(fails: Rscript not available in container)https://chatgpt.com/codex/tasks/task_e_68d0e0af213c832fad22f11d6140dd5d